Roland JX-3P no panel/sound/kbd

[belzrebuth]

Hi all,
Just got a JX-3P which does not seem to respond to anything.
It does not output sound, respond to keyboard or panel keypresses or even enter test mode.
The LEDs that are lit when it turns on are correct (CHORUS, PRESET A, 1) and I can see address activity between RAM,CPU and various other test points (TP10 as well) but it seems dead otherwise.
The fact that the LEDs that light up upon power on are correct, tells me that the CPU and the EPROM work to some extend.
The voltages are okay and I already fixed some of the usual PCB panel cracks.I initially thought that the panel not responding had something to do with those destroyed traces but of course it's irrelevant since there is no sound as well.

I socketed the CPU to possibly try another one, reseated the EPROM and spent several hours probing around for anything unusual but I haven't yet noticed anything really out of the ordinary.

CPU reset goes briefly HI and then LOW upon power up which seems correct.
Various bits and signals are floating around the logic ICs and I can see high speed signals on TP12, 13 marked as clock 1,2 (these probably indicate that the DCO is running).
I don't see any activity on PANEL READ and KEYBOARD READ test point no matter what buttons/keys are pressed.

The fact that the instrument does not enter test mode is a bit suspicious.
The currently installed EPROM is a 27c64 with an unknown ROM revision.
It has a Roland sticker and the number 1 circled.
I read the EPROM image succesfully so the IC itself is not damaged but I can't really be sure about its contents.

The only giveaway of possible damage was this:

The synth had the panel POWER connector connected to the address lines of the panel analog MUX receptacle and where the POWER should go it had the analog MUX signals coming straight from the main board.
I don't know what damage could have been caused by this since the analog mux select lines are inputs.
So instead of the ABC select lines there were -15, 15 and 5.
This could have probably destroyed a 4051 which is responsible for the sense potentiometer which I replaced as soon as I noticed but it's not a major part of the instrument, and possibly not enough to cause the current issue.
The address lines connector had some signals going to ground though so this could have probably caused a certain damage since logic ICs may not have output short circuit protection.
The signals that are going to the analog MUX are still vivid though so probably no harm has taken place.
Nothing else seems out of place from what I've seen.

I'm thinking of replacing the CPU regardless of its mostly healthy(?) activity since I can easily find a working one from another instrument.
The EPROM is going to be trickier since I don't have any 27c128s in stock to write the v4.00 firmware.

What puzzles me is that all the keyboard connector pins are stuck high regardless of any keypresses.
I haven't seen *any* activity when pressing panel buttons as well..
MIDI in/MIDI out are not working as well.

Any ideas would be appreciated.

[belzrebuth]

I went a bit further on this and replaced the CPU with another 8051 that should be working fine.
No change.
I also went ahead and saved an eprom dump of the current revision and flashed v4.00 eprom to the 27c64 which was already there.
I don't know why I've read that I should use a 27c128 since the image wrote fine on the 27c64.
No change with the new eprom as well.
I also measured the A,B,C,D,E,F,G,H test signals that are on the service manual.

A. CPU ALE
Should be a 0 to 4V 2MHz (500nS period) square and I measured a siney looking signal (with two lobes on the low side) 2MHz almost 0 to 6V (a little high).
Kind of higher than what it should be so probably *not* okay.

B. CPU RESET
Goes low as soon as I power up.
ok.

C. RAM RESET.
Goes high at power up.
ok as well.

D. MASTER CLOCK
siney to triangle-y looking waveform around 8V amplitude and about 6.13MHz.
It's there but I don't think it's correct as it should be a 125nS period (8MHz) 0 to 5V square.
I'll call that *not* okay.

E. NOISE

This should be a 0 to 4V signal but I get around 4.7V on the positive and -10V on the negative side.
Total amplitude is almost 14Vpp and it seems to be clipping the BA662 since it's touching the negative rail.
I think that's definitely *not* okay.

F. DCO OUT.
0 to 5V square pulse.
I got a 0 to 5V square at around 11Hz frequency.
The manual does not specify any specific frequency so I'll call that okay based on amplitude alone.

G. DCO OUT
Should be a -12V to 0V inverse sawtooth and I get -14 to 0V sync pulses at around 11Hz (same as above).
So that's *not* okay.

H. DCO
Should be a -5 to 0V sync pulse and I've got it and amplitude is -5 to 0 as stated.
Same frequency I believe as DCO (F) so that's okay.

So out of 8 test signals I only got 4 right, could be worse maybe

Another thing I've noticed as time goes by the S/H TL082s are getting very hot.
And the power supply is also getting hot but no short circuits or anything.
I didn't try to measure power consumption or anything but I thought that may be something to look into.

I will try yet another CPU IC but I don't think it's the CPU that's faulty and I don't think it's the EPROM either since I checked them both.

One of these counter ICs may be faulty or misbehaving but my current guess would be that a logic error is going on.
There are a lot of logic ICs in this board there's a great chance that one or more went bad over 40something years.
One thing I don't understand is how most of the above signals got so off with their levels.
Could it be that the noise VCA is bad?
But what about the master clock or the CPU ALE signal?

I also suspect a multiplexing or maybe a S/H issue since these ICs are getting so hot..