My Collins KWM-2

© Copyright Alexander D. Schapira, 2006-2014, All rights reserved

With thanks to all who provided suggestions and responses to my original request for help,
 I am providing this account of how I got my newly acquired KWM-2 back on the air.
I hope this will be of some interest to other KWM-2 owners.

I bought this rig from its original owner, a retired Collins engineer, who had it in storage for several decades.
It was exactly as delivered by Collins, and according to the seller, had never been out of its case.

With any vintage equipment, there are several things that need to be done before applying power for the first time. After a visual inspection and a little cleaning, a resistance check was in order. Following the chart in the KWM-2 manual, I checked every tube pin resistance to ground. Everything was reasonable except for V11 pin 2. This is the BFO grid. I measured 1.1 Meg (roughly agreeing with the schematic) but the chart in the 1962 KWM-2 manual said 95k. This turned out to be a typo -- the 1978 manual specifies 1 Meg.

Satisfied that there were no dead shorts, I powered the rig up for the first time, and the receiver came to life. With a short wire antenna, I was able to hear WWV on 15 MHz.

The Dummy Load:
In preparation for transmitter testing, I examined the dummy load that came the KWM-2.  It was built by the original KWM-2 owner and consisted of fifty 100 ohm 2 watt resistors in series/parallel to comprise a nominal 50 ohm 100 watt load (five series banks of ten in parallel.) Some of the many solder joints had come apart resulting in an intermittent resistance. I re-flowed the solder joints to solidify the load. It now measures a solid 47 ohms -- a little low, but close enough.

The Power Supply:
The PS was also built by the seller in 1962, using Collins parts, and is similar in circuitry to a PS-2.
Original power supply
While listening to the rig, after about an hour, one of the 40 uF/450V dry electrolytics in the power supply began to snap, crackle, pop, sizzle, and overheat. Uh-Oh.
Inside the original power supply
If you look closely, you can see the full "One Year" guarantee on the capacitors.

I replaced all ten electrolytics with modern equivalents from Mouser. While I was at it, I added a pilot lamp and a jumper to conveniently disable the high voltage. The rectifiers are solid state diodes so "removing the rectifier tube", as the Collins manual suggests, to disable the HV is not an option. I also added a dropping resistor to reduce the B+ slightly, since it was about 50 volts too high. I plan to enclose the PS in a perforated metal cage. Perhaps I'll look for a 516F-2.

The Rig:
After repairing the power supply, the receiver seemed to be working properly, but not the transmitter. I originally observed two problems. Although the KWM-2 did appear to tune up properly into the dummy load:
  1. There was hardly any audio gain, barely enough to budge the ALC meter even with the mic gain all the way up, and
  2. The VOX either pulled in too late or dropped out too early.
I ruled out the microphone by trying four different mics which exhibited similar behavior. In order to help characterize the transmitter problem, I hooked up an RF ammeter in series with the dummy load, and connected the probe of an HP 410B RF VTVM across the load to monitor the output of the rig. When tuned up in the LOCK position I measured about 60 volts RMS and 1.3 amps RMS into the load, for a power output of 78 watts (continuous). These measurements are consistent with the 47 ohm dummy load (I^2R and E^2/R). (The load remained cool for the short time I was driving it.) I then knew, much to my relief, that the finals, driver, and most of the preceding stages were okay.

While working into the dummy load, I listened to the signal on another receiver and heard severely distorted and very weak audio. Since the rig tuned up properly, but had very low and distorted audio, I concentrated on the audio chain.

Time to do some homework:
I downloaded copies of all of the KWM-2 Service Letters and Service Bulletins, from the CCA website. (This alone was reason enough to join the CCA.) I also searched the CCA archives for 'audio' and 'VOX' and printed the relevant articles. At the suggestion of a reader of my original post, I downloaded the latest KWM-2 manual, even though I had the original 1962 manual. I was pleasantly surprised to find all the modifications from the service bulletins listed in the revised manual and indicated on the schematics. (The original service bulletins sometimes provide step-by-step instructions and parts locations for making the modifications; the updated manual only lists the changed components, without instructions or locations. So it is good to have both.)

