Collins S-Line AGC System Explained
Delayed AGC with Separate Threshold, Dual Time-Constant Network & S-Meter Drive
Understanding the 75S-3B/3C AGC loop is essential for diagnosing sensitivity and overload complaints — the AGC threshold, attack/decay time constants, noise clipper, and S-Meter transfer function explained from first principles with the Don Jackson analysis as primary reference
The AGC (Automatic Gain Control) system is the single circuit most responsible for how the 75S-3B/3C “feels” on the air — whether SSB signals sound smooth or pumpy, whether strong signals overload before the AGC can react, whether the S-Meter is stingy or generous, and whether the receiver’s sensitivity appears to change from band to band. Nearly every sensitivity or overload complaint about the 75S-3 family traces back to either a component failure in the AGC circuit or a misunderstanding of how the AGC is designed to work.[1]
The Collins S-Line AGC is a delayed AGC with a separate threshold adjustment. “Delayed” means that the AGC does not begin to reduce receiver gain until the input signal exceeds a threshold level — below that level, the receiver operates at full gain. This threshold is set by potentiometer R57 (IF GAIN ADJUST, accessible through the top of the chassis). The AGC uses a dual time-constant network with FAST and SLOW decay settings selected by the front-panel AGC switch. The S-Meter is driven by a separate circuit that measures AGC-controlled stage currents, not the AGC voltage directly.[2]
The 75S-3B/3C AGC loop operates as follows, starting from the IF output and working back to the gain-controlled stages:[2]
1. IF Signal Sampling: The IF signal voltage from the secondary of transformer T6 is coupled to one of the diode plates in V9 (the AGC rectifier). This is the point where the AGC system “sees” the signal level.
2. AGC Rectification (V9): The diode section of V9 rectifies the IF signal, producing a negative DC voltage proportional to the signal strength. This rectified voltage passes through the filter network R50 and C49 before reaching the AGC time-constant network.[2]
3. AGC Delay (Threshold): Generation of AGC voltage is delayed until the signal voltage at the V9 diode plate exceeds the cathode bias on V9. This bias is set by R57 (IF GAIN ADJUST). R57 is normally adjusted so that AGC action is initiated with a receiver input signal of approximately 2 µV (−101 dBm). Below this threshold, the receiver operates at full gain and the AGC voltage is zero — the S-Meter reads S0.[2]
4. Noise Clipper (V9 second diode section): The other half of V9 (or V16 in the 75A-4) acts as an AVC noise clipper. This clips sharp noise impulses from the AGC voltage, preventing short noise spikes from desensitising the AGC circuit. Without this clipper, impulse noise would charge the AGC time-constant capacitors to a high voltage and then take a long time to discharge — making the receiver “deaf” for an extended period after each noise pulse.[2]
5. Dual Time-Constant Network: The AGC voltage passes through a network of resistors and capacitors that establish the attack and decay time constants. This is the heart of the AGC system and where the FAST/SLOW behaviour is determined.
The 75S-3B/3C uses a stacked dual time-constant architecture consisting of a fast section and a slow section in series:[2]
The fast section (R88 and C153 in parallel) provides a fast charge/discharge rate designed to eliminate short-duration interference — noise pulses are absorbed by C153 without being passed through to the slow section. The slow section (R24 and C50, with C137 switched in parallel in the SLOW position) provides the longer RC time constant that develops a smoothly varying AGC voltage tracking the average signal level.[2]
⚠ The R88 Problem: R88 (680 kΩ) is the critical component that determines how quickly the AGC can charge — and it is too high. At 680 kΩ, C50 and C137 cannot charge fast enough to track SSB signal peaks, so the AGC SLOW mode behaves almost identically to FAST. This is the most common AGC complaint in the 75S-3 family. See the 75S-3 Known Issues guide on this site for the R88 → 68 kΩ modification and the enhanced C50 → 0.47 µF modification.[3]
| Parameter | Stock 75S-3B/3C | R88 → 68 kΩ Mod | Enhanced Mod (R88 ∥ 220 kΩ + C50 → 0.47 µF) |
|---|---|---|---|
R88 effective | 680 kΩ | 68 kΩ | ~166 kΩ (680 kΩ ∥ 220 kΩ) |
C50 | 0.1 µF | 0.1 µF | 0.47 µF |
FAST decay | Too fast — pumps on SSB | Improved | Comparable to Drake R-4C medium |
SLOW decay | Nearly same as FAST | Noticeably slower | Comparable to Drake R-4C slow |
Practical difference | FAST ≈ SLOW (no useful difference) | Clear 2-position difference | 3–4× speed ratio between positions |
R57 sets the AGC threshold — the minimum signal level at which the AGC begins to reduce receiver gain. The manual specifies a threshold of approximately 2 µV (−101 dBm). Below this level, the receiver operates at maximum gain and the S-Meter reads S0. Above this level, the AGC voltage progressively reduces IF and RF stage gain to maintain a relatively constant audio output level.[2]
