View AD1843JS_927051.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

a serial-port 16-bit soundcomm codec ad1843 features single chip integrated speech, audio, fax and modem codec highly configurable stereo ?d adcs and quad ?d dacs supports v.34, v.32bis, and fallback modem standards as well as voice over data dual digital resamplers with programmable input and output phase and frequency three on-chip phase lock loops for synchronization to external signals, including video thirteen analog inputs and seven analog outputs advanced analog and digital signal mixing and digital- to-digital sample rate conversion programmable gain, attenuation and mute on-chip signal filters digital interpolation and decimation analog output low pass 1 hz resolution programmable sample rates from 4 khz to 54 khz derived from a single clock input 80-lead pqfp and 100-lead tqfp packages operation from +5 v or mixed +5 v/+3 v supplies fifo-buffered serial digital interface compatible with adsp-21xx fixed-point dsps advanced power management vhdl model of serial port available; evaluation board and mafe board available general product description the ad1843 soundcomm? codec is a complete analog front end for high performance dsp-based telephony and audio ap- plications. the device integrates the real-world analog i/o re- quirements for many popular functions thereby reducing size, power consumption, and system complexity. the ad1843 soundcomm is the worlds first codec which can support four different sample rates simultaneously, without any beat fre- quency noise issues. this is essential for highly integrated audio/ modem/fax products since the sample rates associated with au- dio are very much distinct from the sample rates associated with telephony-oriented data communication. it is also the first codec to offer on-chip digital phase lock loops for sample rate synchro- nization to external clock signals. this sample rate flexibility is enabled through analog devices continuous time oversampling (cto) technology. the main elements of the ad1843 are its extensive input and mix- ing section, its two channels of sigma-delta ( ?d ) analog-to-digital conversion, its four channels of ?d digital-to-analog conversion, its digital filters, and the clock and control circuitry for implementing the devices different modes. the ad1843 permits flexible sample- rate selection through programming and external synchronization, many input and output options, and many mixing options. (continued on page 11) soundcomm is a trademark of analog devices, inc. rev. 0 information furnished by analog devices is believed to be accurate and reliable. however, no responsibility is assumed by analog devices for its use, nor for any infringements of patents or other rights of third parties which may result from its use. no license is granted by implication or otherwise under any patent or patent rights of analog devices. ? analog devices, inc., 1996 one technology way, p.o. box 9106, norwood, ma 02062-9106, u.s.a. tel: 617/329-4700 fax: 617/326-8703 simplified functional block diagram 20 db 4 3 gnda v cc ?d adc gam = gain attenuation mute gam gam fifo adc dac1 dac2 d i g i t a l i n t e r f a c e 2 control registers left and right channels left and right channels gndd v dd filtl cmout v ref filtr 4 2 hpoutr hpoutc hpoutl lout1 2 mout 2 2 2 min aux3 aux2 aux1 mic lin 3 sync xtal conv bit clkout 4 89 lout2 aafiltl aafiltr 1 sum reset pwrdwn 2 pdmnft mute gam gam gam gam gam gam mute mute mute attn 2 3 attn mute s e l e c t o r mute sclk sdfs sdi sdo bm cs tso tsi xctl [1:0] ?a law 3 clock generation attn m u t e mute driver ?a law fifo ?a law s e l e c t o r s e l e c t o r pga ?d dac ? ? ? ? ? ?d dac mute ? ? ? m u t e attn ? voltage reference ad1843
analog input min typ max units full-scale input voltage (rms values assume sine wave input) all inputs with adrflt & adlflt = 0 and linlsd & linrsd = 0 1 v rms (linlp, linrp, aux1l, aux1r, aux2l, 2.55 2.828 3.1 v p-p aux2r, aux3l, aux3r, min) all inputs with adrflt & adlflt = 0 and linlsd & linrsd = 1 2 v rms (linlp & linln, linrp & linrn) 5.1 5.656 6.2 v p-p all inputs with adrflt & adlflt = 1 and linlsd & linrsd = 0 1.127 v rms (linlp, linrp, aux1l, aux1r, aux2l, aux2r, aux3l, 2.8 3.156 3.5 v p-p aux3r, min) all inputs with adrflt & adlflt = 1 and linlsd & linrsd = 1 2.254 v rms (linlp & linln, linrp & linrn) 5.6 6.312 7.0 v p-p mic with +20 db gain (lmge & rmge = 1 and adrflt & adlflt = 0) 0.1 v rms (micl, micr) 0.25 0.2828 0.31 v p-p mic with 0 db gain (lmge & rmge = 0 and adrflt & adlflt = 0) 1 v rms (micl, micr) 2.55 2.828 3.1 v p-p aux, sum and min input impedance* 10k w (aux1l, aux1r, aux2l, aux2r, aux3l, aux3r, suml, sumr, min) lin input impedance* (linlp, linln, linrp, linrn) 40k w mic input impedance* (micl, micr) 20k w input capacitance* (all inputs) 15 pf programmable gain amplifierCadc min typ max units step size (0 db to 22.5 db) (all steps tested) 1.3 1.5 1.7 db pga gain range span* 21.5 22.5 23.5 db input (aux1, aux2, aux3, min, mic) analog amplifiers/attenuators min typ max units step size (+12.0 db to C30 db) (all steps tested) 1.25 1.5 1.75 db step size (C31.5 db to C34.5 db) (all steps tested) 1.1 1.5 1.9 db input gain/attenuation range* 45.5 46.5 47.5 db mute attenuation* C80.0 db ad1843Cspecifications rev. 0 C2C standard test conditions unless otherwise noted temperature 25 c digital supply (v dd ) 5.0 v analog supply (v cc ) 5.0 v sample rate (f s ) 48 khz input signal 1008 hz analog output passband 20 hz to 20 khz adc fft size 2048 dac fft size 8192 v ih 2.0 v v il 0.8 v v oh 2.4 v v ol 0.4 v i oh C2 ma i ol 2ma adc input conditions mic 20 db gain disabled lin single-ended (linlsd & linrsd = 0) autocalibrated 0 db pga gain C1.0 db relative to full scale line input 16-bit linear mode dac conditions autocalibrated 0 db attenuation 0 db relative to full scale 16-bit linear mode no output load mute off dac1 single-ended dac2 differential
ad1843 rev. 0 C3C digital decimation and interpolation filtersCaudio mode* min max units passband 0 0.40 f s hz passband ripple 0 C0.016 db transition band 0.4 f s 0.6 f s hz stopband 1 0.6 f s hz stopband rejection 91.8 db group delay 15/f s s group delay variation over passband 0.0 m s digital decimation and interpolation filtersCmodem mode* min max units passband 0 0.442 f s hz passband ripple 0 C0.220 db transition band 0.442 f s 0.542 f s hz stopband 2 0.542 f s hz stopband rejection 75.7 db group delay 19/f s s group delay variation over passband 0.0 m s sample rate 24 khz digital decimation and interpolation filtersCresampler mode* min max units passband 0 0.4 f s hz passband ripple 0 C0.035 db transition band 0.4 f s 0.5 f s hz stopband 3 0.5 f s hz stopband rejection 92.2 db group delay 25/f s s group delay variation over passband 0.0 m s analog-to-digital converters min typ max units audio dynamic range (C60 db input, thd+n referenced to full scale, a-weighted, adrflt & adlflt = 0) 80 85 db modem dynamic range (C60 db input, thd+n referenced to full scale, 300 hz to 4 khz analog output passband, linrsd & linlsd = 1, adrflt & adlflt = 1, f s = 12.8 khz) 87 90 db audio thd+n (referenced to full scale) 0.03 % C74 C70 db modem thd+n (C3.0 db referenced to full scale, 300 hz to 4 khz analog output passband, linrsd & linlsd = 1, adrflt & adlflt = 1, f s = 12.8 khz) 0.02 % C78.5 C74 db audio signal-to-intermodulation distortion* (ccif method) C94 C80 db adc crosstalk* lin inputs (input l, ground r, read r; input r, ground l, read l) C80 db line to mic (input lin, ground and select mic, read both channels) C80 db line to aux1, aux2, aux3, min C80 db interchannel gain mismatch (difference of gain errors) 0.5 db adc offset error 10 50 mv
rev. 0 C4C ad1843 dac1 digital-to-analog converters min typ max units audio dynamic range (C60 db input, thd+n referenced to full scale, a-weighted, da1flt = 0) 77 80 db audio thd+n (referenced to full scale, da1flt = 0) 0.03 % C74 C70 db audio signal-to-intermodulation distortion* (ccif method) C92 C80 db interchannel gain mismatch (difference of gain errors) 0.5 db dac crosstalk* (input l, zero r, measure lout1r; input r, zero l, measure lout1l) C77 db total out-of-band energy* (measured from 0.6 f s to 100 khz in audio mode) C60 db audible out-of-band energy* (measured from 0.6 f s to 22 khz in audio mode, tested at f s = 8.0 khz) C72 db dac2 digital-to-analog converters min typ max units audio dynamic range (C60 db input, thd+n referenced to full scale, a-weighted, da2flt = 0) 78 80 db modem dynamic range (C60 db input, thd+n referenced to full scale, 300 hz to 4 khz analog output passband, da2flt = 1, rda2g5:0 & lda2g5:0 = 000101 [4.5 db], f s = 12.8 khz) 87 90 db audio thd+n (referenced to full scale, da2flt = 0) 0.03 % C77 C70 db modem thd+n (C3.0 db referenced to full scale, 300 hz to 4 khz analog output passband, da2flt = 1, rda2g5:0 & lda2g5:0 = 000101 [4.5 db], f s = 12.8 khz) 0.016 % C81 C76 db audio signal-to-intermodulation distortion* (ccif method) C86 C80 db interchannel gain mismatch (difference of gain errors) 0.5 db dac crosstalk* (input l, zero r, measure lout2r; input r, zero l, measure lout2l) C80 db total out-of-band energy* (measured from 0.6 f s to 100 khz in audio mode) C60 db audible out-of-band energy* (measured from 0.6 f s to 22 khz in audio mode, tested at f s = 8.0 khz) C72 db dc offset 525mv dac1 and dac2 analog amplifiers/attenuators min typ max units step size (+12.0 db to C30.0 db) (all steps tested) 1.25 1.5 1.75 db step size (C31.5 db to C34.5 db) (all steps tested) 1.1 1.5 1.9 db step size (C36.0 db to C82.5 db)* 1.3 1.5 1.7 db output attenuation span* 81.5 82.5 83.5 db mute attenuation* C80 db digital mix attenuators min typ max units step size (0 db to C94.5 db)* (all steps tested) 1.3 1.5 1.7 db output attenuation span* 93.5 94.5 95.5 db mute attenuation* C90 db
ad1843 rev. 0 C5C figure 1. timing diagrams analog output min typ max units lout1 full-scale output voltage 0.707 v rms (rms values assume sine wave input) 1.8 2.0 2.2 v p-p lout2 full-scale single-ended output voltage 0.707 v rms (rms values assume sine wave input) 1.8 2.0 2.2 v p-p lout2 full-scale differential output voltage 1.414 v rms (rms values assume sine wave input) 3.6 4.0 4.4 v p-p lout1 output impedance* 600 w lout2 output impedance* 1 w lout1 external load impedance* 10 k w lout2 external load impedance* 2 k w mout external load impedance* 10 k w hpout external load impedance* 16 32 w hpout thd+n (referenced to full scale, 32 w external load impedance) 0.10 % C60 db output capacitance* 15 pf external load capacitance* 100 pf cmout 2.10 2.25 2.40 v external cmout load current* 10 m a cmout output impedance* 4 k w mute click* (muted output minus unmuted midscale dac1 and dac2 outputs) 5mv system specifications max units system frequency response ripple* (line-in to line-out) 1.0 db differential nonlinearity* 1 bit phase linearity deviation* 5 degrees static digital specifications min max units high-level input voltage (v ih ) digital inputs, except sclk 2.0 v dd + 0.3 v xtali and sclk 2.4 v dd + 0.3 v low-level input voltage (v il ) C0.3 0.8 v high-level output voltage (v oh ) 2.4 v low-level output voltage (v ol ) 0.4 v input leakage current (go/nogo tested) C10 10 m a output leakage current (go/nogo tested) C10 10 m a timing parameters (guaranteed over operating temperature and digital supply range) min typ max units serial data frame sync [sdfs] period (t 1 ) (master mode, frs = 1 [16 slots per frame], scf = 0 [sclk = 12.288 mhz]) 20.833 m s frame sync [sdfs] hi pulse width (t 2 )80ns clock [sclk] to frame sync [sdfs] propagation delay (t pd1 )15ns data [sdi] input setup time to sclk (t s )10ns data [sdi] input hold time from sclk (t h )10 ns clock [sclk] to output data [sdo] valid (t dv )15ns clock [sclk] to output data [sdo] three-state [high-z] (t hz )15ns clock [sclk] to time slot output [tso] propagation delay (t pd2 )15ns reset and pwrdwn lo pulse width (t rpwl ) 100 ns t 2 bit 0 bit 14 bit 15 bit 15 bit 14 bit 0 t pd1 sclk sdfs sdi sdo t s t h t dv t hz reset pwrdwn t rpwl 15 14 13 3 2 1 0 15 14 13 t pd2 t pd1 sclk sdfs sdi or sdo tso last valid time slot t 1 15 1413
rev. 0 C6C ad1843 power supply (33 w hpout load) min typ max units power supply rangeanalog v cc 4.75 5.25 v power supply rangedigital v dd 2.85 5.25 v total power supply current5.0 v cc and v dd operating (5.0 v cc and v dd supplies) 210 250 ma total power supply current5.0 v cc /3.0 v dd operating* (5.0 v cc analog/3.0 v dd digital supplies) 150 175 ma analog supply current5.0 v cc operating 60 75 ma digital supply current5.0 v dd operating 150 175 ma digital supply current3.0 v dd operating* 90 100 ma digital power supply currentv dd power down ( pwrdwn lo) 1 ma analog power supply currentv cc power down ( pwrdwn lo) 0.5 ma power dissipation5.0 v cc and v dd operating (current nominal supply) 1250 mw power dissipation5.0 v cc /3.0 v dd operating* (current nominal supply) 875 mw power dissipation5.0 v cc and v dd power down ( pwrdwn lo) (current nominal supply) 7.5 mw power dissipation5.0 v cc /3.0 v dd power down* ( pwrdwn lo) (current nominal supply) 5mw power supply rejection (100 mv p-p signal @ 1 khz)* 40 db (at both analog and digital supply pins, for adc, dac1 and dac2) clock specifications* min typ max units input crystal/clock frequency 24.576 mhz input clock duty cycle (when an external clock is used instead of a crystal) 25/75 75/25 % initialization sample rate change time 0 ms package characteristics typ units pqfp q ja (thermal resistance [junction-to-ambient]) 96 c/w pqfp q jc (thermal resistance [junction-to-case]) 8.75 c/w tqfp q ja (thermal resistance [junction-to-ambient]) 30.6 c/w tqfp q jc (thermal resistance [junction-to-case]) 4.6 c/w notes 1 the stopband repeats itself at multiples of 64 f s , where f s is the sampling frequency. thus the audio mode digital filter will attenuate to C91.8 db or better across the frequency spectrum except for a range of 0.6 f s wide at multiples of 64 f s . 2 the stopband repeats itself at multiples of 64 f s , where f s is the sampling frequency. thus the modem mode digital filter will attenuate to C75.7 db or better across the frequency spectrum except for a range of 0.542 f s wide at multiples of 64 f s . 3 the stopband repeats itself at multiples of 64 f s , where f s is the sampling frequency. thus the resampler mode digital filter will attenuate to C92.2 db or better across the frequency spectrum except for a range of 0.5 f s wide at multiples of 64 f s . *guaranteed, not tested. specifications subject to change without notice. absolute maximum ratings* min max units power supplies digital (v dd ) C0.3 6.0 v analog (v cc ) C0.3 6.0 v input current (except supply pins) 10.0 ma analog input voltage (signal pins) C0.3 v cc + 0.3 v digital input voltage (signal pins) C0.3 v dd + 0.3 v ambient temperature (operating) 0 +70 c storage temperature C65 +150 c esd tolerance (human body 1000 v model per method 3015.2 of mil-std-883b) warning! esd sensitive device caution esd (electrostatic discharge) sensitive device. electrostatic charges as high as 4000 v readily accumulate on the human body and test equipment and can discharge without detection. although the ad1843 features proprietary esd protection circuitry, permanent damage may occur on devices subjected to high energy electrostatic discharges. therefore, proper esd precautions are recommended to avoid performance degradation or loss of functionality. *stresses greater than those listed under absolute maximum ratings may cause permanent damage to the device. this is a stress rating only and functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. exposure to absolute maximum rating conditions for extended periods may affect device reliability. ordering information temperature package package model range description option AD1843JS 0 c to +70 c 80-lead pqfp s-80 AD1843JSt 0 c to +70 c 100-lead tqfp st-100
ad1843 rev. 0 C7C pin configurations 80-lead pqfp 79 78 77 76 75 74 73 72 71 70 69 68 67 66 65 64 63 62 80 61 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 ad1843 top view (not to scale) sdi sclk gndd v dd clkout conv3 v dd v dd v dd v dd bit3 gndd conv2 conv1 bit2 gndd gndd bit1 xtalo xtali gndd xctl1 xctl0 sync3 sync2 sync1 gndd v dd reset pwrdwn v dd pdmnft gnda hpoutl hpoutc hpoutr v cc suml sumr v cc v dd v cc sdo sdfs gndd tsi tso gndd v dd cs bm aux3r aux3l aux2r aux2l aux1r aux1l micr micl min gnda aafiltr filtr aafiltl filtl linrp linrn linlp linln lout2rp lout2rn lout2lp lout2ln mout lout1l gnda cmout v ref gnda lout1r pin description serial interface pin name pqfp tqfp i/o description sclk 79 99 i/o serial clock. sclk is a bidirectional signal that supplies the clock as an output to the serial bus when the bus master (bm) pin is driven hi and accepts the clock as an input when the bm pin is driven lo. when the ad1843 is configured in master mode, the sclk frequency may be set to either 12.288 mhz or 1 6.384 mhz with the scf bit in control register address 26. sdfs 2 2 i/o serial data frame sync. sdfs is a bidirectional signal that supplies the frame synchronization signal as an output to the serial bus when the bus master (bm) pin is driven hi and accepts the frame synchronization signal as an input when the bm pin is driven lo. sdi 80 100 i serial data input. sdi is used by peripheral devices such as the host cpu or a dsp to supply control and playback data information to the ad1843. all control and playback transfers are 16 bits long, msb first. sdo 1 1 o serial data output. sdo is used to supply status/control register readback and capture data information to peripheral devices such as the host cpu or a dsp. all status/control register readback and capture data transfers are 16 bits long, msb first. a three-state output driver is used on this pin. bm 10 12 i bus master. when bm is tied hi the ad1843 is the serial bus master. the ad1843 will then supply the serial clock (sclk) and the frame sync (sdfs) signals for the serial bus. no more than one device (ad1843/cpu/dsp) should be configured as the serial bus master. when bm is tied lo, the ad1843 is con- figured as a bus slave, and will accept the sclk and sdfs signals as inputs. the logic level on this pin must not be changed once reset is deasserted (driven hi).
