PolySix Repairs

[The Old Crow]

On Fri, 8 Sep 2000, Drew Smith wrote:

>         Scott's site says that if all of the program buttons are lit,
> then the problem is probably a blown CPU, and that I should mail him
> if this is the case.  Well, that's the case - all the LED's are lit
> up.  I've mailed him, last night, but haven't received a reply yet.  
> I'm *itching* to find out how this synth feels in my setup, so I'm
> going to ask here, if that's ok with everyone.

  The first thing to look for is bad IC pins or corroded tracks around
IC30 and IC31 as well as the resistor SIP R17.  The IC30/31 pair are bus
buffers that facilitate the connection between the front panel logic as
well as the patch RAM, etc.  Losing any one of these signals will cause
strange things to happen, like stuck LEDs.

  Here is the section of the KLM-367 trace map that will help you check
the tracks and pins of the above-mentioned parts:

http://www.oldcrows.net/~oldcrow/synth/korg/polysix/battfix/klm367_tracemap_sm.jpg


> Symptoms:
> 
>         I turn the synth on.  The LED in the "MG" section (MG == LFO, I
> assume?  Modulation <something>?) flashes slowly, then speeds up until
> it's on fulltime - elapsed time, about five to six seconds.

  OK, so the S/H element for the MG rate is not getting refreshed, and is
slowly losing (or gaining, can't recall offhand) hold cap charge, thus the
rising MG rate.  This begins to suggest the CPU is not running, but until
IC30/31 and their tracks are verified, this is not a certainty.

>         If the machine has been off for a while, when turned on, the
> LED's marked 16, 8, 4, 2, etc - oh yes, this machine has a "Miditec"
> MIDI retrofit in it, hope that's not the problem... - wait a few
> seconds before all coming on at once.  I'm not sure what these buttons
> are originally labeled; the retrofit kit added some nice stickers over
> the original labels.  The program LED's - all eight - are alway all
> on.  All other buttons with LED's are on fulltime as well.

  Hm, the MIDI retofit could be interfering with the KLM-367.  I do not
know how that particular one attaches to the P6.

>         The arpeggiator section - the LED is working.  Twiddling the
> "speed" knob makes the light blink faster or slower appropriately.

  That is a 555 with variable rate--no CPU involved, though the rate
pulses from it do tickle the key assigner CPU (the other CPU on the
KLM-366 board) to advance the key sequence.

>         There is NO SOUND whatsoever, apart from a hum at full volume
> that isn't affected at all by pressing keys or twiddling knobs.

  While pressing a key repeatedly, look carefully at the KLM-366 (the
biggest board in there) near where the SSM2056 chips are located.  Near
each 2056 is a small LED that should light up when the key is pressed.
Each successive key-press should light the next LED.  This will at least
tell you if the key assigner is running.

>         I've still got my fingers crossed that this is just the lack
> of processor, but the only reason I believe that is Scott's site.  
> Thought I'd open it up to the group for discussion, as I'm antsy to
> get this guy running! :)

  If it does turn out to be the CPU (IC22), you will need a new one.
Fortunately one can be made by buying a blank 8748 and programming it with
the code I long ago extracted from the 8048 on a Polysix.  The original
8048s are factory mask-ROM devices, and it is unlikely Korg still has any.
(I asked for one from them in 1990 and had no luck, thus my risking a
known good one to obtain the code and not have to worry about it).

  8748s go for, let's see:

  www.jameco.com: IC,MPU,8748
                  Part No. 53022 
                  Product No. 8748 
                  Price ea. USD: $12.95 (qty. 1)

  They also have the HCMOS version (8748H) for 15.95.  It is a lower-power
device, but either one is fine.

  Now, if the 8749 (key assigner CPU) is also dead, they list it for:

  www.jameco.com: IC,MPU,8749H
                  Part No. 53057
                  Product No. 8749H
                  Price ea. USD: $19.95 (qty. 1)

  It seems they only stock the HCMOS version for this CPU.

  Jameco stocks another useful part for Polysixes: the RAM chip.

  www.jameco.com: IC,SRAM,6514(LH511423/M58981P)
                  Part No. 43238 
                  Product No. 6514
                  Price ea. USD: $2.95 (qty. 1)

  Programming the CPUs must be done on a device programmer, of course.
Hope this helps,

Crow

/**/