The threshold is set during alignment: a 2 µV signal at 28.6 MHz is applied to the antenna jack, the receiver is tuned to it, and R57 is adjusted to the point that produces a just-perceptible increase above the no-signal S-Meter reading. This threshold directly determines the receiver’s sensitivity — setting it too low makes the receiver appear deaf (the AGC is reducing gain before the signal reaches a useful level); setting it too high allows strong signals to overload the IF before the AGC can react.[1]
Don Jackson’s Analysis: The AGC threshold value determines where S0 falls on the absolute power scale. If the threshold is set to 2 µV, then S0 corresponds to approximately −101 dBm. The AGC slope (dB of attenuation per volt of AGC voltage) is remarkably linear in the 75S-3B — Jackson measured an essentially constant dB/Volt characteristic over the usable AGC range. However, the S-Meter transfer function (which converts AGC voltage to meter current via V6 and V7 screen/cathode currents) is markedly nonlinear, degrading the otherwise excellent AGC linearity before it reaches the meter.[1]
The 75S-3B S-Meter circuit measures the AGC voltage indirectly — through variations in the V7 cathode current and the screen currents of V6 and V7 (the two 455 kHz IF amplifiers). Under no-signal conditions, the voltage across R13 (the S-Meter zero adjust) is set equal to the voltage across R21, and the meter reads zero. As AGC voltage increases (stronger signal), the V6/V7 screen and cathode currents change, creating a net current through the S-Meter movement.[2]
Don Jackson’s measurements revealed that the AGC voltage itself has an excellent, nearly constant dB/Volt slope across the usable range — but this linearity is severely degraded by the S-Meter transfer function. The net result is an S-Unit value that varies from approximately 2.9 dB/S-Unit to 5.7 dB/S-Unit depending on where in the S0–S9 range the signal falls, with an average of about 4 dB/S-Unit. The Collins S-Meter is therefore useful for relative comparisons but not reliable for absolute power measurements.[1]
Jackson designed a replacement S-Meter driver circuit that provides a substantially more linear transfer function, producing a much more constant dB/S-Unit value across the meter range. This modification is documented in Part 4 of the CCA Signal AGC series.[4]
Attack time is the time for the AGC to reach its final value after a signal suddenly appears. In the 75S-3B/3C, the attack time is determined primarily by R50, C49, and the forward conduction characteristics of the V9 AGC diode. The stock attack time is relatively slow — slow enough that the first few cycles of a suddenly appearing SSB signal can overload the IF before the AGC catches up.[5]
Decay time is the time for the AGC to release after the signal disappears. This is determined by the R24/R88/C50/C137/C153 time-constant network and the AGC switch position (FAST or SLOW). A longer decay time prevents noise from rising between SSB words or CW characters — but too long a decay makes the receiver feel “sluggish” and can cause the S-Meter to stick at high readings after a strong signal disappears.
Loop stability is a function of the total phase shift around the AGC loop. If the attack time is made too fast (by reducing R50), the loop can become unstable — the AGC will oscillate, producing “pumping” or “motorboating.” The original Collins values represent a conservative stability margin. Community modifications that reduce R88 do not significantly affect loop stability because R88 primarily affects the slow decay path, not the attack path.[5]
Caution on R50 Modifications: Reducing R50 to speed up the AGC attack can be beneficial, but must be done carefully. A value that is too low will increase the AGC attack time at the expense of loop stability. The stock attack time in the 75S-3B is noticeably slow compared to later receivers like the Drake R-4C — some operators consider faster attack a worthwhile improvement.[5]
The AGC voltage is applied to the control grids of the following stages in the 75S-3B/3C, each of which is decoupled from the AGC bus by individual resistor-capacitor networks to prevent instability from inter-stage feedback:[2]
The mechanical filter amplifier (V5A/V6 in the 75S-3B) that immediately follows the mechanical filter is also AGC-controlled. Each controlled stage has its own RC decoupling network on the AGC bus — a failure in any single decoupling capacitor can cause the entire AGC loop to oscillate or behave erratically. These are among the critical mica capacitors identified in the companion Known Issues guide.[3]
| Symptom | Probable Cause | Diagnostic |
|---|---|---|
FAST and SLOW AGC sound the same | R88 (680 kΩ) too high — stock design limitation | Replace R88 with 68 kΩ (or add 220 kΩ in parallel) |
AGC “pumps” on SSB — noise rises between words | Decay time too short; C50 value too small | Increase C50 from 0.1 µF to 0.47 µF; modify R88 |