rev. 0 C8C ad1843 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 99 98 97 96 95 94 93 92 91 90 89 88 87 86 85 84 83 82 81 80 79 78 77 100 76 ad1843 top view (not to scale) sdi sclk nc gndd v dd clkout conv3 nc nc nc nc v dd v dd v dd v dd bit3 gndd conv2 conv1 bit2 gndd gndd bit1 xtalo xtali gndd xctl1 nc xctl0 sync3 sync2 sync1 nc gndd v dd reset pwrdwn nc v dd pdmnft nc gnda hpoutl hpoutc hpoutr v cc suml sumr nc v cc nc nc nc nc nc gnda aafiltr filtr aafiltl filtl linrp linrn linlp linln lout2rp lout2rn lout2lp lout2ln lout1r mout lout1l gnda cmout v ref gnda nc v dd v cc nc nc nc nc sdo sdfs gndd tsi tso gndd v dd cs bm aux3r aux3l aux2r aux2l aux1r aux1l micl micr min nc = no connect serial interface (continued) pin name pqfp tqfp i/o description cs 9 11 i chip select. when cs is set hi, the serial interface i/o pins will be in their normal active states. when cs is reset lo, sclk, sdfs, and sdo are three- stated; sclk, sdfs and sdi inputs are ignored; and tso drives out the logic level received on tsi. tso 6 7 o time slot output. tso is asserted hi by the ad1843 simultaneously with the lsb of the last time slot used by the ad1843. it is used to daisy-chain multiple ad1843s on a common tdm serial bus. if the power-down ( pwrdwn ) pin is asserted or if the chip select pin (cs) is deasserted, tso is set to the logic level on the tsi pin, allowing powered-down or unselected ad1843s on a daisy-chain to be skipped. tsi 5 6 i time slot input. asserting tsi hi indicates to the ad1843 that it should use the next six time slots beginning on the next sclk period. it also enables tso to be asserted at the end of these six time slots. tsi is ignored (but should be tied lo) when the ad1843 is the bus master since the bus master uses the first time slots in a tdm frame. xctl[1:0] 59, 58 72, 74 i/o external control. these signals reflect the status of bits (data 8 and 9) in control register address 28 of the ad1843. they may be used for signaling or controlling external logic. pin configurations 100-lead tqfp
ad1843 rev. 0 C9C analog signals pin name pqfp tqfp i/o description linlp 28 35 i line input left channel positive differential signal. linln 29 36 i line input left channel negative differential signal. linrp 26 33 i line input right channel positive differential signal. linrn 27 34 i line input right channel negative differential signal. micl 18 21 i microphone input left channel. microphone input for the left channel. this signal can be either line level or C20 db from line level. micr 17 22 i microphone input right channel. microphone input for the right channel. this signal can be either line level or C20 db from line level. aux1l 16 20 i auxiliary #1 left channel line input. aux1r 15 19 i auxiliary #1 right channel line input. aux2l 14 18 i auxiliary #2 left channel line input. aux2r 13 17 i auxiliary #2 right channel line input. aux3l 12 16 i auxiliary #3 left channel line input. aux3r 11 15 i auxiliary #3 right channel line input. min 19 23 i monaural (mono) line input. mout 35 44 o monaural (mono) line output. lout1l 36 45 o line output #1 left channel. lout1r 34 43 o line output #1 right channel. hpoutl 47 58 o headphone output left channel. hpoutc 46 57 headphone common return. hpoutr 45 56 o headphone output right channel. lout2lp 32 40 o line output #2 left channel positive differential signal. lout2ln 33 41 o line output #2 left channel negative differential signal. lout2rp 30 38 o line output #2 right channel positive differential signal. lout2rn 31 39 o line output #2 right channel negative differential signal. suml 43 54 i mixer line input left channel. sumr 42 53 i mixer line input right channel. clocks pin name pqfp tqfp i/o description clkout 76 95 o clock output. this signal is a buffered version of xtalo (with a duty cycle restored to at least 60%/40%), the crystal clock output. this pin is enabled by default but can be three-stated by programming a bit in control register address 28. the clkout frequency is 24.576 mhz. sync[3:1] 57, 56, 55 71, 70, 69 i sync inputs. these sync signals are used as the clock source inputs to three receptive plls in the ad1843. these pins accept a clock at, or at a multiple of, the desired sample rate for a-to-d and d-to-a conversions. these inputs are ignored if a sample rate is programmed directly, but should never be left floating. conv[3:1] 75, 71, 67 94, 89, 84 o conversion clock o utputs. th ese output clocks have an average period e qual to (or 128 times) the internal sample rates of the ad1843. these clock outputs are three-stated by default but can be enabled by programming bits in control register address 28. bit[3:1] 74, 70, 66 92, 87, 82 o bit clock outputs. these output clocks can be individually programmed to multiples of the sample rates. support for v.34 or v.32 bit rates is available. these clock outputs are three-stated by default but can be enabled by programming bits in control register address 28.
rev. 0 C10C ad1843 miscellaneous pin name pqfp tqfp i/o description xtali 61 76 i 24.576 mhz crystal input. when using a crystal as the clock source, the crystal should be connected between the xtali and xtalo pins. this crystal should be 24.576 mhz for the normal sampling rate range, i.e., 4 khz to 54 khz. a clock input (perhaps the clkout of another ad1843) may be driven into xtali in place of a crystal. the external clock input must be greater than or equal to 512 times the maximum desired ad1843 sampling frequency. xtalo 62 77 o 24.576 mhz crystal output. when using a crystal as the clock source, the crystal should be connected between the xtali and xtalo pins. if a clock is driven directly into xtali, then xtalo should be left unconnected. pwrdwn 51 64 i power down. pwrdwn is active lo. the assertion of this signal will initialize the on-chip control registers to their default values, and will completely and quietly power down the ad1843. if a crystal is not connected between xtali and xtalo, there must be a 24.576 mhz clock input on xtali for at least 5 ms after this signal is asserted lo for proper operation. the ad1843 will not be completely powered down until after this 5 ms period elapses. the ad1843 always finishes an in-progress power-up sequence before initiating a power-down seque nce, and vice v ersa. if the pwrd wn pin is asserted while a power-up s equence is in progress, the 24.576 mhz clock signal on xtali must persist for a worst case maximum of 479 ms (power up = 470 ms, autocalibration = 4 ms, power down = 5 ms) after pwrdwn is asserted. when init (control register address 0, bit 15) is set to a 1, the power-down sequence is complete. see the power management section for important additional details. reset 52 65 i reset. reset is active lo. the assertion of this signal will initialize the on-chip registers to their default values, and will completely power down the ad1843. reset is similar to pwrdwn , except that when pwrdwn is asserted, power down is quiet and p erformed synchronously to the internal cl ocks. when reset is asserted, power down is noisy and performed asynchronously to the internal clocks. pdmnft 49 61 i power-down mono feedthrough. when the ad1843 mixer is powered down, and pdmnft is asserted hi, the mono input (min, pqfp pin 19) is routed to the mono output (mout, pqfp pin 35), and the signal applied to min will feedthrough to mout. when the ad1843 mixer is powered down and pdmnft is deasserted lo, the feedthrough of min to mout will be muted. when the ad1843 mixer is not powered down, and min to mout feedthrough is desired, the mono input mix mute (control register address 8, bit 15) and the mono output mute (control register address 8, bit 6) must be unmuted. during power-down feedthrough, the signal applied to the min input appears only at the mout output. during normal operation, the signal applied to the min input appears at both the mout and the lout1 outputs. the state of the pdmnft pin should be changed when the ad1843 mixer is powered up. if the state of pdmnft is changed when the ad1843 is in total power-down, audible pops and clicks will likely result. cmout 38 47 o common-mode voltage output. nominal 2.25 volt reference available externally for dc-coupling and level-shifting. cmout should not be used where it will sink or source current. v ref 39 48 i voltage reference filter. voltage reference filter point for external bypassing only. filtl 25 31 i left channel filter. this pin requires a 1.0 m f capacitor to analog ground for proper operation. filtr 23 29 i right channel filter. this pin requires a 1.0 m f capacitor to analog ground for proper operation. aafiltl 24 30 i left channel antialias filter. this pin requires a 1000 pf capacitor to analog ground for proper operation. aafiltr 22 28 i right channel antialias filter. this pin requires a 1000 pf capacitor to analog ground for proper operation.
ad1843 rev. 0 C11C power supplies pin name pqfp tqfp i/o description v cc 20, 41, 44 25, 51, 55 i analog supply voltage (+5 v). gnda 21, 37, 40, 48 27, 46, 49, 59 o analog ground. v dd 4, 8, 50, 53, 5, 10, 62, 66, i digital supply voltage (+5/3 v). 63, 64, 68, 72, 79, 80, 85, 90, 77 96 gndd 3, 7, 54, 60, 4, 9, 67, 75, i digital ground. 65, 69, 73, 78 81, 86, 91, 97 nc 3, 8, 13, 14, no connect. may be left floating. 24, 26, 32, 37, 42, 50, 52, 60, 63, 68, 73, 78, 83, 88, 93, 98 (continued from page 1) the versatility of the device is shown by the following examples of functions it can perform: ? stereo audio input and/or quad output, simultaneously at dif- ferent sample rates ? stereo audio output with simultaneous full duplex modem or fax operation with frequency and phase resampling ? mono audio input and stereo audio output with simultaneous modem receive and transmit for simultaneous voice and data communications ? dual independent audio inputs with audio output for echo- cancelling speakerphones audio functional description the ad1843 soundcomm codec provides a complete audio so- lution with very few external components required. dynamic range of the device exceeds 80 db over the 20 khz audio band and sample rates from 4 khz to 49 khz are supported (up to 54 khz for a single channel if other channels are powered down). the audio functionality of this device is a superset of that found in the analog devices ad1848 soundport ? device which has set the business audio standard throughout the com- puter industry. inputs to the device include a stereo microphone pair, a stereo line pair, a stereo cd input pair (aux1), a stereo synthesized music input pair (aux2), a dual phone line input (aux3), a mono input, and a stereo input from an fm synthesizer (sum). all of these inputs (except sum) are multiplexed to the two ?d a/d converters and are mixable directly as analog signals with the outputs of the d/a converters. all analog input signals (ex- cept sum) can be amplified, attenuated or muted before mix- ing with the outputs of the d/a converters. the device has two pairs of ?d dacs which accept 8- or 16-bit digital data from the serial port. each dac pairs independent sampling rate can either be programmed by control register (with 1 hz resolution) or synchronized to an external input. the second pair of dacs can be used to replace the music syn- thesis dac pair found on many audio products for pcs. out- puts from the ad1843 include a line output, a mono output, a stereo headphone output with its own current return path, and a soundport is a registered trademark of analog devices, inc. differential stereo output for connection to a daa. the line and differential outputs are looped back to the adc input selector. the ad1843s mixing and routing capabilities are extensive. the digital data from both dac channels after interpolation can be routed back to the adc decimators, to support digital- to-digital sample rate conversion (digital resampling). digital data from the adc can also be routed to the two stereo dac pairs, for a digital loopback mode which is helpful for device- level and board-level test. digital data from either stereo dac can be mixed with the digital data feeding the other dac, and the analog signal from dac2 can be mixed with the analog out- put from dac1. sample rates are independently programmable in the range of 4 khz to 54 khz to a 1 hz resolution or sample rates can be synchronized to an external source. up to three different signals can be applied to the devices three digital phase lock loop sync inputs for external synchronization. these sync inputs can also be used in a special mode for au- dio/video synchronization. in this mode, an ntsc or pal de- rived clock signal (approximately 15 khz) is applied to the sync inputs and the device produces one of a variety of stan- dard audio sample rates (32 khz, 44.056 khz, 44.1 khz and 48 khz, and most of these divided by the integers 1 through 8). in this manner, video and audio sample rates which are math- ematically unrelated can be locked together. data communications/telephony functional description the ad1843 includes all data conversion, filtering, and clock generation circuitry needed to implement an echo-cancelling modem with a companion digital signal processor. software- programmable sample rates and clocking modes support all established modem standards including those for the v.34 standard. the ad1843 utilizes advanced ?d technology to move the entire echo-cancelling modem implementation into the digital domain. the device maintains 90 db typical dynamic range throughout all filtering and data conversion across a 9.6 khz passband. purely dsp-based echo cancellation algorithms can maintain robust bit error rates under worst-case signal attenua- tion and echo amplitude conditions. the ad1843s on-chip interpolation filter resamples (both frequency and phase) the re- ceived signal after echo cancellation in the dsp, freeing the pro- cessor for other voice or data communications tasks.
rev. 0 C12C ad1843 figure 2. detailed functional block diagram 8 9 gndd v dd filtl cmout v ref filtr ? s e l e c t o r clock generation hpoutr hpoutc hpoutl lout2lp lout2ln lout2rp lout2rn clkout 2 left right left right gn/at mute s e l e c t o r 20 db left right 43 gnda v cc adc dac1 dac2 d i g i t a l i n t e r f a c e lout1l lout1r aux3l aux2r aux1l aux2l aux1r linrp linrn micr micl gn/at = gain/ attenuation driver control registers ad1843 aafiltl aafiltr suml sumr linlp linln mout reset pwrdwn ? gn/at mute ? gn/at mute ? ? gn/at mute ? ? gn/at mute ? gn/at mute ? ? mute mute gn/at mute mute ?d dac ? ? mute mute attn ? ? mute mute ? ? mute attn mute mute gn/at attn ? ? mute attn fifo ?a law aux3r min fifo ?a law s e l e c t o r ?d adc pga ?a law xtali xtalo sync3 sync2 sync1 conv3 conv2 conv1 bit3 bit2 bit1 sclk sdfs sdi sdo bm cs tso tsi xctl [1:0] pdmnft voltage reference ?d dac
ad1843 rev. 0 C13C on-chip bit and baud clock generation circuitry allows either synchronous or asynchronous operation of the transmit (dac) and receive (adc) paths. each path features independent phase advance and retard adjustments via software control. the ad1843 can also synchronize modem operation to an external terminal band clock. because the device has multiple input and output channels and converters, it is well suited for telephony applications requiring multiple channels for voice and modem. a detailed block diagram of the ad1843 is shown in figure 2. detailed product description the serial-port ad1843 soundcomm codec integrates the key audio and pstn data conversion and control functions into a single integrated circuit. the ad1843 is intended to provide a complete, single-chip audio and fax/modem solution for pc multimedia applications. external circuit requirements are limited to a minimal number of low cost support components. dynamic range exceeds 80 db over the 20 khz audio band. sample rates from 4 khz to 54 khz with 1 hz resolution are supported from a single exter- nal crystal or clock source. the ad1843 soundcomm codec is intended to be interfaced through a dsp chip or an asic to a host bus such as isa, eisa or pci. a general system architecture is shown in figure 3. s y s t e m b u s asic adsp-21xx analog i/o ad1843 figure 3. ad1843 system diagram the soundcomm codec includes a stereo pair of ?d analog-to- digital converters and two stereo pairs of ?d digital-to-analog converters. inputs to the adc can be selected from eight sources of analog signals: stereo line (lin), stereo microphone (mic), stereo auxiliary line #1 (aux1), stereo auxiliary line #2 (aux2), stereo auxiliary line #3 (aux3), mono line (min), mixer output, and dac2 output. a mono output and a stereo headphone driver are included on-chip. a stereo line level input (sum) can be mixed into the output summer. a software-con- trolled programmable gain stage allows independent gain for each adc channel. the adcs output can be digitally mixed with both the dac1 and dac2 inputs. the left and right adc channels can be configured for different sample rates and digital filter function (audio, modem or resampling). the pair of 16-bit outputs from the adcs is available over a se- rial interface that also supports 16-bit digital input to the dacs and control/status information. the ad1843 can accept and generate 16-bit twos-complement pcm linear digital data, 8-bit unsigned magnitude pcm linear data, and 8-bit m -law or a-law companded digital data. the data format is defined indepen- dently for each conversion resource on the ad1843. the ?d dacs are preceded by a four sample deep fifo buffer and a digital interpolation filter. the dac1 and dac2 outputs can be mixed in the digital domain. digital and analog attenua- tors provide independent user volume control (plus mute) over each dac channel. nyquist images and shaped quantized noise are removed from the dacs analog stereo output by on- chip switched-capacitor and continuous-time filters. all of the analog inputs (except the stereo line input) can be mixed with the dac1 output in the analog domain. the dac2 output can also be mixed with the dac1 output in the analog domain. the dac1 and dac2 digital data can be fed back to the digital half of the adc to enable digital resampling operation. dac1 and dac2 can be run at different sample rates and with differ- ent digital filter functions, without any beat frequency problems. functional description this section overviews the functionality of the ad1843 and is intended as a general introduction to the capabilities of the de- vice. as much as possible, detailed reference information has been placed in control registers and other sections. the user is not expected to refer repeatedly to this section. analog inputs the ad1843 soundcomm codec accepts stereo line-level and mic-level inputs. the mono min analog signal input, and lin (differential), mic, aux1, aux2, aux3 and post-mixed dac output analog stereo signals are multiplexed to the inter- nal programmable gain amplifier stage (pga). the pga following the input multiplexer allows left and right independent selectable gains for each channel from 0 db to 22.5 db in +1.5 db steps. the codec can operate either in a global stereo mode or in a global mono mode with left-channel inputs appearing at both channel outputs. analog mixing the min analog mono signal, and the mic, aux1, aux2, aux3 and sum analog stereo signals can be mixed in the ana- log domain with the dac1 output. each channel of each auxil- iary analog input can be independently gained/attenuated from +12 db to C34.5 db in 1.5 db steps or completely muted. the mixer output is available on lout1 externally and as an input to the adcs. even if the ad1843 is not playing back data from its dacs, the analog mix function can still be active. min allows the analog signal intended for the pc speaker to be passed through, attenuated or mixed in the ad1843s analog domain. min can be used to accept other mono input sources. a digital control signal pin pdmnft (power down mono feed through) enables the mono input signal to be fed through to the mono output when the ad1843 mixer is powered down. analog-to-digital datapath the ad1843 ?d adcs incorporate a fourth-order modulator. a single pole of passive filtering is all that is required for antialiasing the analog input because of the adcs high over- sampling ratio. the adcs include linear-phase digital decima- tion filters that low-pass filter the input. adc input overrange conditions will cause bits to be set that can be read. each channel of the mic inputs can be amplified in the analog domain by +20 db to compensate for the voltage swing differ- ence between line l evels and typical condenser microphone levels. digital-to-analog datapath the ?d dacs are preceded by a programmable attenuator and a low-pass digital interpolation filter. the anti-imaging interpo- lation filter oversamples and digitally filters the higher frequency images. the attenuator allows independent control of each dac channel from +12.0 db to C82.5 db in 1.5 db steps plus full mute. the dacs ?d noise shapers oversample and con- vert the signal to a single-bit stream. the dac outputs are then
rev. 0 C14C ad1843 representations in all four formats correspond to equivalent full- scale signals. the eight least-significant bit positions of 8-bit data in 16-bit fields are ignored on input and zeroed on output. the 16-bit pcm data format is capable of representing 96 db of dynamic range. eight-bit pcm can represent 48 db of dynamic range. companded m -law and a-law data formats use nonlinear coding with less precision for large-amplitude signals. the loss of precision is compensated for by an increase in dynamic range to 64 db and 72 db, respectively. on input, 8-bit companded data is expanded to an internal lin- ear representation, according to whether m -law or a-law was specified in the codecs internal registers. note that when m - law compressed data is expanded to a linear format, it requires 14 bits. a-law data expanded requires 13 bits. msb lsb compressed input data 8 7 0 15 msb lsb 3/2 2/1 0 15 expansion msb lsb 3/2 2/1 0 15 dac input 000/00 figure 4. m -law or a-law expansion when 8-bit companding is specified, the adcs linear output is compressed to the format specified. msb lsb 0 15 adc output msb lsb 3/2 2/1 0 15 truncation msb lsb 8 7 0 15 compression 00000000 . figure 5. m -law or a-law compression note that all format conversions take place at input or output. power supplies and voltage reference the ad1843 operates from either +5.0 v analog (v cc ) and digital (v dd ) power supplies or +5.0 v analog and +3.0 v digi- tal supplies. independent analog and digital supplies are recom- mended for optimal performance though excellent results can be obtained in single-supply systems. a voltage reference is incl uded on the codec and its +2.25 v buffered output is available on an external pin (cmout). the reference output can be used for biasing op amps used in single supply systems. the internal ref- erence is externally bypassed to analog ground at the v ref pin. clocks and sample rates the ad1843 operates from a single external clock or crystal source. from a single clock, a wide range of sample rates can be generated. when supplied with a single 24.576 mhz clock, the ad1843 can be programmed to generate any sample frequency between 4 khz and 54 khz with 1 hz resolution. for modem sample rate support, the frequency programmed can also be in- creased by 8/7 using a control bit. all sample rate changes can be made on the fly. the ad1843s sync inputs can be used to synchronize the sampling activity of the four on-chip conversion resources to ex- ternal clock signals, such as video hsync or an isdn network clock. the sync inputs are used by three on-chip digital phase filtered in the analog domain by a combination of switched-capaci- tor and continuous-time filters. they remove the very high fre- quency components of the dac bitstream output. no external components are required. phase linearity at the analog output is achieved by internally compensating for the group delay varia- tion of the analog output filters. changes in dac output attenuation may be programmed to take effect immediately, or only on zero crossings of the digital signal, thereby eliminating zipper noise on playback. each channel has its own independent zero-crossing detector and at- tenuator change control circuitry. a timer guarantees that re- quested volume changes will occur even in the a bsence of an input signal that changes sign. the time-out period is 8 milli- seconds at a 48 khz sampling rate and 48 milliseconds at an 8 khz sampling rate. (time-out [ms] ? 384 ? f s [khz]). digital mixing stereo digital output from the adcs can be mixed digitally with the input to the dacs. digital output from the adcs going out of the serial port is unaffected by this digital mix. along the digital mix datapath, the 16-bit linear output from the adcs is attenuated by an amount specified with control register bits. the level of attenuation applied to the left and right channels is independently programmable. (note that internally the ad1843 always works with 16-bit pcm linear data, digital mixing in- cluded; format conversions take place at the input and output.) sixty-four steps of C1.5 db attenuation are supported to C94.5 db. the digital mix datapath can also be completely muted, pre- venting any mixing of the analog input with the analog output. note that the level of the mixed signal is also a function of the input pga settings, since they affect the adcs output. the sample rate of the adcs and the selected dac pair must be the same for the digital mix function to operate properly. the attenuated digital mix data is digitally summed with the dac input data prior to the dacs datapath attenuators. the digital sum of digital mix data and dac input data is clipped at plus or minus full scale and does not wrap around. because both stereo signals are mixed before the output attenuators, mix data is attenuated a second time by the dacs datapath attenuators. in case the ad1843 is playing back data but input digital dac data fails to arrive in time (dac underrun), then a midscale zero will be added to the digital mix data in place of the unavailable dac data. analog outputs the two mixer line-level outputs are available at external pins. each output channel can be independently muted. when muted, the outputs will settle to a dc value near cmout, the midscale reference voltage. the two dac2 stereo outputs are available at external pins differentially. the full-scale level on these pins is established by programming bits in a control reg- ister. in addition, there is stereo headphone output (with a cur- rent return), and a mono output. both the headphone output and the mono output have a single mute control. digital data types the ad1843 supports four data types: 16-bit twos-complement linear pcm, 8-bit unsigned linear pcm, 8-bit companded m -law, and 8-bit companded a-law, as specified by control register bits. the data type is independently assignable for each conversion resource (i.e., adcl, adcr, dac1 and dac2). data in all four formats is always transferred msb first. eight-bit data is al- ways left-justified in 16-bit fields; said in other words, the msbs of all data types are always aligned; in yet other words, full-scale
