The Strange Case of the Poly 800

There is nothing particularly unusual about us receiving a Korg Poly 800, there are lots of them and they are often destroyed by acid from batteries left in them for years while it was stored in an attic or garden shed, or else the complicated internal power supply damaged by incorrect wall warts. This one was a little bit different though as it turned out. It was totally dead, no lights, no sound, nothing, so it was obviously the power supply.

Except that it wasn’t, the chassis and keyboard were in pretty bad shape, and there was a little battery acid damage on the main pcb, but the power supply was perfect, it had never been damaged and no work had ever been done on it, which for us was a first. So why was it dead ? All power rails were correct and stable, and the Master clock for the CPU was also correct, but nothing Digital seemed to be working.


The above picture is of the Poly 800’s main board, and the majority of the Digital circuitry is in the bottom right hand quadrant of the pcb. The red and yellow dots on some IC’s were added by us as they were all the IC’s that were replaced, and some kind of explanation as to why and the order they were done in.

There was obvious battery damage to one of the keyboard encoding chips which also does the front panel switches, the 800 treats them both the same on a bigger matrix, as it’s legs were green and the solder joints had that matt grey look about them that acid damage does, so IC33 a TC40H240 was replaced with a 74HC240. We got better signals out of the 240 but nothing still worked. With the obvious possible damage dealt with, it was time to dig out the big gun, the newly rebuilt Tektronix scope mentioned over the last couple of entries. The next logical culprit was the 81C55, which is a 40 pin IC that handles the displays and all the DAC stuff, as none of that worked either. The clincher on this was that the Timer Out on pin 6 was dead, master clock went in but no timing clock came out, this signal is the source of the clock for the tone generator chip, so replacing this was next.

The 81C55 we used came from a dead Poly 800 board, whose power supply had been blown to bits, and battery acid put paid to the connector blocks for much of the rest but ironically the rest of the pcb was damage free, a result of the position it was stored in. Another interesting thing is that the chip came out cleanly and easily from this board, while the same chip on the synth board being repaired just wouldn’t shift at all and had to be cut out, and trying to buy one of these chips anywhere these days will be difficult and expensive. We have seen this often before, two identical pcb’s from similar times where one is a joy to work with, and the other a total pain, the reasons are as yet unknown, but it is a real phenomenon, we’ve seen it many times before.

With a new high quality 40 pin socket in the board and the transplanted 81C55 in it, very little had changed except that the Timer clock on pin 6 was now present and correct.

In a complex Digital system, where there are multiple problems in every aspect, diagnosis is very difficult as the closed loop that is a Microcontroller system will produce absolute rubbish everywhere if there is a fault, and instinct and luck are your friends here. The next IC to be replaced was IC16, a TC40H032 as it is used as a glue logic device to handle Enable logic in several parts of the system, and there seemed to be no wish for it to write Data to the Waveform Generator chip, the M5M5232. Replacing this with a 74HC32 didn’t change a thing, this was getting really annoying. Purely because so little of the system was still actually working, we replaced the Address Latch, a 40H373 with a 74HC373 in a nice new socket, and lo and behold the displays started working, and some of the buttons too, but still no keyboard and no sound. The next logical candidate was IC34, a 40H138, which is also part of the keyboard and switch scanning system. The synth still didn’t work but more and more parts of it gradually were creeping into life. The last chip to be replaced was IC29, another 40H138 which is responsible for a lot of the main sections of the synth. We now had most of the synth, keyboard working, displays working, MIDI working and Joystick working but still no sound output.
Although this particular synth had a coin cell battery fitted as standard, and it was still perfectly good, the presets just didn’t exist, and a dead battery scenario in a P800 will switch off all oscillators meaning no sound. What electrical scenario could kill so many components in the CPU section while not being power supply related?

It is still a bit of a mystery, but loading the factory presets from a wav file worked once the synth was finally functional, and all was good, and it has gone back to it’s owner.

Why did we actually take the considerable time it took to complete this repair ? The cost in terms of hours far exceeds the value of the synth, and Poly 800’s are not rare or particularly valuable.

The same client brought us a Korg Poly 61 a few months ago with severe battery acid damage, so severe that corrosion had got as far as the plastic ribbon connector that links the CPU board to the voice card, and destroyed the ribbon and its socket. Attempts were made to breathe some kind of life into it but it stayed resolutely lifeless.

We had a battery damaged Poly 61 of our own, which we had ignored as probably unsalvageable, and couldn’t be made saleable without some new parts. But while we were working on a Poly 61, we dug out our dead one for comparison purposes, and while it was very faulty there were a least signs of life, segments of the displays, random noises when keys were pressed. Some previous owner had replaced the large CV and Gate connector with a Tyco style connector (battery acid had eaten the Korg original one) and while this measured fine, it rendered the CPU board incompatible with the clients one.


Our Poly 61 board still took a fair bit of work, all the passive components in the area were cut out and replaced after scrupulous cleaning of the area, all diodes and transistors were also replaced. All PCB traces in the area were checked for continuity and repaired if necessary, and plated through holes cleared and resoldered. A new coin cell battery was fitted along with a modification to stop the Poly 61 from trying to charge it. A few details relating to the new connector for the gates and CV’s were tidied up and we were ready for switch on. All looked remarkably good, the displays were reading correctly, and the button functions did what they were supposed to do mostly, as dead Tact switches are a common thing in Poly 61’s, as there are not many and they take quite a beating, but with some strong switch bashing, it seemed that all functions were working, so we created and stored a few presets, switched off and waited a while, and then switched back on again.

All the new presets were present and correct and have been since. The only remaining issues were that somebody in the past had replaced the Arpeggiator speed Pot with a totally inappropriate type and value, and the Joystick handle was broken off, so with the consent of the client with the dead 61, we used the Arpeggiator PCB board and the Joystick assembly from the dead Poly 61 to make our one as good as it can be. It still needs and will get a full set of new Tact switches (17 of them) at some point soon, because picking a full set of working switches even from two PCB’s just isn’t worth it when the switches are 30 years old, and they are still being produced by Omron.

This links back to the Poly 800, we have a dead one that will never see the light of day again, in medical terms we have pulled the plug on this one, but it does (or did) have a lot of good parts, and the client’s 800 needed a lot of good parts, so his 800 got a lot from ours (some IC’s and the display and button board and cassette interface sockets which were destroyed by battery acid, and two keys). So in the end honours were even, parts were exchanged from a bunch of synths that were dead to create two that were back to specifications, and the Poly 800 has MIDI, which of course the Poly 61 does not.

From the two pictures above, on the Poly 800 board, the IC’s marked with red squares are the ones that had to be replaced, the ones with yellow markings were replaced but we can’t be sure if it was necessary.

The second picture is of our Tektronix Scope after being rebuilt showing something from the Poly 800, we can’t remember which measurement it was, but it has been all good after the rebuild.


Leave a Reply

Your email address will not be published. Required fields are marked *