T H AT Corporation FEATURES * High Performance VCA, RMS-Level Detector, and three 0pamps in one package Wide Dynamic Range: >105 dB Low THD: <0.09% Low Power: 7 mA typ. Surface-Mount Package 5 VDC Operation Low-voltage, Low-power Analog Engine(R) Dynamics Processor THAT 4311 APPLICATIONS * * * * * * Wireless microphone systems Wireless in-ear monitors Compressors and Limiters Gates De-Essers Duckers * * * * * Description The THAT 4311 Low Power Dynamics Processor combines in a single IC all the active circuitry needed to construct a wide range of dynamics processors. The 4311 includes a high performance, voltage controlled amplifier, a log responding RMS-level sensor and three opamps, one of which is dedicated to the VCA, while the other two may be used for the signal path or control voltage processing. The exponentially-controlled VCA provides two opposing-polarity, voltage sensitive control ports. Dynamic range exceeds 105 dB, and THD is typically 0.09% at 0dB gain. The RMS detector provides accurate RMS to DC conversion over an 80 dB dynamic range. Though originally designed for use in microphone noise reduction systems, the 4311 is a useful building block in a number of analog signal processing applications. The combination of exponential VCA gain control and logarithmic detector response - "decibel-linear" response simplifies the mathematics of designing the control paths of dynamics processors, making it easy to develop audio compressors, limiters, gates, expanders, de-essers, duckers, and the like. The high level of integration ensures excellent temperature tracking between the VCA and the detector, while minimizing the external parts count. Pin Name RMS IN 18 17 16 15 14 13 12 11 VCC 19 ECEC+ SYM OUT DMP20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 IT (ITIME) OA2 -IN RMS OUT CT (CTIME) CLIP OA2 OUT CAP OA1 20 IN VCA OA3 VREF VREF THAT4311 1 IN IT VREF VEE VCC OA3 OUT VCA OUT SYM EC+ RMS CT OUT OA2 VEE 2 3 4 5 6 7 8 9 10 ECVCA IN OA1 OUT OA1 -IN OA1 +IN Figure 1. THAT 4311 equivalent block diagram Table 1. THAT 4311 pin assignments THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Page 2 Low-voltage, Analog Engine(R) Dynamics Processor Preliminary Information SPECIFICATIONS 1 Absolute Maximum Ratings (T A = 25C) Positive Supply Voltage (VCC) Operating Temperature Range (TOP) Max DEC EC+ - (EC-) +15 V -20 to +70C 1V Power Dissipation (PD) (TA = 75C) Storage Temperature Range (TST) 700 mW -40 to +125C Recommended Operating Conditions Parameter Positive Supply Voltage Symbol VCC Conditions Min +5 Typ Max +15 Units V Electrical Characteristics 2 Parameter Supply Current Reference Voltage Symbol ICC VREF Conditions No signal; VCC=+7 VDC Min -- 1.8 Typ 7.0 1.95 Max 9.0 2.1 Units mA V Encode and Decode - Companding Noise Reduction ( VCC = +7V encoder, +15V decoder) Encode Level Match Encode Gain Accuracy LMe Encode mode; f = 1kHz Encode mode, f = 1kHz Vin = LMe + 10dB Vin = LMe - 40dB Decode mode; f = 1kHz Decode mode; f=1kHz Vin = LMd + 5dB Vin = LMd - 20dB Encode mode; THD = 3%; f = 1kHz Decode mode; THD = 3%; f = 1kHz End-to-end; Vin = LMe; f = 1kHz -25.3 -23.0 -20.7 dBV GAe1 GAe2 LMd +3.5 -23 -18.3 +5 -20 -16.0 +6.5 -17 -13.7 dB dB dB Decode Level Match Decode Gain Accuracy GAd1 GAd2 Vime Vomd THDtrim +8.5 -43 3 10.7 -- +10 -40 5 13.7 0.025 +11.5 -37 -- -- -- dB dB dBV dBV % Max Input Voltage Max Output Voltage Total Harmonic Distortion (with trim) Total Harmonic Distortion (no trim) Output Noise THDnotrim End-to-end; Vin = LMd; f = 1kHz -- 0.15 0.7 % Vnod End-to-end ; Vin = short; A-weighted -- 7 -- Vrms 1. All specifications are subject to change without notice. 