ad1843 rev. 0 C15C lock loops, which can be arbitrarily assigned to the conversion resources. the lock range of these digital plls is 4 khz to 54 khz, which is the same range supported by the register- controlled clock generators. if a sync input stops after its associated phase lock loop has had a chance to initially lock, the ad1843 will continue to gen- erate a sample clock (as well as bit clock and conv clock) very similar to the initial frequency, but off by at most 1%. the three sync inputs feed three on-chip digital phase lock loops (dplls) which utilize a first-order loop filter with a 20 hz corner frequency. jitter frequencies above 20 hz are attenuated, and jitter frequencies below 20 hz are interpreted as time base drift, and are tracked. the dpll provides 12 db per octave of jitter rejection. the dplls have been designed to tol- erate at least 2% unit interval (ui) of sync clock jitter. the dplls are critically damped at all input frequencies. power management the ad1843 soundcomm codec has extensive power manage- ment capabilities. hardware power down is performed using the pwrdwn pin. software power management is programmed us- ing control register address 27 and 28. several elements of the ad1843 can be powered down on a selective basis. these blocks include: the dac2 to dac1 analog mixer; the entire dac1 con- version channel; the entire dac2 conversion channel; the analog half of the adc, dac1 and dac2; the headphone driver; the en- tire analog mixer; the right adc channel; the left adc channel; all four conversion channels; clock generator 1; clock generator 2; clock generator 3; conversion clock outputs 1 through 3; bit clock outputs 1 through 3; and the nominal 24.576 mhz clock output. refer to the descriptions of control register address 27 and 28 for further information. for proper operation, the ad1843 must be calibrated following power-up. this initial calibration occurs automatically without any user intervention or programming. subsequent to this initial power-up autocalibration, there is no requirement to recalibrate the soundcomm codec following software power-down sequences. the entire ad1843 or selected portions of the device may be powered down, allowed to idle indefinitely, then powered up and used immediately, without the need for repeated auto- calibration. the digital information obtained during the initial power-up calibration is retained and valid unless the reset or pwrdwn pin is asserted, forcing a hardware reset. (if desired, the user can specify that a calibration cycle occur when leaving the software power-down state by setting acen (control reg- ister address 28, bit 14) to 1.) a hardware reset or power- down clears the calibration information, and therefore a fresh autocalibration cycle is performed by the ad1843 following this event. autocalibration takes approximately 4 ms to complete. the following table provides an indication of the power savings associated with powering-down the various resources in the ad1843. note that the power savings is somewhat order- table i. ad1843 power-down savings +5 v digital, +5 v analog supplies total active operation current: 200 ma average, typical average, typical absolute i dd + normalized software power down control register bit(s) i cc current power savings clkout output enclko bit = 0 8 ma 4% all bit clocks and enbt3, enbt2, enbt1 bits = 0 all conversion clocks encv3, encv2, encv1 bits = 0 2 ma 1% clock generator 1 c1en bit = 0 6 ma 3% clock generator 2 c2en bit = 0 6 ma 3% clock generator 3 c3en bit = 0 6 ma 3% all clock generators c1en, c2en, c3en bits = 0 20 ma 10% headphone driver hpen bit = 0 8 ma 4% dac2 to dac1 mix ddmen bit = 0 2 ma 1% analog input to analog output mix aamen bit = 0 8 ma 4% adc left channel adlen bit = 0 8 ma 4% adc right channel adren bit = 0 8 ma 4% adc left and right channels adlen, adren bits = 0 38 ma 17% dac2 (left and right channels) da2en bit = 0 30 ma 15% dac1 (left and right channels) da1en bit = 0 24 ma 12% dac2 and dac1 (left and right chs) da2en, da1en bits = 0 60 ma 30% adc and dac2 and dac1 adlen, adren, da2en, da1en bits = 0 108 ma 54% analog channel anaen bit = 0 54 ma 27% all control register 27 hpen, ddmen, aamen, adlen, adren, da2en, da1en, anaen bits = 0 134 ma 67% converter pdni bit = 1 140 ma 70% all of the above (register 27 and enclko, enbt3, enbt2, enbt1, encv3, clocks and pdni) encv2, encv1, c1en, c2en, c3en, hpen, ddmen, aamen, adlen, adren, da2en, da1en, anaen bits = 0, pdni bit = 1 176 ma 88%
rev. 0 C16C ad1843 dependent; depending upon the sequence in which the hardware resources are powered down, the savings may be more or less than the typical numbers given. mode changing in general, there are very few restrictions with respect to chang- ing the operating mode of the ad1843. because of the advanced continuous time oversampling technology, the waiting period associated with changes to the sample rate of the data converters (mode change enable resynchronization delay) is eliminated. the only waiting periods associated with the ad1843 occur at start -up, and are documented in the start-up se quence section below. following the start-up se quence, the sample rate of the four data conversion resources on the ad1843 may be changed at any time, on-the-fly (presuming that they are enabled). all gain, mute and attenuation settings of enabled resources may also be changed at any time. channel synchronization if multiple ad1843s are used in a daisy-chained system, and it is desired to synchronize data conversion activity among the multiple codecs, the clock generator blocks of the ad1843s must be enabled on the same frame (see step 5 in the start- up sequence section below). a dac channel does not actually start processing samples until the first rising edge of the conversion clock (conv pin) after the sixth rising edge of frame sync (sdfs pin) after the channel is enabled (via a write to da1en or da2en in control regis- ter address 27). the wait until the sixth rising edge of frame sync is necessary to allow the four deep dac fifo to be filled before conversion commences. the subsequent wait until the rising edge of the conversion clock is necessary to synchronize the serial interface based dac channel enable command with a conversion clock that is potentially already running (which is particularly likely if the sync pin inputs and lock mode are in use). the adc channels behave very similarly to the dac channels. an adc channel does not actually start taking samples until the first rising edge of the conversion clock (conv pin) after the sixth rising edge of frame sync (sdfs pin) after the channel is enabled (via a write to adlen or adren in control register address 27). the wait until the sixth rising edge of frame sync is present so that the adc startup is similar to that of the dac startup, as well as to allow some time for stale adc data inside the ad1843 to be cleared. the subsequent wait until the rising edge of the conversion clock is necessary to synchronize the serial interface based adc channel enable command with a conversion clock that it potentially already running (which is par- ticularly likely if the sync pin inputs and lock mode are in use). supported conversion rates with all conversion channels operating (i.e., adc left, adc right, dac1 and dac2), the ad1843 is able to support sam- pling rates up to 49 khz, which 2.1% higher than the nominal maximum audio standard of 48 khz, to accommodate timebase drift while configured in slave mode. if either one dac (i.e., either dac1 or dac2) or both adc channels (i.e., adc left and adc right) are shut down, then the ad1843 can support sampling up to 54 khz on all channels of the remaining conver- sion resources, as long as the dfree bit (control register ad- dress 27) is asserted (i.e., set to 1). if dfree is not asserted, then the maximum sampling rate for the remaining conversion resources is 49 khz. digital filter selection the operative digital filter modes for the four conversion re- sources on the ad1843 soundcomm are programmed using control register address 25. adlflt (bit 0) selects the digi- tal filter mode for the adc left channel and adrflt (bit 1) selects the digital filter mode for the adc right channel. note that these bits also establish the full-scale input voltage range for these channels as well. da1flt (bit 8) selects the dac1 digi- tal filter mode, and da2flt (bit 9) selects the dac2 digital filter mode. note that these bits also establish the full-scale out- put voltage for these channels as well. the three digital filter modes are audio, modem and resampler. the specifications for these modes are given in the description of control register address 25, as well as in the specifica- tions section of this data sheet. the specifications have been made to satisfy the demands of the applications which the ad1843 can serve. the audio mode provides decimation and interpolation characteristics sufficient for high quality cap- ture and playback of material from 20 hz to 20 khz. the mo- dem mode provides characteristics sufficient for modulation standards up to v.34 quality. the resampling mode provides optimal characteristics for high quality sample rate conversion. while in the resampling mode, all images in the resampled data stream (including those in the transition band) are attenuated to below the quantization noise floor. note that the maximum sample rate for modem mode is 24 khz. digital resampling digital resampling is best achieved by routing the digital output of one of the dacs back to the digital input of one of the adcs. this bypasses the analog portion of the dac and adc, eliminating their noise and signal delay contributions. this fea- ture is enabled by bits daadr1:0 (digital adc right channel source select) and daadl1:0 (digital adc left channel source select) in control register address 25. if the digital resampler filter mode (drsflt bit = 1, control register address 25) is enabled, the dac2 pair is sacri- ficed, but the remaining four channels (adc left and right, dac1 left and right) can still be used in any way they could have been when not in digital resampler filter mode. when in this mode, internal ad1843 hardware normally devoted to dac2 is reallocated to the other four channels, allowing these channels to realize superior digital filtering. note that the ad1843 does not actually have to be in digital resampler filter mode to perform digital resampling, however the superior digital filters in this mode allow for a much higher quality digital resampling. using the ad1843 in a modem application the ad1843 analog performance is sufficient to support the modem analog front end (afe) function, for data modulation standards up to and including the 28.8 kbps v.34 itu stan- dard. the data pump function is performed in a companion dsp, such as the adsp-2181, for which several v.34 algo- rithms (from third party independent algorithm vendors) exist.
ad1843 rev. 0 C17C line in mic in right mic in left aux1 in aux2 in aux3 in mono in sum in line1 out headphone out line2 out right line2 out left mono out sync2 conv1 bit1 serial interface daa serial interface adsp-21xx dsp or asic ad1843 soundcomm codec idma port or parallel port ntsc horizontal sync signal host bus isa or pci external powered multimedia speakers speakerphone audio from dat or cassette pc speaker pstn audio from external mpeg decoder pc attention ?eeper?signal audio from cd-rom external waveform synthesizer figure 6. typical configurations 321015141312 321015141312 321015141312 321015141312 512 bits msb msb msb msb 512 bits 256 bits 256 bits slot 15 slot 0 slot 15 slot 0 slot 31 slot 16 slot 16 256 bits 256 bits sample period n sample period n+1 sample period n+2 sample period n+3 frame m frame m+1 sdi or sdo sclk sdfs frs = 0 [default 32 slots per frame, 2 samples per frame sync] master mode note that ad1843 frame rate is not related to sample rates figure 7. frs = 0, master mode timing 321015141312 321015141312 321015141312 321015141312 256 bits msb msb msb msb 256 bits slot 15 slot 0 slot 15 slot 0 slot 15 slot 0 slot 0 sample period n sample period n+1 sample period n+2 sample period n+3 frame m sdi or sdo sclk sdfs frs = 1 [16 slots per frame, 1 sample per frame sync] master mode note that sclk can be programmed for either 12.288 mhz or 16.384 mhz when in master mode 256 bits frame m+1 frame m+2 frame m+3 figure 8. frs = 1, master mode timing
rev. 0 C18C ad1843 15 14 13 12 321015141312 3210 15141312 321015 512 bits msb msb msb msb 512 bits 256 bits 256 bits slot 15 slot 0 slot 15 slot 0 slot 31 slot 16 256 bits sample period n sample period n+1 sample period n+2 frame m frame m+1 sdi or sdo sclk tsi frs = 0 [default 32 slots per frame, 2 samples per frame sync] example showing gaps between frames slave mode gap gap figure 9a. frs = 0, slave mode timing 321015141312 512 bits msb msb msb msb 512 bits 256 bits 256 bits slot 15 slot 0 slot 15 slot 0 slot 31 slot 16 256 bits sample period n sample period n+1 sample period n+2 frame m frame m+1 sdi or sdo sclk tsi frs = 0 [default 32 slots per frame, 2 samples per frame sync] example showing no gaps between frames slave mode 321015141312 321015141312 32101514 figure 9b. frs = 0, slave mode timing frs = 1 [16 slots per frame, 1 samples per frame sync] example showing gaps between frames slave mode gap gap 15 14 13 12 3210 15 14 13 12 3210 256 bits msb msb 256 bits slot 15 slot 0 slot 15 slot 0 sample period n sample period n+1 frame m frame m+1 sdi or sdo sclk tsi gap figure 10a. frs = 1, slave mode timing frs = 1 [16 slots per frame, 1 samples per frame sync] example showing no gaps between frames slave mode 15 14 13 12 3210 15 14 13 12 3210 256 bits msb msb 256 bits slot 15 slot 0 slot 15 slot 0 sample period n sample period n+1 frame m frame m+1 sdi or sdo sclk tsi 3210 15 14 13 12 msb figure 10b. frs = 1, slave mode timing
ad1843 rev. 0 C19C modem data access arrangement (daa) devices are generally differential on the transmit side, and single-ended on the receive side. the daa transmit input (generally differential) should be connected to the dac2 output, pins lout2lp and lout2ln, or lout2rp and lout2rn. the daa receive output (generally single-ended) should be connected to one of the adc line inputs, linlp or linrp. see the application circuits section below for more detail on the electrical connections. there are several software driver steps that are re- quired to configure the soundcomm codec for use as a modem afe. configure dac2 1. set the da2flt bit (control register address 25, bit 9) to 1, to select the digital modem filter mode. the dac2 out- puts can be used either as differential outputs or single-ended outputs depending on how the pins are connected electrically; no control register writes are required to configure the dac2 outputs as either differential or single- ended. 2. program lda2g5:0 (control register address 10, bits 8 through 13) to 00 0101 (i.e., +4.5 db) or rda2g5:0 (control register 10, bits 0 through 5) to 00 0101 (i.e., +4.5 db), depending on whether the daa transmit input is connected to the left channel dac2 output (use lda2g5:0) or the right channel dac2 output (use rda2g5:0). this code establishes the dac2 nominal analog output swing at 3.156 v p-p single-ended, or 6.312 v p-p differentially. the 3.156 v p-p level is equivalent to 3.17 dbm. configure adc 1. set the adlflt bit (control register address 25, bit 0) to 1, or the adrflt bit (control register address 25, bit 1) to 1, to select the digital modem filter mode. set adlflt if the daa receive output is connected to the ad1843 linlp input; set adrflt if the daa receive output is connected to the ad1843 linrp input. set the linlsd bit (control register address 28, bit 0) to 1 if the daa is connected to the ad1843 linlp input; set the linrsd bit (control register address 28, bit 1) to 1 if the daa is connected to the ad1843 linrp input. 2. program lig3:0 (control register address 2, bits 8 through 11) to 0000 (i.e., 0.0 db) or rig3:0 (control register address 2, bits 0 through 3) to 0000 (i.e., 0.0 db) de- pending on whether the left or right adc input channel is being used for the modem function. this code maps an ana- log input swing of 3.156 v p-p to the full dynamic range of the 16-bit digital sample (i.e., 2 15 ). the 3.156 v p-p level is equivalent to 3.17 dbm. note that if the ad1843 is to be reconfigured dynamically, the affected converter must be powered down before its associated digital filter can be changed. in other words, if the digital filter for the adc left channel is being changed from audio mode to modem mode, the adc left channel must be powered down first (using the adlen bit in control register address 27). use the adren bit in control register address 27 for the adc right channel, the dac1en bit in control register address 27 for dac1, and the dac2en bit in control regis- ter address 27 for dac2. typical configurations figure 6 below illustrates example connections between the ad1843 soundcomm codec and other system resources. the rich analog input and output connectivity of the ad1843 allows a wide variety of configuration possibilities. note that the level of modem, speakerphone and external speaker concurrency is application and dsp resource dependent. serial interface the ad1843 soundcomm codec transmits and receives both data and control/status information through its serial port. the ad1843 can be configured as either master or slave of the serial interface. this is selected by using the bm pin. when bm is tied hi, the ad1843 serves as bus master and supplies the frame sync and the serial clock. when bm is tied lo, the ad1843 serves as bus slave and receives the frame sync and the serial clock. the level on bm should not be altered unless the reset pin ( reset ) is asserted. the ad1843 has six pins devoted to the serial interface: sdi, sdo, sclk, sdfs, tsi and tso. the sdi pin is for serial data input to the ad1843 and the sdo pin is for serial data output from the ad1843. the sclk pin is the serial interface clock. communication in and out of the ad1843 requires bits of data to be transmitted after a rising edge of sclk, and sampled on a falling edge of sclk. when the ad1843 is bus master (bm pin tied hi), the sclk frequency driven by the ad1843 will be 12.288 mhz by default, but this can be in- creased to 16.384 mhz by setting the scf bit in control regis- ter 26. when the ad1843 is bus slave (bm pin tied lo), the sclk frequency driven to the ad1843 may be as high as 24.576 mhz, but must not be any higher than the frequency on the xtali pin. the sdfs pin is for the serial interface frame sync. when bus master, new frames are marked by a hi pulse driven out on sdfs one serial clock period before the frame begins. when bus slave, new frames must be marked by a lo to hi transition driven in on sdfs one serial clock period before the frame be- gins, but the transition back from hi to lo may occur at any time provided the hi and lo times of sdfs are at least one sclk period in duration each. when the ad1843 is bus master, frame size is controlled by the frs bit in control register 26. when frs is set to 1, each frame is divided into 16 slots of 16 bits. when frs is reset to 0, each frame is divided into 32 slots of 16 bits. in 32 slot configuration, the second 16 slots of a frame must have slot as- signments that are identical to the first 16 slots of the frame; 32 slot configuration is essentially 16 slot configuration with every other sdfs pulse missing. although these are the frame sizes
rev. 0 C20C ad1843 for control register write data input and read data output. the remaining slots are used for playback (dac) data input and capture (adc) data output, where each channel has an assigned slot. table ii and figure 11 illustrate these slot assignments. since the conversion channels of the ad1843 can be pro- grammed to run at different sample rates, a communication mechanism indicates when playback channels request data, when playback data is actually sent to the ad1843, and when transmission of capture data from the ad1843 becomes neces- sary. this is facilitated by the control and status words lo- cated in the first slot. the control word indicates which slots in the current frame contain valid playback data. the status word indicates if playback data can be sent to the ad1843 dur- ing the next frame, and which slots in the current frame contain valid capture data. see the descriptions of the control word and the status word below for additional detail. four word fifo buffers are used on the inputs of each of the dacs to allow data to be transferred in small bursts. this re- duces the required response time to playback data requests, and also buffers differences between the frame sync rate and the channel sample rate. the status word indicates that playback data can be sent if there is any room in the buffers, thus tending to keep the input buffers full. underrun flags are available in control register 1, which indicate if an input buffer ran out of data. if an underrun occurs, a zero is used in place of the un- available data. to ensure underruns do not occur, playback data must be sent to the ad1843 within two sample periods after the status word indicates that the dac fifo is not full. note that the dac not full status bits (da2rq and da1rq in the status word output) are updated immediately (i.e., in the same frame as a valid write to the dac fifos). if the dac input valid flags (da2v and da1v in the control word input) are set (i.e., dac data is valid) and only one location in the dac1 and dac2 input fifos is available, then the da2rq and da1rq status bits will reflect this valid write, and will be reset to 0. this is possible because the da2v and da1v bits are in the most significant bits of the control word and the da2rq and da1rq bits are in the least significant bits of the status word, and the ad1843 uses this intervening time table ii. ad1843 slot assignment 32 slot mode (frs reset to 0) slot sdi pin sdo pin 0 & 16 control word input status word output 1 & 17 control register data input control register data output 2 & 18 playback data inputdac1 left capture data outputadc left 3 & 19 playback data inputdac1 right capture data outputadc right 4 & 20 playback data inputdac2 left reserved (unused) 5 & 21 playback data inputdac2 right reserved (unused) 16 slot mode (frs set to 1) slot sdi pin sdo pin 0 control word input status word output 1 control register data input control register data output 2 playback data inputdac1 left capture data outputadc left 3 playback data inputdac1 right capture data outputadc right 4 playback data inputdac2 left reserved (unused) 5 playback data inputdac2 right reserved (unused) produced by an ad1843 serving as bus master, an ad1843 serving as bus slave does not actually require these frame sizes. when frs is set to 1, a slave will operate correctly with any number or fraction of slots, provided there are enough slots for it to complete its necessary communication (see below). when frs is reset to 0, a slave can also operate correctly with a wide range in the number of slots per frame, however it will au- tomatically retake ownership of the serial interface bus 16 slots after it is first given ownership of the bus in a frame. the nominal minimum number of slots when the ad1843 is configured in slave mode is six. the codec must be supplied with at least 6 16 = 96 sclk periods (both rising and falling edges); sclk may be gated (i.e., no need to be continuous) between valid slots. while sdfs marks the beginning of frames, ad1843 bus own- ership during a frame is controlled by the tsi (time slot in) and tso (time slot out) pins. when bus slave, a level hi on tsi grants the ad1843 bus ownership beginning with the next sclk period. the tsi pin is monitored only when an ad1843 does not already own the bus; once an ad1843 is given owner- ship of the bus, the level on tsi is ignored until one sclk period before bus ownership is relinquished. bus ownership will last for six slots. coincident with the final sclk period of the final slot owned, the ad1843 asserts tso hi. this allows chaining of ad1843s onto a common serial bus by connecting the tso pin of one ad1843 to the tsi pin on the next later ad1843 in a chain. in single codec systems where the soundcomm is configured as bus slave, connect the ad1843 sdfs and tsi signals together. when an ad1843 is bus mas- ter, its function is identical to that just described for the slave, except a bus master always owns the first six slots and its tsi pin is ignored (but should be tied lo). whenever an ad1843 does not own the bus, its sdo pin will be three-stated and its sdi pin is ignored. figures 7 through 10 illustrate the signal, slot, sample and frame relationships for the four basic operating modes of the ad1843 serial interface. the ad1843 requires slots of communication each time it takes ownership of the serial bus. the first slot is used for a control word input and a status word output. the second slot is used
ad1843 rev. 0 C21C interval to update the dac not full status bits. therefore driver software does not have to make provision for frame-to- frame delays between control and status information; the infor- mation in each frame is always up to date. the ad1843 supports locking to an external clock which may result in a sample rate that is marginally higher than the nominal audio standard maximum sample rate of 48 khz. this is neces- sary since two crystal-based clock sources are never perfectly matched to one another. one is always at a slightly higher fre- quency. the ad1843 conversion channels have been designed to support sample rates that are up to 2.1% higher than 48 khz (i.e., 49 khz), referenced to the ad1843s clock input on xtali, when all conversion channels are simultaneously run- ning. even higher rates can be supported when all channels are not running simultaneously. see the conversion rates sec- tion above for additional details. the serial interface must also allow for this higher sample rate with a frame sync (sdfs) rate that is at least as high as the sample rate in 16 slot per frame mode, or half as high as the sample rate in 32 slot per frame mode. when the ad1843 is bus master, the sclk frequency is either 12.288 mhz or 16.384 mhz, which allows for up to 48 khz or 64 khz sampling rates, respectively. the ad1843 control registers are read and written by trans- mitting a read/write request bit along with the control register address in the slot 0 control word. when a read is requested, the contents of the control register addressed is transmitted out during slot 1 of the following frame. when a write is re- quested, data to be written must be transmitted to the ad1843 in slot 1, and the former contents of the control register are transmitted out during slot 1 of the following frame. unless oth- erwise noted, control register writes do not take effect until the current round of six communication tdm time slots concludes. equivalently, unless otherwise noted, control register writes do not take effect until the subsequent falling edge of the tso signal. the following sections describe the bit assignments for all time slots. ignored 3-stated control word status word register data register data dac1 left adc left dac1 right adc right dac2 left 0s dac2 right 0s ignored 3-stated sdi sdo 16-bit stereo don't care ignored 3-stated control word status word register data register data ignored 3-stated 0s dac1r don't care dac2l dac2r don't care 0s dac1l don't care 0s 0s adcl adcr sdi sdo 8-bit stereo ignored 3-stated control word status word register data register data dac1 adc left don't care adc right dac2 0s don't care 0s ignored 3-stated sdi sdo 16-bit mono dac1, dac2 ignored 3-stated control word status word register data register data don't care don't care 0s dac2 0s don't care 0s don't care 0s dac1 adcl adcr ignored 3-stated sdi sdo 8-bit mono dac1, dac2 the diagram above is intended to be illustrative of the more commonly used configurations. adc left, adc right, dac1 and dac2 can be individually assigned to 8-bit or 16-bit sample width, and dac1 and dac2 can be individually assigned to stereo or mono mode. each ad1843 consumes 6 tmd slots (requiring 96 sclk periods), leaving 10 tdm slots unused in a 16 slot frame. note that because the serial interface and the adc and dacs are in general asynchronous, not every capture or playback time slot will contain valid data. the host processor must poll the status word to determine whether the adc data is valid and whether the dac is requesting additional samples. time slot 0 time slot 1 time slot 2 time slot 3 time slot 4 time slot 5 sclk sdfs 16 bits figure 11. ad1843 slot assignments
rev. 0 C22C ad1843 serial interface input note that the references to slot numbers are valid only when the ad1843 is configured in master mode. for slave mode, bus owner- ship does not necessarily start on slot 0. control word input (slot 0 or 16) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res da2v da1v data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 r/w res res ia4 ia3 ia2 ia1 ia0 da2v dac2 input valid flag. when in dac2 stereo mode, setting this bit to 1 indicates that slots 4 and 5 contain valid playback data for the left and right channels of dac2, respectively. when in dac2 mono mode, setting this bit to 1 indicates that slot 4 contains valid playback data for both left and right channels of dac2. slot 5 is ignored in dac2 mono mode. when this bit is reset to 0, data in slots 4 and 5 is ignored. this bit is ignored if the ad1843 did not request data for dac2 in the last frame (see the da2rq bit in the status word output). da1v dac1 input valid flag. when in dac1 stereo mode, setting this bit to 1 indicates that slots 2 and 3 contain valid playback data for the left and right channels of dac1, respectively. when in dac1 mono mode, setting this bit to 1 indicates that slot 2 contains valid playback data for both left and right channels of dac1. slot 3 is ignored in dac1 mono mode. when this bit is reset to 0, data in slots 2 and 3 is ignored. this bit is ignored if the ad1843 did not request data for dac1 in the last frame (see the da1rq bit in the status word output). r/w read/write request. either a read from or a write to a control register occurs every frame. setting this bit to 1 indicates a control register read while resetting this bit to 0 initiates a control register write. bits ia4:0 define the control register address. when reading, the contents of the control register addressed are transmitted during slot 1 of the following frame. when writing, the data to be written is taken from slot 1 and the former contents of the control register are transmitted during slot 1 of the following frame. ia4:0 control register address for read or write. res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. control register write data input (slot 1 or 17) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 format for the input data to written to the addressed control register. msb is first. dac1 left sample input (slot 2 or 18) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 data format may be 8-bit unsigned linear pcm, 16-bit signed linear pcm, 8-bit m -law companded, or 8-bit a-law companded. msb is first. data7:0 are ignored in 8-bit linear or 8-bit companded modes.
