Restoring a 40 year old Soundcraftsmen PM860 amplifier

Back in the mid 80s I was experimenting with DIY loudspeakers a lot. I wanted an amplifier that produced a lot of power and was indestructible, so I bought a Soundcraftsmen PM860 for about $400. It served me well, and never complained about anything I connected to it, including DIY electrostatic speakers and woofers. It sounded great, too!

Basic specs: The PM860 is a two-channel, 205 watt amplifier (each channel, 8 Ohms) with a power bandwidth of 20 Hz-20 kHz. IM and THD is 0.05% at 1/4 watt and full rated power. Frequency response is 20 Hz-20 kHz, ±0.1 dB; s/n ratio is 105 dB, and high level sensitivity is 1.5V. Features include a 2-speed fan and clip indicators for each channel. The output stage has 6 power MOSFETs in each channel. Dimensions are 5 x 8.5 x 14- weight is 22 lbs.

Amplifier front before cleaning and recapping.

I used this amp all through the 90s, even after I stopped building speakers, and then it ended up in a storage box. I recently pulled it out to see if it was still working and ran into those darned 1/4" TRS input jacks (why'd they do that?) and couldn't find the adapters that I used to use, so I replaced them with some gold plated phono sockets. I measured the DC offset at the outputs and found -7 mV on channel A and +11 mV on channel B - no problems there, so safe to connect speakers. It worked perfectly and sounded great, however, I don't trust 40 year old electrolytic caps, so it's time for another recap job..,

Rear of the amp after replacing the TRS jacks with phono connectors, but before cleaning and recapping. Normally the fan turns slowly and almost (you can't hear it from more than about 6" away) silently, but if the heatsink or power transformer heat up, the fan speeds up. 

Here's the deal with electrolytic caps: they have a limited lifetime based on operating temperature, voltage, and current. If you operate the cap below any of those specified values, the lifespan increases. As with almost all things electronic, temperature is a huge factor in the lifetime. A basic rule of thumb is that for every 10C drop in operating temperature, you double the life of the capacitor. So a capacitor that is rated for 3000 hours at 100V @ 5A @ 85C should last 6000 hours at 75C, 12000 hours at 65C, 24000 hours at 55C, etc. So when you see capacitor specs with a lifetime of 2000 hours, don't be alarmed (unless you intend to abuse that capacitor).

When electrolytic caps sit unused for a long time (like those in this amplifier had done when it was in storage) it is best to power the device up slowly using a variable transformer to ramp the voltage up and allow the caps some time to "form" the dielectric. I didn't have a variable transformer so I just powered it up. Fortunately, there were no problems.

I hunted down a set of schematics online (that you can DL here). Pages 3 and 6 apply to this amplifier. This amp uses something called phase control regulation to limit inrush current to the power supply and to keep the DC rails at +/-70V. This video has a good explanation of the power circuit at about 13 minutes in:

When you first turn the power on, you can hear the PCR circuit working- it charges up the main filter caps by turning the rectifiers on and off so you hear a quiet thump thump thump sound coming from the power transformer until the caps are fully charged.

PM860 power supply schematic with electrolytic caps highlighted in yellow. The lower right quarter of the schematic is just controlling the fan speed based on the temperatures at three sensors (mounted on the two heatsinks and transformer) at the middle of the page. 

One channel of the PM860 amplifier schematic with electrolytic caps highlighted in yellow. The other channel is identical. Note the input capacitor C1 (far left side of schematic) is a nonpolar electrolytic.

I started the usual way, by making a list of all the electrolytic capacitors in the amp, then checked against the schematic, measured the sizes of existing caps to ensure the new ones are going to fit, and ordered from Mouser. I used Nichicon UKW series parts where I could, but a few weren't available in that series so I selected others that were good, low ESR, long-life subs for the originals. 

For most of them I ordered higher voltage caps than the originals. The main filter caps were 11,000 uF at 75v with amplifier rails that sit at about 72V. That's not a lot of headroom. One thing that kills electrolytic caps quickly is operating at too high voltage, even for a few seconds. I ordered 13,000 uF caps rated for 100V that are rated for 18,000 hour lifetime at >7A ripple current at 85C. This amp normally runs cool, so I expect those caps will last a many years longer than I will. Assuming the operating temperature is 55C (it's actually probably closer to room temperature most of the time), those caps should last 8x18,000= 144,000 hours. That's 24 hours per day for 16.4 years (at rated voltage and current). At a more typical 500 hours per year, that's 288 years! Of course I wouldn't really trust them to last that long, so I'm planning on replacing them again in a hundred years or so. The rest of the caps will have to be replaced again much sooner. Here are the parts I ordered, a little over $70 with tax and shipping:

The capacitors ordered from Mouser Electronics

Step 1: Clean it up!