2. Unless otherwise noted, TA=25C, test circuit as shown in Fig 2. THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Rev. 08/30/01 Preliminary Information Page 3 Electrical Characteristics (con't) Parameter Symbol Conditions Min Typ Max Units Op amp OA1 Offset Voltage Equivalent Input Noise Total Harmonic Distortion Open Loop Gain Gain Bandwidth Product Slew Rate VIO VnOA1 THDOA1 AVO-OA1 GBWOA1 SROA1 RL = 2kW A-weighted 1kHz, AV=1; RL = 10kW RL = 10kW at 50kHz -- --- -- -- 0.5 6.5 0.0007 115 5 2 6 10 0.003 --- -- mV nV Hz % Op amp OA2 Offset Voltage Equivalent Input Noise Total Harmonic Distortion Open Loop Gain Gain Bandwidth Product Slew Rate VIO VnOA1 THDOA1 AVO-OA1 GBWOA1 SROA1 RL = 2kW A-weighted 1kHz, AV=1; RL = 10kW RL = 10kW at 50kHz -- --- -- -- 0.5 7.5 0.0007 110 5 2 6 12 0.003 --- -- mV nV Hz % +20dB V+ Input XLR1 XLR-F 2 3 1 Sym R7 50k R8 cw VVREF 5423 1 C4 300k 5% R9 51R 5% C5 47p NPO R6 + C15 1000u SW1F +40dB +60dB R24 R23 R22 2k80 30R1 280R V+ 11 6 12 5 SW1E R19 C16 + 47u R18 10k0 100k + R27 10k0 R14 31k6 C17 47p 19 18 20 OA1 R25 Output XLR2 XLR-M 1 2 C18 47u 100R U1D 5% THAT4311 3 R1 100k R2 10k0 15 14 20k0 R5 17 EC+ SYM U1A R10 OUT 13 IN OA3 EC20k0 12 47u 100R 5% VCA 16 THAT4311 R3 R4 VREF SW1A 1 3 245 + C19 1000u SW1C 16 1 R17 31k6 C8 47p +40dB R16 280R 3 OA2 R28 100R 5% External Control Input 3 2 1 CN1 TP1 RMS Input C6 10k0 C12 3u3 R20 10k0 C9 3u3 1k33 2 3 18 6 CONTROL-VOLTAGE U1 OP-27 0dB SW1B 62 7 15 R21 10k0 R29 1k33 VREF Bypass Capacitors TP2 RMS Output Power Input 3 2 1 CN2 + C10 22u + C13 22u V+ R15 31k6 U1C VREF THAT4311 V+ C1 10u + 11 10 R11 1 U1B R26 IN OUT 4 + 23k2 2 RMS 5 100R 47u Iset TC 5% THAT4311 + C7 10u R12 261k VVREF D3 1N4004 C11 100n (U1) C14 100n (U1) V- U1E VCC VEE VREF CAP 9 8 VREF VREF RMS Output 3 2 1 CN1 D4 1N4004 THAT4311 C2 10u + + C3 22u Fig 2. THAT 4311 test circuit THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Page 4 Low-voltage Analog Engine(R) Dynamics Processor Preliminary Information Representative Data (Stand-alone) Fig 3. VCA Gain vs. Control Voltage (Ec-) at 25C Fig 4. VCA 1kHz THD+Noise vs. Input, -15 dB Gain Fig 5. VCA 1kHz THD+Noise vs. Input, +15 dB Gain Fig 6. VCA 1kHz THD+Noise vs. Input, 0 dB Gain Fig 7. VCA THD vs. Frequency, 0 dB gain, 1 Vrms Input Fig 8. RMS Output vs. Input Level, 1 kHz & 10 kHz Fig 9. Departure from Ideal Detector Law vs. Level Fig 10. Detector Output vs. Frequency at Various Levels THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Rev. 08/30/01 Preliminary Information Page 5 Representative Data (Companding Noise Reduction) Fig 11. End-to-End Transfer Function, 1kHz Fig 12. End-to-End THD+N Fig 13. Encoder Transfer Function, 1kHz Fig 14. Encoder Frequency Response 20-20kHz +5 Vref V+ R1 200k 20 19 R9 51R R8 50k C5 R7 20k 22kHz 3 pole BW filter R10 6k19 C9 570p 6 Encoder C2 In + 3u3 R2 200k 15 14 13 U1D THAT4311 + _ 18 R3 C4 + 3u3 R5 15k 270p R6 EC+ SYM 17 IN VCA OUT EC- 30k _ R4 12 OA3 + 6k19 U1A THAT4311 R31 3_ 8k06 R13 2k C1 10n V+ Vref U1B 4 OUT IN Iset 1 2 OA2 4k32 + C10 Vref U1C 4n7 THAT4311 R11A 20k VR1 20k 10k optional R11B 7 Encoder Out 16 C6 + 10u R12 261k U1E C16 10u + 11 Vcc 10 Vee Vref Cap 9 8 Vref C7 10u 5 TC RMS Vref + THAT4311 THAT4311 + C3 10u + C8 22u Fig 15. THAT 4311 Noise Reduction Encoder Schematic THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Page 6 Low-voltage, Analog Engine(R) Dynamics Processor Preliminary Information +15 R8 150k R9 51R C6 + 10u R11 15k R12 261k U1B 1 2 IN OUT 4 5 R7 20k R2 56k R3 R1 6k04 R6 24k3 19 RMS TC Iset THAT4311 + C5 10u 8k87 C1 3n3 Decoder In V+ 11 10 C4 + 3u3 U1E Vcc Vee 15 14 13 R5 17 EC+ SYM IN VCA OUT EC24k3 16 9 8 20 C16 10u + Vref Cap _U1A 12 OA3 + THAT4311 Vref R4 7k5 _ Decoder Out 18 OA1 + U1D THAT4311 Vref THAT4311 + C2 10u + C3 22u + C7 10u Fig 16. THAT 4311 Noise Reduction Decoder Schematic Theory of Operation The THAT 4311 Analog Enginea Dynamics Processor combines THAT,s proven Voltage-Controlled Amplifier (VCA) and RMS-Level Detector designs with three opamps to produce a multipurpose dynamics processor useful in a variety of applications. For details of the theory of operation of the VCA and RMS Detector building blocks, the interested reader is referred to THAT Corporation's data sheets on the 218x Series VCAs and the 2252 RMS-Level Detector. Theory of the interconnection of exponentially-controlled VCAs and log-responding level detectors is covered in THAT Corporation's application note AN101, The Mathematics of Log-Based Dynamic Processors. Input signals are currents to the VCA IN pin. This pin is a virtual ground biased at VREF so in , normal operation an input voltage is converted to input current via an appropriately sized resistor (R5 in Fig 2, Page 3). Because the current associated with DC offsets relative to VREF present at the input pin and