Category Archives: Synthesizer

Poly-61 individual voice tuning

When I recalibrated my Poly-61 I had two voices significantly out of tune. Now since all oscillators are digitally controlled and therefore supposed to be in tune, KORG omitted attenuators for individual voices.

Here´s what I found as probable cause and fix: The parts most likely affected by the aging process in the area of voice CV appear to be the Cn02 (n=1..6) capacitors. I could not identify their exact type, but the marking sais they are 10nF with only 1% tolerance, so precision circuitry could be assumed. If I read the schematics correctly then together with the Rn02 resistors they form the input of the antilog. So there are two possible fixes. The first being to replace the capacitors with new high precision 10nF caps. The other fix is what I went for, firstly because I did not have the appropriate capacitors at hand and secondly because I felt that having control over the tuning would be handy. So I replaced R402 and R502 with precision trimpots.

Poly-61 trimpots for inividual voice tuning

Since Rn02 are 15k the best trimpot value would be 20k. I only had 100k at hand so it was a little fiddely to set the right value but it works for now. I still have to get a couple of replacement parts for proper fixes anyway so I will eventually replace all Cn02 with new ones and all Rn02 with 20k trimpots.

Poly-61 repair

Ressurection of a Poly-61

Like some other 80´s polysynths with memory, many of today´s Poly-61s suffer from battery leakage. The problem for most owners without knowledge of the technicalities is that the symptoms come on slowly and do not neccessarily hint at memory. While leaking, the battery may still provide enough power to retain memory while starting to damage the PCB area around it. The device starts acting weird up to doing nothing at all soundwise. The symptoms depend a little on how the synth has been stored and therefore how the acid flowes across the board. If you see a Poly-61 on the used market described as ‘functional’ but ‘holding a steady tone’ it´s most likely a battery dammage.

I purchaised such a unit with the risk of not knowing the exact amount of dammage it had taken. But I was determined to make it fully operable again. After all I thought it couldn´t be worse than the Opera-6 fix.


If you use the information presented here you are doing this at your own risk. I cannot be held responsible for any damage done to your device or yourself caused by your actions! You need the appropriate skills to do work on electrical devices.

The prequel

I looked around for information on the net but could not find much that was useful. Like others I remebered coming across a pretty extensive repair video, but it has been taken down by its author. The rest that is to be found are forum posts that describe the symptoms and take guesses and if they provide a solution it is only the general ‘re-crimp the connectors’ one. Some say that it helped, but never say why. And I wanted to know why …

There are plenty of download sources for the service manual, but the schematics contained are barely readable. It had to do though.


The device initially showed the following faults:

  • Outputs a constant tone
  • Most keys not working
  • DCO2 not working on all voices
  • Number 6 button not working

CPU board fix

First step was to open up the unit and examine the damage on the CPU board.

KLM-509A CPU board battery damage

Interestingly enough the board was missing the battery entirely. The two wire stumps seemed to be the remainders of a first repair attempt.
It would not have made sense to look for certain errors before the general dammage on the board wasn´t fixed. So first step was to remove all parts that where affected by the leak.

KLM-509A CPU board stripped

It should be noted that I also removed CN27B and IC2 after the picture was taken. Next was to clean the affected area. First I used 70% Isopropylic Alcohol and applied a big load of baking soda to neutralize the remaining acid.

KLM-509A CPU board sodarized

After removing the dried up soda with a tooth brush the damaged area looks alot better.

KLM-509A CPU board cleaned

From here we can look for broken traces. With a strong lightsource we can x-ray the board for a visual check. Notice the broken traces in the red marked area.

KLM-509A CPU board x-rayed

Visual examination is not enough though. It is best to verify all traces using a Digital Volt Meter in continuity mode. If at this point one has to be told to use insulated wire to fix the traces then one probably lacks the basic skills to take on this task at all …

As you can see on the picture already I used sockets instead of soldering the ICs into place directly. This helps in finding and correcting faults without putting stress on the chips.

Replacement parts

Sourcing replacement parts was not that hard. Basically all HD14xxx had to be replaced by MOS 40xxx equivalents. Only one transistor had to be sought out of a variety of possible replacements.

Part Original Type Replacement Type Comment
Q11 (2S)C2785 2SC1815 NPN, TO-92, 50V, 150mA
Q10 (2S)A1175 Still available at the shop of my choice
IC HD14xxx MOS40xxx

Battery replacement

I put in a lithium cell insead of the former reloadable type. This requires the load current resistor R70 to be replaced by a diode (1n4148 will do).

The general crimp fix

A heads up in advance: Yes, the board connectors can be the source of many problems on the Poly-61.

However it depends on the general damage, especially in the case of a leaked battery, whether it is enough to just re-crimp the connectors or whether more action is required. In my case the sockets of CN12 and CN27B where badly affected as well and needed to be cleaned to regain full connectivity. The crimps of the connectors where literally falling apart.