I cross-checked the relevant modifications listed in the 1978 manual against the service bulletins looking for changes that might affect either the audio or the VOX. Several numbered service bulletins were relevant, so I made the indicated changes:

#3      C6,   changed from 0.47uF to 1uF, replaced with 600V Orange Drop
#11     C47,  changed from 0.047uF to 0.068uF, replaced with 600V Orange Drop
#7      R201, 220k, added
#8      R199, 2.2M, added
#36B    C227, 0.01uF, added
 -      C160, replaced with 0.47/100V Orange Drop

Changes #7 and #8 require the addition of a terminal post to allow for the insertion of a series resistor.  I tried to be as faithful to the quality of the original Collins work as possible.

R201 Before                        R201 During                        R201 After
R201 Before    R201 During    R201 After

These changes did not cure the audio problem, but did solve the VOX problem. So audio was getting through to the VOX circuits but not to the modulator. (I later removed change #8, R199, since it made the minimum VOX time constant much too long.)

In the course of a very thorough visual inspection, I spotted a cracked resistor, R169, 1k! (No, it wasn't in an area I had worked on.) I replaced it, but that was not relevant to the audio problem.

To troubleshoot the audio problem, I wanted to hook up an audio source to the mic (or phone patch) input, but what audio level to use? I measured the audio output from each microphone (while whistling at a constant amplitude!) and obtained results as follows:
The first two mics came with the KWM-2.

  # Microphone              Milli-Volts (P-P)
 - -------------------     -----------------
 1 Euphonic C-47 Hand Mic  40-50
 2 Stromberg-Carlson #10   40-45
 3 GE/Shure 97U29          120-140
 4 Astatic D-104           120-140

I set my audio source for 1200 Hz, 80 millivolts peak-to-peak (an intermediate value) and connected it to the phone patch (mic) input. I then began to carefully scope out the audio chain starting from the microphone input and working inward.

Just to make things interesting, the audio chain is fully energized only when in TX mode. So I disabled the PA screen and HV, and keyed the PTT line.

I found the audio signals to be normal until I got to the *input* to the balanced mixer, where the audio signal was severely clipped and distorted. Ah-Ha! Using a DMM with a diode test range, it looked like two of the four 1N34A diodes in the balanced mixer ring were bad. (One open, one shorted.) I unsoldered one end of each diode to confirm this. Yes! I replaced all four diodes CR1,2,3,4 with the modern 1N4454 equivalent.

CR1, 2, 3, 4

Voila! Plenty of clean audio right through. This was the main problem all along!

I quickly readjusted the carrier balance, reset the ALC zero, ran coax from my workbench up the basement stairs, and connected up my antenna, a trap dipole for 80/40/20. I heard W8ERN in Michigan on 7.184 MHz, called him, and he came back with a 5x5 report. The band was extremely noisy, but I was ecstatic! I asked W8ERN to indulge me while I switched from the hand microphone to the D-104. The report was "Very, very good audio". K3MRG on frequency also heard me and said I was 30 over S-9, with great audio (using the D-104) .

Having seen the failure of the 50-year-old dry paper electrolytics in the power supply, I decided to replace the paper electrolytics in the KWM-2. Referring to the Collins 1978 KWM-2 manual, in change #3, C264 was changed from 4uF/350V to 20uF/350V, so I used the new value. I also replaced C254, 4uF/350V, another dry paper electrolytic, C102, 100uF/16V, and C259, 8uF/25V.

I decided to leave C106 the 3-section metal can (wet) electrolytic alone since there was no hint of trouble (hum or excessive ripple) with it.

I noticed that C102 had not been the proper value according to the schematics, and didn't look to me like it was a Collins part. And it had no red glyptol on its connections.