any DC offset in preceding stages will be modulated by gain changes (thereby becoming audible as thumps), the input pin is normally AC-coupled (C4 in Fig 2). The VCA output signal is also a current, inverted with respect to the input current. In normal operation, the output current is converted to a voltage via inverter OA3, where the ratio of the conversion is determined by the feedback resistor (R6, Fig 2) connected between OA3's output and its inverting input. The signal path through the VCA and OA3 is non-inverting. The gain of the VCA is controlled by the voltage applied between EC- and the combination of EC+ and SYM. Gain (in decibels) is proportional to EC-, provided that EC+ and SYM are at VREF The con. stant of proportionality is -6.1mV/dB (for 5V supplies) for the voltage at EC-, and 6.1mV/dB for the voltage at EC+, and SYM The VCA - in Brief The THAT 4311 VCA is based on THAT Corporation's highly successful complementary log/anti-log gain cell topology, as used in THAT's 218x and 215x-Series IC VCAs. The THAT 4311 is integrated using a fully complementary, BiFET process. The combination of FETs with high-quality, complementary bipolar transistors (NPNs and PNPS) allows additional flexibility in the design of the VCA over previous efforts. THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Rev. 08/30/01 Preliminary Information Page 7 R5 20k +5 R4 51k R3 51R C1 + Signal In C3 + 47u R6 10k R7 264k 1 IN RMS 5 2 Iset TC 13 15 14 EC+ SYM IN VCA OUT 17 EC16 R2 20k C2 47p U1A OA3 12 R1 Vref 20k Signal Out U1B RMS Out 4 OUT + C4 10u 47u THAT 4311 THAT 4311 +5 C6 + 10u C7 Control Port Drive Vref U1C 6 3 U1D 19 20 OA1 100n 9 11 Vcc Vref Vref 10 8 Cap Vee U1E 18 OA2 + C5 10u + C8 22u Vref 7 THAT 4311 Vref THAT 4311 THAT 4311 Fig 17. Circuit showing gain control at EC- As mentioned, for proper operation, the same voltage must be applied to EC+ and SYM, except for a small (2.5 mV) DC bias applied between these pins. This bias voltage adjusts for internal mismatches in the VCA gain cell which would otherwise cause small differences between the gain of positive and negative half-cycles of the signal. The voltage is usually applied via an external trim potentiometer (R7 in Fig 2), which is adjusted for minimum signal distortion at unity (zero dB) gain. The VCA may be controlled via EC-, as shown in Fig 17, or via the combination of EC+ and SYM. This connection is illustrated in Fig 18. Note that this latter figure shows only that portion of the circuitry needed to drive the positive VCA control port; circuitry associated with OA1, OA2 and the RMS detector has been omitted. While the 4311's VCA circuitry is very similar to that of the THAT 2180 Series VCAs, there are several important differences, as follows: 1. Supply current for the VCA is fixed internally. Approximately 500mA is available for the sum of input and output signal currents. 2. The signal current output of the VCA is internally connected to the inverting input of an on-chip opamp. In order to provide external feedback around this opamp, this node is brought out to a pin. 3. The input stage of the 4311 VCA uses integrated P-channel FETs rather than a bias-current corrected bipolar differential amplifier. Input bias currents have therefore been reduced. The RMS Detector - in Brief The THAT 4311's detector computes RMS level by rectifying input current signals, converting the rectified current to a logarithmic voltage, and applying that voltage to a log-domain filter. The output signal is a DC voltage proportional to the decibel-level of the RMS value of the input signal current. Some AC component (at twice the input frequency) remains superimposed on the DC output. The AC signal is attenuated by a log-domain filter, which constitutes a single-pole roll-off with cutoff determined by an external capacitor and a programmable DC current. As in the VCA, input signals are currents to the RMS IN pin. This input is a virtual ground biased at VREF so a resistor (R11 in Fig 2) is normally used to , convert input voltages to the desired current. The level detector is capable of accurately resolving signals well below 10mV (with a 10kW input resistor). However, if the detector is to accurately track such