[sdiy] seek help on Jupiter 6 repair

[Anderson, Robert O]

After much detailed poking around on the broken Jupiter 6 I've made a bit of
progress.  The bender had an offset of 300mV which I nulled with the two
pots at the top of the bender panel.  I also found that both the 4 and 2
voice boards had high TX signal indicating board faults.  These two signals
are OR'd through 100k resistors to a tranny on the control board.  By
removing the two MC4 connectors on the control board from both voices the
tune light went out and voila, all control functins appear to work fine.  I
can recall patches, play through the MIDI out and no blasted tune light.
Controller board seems fine.

So now I had it pinpointed.. two bad voice cards.. uggh.  I checked the S/H
outs and they seemed funky and one drifted from 600mV to a negative voltage
over time.  I checked the scanning circuitry and there was no strobing.  I
checked all the address lines from the 8031 and they were all blinking.  If
the thing was executing a tune procedure it seemed unlikely that all address
lines would be toggling.

Pulled all three EPROMS, rigged my homemade PROMMER for an 2732 and read
away.  Long story short, the controller EPROM looks fine but both voice card
EPROMs are full of 0xff.
This value is a benign MOV R7,A in 8031, so both voice cards were spinning
their wheels and doing absolutely nothing to the outside world.

I wonder if these EPROMs naturally decayed.  Do they decay to 0xffs?  I
forgot the physics of it but it seems you store charge on an insulator and
eventually the charge leaks away.
I remember the discussion on the imminent mortality of EPROMs from the 1980s
and about forming an EPROM library.  If these just decayed, can I simply
reprogram them for another 20 years?  Does anyone have the voice card Intel
hex file I can use?  Are there different versions like the one for the
controller board?
Regards,
Bob A