Korg poly 61 voice retuning

[Circuitbenders]

I've just pulled my old Korg Poly 61M out of the cupboard to try some mods on it and it seems that each of the 6 voices of polyphony are now out of tune with each other. This means that each note you play is slightly out of tune with the previous note and successive chords just sound a bit ill.

Each voice is clearly labelled on the synth board and there are trimmer pots for VCA, EG. FC (cutoff) and resonance for each voice, but no kind of trimmers for tuning. I can't believe Korg would have made no provision for retuning the 6 polyphony voices to each other, unless of course the tuning was set dead on at the factory and its just components failing at different rates that has caused the detune problems.

Has anyone had any experience of working with one of these things? I've done a fair bit of searching on the net but haven't found any clues of how to sort this out. I've found a service manual for the poly61 (exactly the same machine without midi) HERE but it doesn't include details of retuning and the schematics aren't very good copies.

[nochtanseenspecht]

perhaps there's a trimmer for "scaling" ?

[Circuitbenders]

unfortunately not. There doesn't appear to be any means of manually retuning the polyphony voices to each other on the board that i can see.
Theres three trimmers labelled high, mid and low but they just tune every voice at the same time, which isn't much good if the voices are out of tune with each other in the first place.

[Gordonjcp]

stick it in chord memory and play the same note six times, then watch the outputs of the 4051 demux.  If you've got one voice that's particularly bad but the rest are okayish then chances are you'll see one output not behave the same as the rest.  Replace the bloody awful 4051.

[Circuitbenders]

i was talking to the guy that wrote this incredibly useful guide on the poly61 about this.

http://www.mik-music.org/tiki-index.php?page=Korg%20Poly-61%20Repair&pagenum=1

Because DCO1 is a digitally controlled VCO theres no way to individually tune the voices but his idea was that i was just getting the raw VCO and it had lost its digital control. I was getting nothing on the scope where i was meant to be seeing the TTL signals that controlled the pitch for the VCO.

I'm using the past tense here as i was just tracking down the source of the clock signals when i went to make a cup of tea, and when i came back and turned the Poly back on it was dead. Now when i turn it on all the button LED's come on at once, the numerical LED displays stay dark and it won't respond to any button or key presses.

I'm guessing this is going to be down to some crappy logic IC dying for no apparent reason but its a pain in the arse trying to track it down. I think i'll start with the 4051's though.

[Gordonjcp]

Okay, if it's totally dead like that, forget the 4051s for now and concentrate on the digital board.  Check that the nicad hasn't leaked goo all over the board - that's a nightmare to sort out, but possible with extreme patience and a speedwire pen.  Not only does the corrosive electrolyte dissolve the copper under the lacquer, but it wicks up inside wires and attacks those too!

Since Poly-61s are often cannibalised for their (really expensive) filter ICs, you can often pick up cheap non-runners for parts.

[Circuitbenders]

The battery was replaced a while back. There was a small amount of leakage but that was all cleaned up and fixed at the time and its been working ever since, so unless i've disturbed something i don't think its going to be down to that.

I have noticed on my scope that the signals coming from the clock board are no longer even remotely squarewaves and are a little unstable. They are more like sawtooth waves with a bent upwards slope. Could that be why the system just doesn't seem to be starting up properly?

[Gordonjcp]

That could very well be it.  It's a bit late for me to try to puzzle out the circuit diagram of the clock, but start looking for dry joints, resistors going high in value and possibly dribbly electrolytics.  Also, don't forget that ceramic disc caps can go leaky (electrically, not physically) and cause all sorts of weirdness.

[Circuitbenders]

This thing is just winding me up now.   

I can't see anything obviously wrong component-wise although on closer inspection there is a little more battery leak damage than i thought. I can't find any track breaks although theres always a posibility i've missed something hidden under a chip or something.

The clocks for the MCU's come from what look like a pair of ceramic resonators which both seem to be putting out identical sine waves so i guess thats not the problem. The signals from the clock board seem to be just for digitally tuning the oscillators and have nothing to do with the system otherwise. They still don't look right on the scope but i'm thinking thats something to sort out when i've actually got the thing to start up again.

Both the MCU's and the four 8253 timers that supply the clocks for the oscillator tuning seem to be getting rather hot, although everything else on the board is fine.

[Gordonjcp]

Both the MCU's and the four 8253 timers that supply the clocks for the oscillator tuning seem to be getting rather hot, although everything else on the board is fine.

That's not particularly good!  Check the 5V rail is actually 5V.  If the clock is seriously mis-shapen then it could cause weirdness that might include overheating.

[Circuitbenders]

the 5V rail is at a steady 4.85v and the power supply is working fine.

The clock signal coming from the clock board looks like the picture below althougth mysteriously it did look a bit more like a sawtooth earlier today.

What should i be seeing on the reset pins of the MCU's if anything?

[hubris]

hiya.
to answer your first question...
here's the 61 MIDI service manual supplement with schematics for the 605, and the missing page from the owners manual here.

you may want to reset your poly with .wavs of poly-61 programming tape and try walking through ch.8. Chapter 7 I believe is the factory presets. haven't got that far yet...

The clock signal coming from the clock board looks like the picture below althougth mysteriously it did look a bit more like a sawtooth earlier today.

What should i be seeing on the reset pins of the MCU's if anything?

like this. try the other troubleshoots out too.

Check the pins on the CPU board that connect to the voice board *and the wires!!* for corrosion. I

stick it in chord memory and play the same note six times, then watch the outputs of the 4051 demux.  If you've got one voice that's particularly bad but the rest are okayish then chances are you'll see one output not behave the same as the rest.  Replace the bloody awful 4051.

my poly only plays the sixth note. what's special about the sixth note in chord memory?

[Circuitbenders]

It turned out that the problem with my Poly61M not starting up properly was with the MB3761 voltage detector at the bottom of the board. Finding a replacement in a DIL package was a little tricky but i finally tracked one down from www.dalbani.co.uk .Well, in fact they initially sent me a surface mount one but made a special order from their supplier for me when i complained.

That sorted the main board issues but it appears that the tuning issues were on the synth board. Luckily i got hold of another 61M with a completely destroyed main board but a mint condition synth board, and managed to bodge together a fully working machine out of the two.

[mikekvan]

I made a video that might be helpful for tuning the Korg Poly-61

https://www.youtube.com/watch?v=oG_I3-Ztu0Y