Heathkit SB-101 Rebuild and Restoration Notes
By Chris Codella, W2PA
November 2006
My old Hot Water 100 (as the Heathkit HW-100 transceiver was once called) sat untouched for at least 15 years. I had bought and assembled it in 1970 and used it intensely as a teenager and in my early 20s - it was my first SSB rig. I had last used it regularly back around 1980, then only sporadically as I started work and a family - and moved on to newer equipment. In 2003, after being interested in vintage gear for a few years, I began to think about restoring the old rig and got it working again. Realizing that parts would be increasingly scarce, I decided to buy a “junker” rig as a parts source. Since the HW-100, 101, SB-100, 101, 102 are identical at a basic level, those are the rigs I began to look for. In 2005 I found this SB-101 on eBay, listed as a “parts rig”. It looked pretty rough, missing the mode switch, scratched cabinet, cracked meter - but since I was only interested in parts, I bought it for $35.
The rig was in as bad a condition as I expected. Besides what I saw in the eBay pictures, it was dirty, with a layer of sticky dust, and some kind of yellow-brownish film on some parts. The wiring harness felt sticky too and it had whitish spots all over it, which I surmised was some kind of mold. Green deposits on some of the wire connections indicated that it probably had been stored in a damp environment, maybe someone’s basement, for quite a while. Naturally, the rubber belts linking the Driver/Preselector and Load controls had long ago become dried out and were useless. The main tuning knob didn’t rotate even one full revolution before slipping, and the main kHz dial was cracked right through the “15” mark to the edge, a common cracking point in the SB-series dials (see photo below).
On the positive side, the front panel looked pretty good with hardly a blemish. All the tubes were there, and as a pleasant surprise, it had both the SSB and CW filters! I originally had planned that the first thing I would do would be to retrofit my HW-100 to use the filter switching mechanism from this ‘101 along with a CW filter I had purchased earlier. I now had an extra CW filter.
So I removed both filters and all the tubes, stored them away in a box, and set the rig aside to be used when and if I needed parts.
During the next year or so I removed exactly two more parts: the cover shield from the bandswitch boards, and the 29.5 MHz heterodyne oscillator coil. I sent these, free of charge, to two other hams who were looking for them to revive their own transceivers. I figured I probably wouldn’t need those parts anyway and I might as well help someone else.
Getting the HW-100 up and running got me thinking about the junker SB-101. It was nearly all there – I wondered if I could get it working too. If I did, I’d first have to replace the parts I had removed or given away, in addition to finding replacements for the missing switch, the rubber belts, and broken meter and dial. I began looking again. Within a couple of weeks I had bought a collection of Heathkit transceiver chassis parts, including the missing shield, for a few dollars. I still needed that coil and the mode switch. I ordered replacements from “oldheathkitparts.com”. This was enough to get me going.
I started by removing the knobs, controls and front panel to get to the interior more easily. I had intended to just do some cleaning, replace a few parts and start troubleshooting. But the more I dug, the more digging I did. Whoever had originally assembled the rig had done an awful job. I removed the bandswitch boards and the final cage. I found burned insulation, crooked parts, bad solder joints. It looked as if someone had done some repair work at one time, replacing some parts on the audio board, where there were traces of some sort of thermal catastrophe. Incredibly, there were replacement components that someone had soldered directly to remnant leads of old parts instead of replacing them at the original connection points. What a mess! I continued removing things. I took off the dial assembly and the LMO. I started to clean out the grimy chassis but it was just too difficult to get to where the grime lived. Recognizing I was crossing a threshold, a point of no return, I decided to completely remove the wiring harnesses since there was no way I’d be able to really clean them in place. I also didn’t like the idea of spending time getting this rig on the air with a moldy wiring harness.
It was important to maintain the harness wires in as close to their original length as possible, to make sure they would go back in properly. I therefore unsoldered all the connections instead of simply cutting them off, which would have also left little pieces of wire soldered in all over the place. This took a while. I also saved the individual pieces of thin coax that were not part of the coax harness.
With the harnesses now out, I could see everything a lot better. I kept going. Most of the components in this rig are on the PC boards. Removing all the remaining discrete components from the chassis – resistors, capacitors, chokes, diodes – filled only one small plastic cup. Why stop now? I proceeded to remove the boards, the remaining connectors, terminal strips, relays, final amplifier components, and chassis separators, carefully saving everything except for short runs of interconnecting wire.
