Collins 32S-3 & 32S-3A Known Issues

The Collins 32S-3 (1962–1968) and 32S-3A (1968–1978) are the standard S-Line transmitters, producing 100 watts nominal output on SSB and CW across 80–10 meters. Like the companion 75S-3 receiver, the 32S-3 exhibits a well-characterised set of age-related failures and design-era limitations that the community has thoroughly documented over decades. This guide consolidates the most significant transmitter-specific issues — carrier balance drift, ALC instability, PA neutralisation degradation, CW chirp, VOX faults, and tone oscillator problems — with their root causes and proven fixes.[1]

The separate guides on this site covering Carrier Balance Optimisation, ALC System Theory, and PA Neutralisation Step-by-Step provide in-depth treatment of those individual topics. This document provides the consolidated “known issues” overview with cross-references to those detailed guides where applicable.

Symptom: Transmitted SSB audio sounds “rough,” “buzzy,” or “harsh.” Receiving stations report carrier feedthrough. The carrier null obtained with R2/C21 is shallow (>0.5V at antenna output) or degrades after warm-up. Audio that was acceptable a year ago has progressively deteriorated.[2]

Root Cause: The 1N34A germanium diodes (CR1–CR4) in the balanced modulator ring degrade with age — they develop mismatched forward characteristics, open circuits, or internal shorts. Even one failed diode in the quad severely degrades carrier suppression. The balanced modulator provides ~30 dB suppression when healthy; a degraded quad may provide only 10–15 dB, reducing total suppression (modulator + mechanical filter skirt) from 50 dB to as low as 30 dB — audibly poor.[2]

Fix: Replace all four CR1–CR4 diodes as a matched set with 1N4454 (modern equivalent). Also verify C25 (should be 6 pF per ASAB, not the original 3 pF in early units). Then perform the full iterative R2→C21→R2 carrier null procedure described in the companion Carrier Balance Optimisation guide on this site. Target ≤0.2V RF at the antenna output.[3]

Symptom: The ALC ZERO control cannot hold a stable setting — the meter drifts during warm-up or oscillates around zero. The ALC meter reads below zero in SSB. ALC compression sounds “pumpy” or over-compressed when a 30L-1 is in the chain.[4]

Root Cause: C142, a Black Beauty paper capacitor in the ALC circuit, develops DC leakage that introduces an offset onto the ALC bus, causing the meter to read below zero and the ALC action to behave erratically. Temperature-sensitive components in the ALC zero-set network drift with warm-up. The ALC dual time-constant (fast on V6, slow on V3) degrades when the timing capacitors develop leakage, altering the attack/release profile.[4]

Fix: Replace C142 with a modern film capacitor as a first priority. For persistent zero-set instability, refer to the Don Jackson analysis “Dealing with 32S-3 ALC Zero Adjust Instability” published through the CCA. Replace all Black Beauty capacitors in the ALC timing network. If the ALC meter will not zero at all, R170 (220Ω) across the ALC zero potentiometer can be removed to extend the adjustment range.[5]

Symptom: Plate current dip and maximum output no longer coincide at the same PA TUNING setting. Erratic high plate current. VHF parasitic oscillation (invisible on HF wattmeter). Glowing 6146 plates at idle. In severe cases, destruction of the neutralising capacitor and 6146 tubes.[6]

Root Cause: The original Erie ceramic trimmer neutralising capacitor (C184, 1.8–8.7 pF in early units) degrades with age and repeated thermal cycling. The ceramic dielectric develops micro-cracks that alter its capacitance unpredictably. If 6146B or 6146W tubes have been substituted for original 6146 types, their higher plate-to-grid capacitance may exceed the adjustment range of the original trimmer entirely.[6]

Fix: Upgrade to the air-variable neutralisation circuit (Johnson capacitor, 6 kV series cap, 2W resistor) per the late-production 32S-3A / 9th edition KWM-2/2A configuration. Then perform the full neutralisation procedure with 6146 tubes removed and an oscilloscope or RF voltmeter monitoring feedthrough. Neutralisation must be rechecked whenever 6146 tubes are replaced.[7]