ad1843 rev. 0 C23C dac1 right sample input (slot 3 or 19) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 data format may be 8-bit unsigned linear pcm, 16-bit signed linear pcm, 8-bit m -law companded, or 8-bit a-law companded. msb is first. data7:0 are ignored in 8-bit linear or 8-bit companded modes. dac2 left sample input (slot 4 or 20) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 data format may be 8-bit unsigned linear pcm, 16-bit signed linear pcm, 8-bit m -law companded, or 8-bit a-law companded. msb is first. data7:0 are ignored in 8-bit linear or 8-bit companded modes. dac2 right sample input (slot 5 or 21) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 data format may be 8-bit unsigned linear pcm, 16-bit signed linear pcm, 8-bit m -law companded, or 8-bit a-law companded. msb is first. data7:0 are ignored in 8-bit linear or 8-bit companded modes.
rev. 0 C24C ad1843 serial interface output status word output (slot 0 or 16) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res adrv adlv data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res res res da2rq da1rq adrv right adc channel output valid flag. set to 1 if slot 3 (or 19) contains valid right adc capture data. adlv left adc channel output valid flag. set to 1 if slot 2 (or 18) contains valid left adc capture data. da2rq dac2 input request flag. set to 1 if dac2 is enabled and its four word stereo input buffer is not full. da1rq dac1 input request flag. set to 1 if dac1 is enabled and its four word stereo input buffer is not full. res reserved for future expansion. read back as 0. should be ignored to ensure future compatibility. control register data output (slots 1 or 17) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 contents of control register addressed in previous frame. adc left sample output (slots 2 or 18) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 data format may be 8-bit unsigned linear pcm, 16-bit signed linear pcm, 8-bit m -law companded, or 8-bit a-law companded. msb is first. data7:0 are ignored in 8-bit linear or 8-bit companded modes. adc right sample output (slots 3 or 19) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 data15 data14 data13 data12 data11 data10 data9 data8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 data7 data6 data5 data4 data3 data2 data1 data0 data format may be 8-bit unsigned linear pcm, 16-bit signed linear pcm, 8-bit m -law companded, or 8-bit a-law companded. msb is first. data7:0 are ignored in 8-bit linear or 8-bit companded modes.
ad1843 rev. 0 C25C reserved output (slots 4 or 20) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res res res res res res reserved for future expansion. read back as 0. should be ignored to ensure future compatibility. reserved output (slots 5 or 21) data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res res res res res res reserved for future expansion. read back as 0. should be ignored to ensure future compatibility. control registers control register architecture all control registers are cleared to their default state when either the reset pin or the pwrdwn pin is asserted. all conversion related control registers (addresses 1C15, 25 and 27) are cleared to defaults when the conversion resources of the ad1843 are pow- ered down through the pdni bit in control register address 28. control registers which manage analog dac gain/attenuation (addresses 3C10) are cleared to defaults whenever the resource they manage is powered down. finally, the three clock generator sample phase shift control registers (addresses 18, 21 and 24) are cleared to defaults whenever their associated clock generator is powered down. writes to control registers are blocked until a clearing condition no longer exists. writes to control registers are also blocked immediately after the reset pin or the pwrdwn pin is asserted until the ad1843s internal clocks have settled, which is indicated by the init bit in control register address 0. control registers may always be read. figures 12 and 13 associate the control registers to the ad1843 block diagram.
rev. 0 C26C ad1843 4 cmout v ref ? hpoutr hpoutc hpoutl lout2lp lout2ln lout2rp lout2rn left right 20 db aux3l aux2r aux1l aux2l aux1r linrp linrn micr micl gn/at = gain/ attenuation driver ad1843 aafiltl aafiltr suml sumr linlp linln mout ? ? ? ? ? ? ? ? ? mute mute ?d dac mute aux3r min ?d adc pga voltage reference cra2 gn/at mute cra7 gn/at mute cra4 gn/at mute cra5 gn/at mute cra6 gn/at mute cra8 cra8 cra8 lout1l lout1r gn/at mute cra9 s e l e c t o r gn/at mute cra3 mute cra8 gn/at mute cra10 crax = control register address ? gnda filtr filtl cra2 ?d dac address register description 0 codec status and revision identification (read only) 1 channel status flags 2 input controladc source and gain/attenuation 3 mix controldac2 to mixer 4 mix controlauxiliary 1 to mixer 5 mix controlauxiliary 2 to mixer 6 mix controlauxiliary 3 to mixer 7 mix controlmic to mixer 8 mix/misc. controlmono in to mixer and miscellaneous settings 9 output controldac1 analog gain/attenuation 10 output controldac2 analog gain/attenuation 11 output controldac1 digital attenuation 12 output controldac2 digital attenuation 13 digital mix controladc to dac1 14 digital mix controladc to dac2 15 codec configurationchannel sample rate source select figure 12. ad1843 control registers associated with block diagram
ad1843 rev. 0 C27C 8 9 gndd v dd s e l e c t o r 2 left right left right 3 v cc adc ?a law d i g i t a l i n t e r f a c e control and status registers cra0, cra1, cra25, cra26, cra27, cra28 reset pwrdwn ? ? mute attn ? ? mute ? ? ? ? fifo s e l e c t o r ?a law sclk sdfs sdi sdo bm cs tso tsi xctl [1:0] pdmnft clock generation clkout xtali xtalo sync3 sync2 sync1 conv3 conv2 conv1 bit3 bit2 bit1 cra15, cra16, cra17, cra18, cra19, cra20, cra21, cra22, cra23, cra24 ad1843 crax = control register address ? cra25 cra25 cra25 cra25 cra12 cra11 attn mute cra13 attn mute cra14 cra26 cra26 dac1 dac2 cra26 ?a law fifo attn address register description 16 clock generator 1 controlmode 17 clock generator 1 controlsample rate 18 clock generator 1 controlsample phase shift 19 clock generator 2 controlmode 20 clock generator 2 controlsample rate 21 clock generator 2 controlsample phase shift 22 clock generator 3 controlmode 23 clock generator 3 controlsample rate 24 clock generator 3 controlsample phase shift 25 codec configurationdigital filter and mode select 26 codec configurationserial interface 27 codec configurationchannel power down 28 codec configurationfundamental settings 29 reserved for future expansion 30 reserved for future expansion 31 reserved for future expansion figure 13. ad1843 control registers associated with block diagram
rev. 0 C28C ad1843 address 0 codec status and revision identification data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 init pdno res res res res res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res id3 id2 id1 id0 this register is read only. init clock initialization flag. this bit is set to 1 if the ad1843s internal clocks generated from the crystal input pin xtali have not yet stabilized. when set to 1, control registers may be read if the ad1843 is configured as a bus slave (see the definition for the bm pin), but control register writes will be blocked. this bit is set to a 1 after the reset pin is asserted, or when the power-down sequence (initiated by asserting the pwrdwn pin) has completed. the bit is reset to 0 usually within 400 to 800 m sec after the reset or the pwrdwn pin is deasserted, depending upon parasitic capacitance on the xtali and xtalo pins outside the ad1843. because these parasitics are a func- tion of the board design and layout, the exact amount of time required for the crystal to start to oscillate, and the in- ternal clocks to stabilize cannot be known exactly in advance. pdno converter power-down flag. this bit is set to 1 if the ad1843 conversion resources are powered down, or are in the process of entering or exiting power down. conversion resources are all resources in the ad1843 with the excep- tion of the three clock generators and the serial interface. conversion power-down is entered if either the power-down pin ( pwrdwn ) is driven lo, or if the pdni (converter power down) bit in control register address 28 is as- serted. power down is exited only if both pin ( pwrdwn ) and bit (pdni) are deasserted. when this bit is set to 1, control registers may still be read, but only control registers 16C24, 26 and 28 which do not manage conver- sion resources may be written. see the power management section for further details. this bit is set to 1 imme- diately after a reset since the pdni bit is initially asserted after a reset. id3:0 revision identification. these bits define the revision level of the ad1843. the first version of the ad1843 is 0001. res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1100 0000 0000 0001 (c001 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. address 1 channel status flags data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res su2 su1 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res ovr1 ovr0 ovl1 ovl0 this register is cleared to all 0 after any read or write access. su2 dac2 sample underrun. when set to 1, this bit indicates that the four word stereo input buffer for dac pair 2 ran out of samples. if this occurs, zeros are used in place of the unavailable data. this bit is sticky and is cleared after any write to this register. it is also cleared by powering down dac2 (see the da2en bit in control register address 27). su1 dac1 sample underrun. when set to 1, this bit indicates that the four word stereo input buffer for dac pair 1 ran out of samples. if this occurs, zeros are used in place of the unavailable data. this bit is sticky and is cleared after any write to this register. it is also cleared by powering down dac1 (see the da1en bit in control register address 27). ovr1:0 adc right overrange detect. these bits record the largest output magnitude on the adc right channel and are cleared to 00 after any write to this register. the peak amplitude as recorded by these bits is sticky, i.e., the larg- est output magnitude recorded by these bits will persist until these bits are explicitly cleared. they are also cleared by powering down the adc right channel (see the adren bit in control register address 27). 00 = greater than C1.0 db underrange 01 = between C1.0 db and 0 db underrange 10 = between 0 db and 1 db overrange 11 = greater than 1.0 db overrange
ad1843 rev. 0 C29C ovl1:0 adc left overrange detect. these bits record the largest output magnitude on the adc left channel and are cleared to 00 after any write to this register. the peak amplitude as recorded by these bits is sticky, i.e., the larg- est output magnitude recorded by these bits will persist until these bits are explicitly cleared. they are also cleared by powering down the adc left channel (see the adlen bit in control register address 27). 00 = greater than C1.0 db underrange 01 = between C1.0 db and 0 db underrange 10 = between 0 db and 1 db overrange 11 = greater than 1.0 db overrange res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled). address 2 input control?dc source and gain/attenuation data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lss2 lss1 lss0 lmge lig3 lig2 lig1 lig0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rss2 rss1 rss0 rmge rig3 rig2 rig1 rig0 lss2:0 left adc source select 000 = left line input 001 = left mic input 010 = left auxiliary 1 input 011 = left auxiliary 2 input 100 = left auxiliary 3 input 101 = mono input 110 = left dac1 output 111 = left dac2 output lmge left adc microphone gain enable 0 = 0 db gain 1 = +20 db gain lig3:0 left adc input gain. least significant bit represents +1.5 db. 0000 = 0.0 db gain 1111 = +22.5 db gain rss2:0 right adc source select 000 = right line input 001 = right mic input 010 = right auxiliary 1 input 011 = right auxiliary 2 input 100 = right auxiliary 3 input 101 = mono input 110 = right dac1 output 111 = right dac2 output rmge right adc microphone gain enable 0 = 0 db gain 1 = +20 db gain rig3:0 right adc input gain. least significant bit represents +1.5 db. 0000 = 0.0 db gain 1111 = +22.5 db gain initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register 0 is set to 1 (all conversions disabled).
rev. 0 C30C ad1843 address 3 mix control?ac2 to mixer data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 ld2mm res res ld2m4 ld2m3 ld2m2 ld2m1 ld2m0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rd2mm res res rd2m4 rd2m3 rd2m2 rd2m1 rd2m0 ld2mm left dac2 mix mute 0 = mix enabled 1 = mix muted ld2m4:0 left dac2 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac2 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation rd2mm right dac2 mix mute 0 = mix enabled 1 = mix muted rd2m4:0 right dac2 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac2 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 1000 1000 (8888 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; when the pdno bit in control register address 0 is set to 1 (all conversions disabled); or when the ddmen bit (dac2 to dac1 analog mix disabled), the da2en bit (dac2 powered down), or the anaen bit (analog channels powered down) in control register ad- dress 27 is reset to 0. address 4 mix control?uxiliary 1 to dac1 data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lx1mm res res lx1m4 lx1m3 lx1m2 lx1m1 lx1m0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rx1mm res res rx1m4 rx1m3 rx1m2 rx1m1 rx1m0 lx1mm left auxiliary 1 mix mute 0 = mix enabled 1 = mix muted lx1m4:0 left auxiliary 1 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation rx1mm right auxiliary 1 mix mute 0 = mix enabled 1 = mix muted rx1m4:0 right auxiliary 1 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation
ad1843 rev. 0 C31C res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 1000 1000 (8888 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; when the pdno bit in control register address 0 is set to 1 (all conversions disabled); or when the aamen bit (analog input to analog mix disabled), or the anaen bit (analog channels powered down) in control register address 27 is reset to 0. address 5 mix control?uxiliary 2 to mixer data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lx2mm res res lx2m4 lx2m3 lx2m2 lx2m1 lx2m0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rx2mm res res rx2m4 rx2m3 rx2m2 rx2m1 rx2m0 lx2mm left auxiliary 2 mix mute 0 = mix enabled 1 = mix muted lx2m4:0 left auxiliary 2 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation rx2mm right auxiliary 2 mix mute 0 = mix enabled 1 = mix muted rx2m4:0 right auxiliary 2 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 1000 1000 (8888 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; when the pdno bit in control register ad- dress 0 is set to 1 (all conversions disabled); or when the aamen bit (analog input to analog mix disabled), or the anaen bit (analog channels powered down) in control register address 27 is reset to 0. address 6 mix control?uxiliary 3 to mixer data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lx3mm res res lx3m4 lx3m3 lx3m2 lx3m1 lx3m0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rx3mm res res rx3m4 rx3m3 rx3m2 rx3m1 rx3m0 lx3mm left auxiliary 3 mix mute 0 = mix enabled 1 = mix muted lx3m4:0 left auxiliary 3 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation
rev. 0 C32C ad1843 rx3mm right auxiliary 3 mix mute 0 = mix enabled 1 = mix muted rx3m4:0 right auxiliary 3 mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 1000 1000 (8888 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; when the pdno bit in control register address 0 is set to 1 (all conversions disabled); or when the aamen bit (analog input to analog mix disabled), or the anaen bit (analog channels powered down) in control register address 27 is reset to 0. address 7 mix control?ic to mixer data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lmcmm res res lmcm4 lmcm3 lmcm2 lmcm1 lmcm0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rmcmm res res rmcm4 rmcm3 rmcm2 rmcm1 rmcm0 lmcmm left microphone mix mute 0 = mix enabled 1 = mix muted lmcm4:0 left microphone mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation rmcmm right microphone mix mute 0 = mix enabled 1 = mix muted rmcm4:0 right microphone mix gain/attenuation select. least significant bit represents C1.5 db. referred to 2.0 v p-p dac1 output level. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 1000 1000 (8888 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; when the pdno bit in control register address 0 is set to 1 (all conversions disabled); or when the aamen bit (analog input to analog mix disabled), or the anaen bit (analog channels powered down) in control register address 27 is reset to 0. address 8 mix/miscellaneous control?ono in to mixer and miscellaneous settings data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 mnmm res res mnm4 mnm3 mnm2 mnm1 mnm0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 allmm mnom hpom hpos summ res dac2t dac1t mnmm mono input mix mute 0 = mix enabled 1 = mix muted
ad1843 rev. 0 C33C mnm4:0 left and right mono mix gain/attenuation select. least significant bit represents C1.5 db. 00000 = +12.0 db gain 01000 = 0.0 db 11111 = C34.5 db attenuation allmm all mix mute. mutes all mixing (mic, aux1, aux2, aux3, and dac2) to left and right dac1, overriding the independent mute control bits in control register addresses 3 through 8. 0 = mix to dac1 enabled 1 = mix to dac1 muted mnom mono output mute 0 = mono output enabled 1 = mono output muted hpom headphone output mute (left and right) 0 = headphone output enabled 1 = headphone output muted hpos headphone output voltage swing (left and right) 0 = 2 volts peak-to-peak 1 = 4 volts peak-to-peak summ sum left and right mute. mutes mixing from suml and sumr pins to dac1. 0 = suml and sumr mix to dac1 enabled 1 = suml and sumr mix to dac1 muted dac2t dac2 gain/attenuation change timing. this bit controls when changes to the dac2 gain/attenuation setting (control register 10) take effect. when set to 1, changes take effect immediately. when reset to 0, changes are delayed until either the output level on dac2 crosses zero (midscale), or until after a 10 to 12 ms time-out period is reached. delaying gain/attenuation changes until zero crossings reduces instantaneous output voltage changes, which reduces audible clicks. 0 = gain/attenuation changes applied on signal zero crossing or after a 10-12 ms time-out 1 = gain/attenuation changes applied immediately dac1t dac1 gain/attenuation change timing. this bit controls when changes to the dac1 gain/attenuation setting (control register 9) take effect. when set to 1, changes take effect immediately. when reset to 0, changes are delayed until either the output level on dac1 crosses zero (midscale), or until after a 10 to 12 ms time-out period is reached. delaying gain/attenuation changes until zero crossings reduces instantaneous output voltage changes, which reduces audible clicks. 0 = gain/attenuation changes applied on signal zero crossing or after a 10-12 ms time-out 1 = gain/attenuation changes applied immediately res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 0110 1000 (8868 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions powered down). the msb half (data 15 through data 8) of this control reg- ister is cleared to default and cannot be written to also when: the aamen bit (analog input to analog mix disabled), or the anaen bit (analog channels powered down) in control register 27 is reset to 0. address 9 output control?ac1 analog gain/attenuation data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lda1gm res lda1g5 lda1g4 lda1g3 lda1g2 lda1g1 lda1g0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rda1gm res rda1g5 rda1g4 rda1g3 rda1g2 rda1g1 rda1g0 lda1gm left dac1 analog mute 0 = left dac1 enabled 1 = left dac1 muted
rev. 0 C34C ad1843 lda1g5:0 left dac1 analog/digital gain/attenuation select. least significant bit represents C1.5 db. note that the implementation of the attenuation is mixed analog and digital. 0 00000 = +12.0 db: +12.0 db analog, +0.0 db digital 0 01000 = 0.0 db: +0.0 db analog, +0.0 db digital 0 11111 = C34.5 db: C34.5 db analog, +0.0 db digital 1 00000 = C36.0 db: C34.5 db analog, C1.5 db digital 1 11111 = C82.5 db: C34.5 db analog, C48.0 db digital rda1gm right dac1 analog mute 0 = right dac1 enabled 1 = right dac1 muted rda1g5:0 right dac1 analog/digital gain/attenuation select. least significant bit represents C1.5 db. note that the implementation of the attenuation is mixed analog and digital. 0 00000 = +12.0 db: +12.0 db analog, +0.0 db digital 0 01000 = 0.0 db: +0.0 db analog, +0.0 db digital 0 11111 = C34.5 db: C34.5 db analog, +0.0 db digital 1 00000 = C36.0 db: C34.5 db analog, C1.5 db digital 1 11111 = C82.5 db: C34.5 db analog, C48.0 db digital res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 1000 1000 (8888 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; when the pdno bit in control register address 0 is set to 1 (all conversions disabled); when the anaen bit in control register address 27 is reset to 0 (analog channels powered down); or when the da1en bit in control register address 27 is reset to 0 (dac1 disabled). address 10 output control?ac2 analog gain/attenuation data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lda2gm res lda2g5 lda2g4 lda2g3 lda2g2 lda2g1 lda2g0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rda2gm res rda2g5 rda2g4 rda2g3 rda2g2 rda2g1 rda2g0 lda2gm left dac2 analog mute 0 = left dac2 enabled 1 = left dac2 muted lda2g5:0 left dac2 analog/digital gain/attenuation select. least significant bit represents C1.5 db. note that the imple- mentation of the attenuation is mixed analog and digital. 0 00000 = +12.0 db: +12.0 db analog, +0.0 db digital 0 01000 = 0.0 db: +0.0 db analog, +0.0 db digital 0 11111 = C34.5 db: C34.5 db analog, +0.0 db digital 1 00000 = C36.0 db: C34.5 db analog, C1.5 db digital 1 11111 = C82.5 db: C34.5 db analog, C48.0 db digital rda2gm right dac2 analog mute 0 = right dac2 enabled 1 = right dac2 muted rda2g5:0 right dac2 analog/digital gain/attenuation select. least significant bit represents C1.5 db. note that the imple- mentation of the attenuation is mixed analog and digital. 0 00000 = +12.0 db: +12.0 db analog, +0.0 db digital 0 01000 = 0.0 db: +0.0 db analog, +0.0 db digital 0 11111 = C34.5 db: C34.5 db analog, +0.0 db digital 1 00000 = C36.0 db: C34.5 db analog, C1.5 db digital 1 11111 = C82.5 db: C34.5 db analog, C48.0 db digital res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 1000 1000 1000 (8888 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; when the pdno bit in control register address 0 is set to 1 (all conversions disabled); when the anaen bit in control register address 27 is reset to 0 (analog channels powered down); or when the da2en bit in control register address 27 is reset to 0 (dac2 disabled).