I now had an SB-101 kit, disassembled except for the parts on the PC boards. This should be a lot more fun than just troubleshooting! But first I had a lot of cleaning to do. The parts, boards, and hardware were as grimy as the harness.
The first cleaning job I tackled was the wire and cable harnesses. I decided to try soaking them in a plastic tub filled with a solution of warm water and dish detergent. After soaking for a few minutes I scrubbed the harness with a toothbrush from one end to the other, underwater, then rinsed it off thoroughly. Although the colors had faded over time (more about this later), it came out much cleaner and not sticky any more – and the mold spots were completely gone.
Next, I went after the PC boards, which were quite dusty and grimy. Various soaps and detergents on Q-tips worked partially but were not very effective. I tried isopropyl alcohol, but that simply seemed to dry out the goop leaving behind a whitish film. On the advice of my old friend W1JA of Radiophile.com fame I tried lacquer thinner, which worked like a charm. You have to do this outside because the fumes are quite potent. I poured a little thinner into a ceramic cup, then used a small paintbrush to scrub the PC boards and their parts. The grime magically disappeared after two such applications, rinsing the brush out in between. I’m not sure if it evaporated off, ran off (I didn’t see much dripping), or came all off on the brush, but it was gone, and the markings on the components were left untouched, except for the ones on a tubular capacitor on the audio board. I scrubbed the foil side too. This didn’t make it look like new – there were still spots of corrosion here and there, but the sticky stuff was gone. I decided not to try anything harsher because I didn’t want to risk damaging the foil or components. I did all this outside on a warm day and the boards dried completely in minutes.
I then used lacquer thinner on the cup full of parts and the cup full of screws, nuts, and lockwashers, with similar good results. Swishing the parts around in a cup with the brush made the liquid turn brownish and the parts came out clean. I was careful not to do this with any of the non-metallic parts as I wasn’t sure what the effect would be. For example, the thinner might have attacked the rubber and plastic grommets.
The bare chassis was in fairly good shape. I cleaned it by scrubbing lightly with a kitchen steel-wool soap pad and hosing it off. This left the aluminum with a new “brushed” look. I could have used a polish, I suppose, but I was satisfied with this look. You have to be careful not to scrub the rear panel, however, to preserve the labels.
The rig was now ready for re-assembly. I bought a manual reprint from W7FG, who made very nice ones. It was nearly perfect – all the little check-off spaces were unchecked.
I started at the very beginning (a very good place to start). Since I hadn’t disassembled the PC boards, I followed the steps anyway and checked that the parts were in good condition, reasonably within tolerance, and soldered properly. In the process I found 4 components that were of the wrong value – not out of spec but apparently substituted by someone. I removed these and replaced them with parts of the right values. After I “completed” each board, I removed the solder from all the holes where the harnesses and other wires would be connected later.
Before starting to reassemble the chassis, I also removed the excess solder from all the controls, switches, connectors, and terminal strips, so they’d be ready to accept parts as almost-new (the terminal strips also got the lacquer thinner dunk treatment). I also cleaned the switches and controls with DeoxIt.
Mechanical reassembly went mostly without a hitch. I cleaned the other chassis pieces as I got to them. My old Heathkit red plastic nut-starters came in handy, especially when it came time to remount the PC boards with all those little #3 screws, nuts, and lockwashers.
As I began to reconnect the wire harness I noticed the effects that 37 years or so of aging had on the insulation colors. For one thing, the white background color of the striped wires had turned a light tan color. The other colors had faded a bit too. The worst ones were the yellow wires, which had faded so badly they looked just like the white wires.
Luckily, there are no completely white wires in the harness, so I was able to deduce which wires were supposed to be yellow. The blue and gray wires were beginning to look somewhat alike too – the blue ones had all turned dark gray and the gray ones were still gray but a lighter shade. All of the wires with color stripes were easy to distinguish since the color stripes themselves had not faded at all. The harness wires went back into place without much trouble.
Having been careful about unsoldering the harness wires during removal, there were very few with ends I needed to strip, but many wire ends needed a bit of cleaning to ensure that the solder would wet them. I used very fine finishing sandpaper, pinching the wire end between two sides of a small folded piece, with two fingers applying pressure and a rotating motion axially around the wire end.