Back on the air, a few more contacts confirmed that nothing had gone wrong while changing the electrolytics.

While changing bands, I noticed that I had to 'diddle' the band switch to get it to make contact reliably in the three lowest (3 MHz) band positions. An inspection showed that the 'rotor' of the wafer switch came to rest with its sweeping contact not quite centered under the outer contacts.

In this switch, the shaft is removable (in order to allow the aluminum coil covers to be installed and removed) and so the shaft angle is not fixed relative to the mechanical detent. So I loosened the coupling of the shaft to the knob detent, rotated the shaft ever-so-slightly to center the rotor contact under the nearest fixed contact, and re-tightened the coupling. Now the band switch engages reliably in all positions.

Although the band switch contacts look tarnished, I resisted the temptation to clean them just to make them look pretty; but I did apply a little DeOxit.

The KWM-2 service instructions suggest a complete alignment if the rig hasn't been used in a long time, so I proceeded to do the 'Laboratory Alignment'. I followed every step of the procedure. The rig was now calibrated, aligned, neutralized and back on the air.

Several KWM-2 owners recommended adding a fan to keep the rig cool. This sounded like a good idea. Following some of the published articles, I fabricated a bracket and mounted a 3" 12 volt DC fan in a completely non-destructive and reversible way.

It is powered with a full wave bridge and filter from the 6.3 VAC available on filament pins on the noise blanker socket. It runs fast enough to move hot air, but slowly enough not to be too noisy.

Who let the smoke out?

A few days after moving the rig off the bench and into the shack, I was listening to a QSO while working on something else nearby. Then suddenly -- a whiff of something burning and smoke pouring out of the KWM-2. I immediately shut it off. Another "Oh no." Back down in the shop, and on the bench, there were no obvious signs of anything burned.  I knew that the problem must be in the receiver side, so I disabled the high voltage and screens by popping out the jumpers. Since the power supply was solid state, I gradually brought up the line voltage using a variac and soon found that the source of the smoke was inside the enclosed relay compartment. The 68 ohm 1 watt, series resistor, R157, in the receiver B+ line, had burned up.


But why? It was clear that something had shorted out causing the overload that burned out R157, but what? I turned again to the pin-by-pin resistance chart in the KWM-2 manual, and proceeded to check each tube pin's resistance to ground. (This was the second time I did this. The first time was before I powered the rig on for the first time, at which time everything checked out within reasonable tolerances.) This time, everything checked out okay until I came to V12, pin 1, which was a almost a dead short to ground instead of 10k. This pin is connected to R159, R160, and C237. It took me a long time just to find R159 hidden behind several wires. I carefully moved the wires to get a better look and found that R159, a 1k 1/2 watt, had burned up.


But why? R160, 120k was good, so that left only C237, half of a dual 0.01uF ceramic disk. I clipped out the half of C237 that was connected to V12 pin 1, and found that it read a solid 7 ohms to ground confirming the culprit!  I removed the dual C236/C237 disk completely and replaced it with two 0.01uF, 1000 volt disk ceramics, and I replaced the burned out R159. Now V12 pin 1 had the proper resistance to ground. I inspected R68, 15k, which would have blown if the other half of the dual disk, C236, had shorted, but it was right on. Only one half of the dual 0.01 had shorted. Finally, I replaced R157, the 68 ohm, 1 watt resistor in the relay compartment.

I powered up and the receiver came to life. All the smoke was back inside the components where it belongs.

In the week or so since completing the repairs, I have worked eight countries and about 20 domestic stations using a trap 80/40/20 dipole only thirty feet up. All-in-all, I enjoyed every minute I spent trouble shooting this rig and look forward to meeting you all on the air.


    -Al Schapira, W2ADS

Most of the elapsed time taken to repair the rig was taken up by waiting for parts to arrive from Mouser. They're very fast, but unfortunately, they can't ship parts until *after* you place your order. :-)

Please feel free to contact me at

Page last modified 12/07/15