ad1843 rev. 0 C35C address 11 output control?ac1 digital attenuation data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lda1am res lda1a5 lda1a4 lda1a3 lda1a2 lda1a1 lda1a0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rda1am res rda1a5 rda1a4 rda1a3 rda1a2 rda1a1 rda1a0 lda1am left dac1 digital mute 0 = left dac1 enabled 1 = left dac1 muted lda1a5:0 left dac1 digital attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111111 = C94.5 db attenuation rda1am right dac1 digital mute 0 = right dac1 enabled 1 = right dac1 muted rda1a5:0 right dac1 digital attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111111 = C94.5 db attenuation res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled). address 12 output control?ac2 digital attenuation data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lda2am res lda2a5 lda2a4 lda2a3 lda2a2 lda2a1 lda2a0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rda2am res rda2a5 rda2a4 rda2a3 rda2a2 rda2a1 rda2a0 lda2am left dac2 digital mute 0 = left dac2 enabled 1 = left dac2 muted lda2a5:0 left dac2 digital attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111111 = C94.5 db attenuation rda2am right dac2 digital mute 0 = right dac2 enabled 1 = right dac2 muted rda2a5:0 right dac2 digital attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111111 = C94.5 db attenuation res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled).
rev. 0 C36C ad1843 address 13 digital mix control?dc to dac1 data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lad1mm res lad1m5 lad1m4 lad1m3 lad1m2 lad1m1 lad1m0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rad1mm res rad1m5 rad1m4 rad1m3 rad1m2 rad1m1 rad1m0 restrictions: adc and dac channel mixed must receive conversion rate from same clock generator. serial interface must be run- ning at a rate greater than or equal to the adc/dac conversion rate. for bus master: sdfs frequency 3 adc/dac conversion rate. for bus slave: tsi frequency 3 adc/dac conversion rate. lad1mm digital mix of left adc output with left dac1 input mute. 0 = mix enabled 1 = mix muted lad1m5:0 digital mix of left adc output with left dac1 input attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111110 = C93.0 db attenuation 111111 = full mute rad1mm digital mix of right adc output with right dac1 input mute. 0 = mix enabled 1 = mix muted rad1m5:0 digital mix of right adc output with right dac1 input attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111110 = C93.0 db attenuation 111111 = full mute res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 0000 1000 0000 (8080 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled). address 14 digital mix control?dc to dac2 data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 lad2mm res lad2m5 lad2m4 lad2m3 lad2m2 lad2m1 lad2m0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 rad2mm res rad2m5 rad2m4 rad2m3 rad2m2 rad2m1 rad2m0 restrictions: adc and dac channel mixed must receive conversion rate from same clock generator. serial interface must be run- ning at a rate greater than or equal to the adc/dac conversion rate. for bus master: sdfs frequency 3 adc/dac conversion rate. for bus slave: tsi frequency 3 adc/dac conversion rate. lad2mm digital mix of left adc output with left dac2 input mute. 0 = mix enabled 1 = mix muted lad2m5:0 digital mix of left adc output with left dac2 input attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111110 = C93.0 db attenuation 111111 = full mute rad2mm digital mix of right adc output with right dac2 input mute. 0 = mix enabled 1 = mix muted rad2m5:0 digital mix of right adc output with right dac2 input attenuation select. least significant bit represents C1.5 db. 000000 = +0.0 db attenuation 111110 = C93.0 db attenuation 111111 = full mute res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1000 0000 1000 0000 (8080 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled).
ad1843 rev. 0 C37C address 15 codec configuration?hannel sample rate source select data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res da2c1 da2c0 da1c1 da1c0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res adrc1 adrc0 adlc1 adlc0 da2c1:0 dac2 sample rate source. selects the sample rate clock source for left and right channels of dac2. 00 = conversion sample rate is 48.0 khz 01 = conversion sample rate is from clock generator 1 10 = conversion sample rate is from clock generator 2 11 = conversion sample rate is from clock generator 3 da1c1:0 dac1 sample rate source. selects the sample rate clock source for left and right channels of dac1. 00 = conversion sample rate is 48.0 khz 01 = conversion sample rate is from clock generator 1 10 = conversion sample rate is from clock generator 2 11 = conversion sample rate is from clock generator 3 adrc1:0 adc right sample rate source. selects the sample rate clock source for the right adc channel. 00 = conversion sample rate is 48.0 khz 01 = conversion sample rate is from clock generator 1 01 = conversion sample rate is from clock generator 2 11 = conversion sample rate is from clock generator 3 adlc1:0 adc left sample rate source. selects the sample rate clock source for the left adc channel. 00 = conversion sample rate is 48.0 khz 01 = conversion sample rate is from clock generator 1 10 = conversion sample rate is from clock generator 2 11 = conversion sample rate is from clock generator 3 res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled). address 16 clock generator 1 control?ode data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 c1ref c1vid c1pllg c1p200 c1x8/7 c1c128 res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c1m7 c1m6 c1m5 c1m4 c1m3 c1m2 c1m1 c1m0 c1ref clock generator 1 reference select. selects the fundamental clock reference used by clock generator 1 to synthesize its conversion (sample) and bit clock rates. 0 = clocks are referenced to the input on pin xtali (crystal or master clock input). sample clock frequency is defined by control register address 17 and bit c1x8/7. sample clock phase may be shifted by control register address 18. bit clock frequency is defined by bits c1m7:0 and c1p200. bit c1vid is ignored. 1 = clocks are referenced to the input on pin sync1 (sync 1 clock input). sample clock frequency is defined by c1vid and c1m7:0. sample clock phase is locked to sync1 and cannot be shifted. bit clock frequency is defined by bits c1m7:0 and c1p200 unless in video lock mode (c1vid set to 1) where the bit clocks are not produced. control register addresses 17, 18 and the c1x8/7 bit are ignored.
rev. 0 C38C ad1843 c1vid clock generator 1 video lock mode. this bit is used to select between lock modes when the clock generator 1 is referenced to sync1 (c1ref set to 1). this bit should be reset to 0 if c1ref is reset to 0. when reset to 0, clock generator 1 is in normal lock mode where the conversion clock will be frequency and phase locked to sync1, and the bit clock frequency is chosen using bits c1m7:0 and c1p200. when set to 1, clock generator 1 is in video lock mode, where the conversion clock frequency is selected using bits c1m7:0, and a bit clock is not pro- duced. c1pllg clock generator 1 pll loop gain select. if reset to 0, this bit selects finite pll loop gain, and if set to 1, this bit selects infinite pll loop gain. this bit should nominally be reset to 0. setting it to 1 may enhance the plls ability to lock to certain sync1 inputs, but it may also increase conversion noise. c1p200 clock generator 1 bit clock +200 frequency modifier. when set to 1, the bit clock driven out of pin bit1 will have a frequency that is 200 hertz greater than the frequency selected through bit c1m7:0. this bit is ignored when in video lock mode (c1vid set to 1). c1p200 only modifies the bit clock driven on the bit1 pin. c1x8/7 clock generator 1 conversion clock 8/7 frequency modifier. when set to 1, the conversion clock frequency gen- erated will be 8/7 times the value programmed in control register address 17. this bit is ignored when clocks are referenced to sync1 (c1ref set to 1). c1c128 clock generator 1 conversion clock pin (conv1) frequency select. when set to 1, the frequency driven on to the conv1 pin will be 128 times the conversion rate. when reset to 0, the frequency driven on to the conv1 pin will be the same as the conversion rate. c1c128 only modifies the clock frequency driven on the conv1 pin. res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. c1m7:0 clock generator 1 clock rate modifiers. when not in video lock mode (c1ref and c1vid are not both set to 1): bits c1m7:0 select the bit clock rate which will be driven out on pin bit1. using the following table, the least sig- nificant four bits (c1m3:0) are programmed to the desired bit clock rate, and the most significant four bits (c1m7:4) must be programmed to the conversion clock rate established using control register address 17 and the c1x8/7 bit. if the actual conversion clock differs from the value selected by c1m7:4, then the resultant bit clock will be different from the rate selected by the ratio of the c1m7:4 selected rate to the control register address 17 plus the c1x8/7 bit actual rate. c1m3:0 bit clock frequency* c1m7:4 conversion (sample) rate 0000 2,400 hertz 0000 7,200 hertz 0001 4,800 0001 8,400 0010 7,200 0010 9,000 0011 9,600 0011 9,600 0100 12,000 0100 11,200 0101 14,400 0101 12,000 0110 16,800 0110 12,800 0111 19,200 0111 7,200 8/7 1000 21,600 1000 9,000 8/7 1001 24,000 1001 9,600 8/7 1010 26,400 1010 12,000 8/7 1011 28,800 1011 reserved 1100 reserved 1100 reserved 1101 reserved 1101 reserved 1110 reserved 1110 reserved 1111 see below 1111 see below *bit clock frequencies listed will be increased by 200 hz if c1p200 is set to 1. when c1m7:4 is programmed to 1111 and c1m3:0 is programmed to 1111, the bit clock rate will be 128 times the conversion rate. when in video lock mode (c1ref and c1vid are both set to 1): bits c1m7:0 select the conversion clock rate. the most significant bit (c1m7) must be set to indicate the type of video lock, either ntsc or pal. for an ntsc lock, c1m7 must be reset to 0, and the sync1 pin must receive the ntsc sync frequency (525 lines/frame 30 hz 1000/1001 frame rate ? 15.734 khz). for a pal lock, c1m7 must be set to 1, and the sync1 pin must receive the pal sync frequency (625 lines/frame 25 hz frame rate ? 15.625 khz). the next three most significant bits (c1m6:4) select a desired base conversion clock rate, and the least significant four bits (c1m3:0) select a divisor . the conversion clock created by clock generator 1 will be the base divided by the divisor . the following tables list the possible choices for base and divisor.
ad1843 rev. 0 C39C ntsc (c1m7 = 0): base frequency in hz (c1m6:4) 48,000 32,000 44,100 44,100 2/3 48,000* 44,056 divisor (c1m3:0) (000) (001) (010) (011) (100) (101) 1 (0000) yes yes yes yes yes yes 2 (0001) yes yes yes yes yes yes 3 (0010) yes no yes no yes yes 4 (0011) yes yes yes yes yes yes 5 (0100) yes yes yes yes yes yes 6 (0101) yes no yes no yes yes 7 (0110) no no yes no yes yes 8 (0111) yes no yes no yes yes *when c1m6:4 = 100, base frequency is 48,000 hz only if ntsc sync rate is increased by 1001/1000, or is exactly 15.750 khz. pal (c1m7 = 1): base frequency in hz (c1m6:4) 48,000 32,000 44,100 44,100 2/3 divisor (c1m3:0) (000) (001) (010) (011) 1 (0000) yes yes yes yes 2 (0001) yes yes yes yes 3 (0010) yes no yes no 4 (0011) yes yes yes yes 5 (0100) yes yes yes yes 6 (0101) yes no yes no 7 (0110) yes no yes no 8 (0111) yes no yes no initial default state after reset: 0000 0000 1111 1111 (00ff hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. address 17 clock generator 1 control?ample rate data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 c1c15 c1c14 c1c13 c1c12 c1c11 c1c10 c1c9 c1c8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c1c7 c1c6 c1c5 c1c4 c1c3 c1c2 c1c1 c1c0 c1c15:0 clock generator 1 conversion (sample) rate select. defines the conversion rate produced by clock generator 1 when not referenced to the sync1 pin (control register address 16 bit 15 [c1ref]). one lsb represents exactly one hertz, assuming a 24.576 mhz clock input on the xtali pin. usable range is 4 khz (0x0fa0) to 54 khz (0xd2f0). initial default state after reset: 1011 1011 1000 0000 (bb80 hex), which is 48 khz, assuming a 24.576 mhz clock in- put on the xtali pin. cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. address 18 clock generator 1 control?ample phase shift data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res c1pd data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c1p7 c1p6 c1p5 c1p4 c1p3 c1p2 c1p1 c1p0 c1pd clock generator 1 phase shift direction. this bit controls the direction of sample clock phase shift. 0 = phase advance 1 = phase retard c1p7:0 clock generator 1 phase shift magnitude. these bits control the magnitude of sample clock phase shift. one lsb represents exactly 0.12 degrees. lsbs are processed and decremented at a rate of 3.072 mhz (assuming a 24.576 mhz clock input on the xtali pin). when this register is read, it indicates any phase advance/retard remaining to be processed as of the beginning of slot 0 if bus master, or when tsi was received if bus slave. this register may be
rev. 0 C40C ad1843 res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the g1en bit in control register address 28 is reset to 0 (clock generator 1 disabled). address 19 clock generator 2 control?ode data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 c2ref c2vid c2pllg c2p200 c2x8/7 c2c128 res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c2m7 c2m6 c2m5 c2m4 c2m3 c2m2 c2m1 c2m0 c2ref clock generator 2 reference select. selects the fundamental clock reference used by clock generator 2 to synthesize its conversion (sample) and bit clock rates. 0 = clocks are referenced to the input on pin xtali (crystal or master clock input). sample clock frequency is defined by control register address 20 and bit c2x8/7. sample clock phase may be shifted by control register address 21. bit clock frequency is defined by bits c2m7:0 and c2p200. bit c2vid is ignored. 1 = clocks are referenced to the input on pin sync2 (sync 2 clock input). sample clock frequency is defined by c2vid and c2m7:0. sample clock phase is locked to sync2 and cannot be shifted. bit clock frequency is defined by bits c2m7:0 and c2p200 unless in video lock mode (c2vid set to 1) where the bit clocks are not produced. control register addresses 17, 18 and the c2x8/7 bit are ignored. c2vid clock generator 2 video lock mode. this bit is used to select between lock modes when the clock generator 2 is referenced to sync2 (c2ref set to 1). this bit should be reset to 0 if c2ref is reset to 0. when reset to 0, clock generator 2 is in normal lock mode where the conversion clock will be frequency and phase locked to sync2, and the bit clock frequency is chosen using bits c2m7:0 and c2p200. when set to 1, clock generator 2 is in video lock mode, where the conversion clock frequency is selected using bits c2m7:0, and a bit clock is not produced. c2pllg clock generator 2 pll loop gain select. if reset to 0, this bit selects finite pll loop gain, and if set to 1, this bit selects infinite pll loop gain. this bit should nominally be reset to 0. setting it to 1 may enhance the plls ability to lock to certain sync2 inputs, but it may also increase conversion noise. c2p200 clock generator 2 bit clock +200 frequency modifier. when set to 1, the bit clock driven out of pin bit2 will have a frequency that is 200 hertz greater than the frequency selected through bit c2m7:0. this bit is ignored when in video lock mode (c2vid set to 1). c2p200 only modifies the bit clock driven on the bit2 pin. c2x8/7 clock generator 2 conversion clock 8/7 frequency modifier. when set to 1, the conversion clock frequency gen- erated will be 8/7 times the value programmed in control register address 20. this bit is ignored when clocks are referenced to sync2 (c2ref set to 1). c2c128 clock generator 2 conversion clock pin (conv2) frequency select. when set to 1, the frequency driven on to the conv2 pin will be 128 times the conversion rate. when reset to 0, the frequency driven on to the conv2 pin will be the same as the conversion rate. c2c128 only modifies the clock frequency driven on the conv2 pin. res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. c2m7:0 clock generator 2 clock rate modifiers. when not in video lock mode (c2ref and c2vid are not both set to 1): bits c2m7:0 select the bit clock rate which will be driven out on pin bit2. using the following table, the least sig- nificant four bits (c2m3:0) are programmed to the desired bit clock rate, and the most significant four bits overwritten even if all previously programmed phase advance/retard has not been processed. when written, the con- tents of this register (just prior to the write) are transmitted during slot 1 of the following frame (as with all control register writes).