low-level signals, AC coupling is normally required. THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Page 8 Low-voltage, Analog Engine(R) Dynamics Processor Preliminary Information N/C N/C 1 2 U1B IN Iset OUT 4 5 RMS TC N/C N/C R5 50k +5 THAT 4311 Control Port Drive R4 51k R3 51R R2 20k C2 47p U1A OA3 12 Signal In C1 + 47u R1 20k 13 15 14 EC+ SYM 17 IN VCA OUT EC16 Signal Out THAT 4311 U1C 6 Vref C7 +5 U1E Vee 11 Vcc Vref 9 Cap 8 Vref U1D N/C N/C N/C 19 20 100n 10 Vref N/C + C5 10u + C8 22u 3 OA1 18 N/C THAT 4311 OA2 Vref 7 THAT 4311 THAT 4311 Fig 18. Circuit showing gain control at EC+ The log-domain filter cutoff frequency is usually placed well below the frequency range of interest. For an audio-band detector, a typical value would be 5Hz, or a 32ms time constant (t). The filter's time constant is determined by an external capacitor attached to the CT pin, and an internal current source (ITIME) connected to CT. The current source is programmed via the IT pin: current in IT is mirrored to ITIME with a gain of approximately one. The resulting time constant t is approximately equal to (0.026 CT) / IT. Note that, as a result of the mathematics of rms detection, the attack and release time constants are fixed in their relationship to each other. The DC output of the detector is scaled with the same constant of proportionality as the VCA gain control: 6.1mV/dB. The detector's zero dB reference (Iin0, the input current which causes zero volts output), is determined by IT as follows: Iin0=IT. The detector output stage is capable of sinking or sourcing l00mA. Differences between the 4311's RMS-Level Detector circuitry and that of the THAT 2252 RMS Detector are as follows: 1. The rectifier in the 4311 RMS Detector is internally balanced by design, and cannot be balanced via an external control. The 4311 will typically balance positive and negative halves of the input signal within 1.5%, but in extreme cases the mismatch may reach +20%. However, a 20% mismatch will not significantly increase ripple-induced distortion in dynamics processors over that caused by signal ripple alone. 2. The time constant of the 4311's RMS detector is determined by the combination of an external capacitor (connected to the CT pin) and an internal, programmable current source. The current source is equal to IT. Normally, a resistor is not connected directly to the CT pin on the 4311. 3. The zero dB reference point, or level match, is not adjustable via an external current source. However, as in the 2252, the level match is affected by the timing current, which, in this case, is drawn from the IT pin and mirrored internally to CT. 4. The input stage of the 4311 RMS detector uses integrated P-channel FETs rather than a bias-current corrected bipolar differential amplifier. Input bias currents are therefore negligible, improving performance at low signal levels. The Opamps - in Brief The three opamps in the 4311 are intended for general purpose applications. All are 5MHz opamps with slew rates of approximately 2V/ms. All use bipolar PNP input stages. However, the design of each is optimized for its expected use. Therefore, to get the THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Rev. 08/30/01 Preliminary Information Page 9 +5 R5 20k R4 33k R3 51R Vref 15 14 13 SYM EC+ 17 IN VCA OUT EC16 R16 R2 20k C2 47p U1A OA3 12 In R10 82k R9 +5 C7 R11 51k 1N4148 U1B 1 IN Iset OUT 4 5 6 C1 + 47u +5 R17 82k R1 20k Out THAT4311 4k99 Vref 10k Gain R18 50k C5 100n R14 R15 10k 19 20 Threshold R12 20k Vin0=-10dBu C3 + 47u R6 CR1 100n U1D OA1 18 R8 4k99 + C4 10u 3 28k7 2 R7 267k RMS TC OA2 7 CR2 1k43 R13 10k THAT4311 Vref THAT 4311 Compression 1N4148 U1C Ratio Vref THAT4311 +5 Vref U1E C6 11 + 10u Vcc Vee Vref Cap 9 10 Vref 8 THAT 4311 + C8 10u + C9 22u Fig 19. Simple compressor / limiter using the THAT 4311 most out of the 4311, it is useful to know the major differences among these opamps. OA3, being internally connected to the output of the VCA. is intended for current-to-voltage conversion. Its input noise performance, at 7.5nV / Hz, complements that of the VCA, adding negligible noise at unity gain. Its output section is capable of driving 1mA into a 2kW load. OA1 is the quietest opamp of