Symptom: CW signal has an audible frequency shift (“chirp”) at the beginning of each keydown, heard on a separate monitoring receiver. The sidetone may also sound unstable. The chirp may be present on the transmitter PTO but not when using the receiver PTO in transceive mode.[8]

Root Cause: A weak or gassy 6AU6 tube (V301) in the 70K-2 PTO is the most common cause. The PTO frequency shifts slightly as the tube’s operating point settles after keying. A secondary cause is the CW CAL control being set too high — advancing it beyond approximately the 12 o’clock position can cause keying transients that modulate the PTO frequency. In the 32S-3, the CW carrier is generated by the USB crystal feeding the first mixer directly (bypassing the balanced modulator and mechanical filter), so any PTO instability translates directly to CW chirp.[8]

Fix: Replace the 6AU6 PTO tube with a fresh, low-noise example. Reduce the CW CAL setting to approximately 10–11 o’clock. If chirp persists, check the PTO B+ decoupling capacitors (C306/C309/C310) and the 1N34A switching diode CR301 — these are the same drift-producing components addressed in the PTO overhaul guide.[8]

Symptom: The VOX relay (K2) chatters or “machine-guns” when PTT or VOX is activated. Or: no VOX operation at all — VOX GAIN control has no effect, but TUNE and LOCK KEY work normally. These two symptoms frequently coexist with no SSB output and no ALC meter reading.[9]

Root Cause: The V1 dual-triode (6U8A) provides audio to both the SSB generation chain and the VOX amplifier. If V1 fails or if Black Beauty coupling capacitors in the first two audio stages develop leakage, both SSB audio and VOX are lost simultaneously — the receiver-side audio (Anti-VOX) is unaffected, creating the confusing symptom of “VOX doesn’t work but I can receive fine.” Relay chatter is caused by relay transients coupling through the ground lead of R3 to the emission switch, feeding back into the VOX rectifier.[9]

Fix: Replace V1 (try swapping the two 6U8A tubes first to isolate). Replace Black Beauty coupling capacitors in the V1 audio chain. For relay chatter, add a 0.01 µF capacitor (CPN 913-3013-000) from V1 pin 7 to ground per Collins ASAB documentation. For the separate “phantom tone” on B+ lines issue, see Issue 6 below.[5]

Symptom: CW sidetone quality is unstable — the tone warbles or shifts pitch. A faint tone may be present on transmitted SSB audio even with the emission switch in USB or LSB position. The B+ lines carry a low-level AC component at the tone oscillator frequency.[10]

Root Cause: V11 operates as an R-C phase-shift oscillator. In USB and LSB modes, the cathode bias resistor R66 is disconnected from ground (“floating”), but C6B (25 µF @ 25V) remains connected from cathode to ground and slowly charges toward the plate potential (~150V) — far exceeding its 25V rating. This leakage couples tone oscillator energy onto the B+ lines. Additionally, small B+ voltage fluctuations shift the phase-shift network frequencies, causing pitch instability.[10]

Fix: Add a 5 µF 400V electrolytic from the junction of R63/R62 to ground to absorb the voltage transient during break-in keying. Replace C107 with a high-stability metal-film or polypropylene capacitor of equal value for improved phase-shift stability. Some operators add a switching modification to completely disconnect C6B from the circuit in LSB/USB modes, eliminating the phantom tone entirely.[10]

Symptom: Multiple simultaneous problems: ALC reads below zero, VOX doesn’t work, SSB audio is distorted or absent, excessive PA idle current, hot resistors throughout the chassis. Symptoms that don’t respond to tube substitution or alignment.[5]

Root Cause: Identical to the 75S-3 — Sprague Black Beauty oiled paper capacitors develop progressive DC leakage with age. In the 32S-3, the most consequential failures are C142 (ALC circuit, tied to B+ rail), coupling capacitors in the V1 audio chain, and timing capacitors in the VOX and ALC circuits. Because Black Beauties fail progressively, the symptoms accumulate gradually and are easily misattributed to tubes, alignment, or other components.[5]