I only had one case in which a PC board foil pad came loose. It was a small island used only for a binding spot between two wires, connecting to nothing else on the board, and thus there were no board-mounted components to hold it in place. I was able to keep it in place by feeding its connecting wires through to the component side and bending them over to keep it from pulling loose. With two wires thus inserted it was stable. If I had it to do over again I’d probably put a dab of super glue under the pad too.
When it came time to put discrete, point-to-point components back in, I discovered that it was relatively easy to figure out which components of the same value went where, judging by the length of their leads and sometimes how they were bent. I had not disassembled the capacitor-resistor combinations used for bypass on the final amplifier tube sockets, so they went back in as assembled units.
In the process of reassembling the front panel I became an expert on the SB-series dial mechanism out of necessity. Although the assembly instructions are detailed, mostly clear and well written, the resulting assembly won’t necessarily work very well if you start with a non-working mechanism as I did. To figure out which pieces were out of spec, I had to understand how it worked. Damaged or dirty parts can also cause problems. Mine were very grimy so I started out by cleaning them. Even the “new” kHz dial I had bought needed cleaning.
The mechanism is surprisingly simple. The main kHz dial is mounted directly on the LMO shaft, which can make five complete revolutions covering 500 kHz. There is a spiral track on the back of this dial that progresses radially outwards as the dial rotates clockwise (as viewed from the front). In this spiral track rides a nylon pin, like a turntable needle riding in a vinyl record groove, but moving outwards from the center rather than inwards as the dial rotates from 0 towards 100. The pin, in turn, causes a lever to which it’s attached to move outwards and towards the right as the dial rotates, rather like the tonearm of a turntable (but, again, backwards). This lever, in turn, causes the slide rule dial pointer at the top to move from left to right and mark off 100s of kHz, one for each complete rotation of the kHz dial. That mechanism alone completely maps one-kHz calibration onto the LMO shaft position. However, 100kHz per revolution is too fast a tuning rate for normal operation so it must be geared down.
Instead of using a set of gears to slow down the tuning rate at the knob, Heathkit designed a much simpler arrangement, based to a large degree on the Collins S/Line dial mechanism (see my other article that discusses the comparison). The knob shaft, at its opposite end inside the front panel, has a pair of thin spring-metal disks separated by about 1/8 inch on the shaft. This pair of disks straddles the inner edge of a metal ring mounted on the front of the main dial. Since they touch the inner edge of the ring, the dial turns in the same direction as the knob (whereas if they were touching the outer edge, the two would turn in opposite directions). Even though there are no gear teeth involved, and the pair of metal disks merely touches the dial ring, it’s enough to smoothly rotate the dial. The knob shaft is secured to the front panel with a brass bushing. The pressure the metal disks exert on the dial ring is adjusted by moving this bushing up or down very slightly before tightening it in place. Moving it downward makes the tuning stiffer. If you move it down too much, the pair of metal disks bend apart and cause extra wear on the whole assembly. If you move it up too much, they lose contact with the dial and begin to slip.
The kHz dial hairline, or fiducial mark, is engraved in a clear plastic window attached to the center of the main dial that rotates around its bushing which mounts on the LMO shaft. It’s held in place with a snap ring so it can rotate freely. When the assembly is mounted to the LMO, the hairline window is held in position by an assembly similar to the one for the main tuning knob. A pair of smaller metal disks is positioned to contact the curved upper edge of the hairline window and is attached to a small shaft that extends through the escutcheon to a small metal knob. When you turn this knob, it causes the hairline window to rotate, moving its position relative to the kHz dial. The trick to making this work right is to tighten the knob’s setscrew on the shaft such that it tends to pull the hairline window away from the main dial. That way it doesn’t rub up against the dial as it turns.
Now that you know how it works, I can describe what was wrong with mine (aside from being dirty and cracked). I imagine these problems are common in all SB-series units. All of the problems with the main dial itself arose from misplacement of the various parts and were solved by closely following the assembly instructions. But sometime in the past, the metal disk at the end of the hairline adjustment shaft had become disconnected and someone had tried to solder it back on. But they had not properly pushed it down onto the shaft until it hit the “shoulder” that keeps it spaced properly from the other disk. The spacing was so large that the disks didn’t grip the hairline window tight enough to keep it from rotating on its own and so it would not stay where you set it. A simple fix was to push it down on the shaft with a hot soldering iron tip. It took a couple of tries to get it fixed parallel to the other disk.