ad1843 rev. 0 C41C (c2m7:4) must be programmed to the conversion clock rate established using control register address 20 and the c2x8/7 bit. if the actual conversion clock differs from the value selected by c2m7:4, then the resultant bit clock will be different from the rate selected by the ratio of the c2m7:4 selected rate to the control register address 20 plus the c2x8/7 bit actual rate. c2m3:0 bit clock frequency* c2m7:4 conversion (sample) rate 0000 2,400 hertz 0000 7,200 hertz 0001 4,800 0001 8,400 0010 7,200 0010 9,000 0011 9,600 0011 9,600 0100 12,000 0100 11,200 0101 14,400 0101 12,000 0110 16,800 0110 12,800 0111 19,200 0111 7,200 8/7 1000 21,600 1000 9,000 8/7 1001 24,000 1001 9,600 8/7 1010 26,400 1010 12,000 8/7 1011 28,800 1011 reserved 1100 reserved 1100 reserved 1101 reserved 1101 reserved 1110 reserved 1110 reserved 1111 see below 1111 see below *bit clock frequencies listed will be increased by 200 hz if c2p200 is set to 1. when c2m7:4 is programmed to 1111 and c2m3:0 is programmed to 1111, the bit clock rate will be 128 times the conversion rate. when in video lock mode (c2ref and c2vid are both set to 1): bits c2m7:0 select the conversion clock rate. the most significant bit (c2m7) must be set to indicate the type of video lock, either ntsc or pal. for an ntsc lock, c2m7 must be reset to 0, and the sync2 pin must receive the ntsc sync frequency (525 lines/frame 30 hz 1000/1001 frame rate ? 15.734 khz). for a pal lock, c2m7 must be set to 1, and the sync2 pin must receive the pal sync frequency (625 lines/frame 25 hz frame rate ? 15.625 khz). the next three most significant bits (c2m6:4) select a desired base conversion clock rate, and the least significant four bits (c2m3:0) select a divisor . the conversion clock created by clock generator 2 will be the base divided by the divisor . the following tables list the possible choices for base and divisor. ntsc (c2m7 = 0): base frequency in hz (c2m6:4) 48,000 32,000 44,100 44,100 2/3 48,000* 44,056 divisor (c2m3:0) (000) (001) (010) (011) (100) (101) 1 (0000) yes yes yes yes yes yes 2 (0001) yes yes yes yes yes yes 3 (0010) yes no yes no yes yes 4 (0011) yes yes yes yes yes yes 5 (0100) yes yes yes yes yes yes 6 (0101) yes no yes no yes yes 7 (0110) no no yes no yes yes 8 (0111) yes no yes no yes yes *when c2m6:4 = 100, base frequency is 48,000 hz only if ntsc sync rate is increased by 1001/1000, or is exactly 15.750 khz. pal (c2m7 = 1): base frequency in hz (c2m6:4) 48,000 32,000 44,100 44,100 2/3 divisor (c2m3:0) (000) (001) (010) (011) 1 (0000) yes yes yes yes 2 (0001) yes yes yes yes 3 (0010) yes no yes no 4 (0011) yes yes yes yes 5 (0100) yes yes yes yes 6 (0101) yes no yes no 7 (0110) yes no yes no 8 (0111) yes no yes no initial default state after reset: 0000 0000 1111 1111 (00ff hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo.
rev. 0 C42C ad1843 address 20 clock generator 2 control?ample rate data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 c2c15 c2c14 c2c13 c2c12 c2c11 c2c10 c2c9 c2c8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c2c7 c2c6 c2c5 c2c4 c2c3 c2c2 c2c1 c2c0 c2c15:0 clock generator 2 conversion (sample) rate select. defines the conversion rate produced by clock generator 2 when not referenced to the sync2 pin (control register address 19 bit 15 [c2ref]). one lsb represents exactly one hertz, assuming a 24.576 mhz clock input on the xtali pin. usable range is 4 khz (0x0fa0) to 54 khz (0xd2f0). initial default state after reset: 1011 1011 1000 0000 (bb80 hex), which is 48 khz, assuming a 24.576 mhz clock in- put on the xtali pin. cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. address 21 clock generator 2 control?ample phase shift data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res c2pd data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c2p7 c2p6 c2p5 c2p4 c2p3 c2p2 c2p1 c2p0 c2pd clock generator 2 phase shift direction. this bit controls the direction of sample clock phase shift. 0 = phase advance 1 = phase retard c2p7:0 clock generator 2 phase shift magnitude. these bits control the magnitude of sample clock phase shift. one lsb repre- sents exactly 0.12 degrees. lsbs are processed and decremented at a rate of 3.072 mhz (assuming a 24.576 mhz clock input on the xtali pin). when this register is read, it indicates any phase advance/retard remaining to be processed as of the beginning of slot 0 if bus master, or when tsi was received if bus slave. this register may be overwritten even if all pre- viously programmed phase advance/retard has not been processed. when written, the contents of this register (just prior to the write) are transmitted during slot 1 of the following frame (as with all control register writes). res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the g2en bit in control register address 28 is reset to 0 (clock generator 2 disabled). address 22 clock generator 3 control?ode data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 c3ref c3vid c3pllg c3p200 c3x8/7 c3c128 res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c3m7 c3m6 c3m5 c3m4 c3m3 c3m2 c3m1 c3m0 c3ref clock generator 3 reference select. selects the fundamental clock reference used by clock generator 3 to synthe- size its conversion (sample) and bit clock rates. 0 = clocks are referenced to the input on pin xtali (crystal or master clock input). sample clock frequency is defined by control register address 23 and bit c3x8/7. sample clock phase may be shifted by control register address 24. bit clock frequency is defined by bits c3m7:0 and c3p200. bit c3vid is ignored.
ad1843 rev. 0 C43C 1 = clocks are referenced to the input on pin sync3 (sync 3 clock input). sample clock frequency is defined by c3vid and c3m7:0. sample clock phase is locked to sync3 and cannot be shifted. bit clock frequency is defined by bits c3m7:0 and c3p200 unless in video lock mode (c3vid set to 1) where the bit clocks are not produced. control register addresses 23, 24 and the c3x8/7 bits are ignored. c3vid clock generator 3 video lock mode. this bit is used to select between lock modes when the clock generator 3 is referenced to sync3 (c3ref set to 1). this bit should be reset to 0 if c3ref is reset to 0. when reset to 0, clock generator 3 is in normal lock mode where the conversion clock will be frequency and phase locked to sync3, and the bit clock frequency is chosen using bits c3m7:0 and c3p200. when set to 1, clock generator 3 is in video lock mode, where the conversion clock frequency is selected using bits c3m7:0, and a bit clock is not produced. c3pllg clock generator 3 pll loop gain select. if reset to 0, this bit selects finite pll loop gain, and if set to 1, this bit selects infinite pll loop gain. this bit should nominally be reset to 0. setting it to 1 may enhance the plls ability to lock to certain sync3 inputs, but it may also increase conversion noise. c3p200 clock generator 3 bit clock +200 frequency modifier. when set to 1, the bit clock driven out of pin bit3 will have a frequency that is 200 hertz greater than the frequency selected through bit c3m7:0. this bit is ignored when in video lock mode (c3vid set to 1). c3p200 only modifies the bit clock driven on the bit3 pin. c3x8/7 clock generator 3 conversion clock 8/7 frequency modifier. when set to 1, the conversion clock frequency gen- erated will be 8/7 times the value programmed in control register address 23. this bit is ignored when clocks are referenced to sync3 (c3ref set to 1). c3c128 clock generator 3 conversion clock pin (conv3) frequency select. when set to 1, the frequency driven on to the conv3 pin will be 128 times the conversion rate. when reset to 0, the frequency driven on to the conv3 pin will be the same as the conversion rate. c3c128 only modifies the clock frequency driven on the conv3 pin. res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. c3m7:0 clock generator 3 clock rate modifiers. when not in video lock mode (c3ref and c3vid are not both set to 1): bits c3m7:0 select the bit clock rate which will be driven out on pin bit3. using the following table, the least sig- nificant four bits (c3m3:0) are programmed to the desired bit clock rate, and the most significant four bits (c3m7:4) must be programmed to the conversion clock rate established using control register address 23 and the c3x8/7 bit. if the actual conversion clock differs from the value selected by c3m7:4, then the resultant bit clock will be different from the rate selected by the ratio of the c3m7:4 selected rate to the control register address 23 plus the c3x8/7 bit actual rate. c3m3:0 bit clock frequency* c3m7:4 conversion (sample) rate 0000 2,400 hertz 0000 7,200 hertz 0001 4,800 0001 8,400 0010 7,200 0010 9,000 0011 9,600 0011 9,600 0100 12,000 0100 11,200 0101 14,400 0101 12,000 0110 16,800 0110 12,800 0111 19,200 0111 7,200 8/7 1000 21,600 1000 9,000 8/7 1001 24,000 1001 9,600 8/7 1010 26,400 1010 12,000 8/7 1011 28,800 1011 reserved 1100 reserved 1100 reserved 1101 reserved 1101 reserved 1110 reserved 1110 reserved 1111 see below 1111 see below *bit clock frequencies listed will be increased by 200 hz if c3p200 is set to 1. when c3m7:4 is programmed to 1111 and c3m3:0 is programmed to 1111, the bit clock rate will be 128 times the conversion rate.
rev. 0 C44C ad1843 when in video lock mode (c3ref and c3vid are both set to 1): bits c3m7:0 select the conversion clock rate. the most significant bit (c3m7) must be set to indicate the type of video lock, either ntsc or pal. for an ntsc lock, c3m7 must be reset to 0, and the sync3 pin must receive the ntsc sync frequency (525 lines/frame 30 hz 1000/1001 frame rate ? 15.734 khz). for a pal lock, c3m7 must be set to 1, and the sync3 pin must receive the pal sync frequency (625 lines/frame 25 hz frame rate ? 15.625 khz). the next three most significant bits (c3m6:4) select a desired base conversion clock rate, and the least significant four bits (c3m3:0) select a divisor . the conversion clock created by clock generator 3 will be the base divided by the divisor . the following tables list the possible choices for base and divisor. ntsc (c3m7 = 0): base frequency in hz (c3m6:4) 48,000 32,000 44,100 44,100 2/3 48,000* 44,056 divisor (c3m3:0) (000) (001) (010) (011) (100) (101) 1 (0000) yes yes yes yes yes yes 2 (0001) yes yes yes yes yes yes 3 (0010) yes no yes no yes yes 4 (0011) yes yes yes yes yes yes 5 (0100) yes yes yes yes yes yes 6 (0101) yes no yes no yes yes 7 (0110) no no yes no yes yes 8 (0111) yes no yes no yes yes *when c3m6:4 = 100, base frequency is 48,000 hz only if ntsc sync rate is increased by 1001/1000, or is exactly 15.570 khz. pal (c3m7 = 1): base frequency in hz (c3m6:4) 48,000 32,000 44,100 44,100 2/3 divisor (c3m3:0) (000) (001) (010) (011) 1 (0000) yes yes yes yes 2 (0001) yes yes yes yes 3 (0010) yes no yes no 4 (0011) yes yes yes yes 5 (0100) yes yes yes yes 6 (0101) yes no yes no 7 (0110) yes no yes no 8 (0111) yes no yes no initial default state after reset: 0000 0000 1111 1111 (00ff hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. address 23 clock generator 3 control?ample rate data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 c3c15 c3c14 c3c13 c3c12 c3c11 c3c10 c3c9 c3c8 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c3c7 c3c6 c3c5 c3c4 c3c3 c3c2 c3c1 c3c0 c3c15:0 clock generator 3 conversion (sample) rate select. defines the conversion rate produced by clock generator 3 when not referenced to the sync3 pin (control register address 22 bit 15 [c3ref]). one lsb represents exactly one hertz, assuming a 24.576 mhz clock input on the xtali pin. usable range is 4 khz (0x0fa0) to 54 khz (0xd2f0). initial default state after reset: 1011 1011 1000 0000 (bb80 hex), which is 48 khz, assuming a 24.576 mhz clock in- put on the xtali pin. cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo.
ad1843 rev. 0 C45C address 24 clock generator 3 control?ample phase shift data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res c3pd data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 c3p7 c3p6 c3p5 c3p4 c3p3 c3p2 c3p1 c3p0 c3pd clock generator 3 phase shift direction. this bit controls the direction of sample clock phase shift. 0 = phase advance 1 = phase retard c3p7:0 clock generator 3 phase shift magnitude. these bits control the magnitude of sample clock phase shift. one lsb represents exactly 0.12 degrees. lsbs are processed and decremented at a rate of 3.072 mhz (assuming a 24.576 mhz clock input on the xtali pin). when this register is read, it indicates any phase advance/retard remaining to be processed as of the beginning of slot 0 if bus master, or when tsi was received if bus slave. this register may be overwritten even if all previously programmed phase advance/retard has not been processed. when written, the con- tents of this register (just prior to the write) are transmitted during slot 1 of the following frame (as with all control register writes). res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the g3en bit in control register address 28 is reset to 0 (clock generator 3 disabled). address 25 codec configuration?igital filter and mode select data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 drsflt damix res res res res da2flt da1flt data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 daadr1 daadr0 daadl1 daadl0 res res adrflt adlflt restrictions: modem filter should not be selected when channel sample rate is above 24 khz. drsflt digital resampler filter mode. dac2 is sacrificed so that the remaining four channels (left adc, right adc, left and right dac1) have sufficient resources to realize the more stringent resampling filter requirements. see below for filter specifications. (note: digital resampling can be done without using the resampling filters but performance is not as optimal.) this bit can be altered only if all channels (dac1, dac2, adl, and adr) are powered down. note that this bit enables the digital resampler filters, but does not enable the digital resampler data pathways. use the daadr1:0 and daadl1:0 bits to enable the resampler data pathways. 0 = audio/modem filter modes (defined by da1flt, da2flt, adlflt and adrflt) 1 = digital resampler filter mode (da1flt, da2flt, adlflt and adrflt filter selections are ignored). damix dac digital mix enable. when set to 1, dac1 and dac2 are mixed at the output of the digital interpolation fil- ter. left and right channels are still separated, but dac1 and dac2 outputs are identical. this bit can be altered only if both dac1 and dac2 are powered down. 0 = digital mix disabled 1 = digital mix enabled da2flt dac2 (left and right channels) digital filter select and analog output swing select. digital filter selected is over- ridden when drsflt is set to 1. see below for filter specifications. this bit can be altered only if dac2 is powered down. 0 = digital audio filter. nominal analog swing is 2.000 v p-p when output level is +0.0 db (see control register address 10). 1 = digital modem filter. nominal analog swing is 3.156 v p-p when output level is +4.5 db (see control register address 10).
rev. 0 C46C ad1843 da1flt dac1 (left and right channels) digital filter select. this bit is overridden when drsflt is set to 1. see be- low for filter specifications. this bit can be altered only if dac1 is powered down. unlike da2flt, this bit does not control the analog output swing of dac1, which is fixed at 2.000 v p-p when the output level is set to 0 db by control register address 9. 0 = digital audio filter 1 = digital modem filter daadr1:0 digital adc right channel source select. 00 = input from analog: adc right channel (normal operation) 01 = reserved 10 = input from digital: dac1 right channel (digital resampler) 11 = input from digital: dac2 right channel (digital resampler) daadl1:0 digital adc left channel source select. 00 = input from analog: adc left channel (normal operation) 01 = reserved 10 = input from digital: dac1 left channel (digital resampler) 11 = input from digital: dac2 left channel (digital resampler) adrflt adc right channel digital filter select and analog input full scale mapping select. digital filter selected is over- ridden when drsflt is set to 1. see below for filter specifications. this bit can be altered only if the adc right channel is powered down. 0 = digital audio filter. 2.800 v p-p analog input maps to 2 15 when gain is set to 0.0 db (see control register address 2). 1 = digital modem filter. 3.156 v p-p analog input maps to 2 15 when gain is set to 0.0 db (see control register address 2). adlflt adc left channel digital filter select and analog input full scale mapping select. digital filter selected is over- ridden when drsflt is set to 1. see below for filter specifications. this bit can be altered only if the adc left channel is powered down. 0 = digital audio filter. 2.800 v p-p analog input maps to 2 15 when gain is set to 0.0 db (see control register address 2). 1 = digital modem filter. 3.156 v p-p analog input maps to 2 15 when gain is set to 0.0 db (see control register address 2). res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled). the table below lists the single path (dac or adc only) filter characteristics of the audio, modem, and resampler digital filters. passband ripple and group delay must be doubled if the dac is routed back through the adc for digital resampling. for further information, see the filter selection section and the filter frequency response plots, figures 34 through 36. passband stopband linear filter ripple edge ripple edge group delay audio C0.016 db 0.400 f s < C91.8 db 0.600 f s < 15/f s modem C0.220 db 0.442 f s < C75.7 db 0.542 f s < 19/f s resampler C0.035 db 0.400 f s < C92.2 db 0.500 f s < 25/f s f s is the conversion (sample) rate. address 26 codec configuration?erial interface data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 da2sm da1sm res res da2f1 da2f0 da1f1 da1f0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 scf frs frst adtlk adrf1 adrf0 adlf1 adlf0
ad1843 rev. 0 C47C da2sm dac2 stereo/mono mode select. when in stereo mode, data transmitted to the ad1843 during slot 4 is used by dac2 left and data transmitted to the ad1843 during slot 5 is used by dac2 right. when in mono mode, data transmitted to the ad1843 during slot 4 is used by both dac2 left and right, and data transmitted to the ad1843 during slot 5 is ignored. (note: slot 0 is assumed to be the first slot.) 0 = stereo mode 1 = mono mode da1sm dac1 stereo/mono mode select. when in stereo mode, data transmitted to the ad1843 during slot 2 is used by dac1 left and data transmitted to the ad1843 during slot 3 is used by dac1 right. when in mono mode, data transmitted to the ad1843 during slot 2 is used by both dac1 left and right, and data transmitted to the ad1843 during slot 3 is ignored. (note: slot 0 is assumed to be the first slot.) 0 = stereo mode 1 = mono mode da2f1:0 dac2 (left and right channels) data format select. 00 = 8-bit unsigned linear pcm 01 = 16-bit signed linear pcm 10 = 8-bit m -law companded 11 = 8-bit a-law companded da1f1:0 dac1 (left and right channels) data format select. 00 = 8-bit unsigned linear pcm 01 = 16-bit signed linear pcm 10 = 8-bit m -law companded 11 = 8-bit a-law companded scf sclk frequency select. changes to this bit do not take effect until shortly after the sixth slot (the final slot owned) is completed. relevant when the ad1843 is in master mode only. 0 = 12.288 mhz 1 = 16.384 mhz frs frame size select. selects the number of slots per frame. changes to this bit do not take effect until the time specified by the frst bit. 0 = 32 slots per frame 1 = 16 slots per frame frst frame size change timing. selects the point in time when frs takes effect. 0 = frame size changes once the current frame is complete 1 = frame size changes immediately, beginning with the current frame adtlk adc transmit lock mode select. when this bit is set to 1, adc transmit lock mode is entered. in this mode, left and right adc samples are transmitted during a frame only if both left and right samples are ready to be transmitted. when not in adc transmit lock mode, left and right samples are transmitted as they individually become available. note that even if left and right adcs are programmed to the same sample rate, unless the ad1843 is in transmit lock mode, adc samples will not necessarily be transmitted paired together in a tdm frame. this bit should be set to 1 only if both adcs are powered down and only if both adcs will be programmed to the sample rate once they are enabled. this bit may be reset to 0 at any time. while in adc transmit lock mode, both adcs must be enabled and disabled simulta- neously (write to control register address 27 must set adlen and adren both to 0 or to 1). 0 = adc transmit lock mode disabled 1 = adc transmit lock mode enabled adrf1:0 adc right channel data format select. 00 = 8-bit unsigned linear pcm 01 = 16-bit signed linear pcm 10 = 8-bit m -law companded 11 = 8-bit a-law companded adlf1:0 adc left channel data format select. 00 = 8-bit unsigned linear pcm 01 = 16-bit signed linear pcm 10 = 8-bit m -law companded 11 = 8-bit a-law companded res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo.