Fix: Replace all Black Beauty capacitors with modern film types (Orange Drop or equivalent). This should be the first restoration step, before any alignment or performance measurement. Complete capacitor kits specific to the 32S-3 are available from specialty suppliers.[11]

Symptom: No CW output despite normal SSB operation. Intermittent switching between USB and LSB. No output on TUNE or LOCK KEY. Erratic mode behaviour that changes when the switch is vigorously toggled.[12]

Root Cause: The emission switch (S9, with multiple closely-spaced wafers) develops silver oxide and sulfide tarnish on its contacts. The 32S-3’s CW mode uses a completely different signal path from SSB — the CW carrier comes from the USB crystal and feeds the first mixer directly, bypassing the balanced modulator and mechanical filter. Oxidised contacts on specific emission switch wafers can disable CW while leaving SSB operational, or vice versa.[12]

Fix: Apply DeoxIT D5 using the dropper bottle — a quarter drop per contact — followed by vigorous switch operation through all positions. The closely-spaced wafers make spray application impractical and risk saturating the switch insulators. If contacts are severely corroded, the switch may need to be partially disassembled for direct treatment.[12]

Symptom: Mushy transmit audio; instability in VOX operation; elevated or incorrect bias voltages that don’t match the manual’s voltage table; resistors that look physically normal but have drifted 30–100% above their marked value.[13]

Root Cause: Carbon composition resistors under sustained thermal stress drift upward in value over decades. The most consequential resistors in the 32S-3 include R4 (68 kΩ) and R7 (47 kΩ) in the audio chain — whose drift directly causes mushy transmit audio — R148 (820Ω, 2W) in the PA screen circuit, and R175 (4700Ω) whose total failure kills transmit audio instantly. These failures are invisible to visual inspection.[13]

Fix: Measure every carbon composition resistor in the signal path with a DMM and compare to the manual’s resistance table. Replace any resistor that has drifted more than 20% from its marked value. Where Collins specified ½W resistors in thermally stressed locations, upgrade to 1W replacements. R148 should be upgraded to a 5–7W wirewound type per ASAB recommendation.[13]

Symptom: Spurious signal(s) appearing on frequencies outside the intended operating band, reported by other stations or visible on a spectrum analyzer. May only appear on certain bands or at certain power levels.[1]

Root Cause: Exciter tuning misalignment can produce mixing products that fall outside the passband of the subsequent RF stages. Additionally, certain well-intentioned modifications to the 32S-3 can inadvertently create new spurious mixing products. The CCA has published a specific diagnostic article on this issue.[1]

Fix: Perform a complete exciter alignment per the manual, starting from the inductances (which most operators skip, only adjusting trimmers). Refer to the CCA article “32S-3 Mod Produces Out-Of-Band Spur: How to Find & Fix” for the diagnostic procedure. If modifications have been applied, reverse them and re-test before the spur is assumed to be a design issue.[1]