the three, and with its typical input referred noise of 6.5nV / Hz, is the opamp of choice for input stages. Its output section is nominally capable of driving 3mA into a 5kW load. OA2 is best suited for control voltage processing, though is does have anti-paralleled diodes that can be used to fashion it into a clipper. (However in most applications where a clipper is needed, it's preferable to place it around OA3). OA2's input noise is comparable to OA3 and its output drive is comparably to OA1. The Reference Voltage - In Brief THAT Corporation's log/anti-log VCAs and RMS detectors require a reference voltage between the positive and negative power supplies, and to supply this, the THAT 4311 provides an on-chip, 2V reference about which the VCA, the RMS detector, and OA2 are biased. This reference is a buffered band-gap reference that is amplified to 2V. Pins are provided for filter capacitors at both the input and the output of the buffer, which are labeled CAP and VREF respectively. Application Information As noted previously, the THAT 4311 was originally designed for noise reduction systems, hence the emphasis on those parameters in the specifications. Its low power consumption, integration, and similarity to the THAT 4301, however, extend its utility to a variety of other products and applications. The circuit of Fig 19, shows a typical application for the THAT 4311. This simple compressor/limiter design features adjustable hard-knee threshold, compression ratio, and static gain. The applications discussion in this data sheet will center on this circuit for the purpose of illustrating important design issues. Signal Path As mentioned in the section on theory, the VCA input pin is a virtual ground with negative feedback provided internally. An input resistor (R1, 20kW) is required to convert the AC input voltage to a current THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Page 10 Low-voltage, Analog Engine(R) Dynamics Processor Preliminary Information within the linear range of the THAT 4311. (Peak VCA input currents should be kept under 250 mA for best distortion performance.) The coupling capacitor (Cl, 47 mF) is strongly recommended to block DC current from preceding stages (and from offset voltage at the input of the VCA). Any DC current into the VCA will be modulated by varying gain in the VCA, showing up in the output as "thumps". Note that Cl, in conjunction with R1, will set the low frequency limit of the circuit. The VCA output is connected to OA3, configured as an inverting current-to-voltage converter. OA3's feedback components (R2, 20 kW, and C2, 47 pF) determine the constant of current-to-voltage conversion. The simplest way to deal with this is to recognize that when the VCA is set for unity (zero dB) gain, the input to output voltage gain is simply R2/R1, much like the case of a single inverting stage. If, for some reason, more than zero dB gain is required when the VCA is set to unity, then the resistors may be skewed to provide it. Note that the feedback capacitor (C2) is required for stability. The VCA output has approximately 45pF of capacitance to ground, which must be neutralized via the 47pF feedback capacitor across R2. The VCA gain is controlled via the EC- terminal, whereby gain in dB will be proportional to the negative of the voltage at EC-. In this application the EC+ terminal is tied to VREF though it could be the , driven port, or the control ports could be driven differentially. The SYM terminal is returned nearly to the EC+ terminal (which is in this case VREF) via a small resistor (R3, 51W). The VCA SYM trim (R5, 20kW) allows a small voltage to be applied to the SYM terminal via R4 (33kW). This voltage adjusts for small mismatches within the VCA gain cell, thereby reducing even-order distortion products. To adjust the trim, apply to the input a middle-level, middle-frequency signal (1kHz at 200mVrms is a good choice with this circuit) and observe THD at the signal output. Adjust the trim for minimum THD. the VCA input circuitry, C3 in conjunction with R6 will set the lower frequency limit of the detector. The time response of the RMS detector is determined by the capacitor attached to CT (C4, 10 uF) and the size