rev. 0 C48C ad1843 address 27 codec configuration?hannel power down data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 dfree res res ddmen res res da2en da1en data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 anaen hpen res aamen res res adren adlen dfree digital resource free. when set to 1, this bit reduces the level of power down normally achieved by asserting the power down bits in this control register. however, it allows the digital resources of channels powered down to be partially transferred to the remaining enabled channels and thus permits conversion rates higher than the 48 khz nominal maximum. see the supported conversion rates section. 0 = power down of digital resources is complete 1 = power down of digital resources is incomplete, freeing transferable resources to enabled channels ddmen dac2 to dac1 mix enable/power down. when reset to 0, control register address 3 is cleared to its default and cannot be written. 0 = mix is powered down 1 = mix is enabled da2en dac2 (left and right channels) enable/power down. when this bit is reset to 0, dac2 is powered down and serial interface requests for samples (see slot 0) will cease immediately. when this bit is set to 1, dac2 is enabled and requests for samples resume on the next frame. conversions will not actually begin until the first rising edge of the conversion clock (conv pin) after the sixth rising edge of frame sync (sdfs pin) after this bit is set to 1. this delay allows the four word stereo input buffer for dac2 to be filled before conversions begin and allows the serial in- terface to synchronize with the conversion clock, which is potentially already running. this delay also allows dacs on multiple ad1843s sharing the same frame sync to begin conversions synchronously, providing they are enabled at any time during the same frame. when this bit is reset to 0, control register address 10, which is the output level control register for dac2, is cleared to default and cannot be written. 0 = dac2 is powered down, control register address 10 (dac2 output level control) is cleared 1 = dac2 is enabled da1en dac1 (left and right channels) enable/power down. when this bit is reset to 0, dac1 is powered down and serial interface requests for samples (see slot 0) will cease immediately. when this bit is set to 1, dac1 is enabled and requests for samples resume on the next frame. conversions will not actually begin until the first rising edge of the conversion clock (conv pin) after the sixth rising edge of frame sync (sdfs pin) after this bit is set to 1. this delay allows the four word stereo input buffer for dac1 to be filled before conversions begin and allows the serial in- terface to synchronize with the conversion clock, which is potentially already running. this delay also allows dacs on multiple ad1843s sharing the same frame sync to begin conversions synchronously, providing they are enabled at any time during the same frame. when this bit is reset to 0, control register address 9, which is the output level control register for dac1, is cleared to default and cannot be written. 0 = dac1 is powered down, control register address 9 (dac1 output level control) is cleared 1 = dac1 is enabled anaen analog channel enable/power down. if this bit is reset to 0, all analog conversion circuitry (related to bits da1en, da2en, adlen, adren, ddmen and aamen) will be powered down. when reset to 0, the only conversion channels usable (but must be enabled separately) are the digital resampling paths. since resetting this bit to 0 powers down the analog adc, dac1, dac2 and analog mixers, their related output level control register addresses 4 through 7, half of 8, 9 and 10 are cleared and cannot be written while anaen remains 0. 0 = analog channels are powered down, control register addresses 4 through 7, half of 8, 9 and 10 are cleared to default (digital to digital still possible) 1 = analog dac and adc paths are controlled by da1en, da2en, adlen, adren, ddmen and aamen hpen headphone enable/power down. 0 = headphone output is powered down 1 = headphone output is enabled aamen analog input to analog output mix enable/power down. when reset to 0, control register addresses 4 - 7 and half of control register address 8 are cleared and cannot be written. 0 = mix is powered down 1 = mix is enabled
ad1843 rev. 0 C49C adren adc right channel enable/power down. when this bit is reset to 0, the right adc channel is powered down and serial interface sample output will cease after the current frame. when this bit is set to 1, the right adc channel is enabled and sampling of analog input will begin on the first rising edge of the conversion clock (conv pin) after the sixth rising edge of frame sync (sdfs pin). this delay allows the adc startup to be similar to the dac startup, and allows some time for stale adc data inside the ad1843 to be cleared. it also allows the serial interface to synchronize with the conversion clock, which is potentially already running. multiple adcs on multiple ad1843s which share the same frame sync will be synchronously started if they are enabled at any time during the same frame. 0 = adc right channel is powered down 1 = adc right channel is enabled adlen adc left channel enable/power down. when this bit is reset to 0, the left adc channel is powered down and serial interface sample output will cease after the current frame. when this bit is set to 1, the left adc channel is enabled and sampling of analog input will begin on the first rising edge of the conversion clock (conv pin) after the sixth rising edge of frame sync (sdfs pin). this delay allows the adc startup to be similar to the dac startup, and allows some time for stale adc data inside the ad1843 to be cleared. it also allows the serial interface to synchronize with the conversion clock, which is potentially already running. multiple adcs on multiple ad1843s which share the same frame sync will be synchronously started if they are enabled at any time during the same frame. 0 = adc left channel is powered down 1 = adc left channel is enabled res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 1100 0000 (00c0 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; when the pwrdwn pin is asserted lo; or when the pdno bit in control register address 0 is set to 1 (all conversions disabled). address 28 codec configuration?undamental settings data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 pdni acen c3en c2en c1en enclko xctl1 xctl0 data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 encv3 enbt3 encv2 enbt2 encv1 enbt1 linrsd linlsd pdni converter power down. when set to 1, this bit initiates the process of powering down all conversion channels, overriding the settings in control register address 27. unlike the individual power down bits in control register address 27, asserting this bit allows the dac and v ref output pins to slowly decay to ground. setting this bit to 1 results in lower power consumption than results from powering down all channels individually through control regis- ter address 27, and also reduces speaker click if asserted before the power supply is removed. power down is not complete until approximately 5 ms after this bit is set to 1. during this time, the ad1843 still requires a clock in- put from xtali for proper operation. when reset to 0, this bit initiates the process of preparing the ad1843 for conversions after power down. approximately 470 ms are necessary to exit power down. the power up/down status of the ad1843 may be monitored through the pdno bit in control register address 0. asserting the power down pin ( pwrdwn ) will result in an even greater degree of power down as it also powers down the serial interface and crystal oscillator. once a power up or power down sequence has been initiated, its completion cannot be interrupted. changes made to the state of pdni (i.e., writes to pdni) are ignored until a previously pending pdni state change has been completely processed. 0 = normal (non-power down) operation 1 = initiate power down of all conversion channels acen autocalibration enable. when set to 1, autocalibration will occur each time power down is exited (pdni is changed from a 1 to a 0). autocalibration increases the time required to exit power down by 4 ms (from 470 ms to 474 ms). when reset to 0, autocalibration is disabled. this bit is set to 1 initially after reset and can- not be overwritten until after the ad1843 has come out of power down (pdno bit in control register address 0 reset to 0) at least once. note that an autocalibration cycle is always performed following a hardware reset (i.e., reset pin asserted) or a hardware power down (i.e., pwrdwn pin asserted), regardless of the state of acen. 0 = autocalibration disabled 1 = autocalibration enabled, initiated upon exiting power down c3en clock generator 3 enable/power down. 0 = clock generator 3 powered down 1 = clock generator 3 enabled
rev. 0 C50C ad1843 c2en clock generator 2 enable/power down. 0 = clock generator 2 powered down 1 = clock generator 2 enabled c1en clock generator 1 enable/power down. 0 = clock generator 1 powered down 1 = clock generator 1 enabled enclko clkout pin enable. 0 = clock output is three-stated (powered down) 1 = clock output is enabled xctl1:0 external control. the state of these independent bits is reflected on the respective xctl1 and xctl0 output pins. 0 = ttl logic level lo on xctl1/xctl0 pin 1 = ttl logic level hi on xctl1/xctl0 pin encv3 conv3 pin enable. 0 = conv3 is three-stated (powered down) 1 = conv3 is enabled enbt3 bit3 pin enable. 0 = bit3 is three-stated (powered down) 1 = bit3 is enabled encv2 conv2 pin enable. 0 = conv2 is three-stated (powered down) 1 = conv2 is enabled enbt2 bit2 pin enable. 0 = bit2 is three-stated (powered down) 1 = bit2 is enabled encv1 conv1 pin enable. 0 = conv1 is three-stated (powered down) 1 = conv1 is enabled enbt1 bit1 pin enable. 0 = bit1 is three-stated (powered down) 1 = bit1 is enabled linrsd line input right channel single-ended or differential configuration. 0 = right channel line input is single-ended 1 = right channel line input is differential linlsd line input left channel single-ended or differential configuration. 0 = left channel line input is single-ended 1 = left channel line input is differential res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 1100 0100 0000 0000 (c400 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. address 29 reserved for future expansion data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res res res res res res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo.
ad1843 rev. 0 C51C address 30 reserved for future expansion data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res res res res res res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. address 31 reserved for future expansion data 15 data 14 data 13 data 12 data 11 data 10 data 9 data 8 res res res res res res res res data 7 data 6 data 5 data 4 data 3 data 2 data 1 data 0 res res res res res res res res res reserved for future expansion. to ensure future compatibility, write 0 to all reserved bits. initial default state after reset: 0000 0000 0000 0000 (0000 hex). cleared to default and cannot be written to when: the reset pin is asserted lo; or when the pwrdwn pin is asserted lo. bit and register maps a map of all tdm time slot bit assignments and control register contents is summarized for reference as follows in figure 14 and figure 15: input slot 0 or 16 1 or 17 2 or 18 3 or 19 4 or 20 5 or 21 data 15 res data 15 data 15 data 15 data 15 data 15 data 14 res data 14 data 14 data 14 data 14 data 14 data 13 res data 13 data 13 data 13 data 13 data 13 data 12 res data 12 data 12 data 12 data 12 data 12 data 11 res data 11 data 11 data 11 data 11 data 11 data 10 res data 10 data 10 data 10 data 10 data 10 data 9 da2v data 9 data 9 data 9 data 9 data 9 data 8 da1v data 8 data 8 data 8 data 8 data 8 data 7 r/w data 7 data 7 data 7 data 7 data 7 data 6 res data 6 data 6 data 6 data 6 data 6 data 5 res data 5 data 5 data 5 data 5 data 5 data 4 ia4 data 4 data 4 data 4 data 4 data 4 data 3 ia3 data 3 data 3 data 3 data 3 data 3 data 2 ia2 data 2 data 2 data 2 data 2 data 2 data 1 ia1 data 1 data 1 data 1 data 1 data 1 data 0 ia0 data 0 data 0 data 0 data 0 data 0 output slot 0 or 16 1 or 17 2 or 18 3 or 19 4 or 20 5 or 21 data 15 res data 15 data 15 data 15 res res data 14 res data 14 data 14 data 14 res res data 13 res data 13 data 13 data 13 res res data 12 res data 12 data 12 data 12 res res data 11 res data 11 data 11 data 11 res res data 10 res data 10 data 10 data 10 res res data 9 adrv data 9 data 9 data 9 res res data 8 adlv data 8 data 8 data 8 res res data 7 res data 7 data 7 data 7 res res data 6 res data 6 data 6 data 6 res res data 5 res data 5 data 5 data 5 res res data 4 res data 4 data 4 data 4 res res data 3 res data 3 data 3 data 3 res res data 2 res data 2 data 2 data 2 res res data 1 da2rq data 1 data 1 data 1 res res data 0 da1rq data 0 data 0 data 0 res res figure 14. ad1843 tdm time slot bit assignment map
rev. 0 C52C ad1843 figure 15. ad1843 control register map register readback (on the following frame) matches the data written. writes to these control registers will fail until the ad1843 internal clocks are stabilized. if the system has been designed for 16 slots per frame, it is suggested that this first polling write/readback is to control register address 26, with frs (bit 6) set to 1 and frst (bit 5) set to 1, so that the a d1843 will support the 16 slot per frame mode immediately. 4. put the conversion resources into standby . with the ex- ception of the serial interface, the ad1843 is still completely powered down before this step. the pdni bit (control reg- ister address 28, bit 15) should now be reset to 0 to ini- tiate the process of taking the ad1843 conversion resources out of power down and into standby. this requires approxi- mately 470 ms. an autocalibration cycle always occurs (automatically, without user programming or intervention) following the deassertion of the reset pin or the pwrdwn pin (i.e., hardware reset or hardware power down), adding another approximately 4 ms, for a total of approximately 474 ms. subsequent software power-down cycles, programmed using control register bits, do not ordi- narily require additional calibration cycles. the original cali- bration information is retained during the power down sequence. however, the user can force an autocalibration to occur following a software power down by setting acen (control register address 28, bit 14) to 1. the host cpu or dsp can poll the pdno bit (control register address 0, bit 14) to determine when the conversion resources are out of power down. alternatively, the host cpu or dsp can write to control register address 27 (using a data pattern different from the initial reset default), and proceed when the control reg address 0 address 1 address 2 address 3 address 4 address 5 address 6 address 7 address 8 address 9 address 10 address 11 address 12 address 13 address 14 address 15 address 16 address 17 address 18 address 19 address 20 address 21 address 22 address 23 address 24 address 25 address 26 address 27 address 28 address 29 address 30 address 31 data 15 init res lss2 ld1mm lx1mm lx2mm lx3mm lmcmm mnmm lda1gm lda2gm lda1am lda2am lad1mm lad2mm res c1ref c1c15 res c2ref c2c15 res c3ref c3c15 res drsflt da2sm dfree pdni res res res data 14 pdno res lss1 res res res res res res res res res res res res res c1vid c1c14 res c2vid c2c14 res c3vid c3c14 res damix da1sm res acen res res res data 13 res res lss0 res res res res res res lda1g5 lda2g5 lda1a5 lda2a5 lad1m5 lad2m5 res c1pllg c1c13 res c2pllg c2c13 res c3pllg c3c13 res res res res c3en res res res data 12 res res lmge ld1m4 lx1m4 lx2m4 lx3m4 lmcm4 mnm4 lda1g4 lda2g4 lda1a4 lda2a4 lad1m4 lad2m4 res c1p200 c1c12 res c2p200 c2c12 res c3p200 c3c12 res res res ddmen c2en res res res data 11 res res lig3 ld1m3 lx1m3 lx2m3 lx3m3 lmcm3 mnm3 lda1g3 lda2g3 lda1a3 lda2a3 lad1m3 lad2m3 da2c1 c1x8/7 c1c11 res c2x8/7 c2c11 res c3x8/7 c3c11 res res da2f1 res c1en res res res data 10 res res lig2 ld1m2 lx1m2 lx2m2 lx3m2 lmcm2 mnm2 lda1g2 lda2g2 lda1a2 lda2a2 lad1m2 lad2m2 da2c0 c1c128 c1c10 res c2c128 c2c10 res c3c128 c3c10 res res da2f0 res enclko res res res data 9 res su2 lig1 ld1m1 lx1m1 lx2m1 lx3m1 lmcm1 mnm1 lda1g1 lda2g1 lda1a1 lda2a1 lad1m1 lad2m1 da1c1 res c1c9 res res c2c9 res res c3c9 res da2flt da1f1 da2en xctl1 res res res data 8 res su1 lig0 ld1m0 lx1m0 lx2m0 lx3m0 lmcm0 mnm0 lda1g0 lda2g0 lda1a0 lda2a0 lad1m0 lad2m0 da1c0 res c1c8 c1pd res c2c8 c2pd res c3c8 c3pd da1flt da1f0 da1en xctl0 res res res data 7 res res rss2 rd1mm rx1mm rx2mm rx3mm rmcmm allmm rda1gm rda2gm rda1am rda2am rad1mm rad2mm res c1m7 c1c7 c1p7 c2m7 c2c7 c2p7 c3m7 c3c7 c3p7 daadr1 scf anaen encv3 res res res data 6 res res rss1 res res res res res mnom res res res res res res res c1m6 c1c6 c1p6 c2m6 c2c6 c2p6 c3m6 c3c6 c3p6 daadr0 frs hpen enbt3 res res res data 5 res res rss0 res res res res res hpom rda1g5 rda2g5 rda1a5 rda2a5 rad1m5 rad2m5 res c1m5 c1c5 c1p5 c2m5 c2c5 c2p5 c3m5 c3c5 c3p5 daadl1 frst res encv2 res res res data 4 res res rmge rd1m4 rx1m4 rx2m4 rx3m4 rmcm4 hpos rda1g4 rda2g4 rda1a4 rda2a4 rad1m4 rad2m4 res c1m4 c1c4 c1p4 c2m4 c2c4 c2p4 c3m4 c3c4 c3p4 daadl0 adtlk aamen enbt2 res res res data 3 id3 ovr1 rig3 rd1m3 rx1m3 rx2m3 rx3m3 rmcm3 summ rda1g3 rda2g3 rda1a3 rda2a3 rad1m3 rad2m3 adrc1 c1m3 c1c3 c1p3 c2m3 c2c3 c2p3 c3m3 c3c3 c3p3 res adrf1 res encv1 res res res data 2 id2 ovr0 rig2 rd1m2 rx1m2 rx2m2 rx3m2 rmcm2 res rda1g2 rda2g2 rda1a2 rda2a2 rad1m2 rad2m2 adrc0 c1m2 c1c2 c1p2 c2m2 c2c2 c2p2 c3m2 c3c2 c3p2 res adrf0 res enbt1 res res res data 1 id1 ovl1 rig1 rd1m1 rx1m1 rx2m1 rx3m1 rmcm1 dac2t rda1g1 rda2g1 rda1a1 rda2a1 rad1m1 rad2m1 adlc1 c1m1 c1c1 c1p1 c2m1 c2c1 c2p1 c3m1 c3c1 c3p1 adrflt adlf1 adren linrsd res res res data 0 id0 ovl0 rig0 rd1m0 rx1m0 rx2m0 rx3m0 rmcm0 dac1t rda1g0 rda2g0 rda1a0 rda2a0 rad1m0 rad2m0 adlc0 c1m0 c1c0 c1p0 c2m0 c2c0 c2p0 c3m0 c3c0 c3p0 adlflt adlf0 adlen linlsd res res res figure 16 shows an annotated version of the ad1843 detailed block diagram, indicating the specific control register bit fields which configures the various features of the soundcomm codec. start-up sequence the following paragraphs describe a typical, generic start-up sequence, for the purpose of helping hardware, systems and software driver engineers understand some of the considerations involved in bringing up a system which includes the ad1843. note that it does not exhaustively outline all of the flexible con- figurations and features available in the ad1843. 1. system power supplies stabilize. ideally, the ad1843 reset pin should be held lo during this period. if reset is not asserted, the ad1843 will come up in an un- known state. 2. assert the reset signal (plcc pin 52, tqfp pin 65). once the power supplies have stabilized, the ad1843 re- set must be asserted lo for at least 100 ns with bm (plcc pin 10, tqfp pin 12) tied to the appropriate level to establish whether the codec is serial bus master or slave. 3. deassert the reset signal and enter a wait period to allow the ad1843 internal clocks and the external crys- tal oscillator to stabilize . the wait period duration will be typically 400 m s to 800 m s after reset is deasserted, but sig- nificantly longer time may be necessary depending on xtali and xtalo pin parasitics. if the ad1843 is serial bus master, the host cpu or dsp can poll the init bit (control register address 0, bit 15) to determine when the ad1843 is ready to proceed. alternatively, the host cpu or dsp can write to control register 16 through 24, or 26 or 28 (using a data pattern different from the initial reset default for that control register), and proceed when the control
ad1843 rev. 0 C53C figure 16. ad1843 annotated detailed block diagram with control register address and data values pdmnft 8 9 gndd v dd cmout v ref ? clock generation clkout xtali xtalo sync3 sync2 sync1 conv3 conv2 conv1 bit3 bit2 bit1 hpoutr hpoutc hpoutl lout2lp lout2ln lout2rp lout2rn 2 left right 20db left right 3 v cc adc dac1 dac2 d i g i t a l i n t e r f a c e lout1l lout1r aux3l aux2r aux1l aux2l aux1r linrp linrn micr micl gn/at = gain/attenuation driver hpen[27] control and status registers ad1843 aafiltl aafiltr suml sumr linlp linln mout reset pwrdwn ? ? ? ? ? ? ? ? ? mute mnom[8] mute ? ? mute damix[25] ? ? ? ? ? aux3r min ?d adc adlen[27] adren[27] pga lig[2] rig[2] ?a law sclk sdfs sdi sdo bm cs tso tsi xctl [1:0] voltage reference rmge[2] lmge[2] s e l e c t o r (0) (1) (2) (3) (4) (5) (6) (7) rss[2] lss[2] (2) (3) daadl[25] daadr[25] (0) daadl[25] daadr[25] selector selector ?a law ?a law [ ] indicates control register address ( ) indicates control register data mute hpom[8] mute summ[8] mute lda2gm[10] rda2gm[10] gn/at lda2g[10] rda2g[10] ?d dac da2en[27] gn/at ld2m[3] rd2m[3] mute ld2mm[3] rd2mm[3] allmm[8] gn/at lda1g[9] rda1g[9] mute lda1gm[9] rda1gm[9] ?d dac da1en[27] left right mute mnmm[8] allmm[8] mute lx3mm[6] rx3mm[6] allmm[8] gn/at lx3m[6] rx3m[6] gn/at lx2m[5] rx2m[5] gn/at lx1m[4] rx1m[4] gn/at lmcm[7] rmcm[7] gn/at mnm[8] mute lx2mm[5] rx2mm[5] allmm[8] mute lx1mm[4] rx1mm[4] allmm[8] mute lmcmm[7] rmcmm[7] allmm[8] left right mute lda1am[11] rda1am[11] attn lda1a[11] rda1a[11] mute lda2am[12] rda2am[12] attn lda2a[12] rda2a[12] fifo fifo ? attn lad2m[14] rad2m[14] filtl filtr 4 gnda mute damix[25] mute lad2mm[14] rad2mm[14] attn lad1m[13] rad1m[13] mute lad1mm[13] rad1mm[13]