Inside the amp before disassembly and cleaning. Most of the amp board is under the fan shroud. The power supply board is on the lower right. I don't think they prioritized serviceability when they designed this thing.

The circuit boards and the chassis had quite a bit of dust and flux residue on them, so I decided the best way to proceed would be to take it all apart, clean it up, recap the PCBs, then put it back together. There were a lot of wires running between the boards, and no markings on the wires or boards, so before disassembly I took a lot of photos and made drawings and notes so I'd have a good chance of getting the whole thing back together in working order.

Once I had the PCBs out, I cleaned them with IPA, a toothbrush, and skinny bottle brush. I put IPA in a spray bottle and sprayed, and scrubbed, and sprayed some more. In the end, I got the boards looking better than they did (probably) when the amp was new.

Amplifier board before cleaning. This is the part of the board that sits under the fan. Having a fan blowing air through the amp all the time leads to accumulation of dust.

Amplifier board after cleaning.

This is the side of the amp board closer to the power supply. This is where most of the connections to the power supply board are made. The power supply board overlaps this area so you can't see much of it in the photo of the assembled amp with the cover off.

Power supply board before cleaning. Not too bad, really, but this board was easy to reach with a vacuum cleaner brush during the years I was using the amplifier.

Power supply board after cleaning, before recapping.

Step 2: Recap

I tested most of the new caps before I soldered them down, and didn't find any out of spec. I was not able to test the bigger electrolytics that were outside the limits of my LCR meter.

Ordinarily I would prefer to recap one section of the circuit at a time, test, then move on, but the complexity of the wiring and mechanical layout in this amp made that impractical, so I just shot-gunned it, replacing all caps in one go, starting with the power supply board, and kept my fingers crossed for luck. I was more worried about reconnecting all those wires than I was about getting the caps right.

At each cap, I verified the capacitance and voltage, and used a sharpie to put a dot on the PCB at the negative terminal before removing the original cap. Then I desoldered each cap using a manual solder sucker to remove the solder from the board and installed the new ones, one by one, until all the caps were replaced. The solder sucker is made in Japan and has a soft silicone tip that seals well and almost never fails to suck out all the solder on the first attempt. The silicone tipped tool works much better than the old PTFE tipped tools I used to use.

Power supply board after recapping. Note the dots on the PCB near each cap to indicate where the negative lead of the original caps were placed. Note: all the new caps have pressure relief folds in the tops of the cans- if they ever fail they won't spray corrosive electrolyte all over the PCB. The cement resistor on the left is what keeps the fan running slow and quiet most of the time. That resistor runs HOT and is stood a few mm off the PCB.

Old power supply filter cap, left, and new one, right. There are two important size differences- the overall height of the caps and the height of the terminals over the tops of the caps. The first requires a spacer under the new caps to bring them up to the height of the old caps, and the second requires adding some spacers (washers) between the new caps and the bottom of the PCB to make room for the hardware that holds the caps down in the chassis.

This is how the original caps were held in place in the amp. Who thought this was a good idea? The new cap terminal posts are short, and if I mount the power supply PCB on them as-is the cap hold-down nut will contact the bottom of the PCB. I fixed that problem by soldering stacked copper washers to the bottom of the power supply PCB.

The PCB really needs to be where it was, vertically, with the original caps- any higher and it will bump against the top cover of the chassis and much lower and the PCB will hit the coils of the power transformer. I designed and 3D printed a spacer for the bottom of the caps that lifts them close to their original height so the power supply PCB will clear the top of the power transformer. Then I added three copper washers as spacers on the caps' screw terminals so the PCB would clear the bolt holding the caps down. Sheesh!

If I didn't care about maintaining the original form-factor of this amp, I'd rebuild it in a new chassis with a more reasonable layout (the fan would blow air down the middle of the heat sink, the power supply board would be located a little further from the power transformer so the board wouldn't overlap the coils, etc. 

This is the spacer I designed and printed to lift the power supply filter caps to their original height. It has the added benefit of preventing the caps from sliding on the bottom of the amp chassis.

Here you can see the new filter caps, the 3D printed yellow standoff for them and you can see the copper washers on top of the cap terminals to ensure that the PCB clears the cap hold-down nut.

The Chassis

I didn't bother to do anything but clean the chassis- there was only a minor ding at one small spot on the top edge- hardly worth the trouble to try to refinish the whole thing.

All Done!

Once I finished recapping, I crossed my fingers for luck and powered it up. One small 10uF 100V cap on the amp board instantly exploded! I took it all back apart and discovered that I had put that cap in backwards. Electrolytics don't like being connected with reverse polarity! Fortunately, no other harm was done and after triple checking the part location against the schematic, replacing that cap and reassembling, the amp worked perfectly. It should be good for at least another 20-40 years and will probably outlive me. Now I have to decide if I'm going to keep it or sell it. Hmmmm.