of the current in pin IT (determined by R7, 267 kW and VREF 2V). Since the voltage at IT is , approximately 2V, the circuit of Figure 19 produces 7.5 mA in IT, The current in IT is mirrored to the CT pin, where it is available to discharge the timing capacitor (C4). The combination produces a log filter with time constant equal to approximately 0.026 CT/IT (~35 ms in the circuit shown). The waveform at CT will follow the logged (decibel) value of the input signal envelope, plus a DC offset of about 2VBE plus VREF or about 3.3V. The capacitor used should be a low-leakage, electrolytic type in order not to add significantly to the timing current. The output stage of the RMS detector serves to buffer the voltage at CT and removes the 1.3 VDC (2 VBE) offset, resulting in an output centered around VREF for input signals of about 245 mVrms, or -10 dBu. The output voltage increases 6.1 mV for every 1 dB increase in input signal level. This relationship holds over more than a 60 dB range in input currents. Control Path The primary function of an audio compressor is to reduce a signal's dynamic range. A 2:1 compressor reduces a 100 dB dynamic range to 50 dB. A limiter, or infinite compressor, is a special case of compressor where the dynamic range is reduced to the point where the rms level of the signal is constant. This reduction in dynamic range is accomplished by a) raising the gain when the signal is below some particular level -- often referred to as the 'zero dB reference level' -- and b) reducing the gain when the signal is above that level. In addition, these devices often have a threshold, below which the signal is passed unprocessed and above which compression takes place. This feature keeps the noise floor from rising to noticeable levels in the absence of signal. We previously established that the zero dB reference level of the detector is -10 dBu (zero dB reference level = 7.5 , R6 = 28.7 kilohms). Neglecting the effect of the threshold control (R11 and R12), when the output is below this level the output of OA2 is driven high, forward biasing CR1 and reverse biasing CR2. Since CR2 is not conducting, no signal is passed to the VCA's control port by OA1. When the signal level exceeds -10 dBu, the output of the RMS detector goes positive, and CR2 begins to conduct. In RMS-Level Detector The RMS detector's input is similar to that of the VCA. An input resistor (R6, 28.7kW) converts the AC input voltage to a current within the linear range of the THAT 4311. The coupling capacitor (C3, 47mF) is recommended to block the current from preceding stages (and from offset voltage at the input of the detector). Any DC current into the detector will limit the low-level resolution of the detector, and will upset the rectifier balance at low levels. Note that, as with THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Rev. 08/30/01 Preliminary Information Page 11 this case, OA2's feedback is provided through R9, and the sensitivity at this point is 12.2 mV/dB, since the output of the RMS detector is multiplied by -R9/R8, or a gain of -2. A threshold control is provided to vary the threshold above or below -10 dBu. The output sensitivity of the RMS detector is 6.1 mV/dB. This is converted to a current by R8, and the sensitivity at the summing node of OA2 is 6.1mV dB 4.99 kW If a linear taper pot is used for R13, the compression ratio will be 1:2 at the middle of the rotation. However, 1:2 compression in an above-threshold compressor is not very strong processing, so 1:4 is often preferred at the midpoint. R14 warps the taper of R13 so that 1:4 compression occurs at approximately the midpoint of R13's rotation, The GAIN control (R18) is used to provide static gain or attenuation in the signal path. This control adds between 120mV and -180mV of offset to the output of OA1, which is approximately a -20dB to +30dB change in gain of the VCA. The gain control signal is passed into OA1 via R17, but this signal is also passed back to the threshold amplifier (OA2) via R10. This arrangement results in the threshold being fixed relative to the output. In other words, as the gain is increased, the threshold is lowered to keep the threshold of compression or limiting at the