1. Collins Collectors Association — RX For Your Collins. Master index of S-Line transmitter articles: “32S-3 Chirp Problem in CW mode,” “32S-3 ALC Instability,” “32S-3 Trouble Shooting Voltage Table,” “Dealing with 32S-3 ALC Zero Adjust Instability — Jackson,” “Repairing the 32S-3 — Carns and Bud, K7RMT,” “32S-3 Mod Produces Out-Of-Band Spur: How to Find & Fix.” collinsradio.org — RX For Your Collins
2. Collins Collectors Association — 32S-3 Equipment Profile. Balanced modulator circuit description, ALC dual time-constant (fast V6 / slow V3), RF inverse feedback, CW keying circuits. collinsradio.org — 32S-3
3. Gamma-Five.com. My Collins KWM-2 — Balanced Modulator Diode Quad Failure. 1N34A failure diagnosis (one open, one shorted), severe audio clipping at modulator input, 1N4454 replacement. gamma-five.com — KWM-2 Restoration
4. Antique Radio Forums. Collins 32S-3 Transmitter — ALC and VOX troubleshooting. ALC meter below zero, no VOX, V1 audio chain diagnosis, WQ9E and Norm analysis. December 2021. antiqueradios.com — 32S-3 XTAL OSC
5. VK3KCM. KWM-2/2A Service Information. C142 leakage, C264 ripple suppression, relay transient fix, C25 coupling improvement, R170 ALC zero range extension, compiled ASAB service data. angelfire.com/de/vk3kcm — KWM-2 Service
6. Antique Radio Forums. Collins 32S-3 Grid Current — KB6GM, WQ9E analysis. Marginal neutralising circuit, 6146B incompatibility, VHF parasitic diagnosis, 6CL6 driver shield. March 2019. antiqueradios.com — 32S-3 Grid Current
7. Borowski, Gerd DJ7HS. Collins KWM-2: Installing New Parts for PA Neutralization. W7KSG kit installation, Johnson air variable, photographic step-by-step. qsl.net/dj7hs — KWM-2 Neutralisation
8. Collins Reflector. 32S-3 CW Chirp Fixed — W3YY. Weak 6AU6 PTO tube as root cause; CW CAL setting interaction; chirp present on transmitter PTO but not receiver PTO in transceive. September 2010. Collins Reflector — CW Chirp
9. Carns. KWM-2/2A Keying Circuit — Functional Description and Issues. Complete VOX/keying chain analysis, relay chatter diagnosis, PL-068 connector oxidation, three keying modes explained. CCA publication. collinsradio.org — Keying Issues (PDF)
10. VK3KCM. 32S-1/3 Service Information. Tone oscillator instability (C6B charging, phantom tone), SB listing for 32S-1/32S-3, SIL 1-75 aging instability (C56/C123), carrier null procedures. angelfire.com/de/vk3kcm — 32S-1/3 Service
11. Collins Reflector / QTH.net. 32S-3 No CW Keying — SM2EKM. CW signal path bypasses balanced modulator; emission switch contact oxidation as root cause of CW-only failure. October 2009. Collins Reflector — No CW Keying
12. Antique Radio Forums. Collins 32S-3 Restoration. Field alignment approach, exciter tuning inductance adjustment, carrier and mixer balance using output meter on receiver speaker. May 2020. antiqueradios.com — 32S-3 Restoration
13. VK3KCM. KWM-2/2A Additional Service Information. Carbon composition resistor drift table: R4, R7 (audio — mushy audio), R148 (PA screen — overheating), R175 (4700Ω — sudden transmit audio failure), R50 (tone oscillator). angelfire.com/de/vk3kcm — KWM-2A Info
14. Collins 32S-3 Instruction Book, 7th Edition (June 1969). Circuit description, alignment procedures, keying adjustment, voltage/resistance tables. collinsradio.org — 32S-3 Manual (PDF)

Don Jackson, W5QN — For the definitive ALC zero instability analysis and the S-Line AGC/ALC technical article series published through the CCA.

Carns — For the comprehensive KWM-2/2A keying circuit functional description that explains the three keying modes and documents the most common VOX/relay faults.

Gerd Borowski, DJ7HS — For the photographic PA neutralisation upgrade guide and the detailed KWM-2 T/R switching transient analysis.

Bob W3YY — For documenting the CW chirp root cause (weak 6AU6 PTO tube) with the diagnostic technique of comparing transmitter PTO vs. receiver PTO in transceive mode.

VK3KCM — For compiling and hosting the 32S-1/3 and KWM-2/2A service information including the tone oscillator instability analysis and resistor drift table.

Rodger WQ9E — For extensive troubleshooting contributions on the Antique Radio Forums covering PA neutralisation, grid current diagnosis, and 6CL6 driver shield analysis.

Collins Collectors Association (CCA) — For maintaining the “RX For Your Collins” technical article library, the CCA Signal magazine archives, and the factory service bulletins and manuals.