Finally, I arrived at the bottom of page 78 of the 151 page manual and was greeted by the words “This completes the wiring of the Transceiver.” Cool! The assembly was not quite done yet but the long process of disassembly and reassembly was almost complete.
The resistance checks uncovered a mistake. The bandswitch boards must be assembled with their rotors oriented properly. Their center holes, through which the shaft passes, are not round – each one has two “flats” that match the cross section of the shaft and a little notch on one of the two flats. Thus they can be oriented in two ways, with the notch either up or down. Three of the four boards go in with the notch up and one goes in with it down. I had not read the instructions carefully enough and put them all in pointing up, even though the boards have the proper orientation printed right next to the switch wafer. It was easy to correct.
Next step was to do the initial power-up checks. I immediately knew something else was amiss since upon turning on the power neither the pilot lamps nor the tubes lit up. Also, after just a few seconds of power, the power supply cable was a bit warm. The resistance checks had failed to turn up a short in the filament supply wiring. Since the filament wires go all over the place, it takes some time to locate such a fault and since the schematic only indicates connectivity, not the routing of actual wires, it’s of limited help. The 12.6 V filament supply is split into two tiers of 6.3 V each. Some filaments are connected from tier 1 to tier 2, some from tier 2 to ground, and the 12.6 V ones go across the whole thing. Resistance checks indicated that the top tier – the brown-white wires – was shorted to ground, whereas the second tier – the brown wires – was just fine. The next step was to unsolder the brown-white wires one at a time to try to isolate the short. This one was caused by one of the wires in the harness rubbing too tightly against one of the PC board mounting screws and shorting to it. Again – it was a simple fix.
Next try, everything lit up fine. In fact, all the basic supply voltages seemed fine. I even could hear a hiss from the audio section, and a faint signal from the calibrator. RF and AF gains, mode switch, even preselector, all operated as expected. This meant that most sections of the receiver were confirmed as operating. However, I noticed two new problems before even following more check-out instructions.
The LMO was unstable. I traced this to the output coax lead which made the frequency jump around wildly when I moved it. This was simply a contact problem I cleaned the phono socket collar and the ground shell of the plug with Deoxit and everything was fine.
Next problem: the relays did not energize – neither the PTT line nor the Tune position of the mode switch caused them to operate. I traced the problem to the grey wire power lead going to RL1, which had become disconnected.
Now for alignment. Following the procedure in the manual went fairly smoothly – everything peaked up close to where it was originally set. But I noticed two problems. First, the calibrator signal fell off in strength dramatically on bands above 7 MHz. By the time I got to the 29.5 band it was barely moving the s-meter. To check whether this was a receiver problem or a calibrator problem, I used a signal generator at a fixed level to check sensitivity on all the bands. This confirmed that I was just getting an ever weaker signal out of the calibrator. I’ve not yet tried to figure out why, and will have to go back to it later.
Getting the calibrator set correctly is important because if it’s not, the VFO will not line up well from band to band. Get it right and you can change bands without having to move the main dial pointer very much at all. In fact, I’m very impressed with the SB-101 LMO. It’s very linear – calibration hardly changes from one end of the band to the other – and after warm up it’s surprisingly stable. Heathkit did a good job of design and the manufacturer, TRW for this particular unit, did a great job of picking the right temperature compensating or tracking components.
There was only one hitch in aligning the transmitter. I found that while the manual instructs that you should get full scale grid current after tuning up, on some bands it was significantly lower – even though the output power was normal. I discovered that the position of the driver neutralizing wire had a big effect on this. By changing its position around I was able to get equally reasonable levels of grid current on all bands. By the way, this wasn’t an issue on the HW-100 since there’s no built-in way to measure grid current on that rig.
To finish things off, I installed new feet, ordered on the Internet. I also cleaned, then oversprayed the cabinet with a matching SB color paint I bought on eBay. Heathkit used various shades over the run of the SB-series, so it was a close match (since there was no specific target to match).
It was a fun project that brought back memories of building Heathkits years ago. Below is a picture of my "new" SB-101.
Copyright © 2009 Christopher F. Codella, W2PA. All rights reserved.