March has been an interesting month, with a wide variety of equipment coming through, but amongst the most challenging have been a Rev 3 Prophet 5 and a Korg Polysix.
This particular synth is a Rev 3.3 with 120 Presets, one of the later ones made in July 1982.
It used to belong to a hire company here in Dublin, and was for sale for quite a while through the local internet trade site. Unfortunately during this sales drive, it died totally, and one of our clients took the gamble of purchasing it at a much reduced price in the expectation that it could be repaired.
The electronic repair of the synth was the easy part. An important chip in the microprocessor section had failed but it wasn’t too difficult to diagnose and replace. After that it booted, tuned and played fine, there is no issue with expensive CEM chips and all 5 voices seem to be perfect.
The main issues with this synth turned out to be related a period of non use and storage in a less than ideal environment, probably cold and damp and not in a flightcase. Also the synth had heavy things dropped on some of the pots (or had fallen heavily on something else) as there were dents on the front panel at both ends of the panel and a couple of broken pots.
It was time to dismantle the front panel pcb’s.
Apart from obvious damage to a few pots, the main concern was corrosion of the pot bodies, it took WD40 rust release spray to even unbolt the pot locking washers to release the front panel boards.
The pots themselves were completely coated with a white powder from oxidation of whatever plating they originally had, and the lubricant grease had turned to something resembling candle wax.
As a result, all of the pots had to be desoldered from the circuit board, dismantled, cleaned, regreased and reinstalled onto the panel boards
The hardest and most time consuming part of the operation is removing the pots from the P5’s front panels, as they are plated through holes with very large track pads, no fear of dry joints with a Prophet front panel. Care must be taken to clear the pots’ mounting pins on both sides of the circuit board, as dragging the pots out will break tracks and you don’t want that to happen.
Replacing the pots afterwards is also tricky as they have to be carefully aligned for height and flatness as the pot bodies float in large holes in the front panel pcb’s, they don’t just lie flat on the boards.
The pots themselves are made by a company called Centralab in the US, not a name we’ve come across often, but the pots themselves are a joy to work with. They dismantle easily into 4 parts which can be individually cleaned and greased, and they reassemble neatly and without fuss.
With the rebuild of the front panels complete, it was time to take a hammer to the front panel.
The metal front panel was removed from its walnut sleeve (don’t hammer a Prophet while its wearing the walnut). The dents were around the Polymod Osc pot at one end, and around the Amplifier Release pot at the other. Pieces of 50mm square wood were used to cushion both sides of the panel and to prevent scratches as Harry the hammer (our financial controller!) did his work.
With that done,it was time to reassemble the front panel with its pcb’s which went smoothly, all the pots lined up with their holes and the WD40’d locking nuts tightened up nicely.
As this project moved from being a repair to becoming a restoration, the keyboard itself was next to get some TLC. A set of keyboard bushes were obtained from Arcsound in the US.
We were a little sceptical at first with the claims made about new bushes on a Pratt-Read keyboard, but we were pleasantly surprised by the results.
The old bushes were a shade of grey (originally they were black) as the lubricant on them dried out and turned to a powder like talc. The black and white key tops had to be removed, and a sharp craft knife used to cut out the old bushes.
When all 61 keys were re-bushed, we were impressed by the smooth action of the keyboard, and particularly the absence of key clack during fast releases, a feature of most old P-R keyboards.
Final assembly and test are imminent as soon as the benches can be cleared…
Needless to say this went well and the synth has gone back to its owner
The other intriguing synth of the month is a Polysix.
Bought as a fixer upper, the symptoms all pointed to the leaky battery and acid problem.
The Programming system didn’t work, all the presets sounded the same, and random lights came on in the programming section, and a lot of the programming switches didn’t work.
The obvious first port of call was the pcb near the battery, full cleaning with an Ammonia based cleaner followed by Isopropanol to remove any traces of the original battery acid.
Ironically some previous owner had taken the original battery from the CPU board and put a Lithium battery in a custom holder to the right of the Voice card. Unfortunately the damage was already done and the modifications to the CPU board that are necessary to cope with a Lithium battery were not done which meant that the large new battery was also dead on arrival.
Taking out the KLM367 CPU board after some preliminary diagnosis told us what the internet folklore says. IC31 was faulty and lots of tracks on the topside were open circuit due to the battery acid eating through them.
Careful and thorough point to point measuring of the track resistance, and also checking plated through hole resistances, the replacement of IC31 and the use of Kynar wire links to replace damaged tracks got us to first base (repairwise).
Now the synth booted and the programmer section did most of what would be expected. At this juncture it was obvious that a lot of the tact switches in the Programmer section were faulty, another clue was that the pcb the switches live on was severely bowed inward from endless repeated attempts to make buttons work. A new CR2032 coin cell holder (and battery of course) were installed on the KLM367 to take over the backup duties.