Fixing the faults

Constant tone

If a synth outputs sound constantly, chances are high that the VCA is open. Now in most cases the VCA is driven by an envelope generator. EGs need some sort of gate or trigger. The Poly-61 uses SSM2056 envelope generator ICs and they have a Gate input at pin 6. I checked the pins with a DVM and found ~1.3V at five of the six. Now 1.3V is given in the datasheet as the maximum required threshold voltage for the gate. So all these inputs where active. Looking at the schematics you will eventually find that the gate lines are directly connected to buffer IC outputs on KLM-509 via CN12. With CN12 being severely damaged by the battery acid and having barely any connectivity, the affected gate lines where floating. I assume that this causes the 2056 to pull the inputs high internally for some reason. I cleaned the pins of the board connector with a dremel brush and for the lack of replacement crimps came up with a temporary solution until I manage to source the right crimps.

KLM-509A CN12 custom crimping

DCO2 missing

DCO2 for the six voices are provided by IC5  and IC14 on KLM-509. DCO2 was missing on voices 4-6. This meant that one of the chips was doing its job and the other one wasn´t. It turned out that a trace between IC11 and IC14 was cracked, leaving the CS line non-functional.

Number 6 switch

This was in fact an easy fix, because it was also only a cracked trace on KLM-509, but it was the hardest to find out because of the barely readable schematics. In the end I was lucky that I didn´t depend on it more.


That was about it. One Poly-61 back to life. All that was left where the typical restauration tasks like cleaning pots and key contacts and cosmetics.

S-330 Sampler Repair

I recently bought two Roland S-330 Samplers from a fellow musician at a synthesizer meeting for a reasonable price. Two things made me take them: Firstly, I´ve never owned a Roland S-series sampler and secondly one of them was declared defective. So that rose my ambition to get it fixed. And oh was I surprised about what was the cause of the error!

The previous owner told me that, when switching the unit on, the only thing happening was the LED of the disk drive lighting up. My initial thought was well, at least it still does something, so it can´t be too bad…

Checking it myself I also noticed that the Roll button LED was blinking very fast, and it was not a random flicker. It looked more like a clock-driven reset. My first assumption was that many times there is something wrong with the power supply, in a manner that, at a certain point of the boot process, the unit would draw so much current that a defective PSU would drop in voltage, thus causing a reset and put the unit into an infinite loop. So I was getting ready to recap the PSU but all voltages measured just fine, even under load. Also, all parts really looked okay, so it had to be something else.

So I dismounted the mainboard, performed a visual check of the top side and everything looked fine. Then I flipped it around to check for dull solder pads. And it was not very long until I discovered this:

S-330 mainboard w. troubling wire

S-330 mainboard with troubling wire

If you find it hard to locate it then maybe beauce someone used a blank piece of wire to bridge two points on the board. It was drawn between a pin of the floppy controller to an output of a transistor array. Maybe some kind of pull-up hijacking to fix an unreliable DMA controller. To be honest I did not investigate any further, because it was quite obvious that blank wire across the mainboard touching at least four other pins that where most likely not meant to be part of the equation would cause some sort of trouble… I could imagine that at the beginning this wire was arranged in such a way that it didn´t touch any other pin, but eventually did so, caused by bumps or similar movement. I really can´t believe that someone did this and then said ‘well, I tried something but it didn´t work’!

So at the end a simple fix. Replaced the short-cutter with an isolated wire and, surprise surprise… a working unit! And this one is even in better condition than the unit that was working all along, and to no surprise. Probably the faulty unit has been sitting in a rack unused long enough to prevent some parts from wearing out. The display is still bright like on the first day and the disk drive runs smoothely compared to the ‘ever working’ unit.

But I´m not going to stop here. I am planning on getting one of these MSX to USB adaptors for easy control of the unit. And also there is a schematic for a Roland DIN to VGA adaptor (Roland RGB-25I adaptor cable), so I want to try that out and see if I can hook the unit up to a standard flat screen.

Red Glue Problem

Some older Roland keyboards suffer from a bad choice of material. The metal  plates that give weight to the keys are glued in with a red substance that eventually looses its consistency and drips onto the keybed. Roland did offer a replacement for applicable devices, but it appears that they are now out of stock and can therefor no longer provide that service. Affected devices are mainly JD-800 and D-70. Both my units were affected so I had to come up with a solution. Fortunately not much harm has been done yet so I just had to do a little cleaning.

Here is an image where you can see what it looks like when the glue starts to flow

JD-800 bad key

On the next picture you can see the consistency that it takes on. I used a screwdriver to push back or scrape off the red stuff on seriously affected keys.

JD-800 key scrapeoff

I then took a hot glue gun and sealed the red glue along the edges. This should prevent it from flowing. At least I haven´t experienced any leakage again since then.

JD-800 bad key sealed

The whole procedure took a good hour and about ten sticks of hot glue as the black keys need a little more filling.