same output level. This is particularly important in limiters, since it keeps the gain control from interacting with the threshold. C5 is used to attenuate the noise of OA1, OA2, and the resistors R8 through R16 used in the control path. All these active and passive components produce noise which is passed on to the control port of the VCA, causing modulation of the signal. By itself, the THAT 4311 VCA produces very little noise modulation, and its performance can be significantly degraded by the use of noisy components in the control voltage path. = 12 mA . The wiper of R12 can swing between -2V and +3V relative to the summing node of OA2 which is at VREF If we want the threshold to swing as high as . +30 dB, then the value required for R11 can be calculated as R11 = 2V mA 1.2dB 30 dB 51 kW when rounded to the nearest 5% resistor value. Using this value and knowing that the pot's swing in the other direction is 3V, we can calculate the threshold swing in the negative direction to be 3V 51kW mA 1.2dB -49 dB Since the zero dB reference level of the detector is -10 dBu, the threshold can be adjusted from 20 dBu to -59 dBu. The output of the threshold detector represents the signal level above the determined threshold, at a constant of about 13mV/dB (from [R9/R8] 6.1mV/dB). This signal is passed on to the COMPRESSION control (R13), which variably attenuates the signal passed on to OA1. Note that the gain of OA1, from the wiper of the COMPRESSION control to OA1's output is R16/R15 (0.5), precisely the inverse of the gain of OA2. Therefore, the COMPRESSION control lets the user vary the above threshold gain between the RMS detector output and the output of OA1, from zero to a maximum of unity. The gain control constant of the VCA (6.1mV/dB) is exactly equal to the output scaling constant of the RMS detector. Therefore, at maximum COMPRESSION, above threshold, every dB increase in input signal level causes a 6.1mV increase in the output of OA1, which in turn causes a 1dB decrease in the VCA gain. With this setting, the output will not increase despite large increases in input level above threshold. This is infinite compression. For intermediate settings of COMPRESSION, a 1dB increase in input signal level will cause less than a 1dB decrease in gain, thereby varying the compression ratio. The resistor R14 is included to alter the taper of the COMPRESSION pot to better suit common usage. Overall Result The resulting compressor circuit provides hard-knee compression above threshold with three essential user adjustable controls. The threshold of compression may be varied over a range from about -58dBu to +20dBu. The compression ratio may be varied from 1:1 (no compression) to :1. And, static gain may be added between -20 and +30dB. Audio performance is excellent, with THD running below 0.1% at middle frequencies even with 10 dB of compression, and an input dynamic range of over 105dB. Perhaps most important, this example design only scratches the surface of the large body of applications circuits which may be constructed with the THAT 4311. The combination of an accurate, wide dynamic range, log-responding level detector with a high-quality, exponentially-responding VCA produces a versatile and powerful analog engine. These, along with its on-board opamps, allow a designer to construct, with a single IC and a handful of external components, gates, expanders, de-essers, noise reduction systems and the like. THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com Page 12 Low-voltage, Analog Engine(R) Dynamics Processor Preliminary Information Package Information The THAT 4311 is available in a 20-pin surface mount package. The package dimensions are shown in Fig 20 while the pinout is given in Table 1. 0-10 BC 1 D F A G E J H I Item A B C D E F G H I J Millimeters 10.0 0.3 5.0 0.2 6.8 0.4 0.35 0.1 0.95 0.87 MAX 1.6 0.015 0.15 0.1 0.5 0.2 0.15 +0.1 -0.05 Inches 0.394 0.012 0.197 0.008 0.268 0.016 0.014 0.004 0.037 0.034 MAX 0.063 0.006 0.006 0.004 0.02 0.008 0.006 +0.004 -0.002 Figure 20. -S (DMP20 surface mount) package drawing THAT Corporation; 45 Sumner Street; Milford, Massachusetts 01757-1656; USA Tel: +1 508 478 9200; Fax: +1 508 478 0990; Web: www.thatcorp.com