The Programmer section looked good and the arpeggiator was working well, so time to maybe load the factory presets into it from a Laptop file. The file appeared to load faultlessly but there were still no presets, just the random gibberish from before. More Scope and Logic Analyser analysis said the Memory chip was dead, none of the tracks were, that had been checked thoroughly in Phase 1 of testing.
The old RAM chip was removed , a socket put in and a 6514 RAM chip from a Moog Source put in as a temporary measure. The presets were reloaded and lo and behold they were there, even after a switch off and back on again, and they have been since.
The next major niggle was that only 3 voices made any noise, and they were very different from each other now that the factory presets were present and correct.
Now for an additional surprise, there was acid damage on voices 5 and 6 on the KLM366 voice card.
Much thinking was done and some consultation with other Polysix owners was had to figure out how battery acid could leap from the left hand side of the CPU board to the bottom right hand side of the voice card over 50cm away without damaging anything inbetween. There was no storage scenario which could lead to those effects.
Voice 6 was dead and Voice 5 was flaky, so a thorough investigation of the tracks in that area was undertaken, as it turned out there were more damaged tracks on the voice card than on the CPU card,
After considerable thought it struck us that whoever took the original battery from the CPU board had replaced it with equally dangerous battery technology in a custom holder to the right of the Voice card which did the same damage there. We are thinking Zinc batteries for the replacement as we had to cut out and replace not only the resistors in the Voice 5 & 6 section but also the IC’s and most of the transistors.
Having got this far, it was time for some calibration, the acid test to see if a synth is working properly.
It wasn’t, we now had 6 oscillators working perfectly but still only 4 voices, two of them just produced thumps when activated. (but please see addendum at the foot of this section).
As it surprisingly turned out, 5 out of the 6 filter IC’s were failed, and the 6th wouldn’t calibrate properly, it would make sound and some resonance but not enough. In more detail, 2 of the filter chips (SSM2044’s) had no output at all and 3 others had no resonance control. The 6th IC had output and resonance but ran out of bandwidth far too soon.
Lots of double, treble and quadruple checking was done to see if we had missed something, but no we had not, we also could not believe that all the filter chips were faulty, but they were.
Our client procured a full set of filters on Ebay for a reasonable price and when they arrived we installed them. Immediately there were 6 voices and they all calibrated beautifully.
While waiting for the filter IC’s and a complete set of 19 tact switches for the programmer section to arrive, attention had to be paid to the other issues that the synth had.
The keyboard had multiple faulty keys across its entire span,which obviously required a complete strip and clean of the whole contact strip pcb (several times). Quite commonly with many of these older rubber contact keyboards, if you fix one key problem, another will appear elsewhere and it often takes many attempts before the whole keyboard works properly..
The volume pot was unusably intermittent and scratchy, needing to be removed from its front panel pcb, and then dismantled, cleaned and rebuilt. Surprisingly enough, both output sockets were so corroded and dirty that they too didn’t work. They had to be removed from the rear panel and the contacts cleaned and polished with metal polish.
The addendum is the latest bad news from this Cerberus like synth.
Each time the synth was turned on, Voices 5 and 6 would be out of tune with the others, they could be recalibrated fine and would work for the time the synth was switched on but after a switch off and back on again they were way out of tune again. As the Polysix has no Autotune feature, this is something that has to be rectified.
After removing the voice card yet again, and cutting out the remaining IC’s around Voice 5 and 6 along with the remaining resistors and more transistors (the reason being not so much because the components themselves are faulty but because it is almost impossible to thoroughly clean the pcb with the components in place). Also, the Oscillator section of the Polysix is very densely packed and even getting meter probes in there is almost impossible, added to the fact that the acid corrosion makes the outside of the solder joints an insulator, thus making continuity and resistance measurements at least unreliable, and at worst impossible.
As of early May, voice 6 is fine but voice 5 is still flat and non-linear, not by much but still not right.
Tomorrow is another day…
Another day dawned, and the keyboard was taken out once again to allow test equipment access to the oscillator section. Apart from the K30A O FET, everything in voice 5 was brand new but also of the original type. In a moment of desperation we replaced the TLO72 with a TLO82 just to see what difference it made.
It did make a difference, voice 5 was now sharp, we thought great, trim it down and job done.
Not so, when voice 5 was flat, the tune low trimmer ran out of track before tuning was achieved, when it was sharp the same thing happened at the other end of the trimmer.
We next tried a TLO62, and in a Goldilocks moment, it was just right.
Finally as of May 6th, this one was put to bed.