Receiver overloads on strong signals before AGC reacts | Attack time too slow (stock limitation); or leaky AGC decoupling capacitors | Check mica caps on AGC bus; consider cautious R50 reduction |
S-Meter pinned or reads backward | AGC decoupling cap leaking positive voltage onto AGC bus | Check all mica caps connected to AGC line for DC leakage |
S-Meter zero wanders continuously | R13 pot dirty; thermal drift in V6/V7 screen/cathode circuits | Clean R13; reverse leads to use fresh section of pot; replace weak V6/V7 |
Receiver sensitivity varies band to band | RF alignment drift changes AGC threshold relative to signal level | Realign RF front end; recheck R57 threshold setting at 28.6 MHz |
S-Meter stingy — few signals above S9 | Weak IF tubes; leaky Black Beauties reducing IF gain; AGC threshold set too high | Replace V6/V7; replace Black Beauties; verify R57 per manual |
AGC oscillation / “motorboating” | Shorted or leaky AGC decoupling cap; shield ground fault | Check all AGC bus decoupling caps; verify all can/shield grounds tight |
The AGC architecture described here applies to the entire 75S-3 family (75S-3, 3A, 3B, 3C) with the same R88 problem affecting all models. The KWM-2/2A uses a similar dual time-constant AGC architecture with the same 680 kΩ value for R88 — the R88 → 68 kΩ modification applies equally to the KWM-2/2A. The 75A-4 uses a different AGC circuit topology (the W1LSB AGC modification addresses its specific attack-time limitation) but shares the general principle of delayed AGC with a threshold adjustment. The 51J-4 uses a fundamentally different AVC architecture and is not directly comparable.[3]
- Jackson, Don W5QN. S-Line AGC Theory/Stability & Improvements — Part 3: S-Meter Stability Analysis. CCA Signal magazine Q4 2011. AGC slope linearity measurements, S-Meter transfer function analysis, dB/S-Unit variation (2.9–5.7 dB/S-Unit over S0–S9 range), AGC threshold interaction with S-Meter calibration. collinsradio.org — Signal Q4 2011 (PDF)
- WA3KEY. 75S-3B/3C Receiver — Circuit Description. Complete AGC circuit description: V9 AGC rectifier, delayed AGC via R57 cathode bias, dual time-constant network (R88/C153 fast, R24/C50/C137 slow), noise clipper, AGC distribution to RF/IF stages, S-Meter via V6/V7 screen/cathode currents. wa3key.com — 75S-3B/3C
- VK3KCM. 75S-1/2/3 Service Information. R88 modification (680 kΩ → 68 kΩ), C50 modification (0.1 µF → 0.47 µF), enhanced modification with 220 kΩ parallel, applicability to 75A-4/KWM-2A/51S-1, Drake R-4C comparison benchmark. angelfire.com/de/vk3kcm — 75S Service
- Jackson, Don W5QN. S-Line AGC Theory/Stability & Improvements — Part 4: S-Meter Driver Redesign. CCA Signal magazine Q1 2012. Replacement S-Meter driver circuit design, linear transfer function, clamping function for standby/mute, AGC voltage limiting for amplifier protection. collinsradio.org — Signal Q1 2012 (PDF)
- Collins Reflector. More on 75S-3C AGC Changes. R50 attack time trade-off discussion, loop stability considerations, community AGC modification experience comparing to Drake R-4C attack/decay characteristics. September 2011. Collins Reflector — AGC Changes
- Jackson, Don W5QN. 75S-3B AGC Troubleshooting Correspondence. IF gain pot setting procedure (2 µV threshold), V9 pin 5 voltage measurements at calibrated input levels, IF shield grounding importance, filter insertion loss measurement technique. Collins Reflector, July 2016. Collins Reflector — AGC Troubleshooting
- Collins Collectors Association — RX For Your Collins. Master index: “S-Line AGC Theory/Stability & Improvements — Jackson” (4-part series, Q2 2011 through Q1 2012), “75S-3 & 3B/C S-Meter Stability Analysis and Improvements — Jackson.” collinsradio.org — RX For Your Collins
- QSL.net / VA3IUL. Automatic Gain Control (AGC) in Receivers. General AGC theory: attack/decay time definitions, loop bandwidth vs. gain reduction, AM modulation frequency constraint, delayed AGC threshold concept. qsl.net/va3iul — AGC Theory (PDF)
- Collins 75S-3B/3C Instruction Book. AGC circuit description, R57 IF GAIN ADJUST alignment procedure (2 µV at 28.6 MHz), S-Meter zero adjustment, AGC switch FAST/SLOW time constant selection. wa3key.com — 75S-3B/3C Manual Reference
Don Jackson, W5QN — For the definitive four-part AGC analysis series published in the CCA Signal magazine (Q2 2011 through Q1 2012) that provides the quantitative foundation for understanding the S-Line AGC system. Jackson’s measurements of AGC slope linearity, S-Meter transfer function nonlinearity, and his redesigned S-Meter driver circuit represent the most rigorous analysis of the 75S-3B AGC ever published.
VK3KCM — For compiling the 75S-1/2/3 service information that documents the R88 and C50 modifications with the Drake R-4C comparison benchmark that gives operators an intuitive reference for the modified AGC behaviour.
WA3KEY — For the 75S-3B/3C circuit description that clearly explains the delayed AGC architecture, the R57 threshold, and the dual time-constant network.
Collins Collectors Association (CCA) — For publishing the Jackson AGC series in the Signal magazine and maintaining the “RX For Your Collins” technical library.