rev. 0 C54C ad1843 reset = lo pwrdwn = hi, reset = hi pwrdwn = lo pwrdwn = hi acen = 1 acen = 0 pdni = 1 or pwrdwn = lo pdni = 1 pdni = 0 conversion resources power down init = 0 pdno = 1 all down init = 1 pdno =1 initializing init = 1 pdno = 1 time: 800?? autocal init = 0 pdno = 1 time: < 4ms stby stby en stby en stby en en conversion resources powered up init = 0 pdno = 0 adcl adcr dac1 dac2 conversion resources entering power down init = 0 pdno = 1 time: < 5ms pwrdwn = lo pdni = 0 conversion resources exiting power down init = 0 pdno = 1 time: < 470ms interface: down clock generators: down conversions: down interface: reads only if slave, down if master clock generators: down conversions: down interface: alive clock generators: enableable (control register address 28) conversions: down interface: alive clock generators: enableable (control register address 28) conversions: down interface: alive clock generators: enableable (control register address 28) conversions: enableable (control register address 27) none none 16-24, 26, 28 potentially all, depending on enabled resources (see text) 16-24, 26, 28 ad1843 state activity level writable registers * = state = transmission (will not be interrupted once started unless reset = lo) * registers are always readable, unless reset = lo figure 17. ad1843 start-up sequence state diagram control register readback (on the following frame) matches the data written. writes to control register address 27 will fail unless the conversion resources are out of power down. 5. power up the clock generators and enable clock output pins. when the pdni bit (control register address 28, bit 15) is reset to 0 in step 4 above, the c3en, c2en, c1en, conv[3:1], and bit[3:1] bits, which are also lo- cated in control register address 28, may be set to 1 to power up any clock generators which will be used, and to enable any conversion clock and bit clock outputs needed. while waiting for the conversion resources to enter standby, it is convenient to write to the clock generator control regis- ters (16 through 24) to select sample rates, bit clock frequen- cies, external/internal sample rate lock mode, etc. this is also a convenient time to write to control register address 26 (if it was not already written in step 3) to select the serial interface configuration such as data format, sclk fre- quency, etc. 6. configure conversion resources while they are in standby. once in standby, the conversion resources may be configured. they are forced to their defaults whenever pow- ered down. control register address 2 may be written to se- lect between adc sources and gain levels. control register address 15 may be written to select the assignments of sample clock sources to conversion resource. control regis- ter addresses 13 and 14 may be written to enable digital data flow paths which mix adc output back into the dac input. control register address 25 may be written to select conver- sion resource filter mode (audio, modem or resampler), and to select digital data flow paths which may be used to achieve digital-to-digital sample rate conversion and dac-to-dac digital mixing. note that, with the exception of filter mode and dac-to-dac digital mixing, all of these configuration selections may also be changed while a conversion resource is out of standby and enabled. 7. enable conversion resources . conversion resources are taken out of standby and enabled by writes to control regis- ter address 27. control register address 27 contains indi- vidual enable bits for the conversion resources which are grouped as follows: dac1 (stereo pair); dac2 (stereo pair); adc left channel; adc right channel; dac2 to dac1 mix; and adc input to dac1 output mix. if a dac is enabled, it will begin requesting playback samples on the next tdm frame (i.e., the da2rq and/or the da1rq flags in the status register will be set to 1). conversions will not actually begin until the first rising edge of the conversion clock (assigned to the dac) after the sixth rising edge of the frame sync (sdfs) signal. this delay al- lows the four word deep stereo fifos to be filled before con- versions begin. (six frames are required to fill a four word deep fifo because on the first frame, the dac playback is enabled; on the second frame, the da2rq and/or da1rq flags are set to 1; then four more frames are required to ac- tually fill the fifo.) if an adc is enabled, sampling of the analog input will be- gin on the first rising edge of the conversion clock (assigned to the adc) after the sixth rising edge of the frame sync (sdfs) signal. the data valid flags (adrv and/or adlv) in the status word will be set to a 1, indicating a sample is being transmitted shortly thereafter. the exact tdm frame
ad1843 rev. 0 C55C in which the valid flag is first asserted, and continues to be asserted in future frames, depends on how the ad1843 is configured, i.e., the sample rates selected, the frame size selected, and how the shared digital resources within the ad1843 are internally allocated to process all conversion channels. note that because of this, two adcs which are enabled during the same frame and share the same sample rate clock, will not necessarily transmit data during the same frames thereafter, even though their analog inputs will always be sampled at the same instant in time. the two adcs will transmit data during the same frame if adtlk (control register address 26, bit 4) is set hi. see the description of the adtlk bit for restrictions. if mixing is enabled, either a mix from dac2 to dac1 or a mix from an adc input to a dac output, then the mix will begin during the same tdm frame that it is enabled. 8. configure conversion resources while they are enabled . once enabled, dac output gain/attenuation levels may be changed using control register addresses 3 through 10. these registers are forced to their default of full attenuation whenever the resource they control is either in standby or completely powered down. note that while a conversion re- source is enabled, all configuration selections related to it may be changed except those outlined in step 6. the ad1843 start-up sequence state diagram is shown in figure 17. applications circuits the ad1843 soundcomm codec has been designed to require a minimum of external circuitry. the recommended circuits are shown in figures 18 through 34. analog devices estimates that the total cost of all the components shown in these figures (in- cluding the crystal but not including the daa) to be less than $6 in 10,000 piece quantities. industry-standard compact disc line-levels are 2 v rms cen- tered around analog ground. (for other audio equipment, line level is much more loosely defined.) the ad1843 soundcomm codec is a +5 v analog supply powered device. nominal line level voltage swings for the ad1843 are defined to be 1 v rms (adrflt and adlflt = 0) for a sine wave adc input and 0.707 v rms for a sine wave dac output (da2flt = 0). thus, 2 v rms input analog signals must be attenuated and either cen- tered around the reference voltage intermediate between 0 v and +5 v or ac-coupled. the cmout pin will be at this inter- mediate voltage, nominally 2.25 v. it has limited drive but can be used as a voltage datum to an op amp input. cmout load- ing should be minimized to limit any audible clicks and pops. note that dc-coupled inputs are not recommended, as they pro- vide no performance benefits with the ad1843 architecture. furthermore, dc offsets on inputs create the potential for clicks when changing the input mix gain/attenuation/mute. the reset pin must be asserted at or shortly after power up in order to initialize the ad1843 control registers to their default states. if the ad1843 will not be used immediately, and if it is desired to utilize the mono input to mono output feedthrough feature, it is essential that the mono output is ac-coupled to pre- vent audible pops and clicks. it is recommended that all outputs are ac coupled, as standing current in output loads or in the voltage reference will contribute to audible pops and clicks. a circuit for 2 v rms line-level inputs is shown in figure 18. note that this is a divide-by-two resistive divider, and that the line input is being used in a single-ended configuration. the 1 m f ac-coupling capacitor may be of any type (tantalum is a popular choice). linlp 5.1k 1? 5.1k linrp 5.1k 1? 5.1k 1nf 1nf figure 18. ad1843 2 v rms single-ended line-level input circuit if line-level inputs are already at the 1 v rms levels expected by the ad1843, the circuit shown in figure 19 below should be used. note that when single-ended line inputs are desired, only the linrp and linlp pins are used. do not use the linrn or linln pins if the line input is single-ended. linlp 1? 1? linrp figure 19. ad1843 1 v rms single-ended line-level input circuit the three auxiliary inputs, the sum inputs, and the mono in- put of the ad1843 present an input impedance which is lower than the stereo line input. the circuit shown in figure 20 should be used with 2 v rms swings on these inputs. with 1 v rms swings, the circuit in figure 19 above should be used. 4.1k 1? 3.3k aux1l aux2l aux3l suml 1nf 4.1k 1? 3.3k 1nf min 4.1k 1? 3.3k aux1r aux2r aux3r sumr 1nf figure 20. ad1843 2 v rms aux, sum, and mono input circuits for optimal performance, the ad1843 includes provision for a differential configuration of the lin inputs. figure 21 illus- trates a simple single-ended to differential converter. for the best noise immunity, the circuitry to the left of the dotted line should be located as close to the driving signal source as possible. the 0.9k resistor and the 1000 pf capacitor are manda- tory when using the linrn and/or the linln inputs. furthermore, the 0.9k resistor and the 1000 pf npo capacitor should be located as close to the ad1843 as possible. if the output range of the source does not match that of the ad1843 input, the op amp gain setting resistor values should be scaled to match the source voltage range to the ad1843 input voltage range. inexpensive feedback capacitors can be added (as shown in the figure 21 with dotted lines) for additional noise filtering (6 db per octave). without the additional filter capacitor, there is a single antialiasing pole at approximately 160 khz. there is
rev. 0 C56C ad1843 room in the frequency domain for a second pole, especially in modem applications, where the signal bandwidth is 4.2 khz. a suggested capacitor value is 120 pf when using the 10k op amp feedback resistor as shown, for a C3 db point of approximately 133 khz. note that the combined effect of both filters causes a gain error of approximately 0.1% at 4.2 khz. 1? 0.9k linrn linln 10k 10k 1/2 ssm2135 10k source 10k 1/2 ssm2135 1? 1000pf npo linlp linrp close to source close to ad1843 1? cmbuf cmbuf figure 21. single-ended to differential converter the ad1843 codec contains an optional +20 db gain block to accommodate condenser microphones. particular system re- quirements will depend upon the characteristics of the intended microphone. figure 22 illustrates one example of how an elec- tret condenser mic requiring phantom power could be con- nected to the ad1843. cmout is shown buffered by an op amp; the current drawn from cmout should be as minimal as possible. note that if a battery-powered microphone is used, the buffer and r2s are not needed. the values of r1, r2, and c should be chosen in light of the mic characteristics and in- tended gain. typical values for these might be r1 = 20 k w , r2 = 2 k w , and c = 220 pf. figure 22 shows a voltage follower buffer on the cmout signal. the output of this buffer, cmbuf, can be used in any circuit which needs the common- mode bias point from the ad1843s on-chip voltage reference. the ad820 is a jfet input op amp whose very high imped- ance inputs present essentially no load on the cmout output from the ad1843. 1/2 ssm2135 or ad820 5k 1? cmout micl left electret condenser microphone input right electret condenser microphone input r1 0.33? c 1/2 ad820 r2 r2 1/2 ssm2135 or ad820 5k 1? micr r1 0.33? c cmbuf figure 22. ad1843 phantom-powered microphone input circuit connect unused analog inputs to cmbuf or leave them unconnected; do not connect unused analog inputs to either analog power or ground! figure 23 shows ac-coupled line outputs. the resistors are used to center the output signals around analog ground. if dc- coupling is desired, cmout could be used with op amps as mentioned above, if desired. lout1r lout2rp 47k 1? lout1l lout2lp 47k 1? figure 23. ad1843 (single-ended) line output connections for optimal performance, the ad1843 dac2 output has provi- sion for a differential configuration. a simple differential-to- single-ended conversion circuit is shown in figure 24. the noise rejection of this circuit is limited by the match of the r i and r f resistors. the resistors should be 1% tolerance compo- nents. the equation that determines the output voltage is as follows: v o = ( r f / r i )( v +) C ( r f / r i )( v C) for maximum noise immunity, this circuit should be located in close proximity to where v o is processed. another differential receiver option is the ssm2141, which provides better matching than a discrete implementation. 1/2 ssm2135 r f 1/2 ssm2135 lout2ln (v? 10pf v o r i r f 10pf lout2lp (v+) cmbuf r i 1/2 ssm2135 r f 1/2 ssm2135 lout2rn (v? 10pf v o r i r f 10pf lout2rp (v+) cmbuf r i figure 24. ad1843 differential-to-single-ended converter line output connections a circuit for headphone drive is illustrated in figure 25. the ad1843 headphone output is designed to drive loads of 32 ohms or smaller (i.e., higher impedance). if larger loads are used (e.g., 16 ohms or 8 ohms), the analog output will be dis- torted for large output signals because of current limiting. walkman-type headphone impedances are generally around 32 ohms. telephone handset impedances are typically 150 ohms. hpoutl hpoutr hpoutc figure 25. ad1843 headphone drive connections figure 26 shows an example circuit for the mono output (mout) from the ad1843. the op279 is a single +5 v supply op amp which can drive a small 8 ohm or 16 ohm speaker to modest levels.
ad1843 rev. 0 C57C mout cmbuf 10k 10k 1/2 op279 16 optional to pc speaker 470? figure 26. ad1843 mono output circuit figure 27 illustrates reference bypassing. v ref should only be connected to its bypass capacitors. the 10 m f capacitor should be tantalum, and the 0.1 m f capacitor should be ceramic. figure 27 shows an optional voltage follower buffer on the cmout signal. the output of this buffer, cmbuf, can be used in any circuit which needs the common-mode bias point from the ad1843s on-chip voltage reference. the ad820 is a jfet in- put op amp whose very high impedance inputs present essen- tially no load on the cmout output from the ad1843. v ref 10? 0.1? cmout 10? 0.1? cmbuf 1/2 ad820 optional buffer figure 27. ad1843 voltage reference bypassing figure 28 illustrates the signal-path filtering capacitors, filtl and filtr, connections to analog ground. the 1.0 m f capaci- tors required by the ad1843 can be of any type. 1.0? filtl 1.0? filtr figure 28. ad1843 external filter capacitor connections figure 29 illustrates the antialias filtering capacitors, aafiltl and aafiltr, connections to analog ground. the 1000 pf capacitors must be npo types. 1000pf npo aafiltl aafiltr 1000pf npo figure 29. ad1843 antialias filter capacitor connections the 24.576 mhz crystal shown in the crystal connection cir- cuitry of figure 30 should be fundamental-mode and parallel- tuned. note that using the exact data sheet frequency is not required and that external clock sources can be used in place of the ad1843s internal oscillator. if using an external clock source, apply it to the crystal input pin (xtali) while leaving the crystal output pin (xtalo) unconnected. attention 24.576 mhz xtali xtalo 20?4pf 20?4pf figure 30. ad1843 crystal connections should be paid to providing a low-jitter external input clock. ideally, there should be no greater than 5 ns rms of random (white) phase jitter to ensure that it does not significantly de- grade the snr of the ad1843. figure 31 and figure 32 show example circuits for a data access arrangement (daa). these circuits are shown as examples only; daa circuits are subject to regulatory approval since they connect directly to the public switched telephone network (pstn). 220 w ssi 73m9001 txa txa/ monsum tip ring oh/ rly2/ ag0 ag1 power/ rxa spk spk/ tran1 tran2 ri/ reout1 reout2 vcca vccd agnd dgnd 23 22 15 29 2 8 7 14 17 18 lo hi lo 6 21 12 11 1 30 19 9 13 16 10 20 +5va v cc 56k to ad1843 xctl0 ring indication to host interrupt 1 2 3 4 5 6 j1 rj11 single phone coil 220? to ad1843 linlp to ad1843 min lout2ln lout2lp v cc from ad1843 coil figure 31. silicon systems daa example circuit figure 32. cermetek daa example circuit xmit+ xmit rcv offhk rt tip ring v cc gnd cermetek ch1837 daa c3 100nf c2 100nf c1 100nf lout2lp lout2ln linlp xctl0 8 6 5 3 4 u3b 4 3 sn74hc14 u3a 2 1 sn74hc14 ad1843 c4 100nf dv dd 9 1 2 earth ground vr1 mov vr3 mov vr2 mov f1 current limit f2 current limit c6, 1nf 1.5kv c5, 1nf 1.5kv ring (red) tip (green) 6 5 4 3 2 1 j1 rj-11 jack u3d 8 9 sn74hc14 u3c 6 5 sn74hc14 dv dd r1 100k irq to dsp or host cpu interrupt +c7 4.7? tant (optional) notes: 1. vr1,2,3. over voltage and lightning protection mov sidactors. available as a single unit from teccor, part number p3203ab or p3203aa. 2. c5,6. emi/rfi suppression. high voltage disk ceramic capacitors, 1nf 1.5kv. 3. f1,2. current limiting devices. raychem poly fuse, part number tb600-45; little fuse type 251.250 or bussman type mcr1/4. for non ul applications, a 10 ohm, 1/8w carbon film resistor may be used. 4. l1,2. emi/rfi suppression. ferrite beads, fair-rite part number 2643666611 or 2943666661. 5. j1. fcc approved rj-11 jack. ref fcc public notice #42269 l1 inductor l2 inductor
rev. 0 C58C ad1843 good, standard engineering practices should be applied for power-supply decoupling. decoupling capacitors should be placed as close as possible to package pins. if a separate analog power supply is not available, we recommend the circuit shown in figure 33 for using a single +5 v supply. this circuitry should be as close to the supply pins as is practical. ferrite or suitable inductor 1.6 w v cc 0.1? 1? ferrite or suitable inductor 0.1? 1? v cc v cc 0.1? v dd 0.1? v dd 0.1? v dd 0.1? v dd 0.1? v dd 0.1? v dd 0.1? v dd 0.1? v dd 0.1? v dd +5v supply 0.1? 1? figure 33. ad1843 recommended power supply bypassing figure 34 illustrates an optional circuit for generating the +5 v analog supply for the ad1843. the +12 v rail available on the isa bus can be regulated down to +5 v using an ordinary 7805 three-terminal regulator and the bypassing and decoupling capacitors shown. the digital supply should still be decoupled and bypassed as shown above in figure 33. 10? 220nf 10? 220nf 220nf +12v pc 0 v ferrite bead in out gnd 7805 agnd dgnd +5va 1? figure 34. ad1843 optional analog supply regulation analog devices recommends a split ground plane as shown in figure 35a (pqfp) and figure 35b (tqfp). the analog plane and the digital plane are connected directly under the ad1843. splitting the ground plane directly under the soundcomm codec is optimal because analog pins will be located above the analog ground plane and digital pins will be located directly above the digital ground plane for the best isolation. the digital ground and analog ground should be tied together in the vicin- ity of the ad1843. other schemes may also yield satisfactory results. if the split ground plane recommended here is not pos- sible, the ad1843 should be entirely over the analog ground plane. analog ground plane digital ground plane ad1843 50 49 10 11 pqfp figure 35a. ad1843 recommended pqfp ground plane analog ground plane digital ground plane ad1843 62 61 13 14 tqfp figure 35b. ad1843 recommended tqfp ground plane
C59C khz 0 ?0 04 81216 ?0 ?0 ?00 ?10 ?0 ?0 ?0 ?0 db ?0 ?0 20 24 28 32 36 40 44 a. adc audio full khz 0 ?0 064 32 160 224 288 ?0 ?40 ?60 ?80 ?0 ?0 ?20 ?00 db 96 128 192 256 320 352 b. dac audio 0C384 khz khz 0.0000 02 db 4 6 8 1012141618 ?.0050 ?.0100 c. adc audio passband khz 0.008 0.006 018 4 8 12 16 0.000 0.004 0.002 db ?.002 ?.004 ?.010 ?.006 ?.008 ?.012 ?.018 ?.014 ?.016 2 6 10 14 d. dac audio passband figure 36. ad1843 audio mode frequency response plots (full-scale line-level inputs, 0 db gain) based on 48 khz sample rate; other sample rates will show slightly different behavior. see the specifications section for performance limits. the dac plots do not reflect the additional benefi- cial roll-off of the ad1843 analog filters. out-of-band images will be attenuated by an additional 31.4 db at 100 khz. rev. 0 frequency response plotsCad1843
rev. 0 C60C ad1843 khz 0 ?0 0 0.8 1.6 ?0 ?0 ?00 ?10 ?0 ?0 ?0 ?0 db ?0 ?0 2.4 3.2 6.4 4.0 4.8 5.6 7.2 a. adc modem full khz 0 ?0 0 9.6 4.8 24.0 33.6 43.2 14.4 19.2 28.8 38.4 48.0 52.8 ?0 ?40 ?60 ?80 ?0 ?0 ?20 ?00 db b. dac modem 0C57.6 khz khz ?.020 0.2 0.6 3.0 1.0 2.2 1.4 2.6 1.8 db 0.040 0.020 0.000 ?.040 ?.060 ?.080 ?.100 ?.120 ?.140 ?.160 ?.180 ?.200 ?.220 ?.240 c. adc modem passband khz 0.040 0.000 ?.040 ?.080 ?.120 ?.160 ?.200 ?.240 0.080 0.2 0.6 3.0 1.0 2.2 1.4 2.6 1.8 db d. dac modem passband figure 37. ad1843 m odem mode frequency res ponse plots (full-scale line-level inputs, 0 db gain) based on 7.2 khz sample rate; other sample rates will show slightly different behavior. see the specifications section for performance limits. the dac plots do not reflect the additional beneficial roll-off of the ad1843 analog filters. out-of-band images will be attenuated by an additional 31.4 db at 100 khz.
ad1843 rev. 0 C61C khz 0 ?0 048 24 12 20 16 40 28 36 32 44 48 ?0 ?0 ?00 ?10 ?0 ?0 ?0 ?0 db ?0 ?0 a. adc resampler full khz 0 ?0 064 32 160 224 288 96 128 192 256 320 352 ?0 ?40 ?60 ?80 ?0 ?0 ?20 ?00 db b. dac resampler 0C384 khz khz ?.005 0.005 0.000 ?.010 ?.015 ?.020 ?.025 ?.030 ?.035 024 12 814 10 6 18 db c. adc resampler passband khz 0.004 0.000 ?.004 ?.008 ?.012 ?.016 ?.020 ?.036 0.008 24 18 614 10 16 12 8 0 db ?.024 ?.028 ?.032 d. dac resampler passband figure 38. ad1843 resampler mode frequency response plots (full-scale line-level inputs, 0 db gain) based on 48 khz sample rate; other sample rates will show slightly different behavior. see the specifica- tions section for performance limits. the dac plots do not reflect the additional beneficial roll-off of the ad1843 analog filters. out-of-band images will be attenuated by an additional 31.4 db at 100 khz.
rev. 0 C62C ad1843 table of contents features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 general product description . . . . . . . . . . . . . 1 specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 ordering information . . . . . . . . . . . . . . . . . . . . 6 pin description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 pin configurations . . . . . . . . . . . . . . . . . . . . . . . 7, 8 detailed functional block diagram . . . . . 12 detailed product description . . . . . . . . . . . . 13 functional description . . . . . . . . . . . . . . . . . . . 13 serial interface . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 serial interface input . . . . . . . . . . . . . . . . . . . . 22 serial interface output . . . . . . . . . . . . . . . . . . 24 control registers . . . . . . . . . . . . . . . . . . . . . . . . . 25 bit and register maps . . . . . . . . . . . . . . . . . . . . . . 51 start-up sequence . . . . . . . . . . . . . . . . . . . . . . . . . 52 applications circuits . . . . . . . . . . . . . . . . . . . . . 55 frequency response plots . . . . . . . . . . . . . . . . 59 outline dimensions dimensions shown in inches and (mm) st-100 100-terminal plastic thin quad flatpack (tqfp) seating plane 0.026 (0.65) 0.014 (0.35) 0.061 (1.55) 0.049 (1.25) 12 typ 0.007 (0.177) 0.003 (0.077) 6 4 0 ?10 0.004 (0.102) max lead coplanarity top view (pins down) 1 25 26 51 50 75 100 76 0.012 (0.20) 0.004 (0.10) 0.640 (16.25) 0.620 (15.75) sq 0.555 (14.10) 0.547 (13.90) sq 0.020 (0.50) bsc s-80 80-terminal plastic quad flatpack (pqfp) seating plane 0.096 (2.45) max 0.037 (0.95) 0.026 (0.65) 0.004 (0.10) max 0.083 (2.10) 0.075 (1.90) 0.010 (0.25) min 0.015 (0.38) 0.009 (0.22) 0.690 (17.45) 0.667 (16.95) 0.555 (14.10) 0.547 (13.90) 0.555 (14.10) 0.547 (13.90) 0.690 (17.45 0.667 (16.95) 1 20 21 41 40 60 61 80 0.486 (12.35) typ 0.486 (12.35) typ top view (pins down) 0.029 (0.73) 0.023 (0.57)
C63C
C64C printed in u.s.a. c2097C6C1/96

▲Up To Search▲

Price & Availability of AD1843JS

	To Download AD1843JS Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .