Roland JX-8P Locks Up after Booting - Erratic Behavior

My 8P appears to boot (completes countdown), but then goes berserk and everything appears to be locked up (with no sound). Most of the time, the display flashes “Insert Cartridge,” but occasionally the panel LEDs will flash and there will be a few seconds of garbled ROM data flashing rapidly on the display. For a while I thought this behavior could be the result of panel switches being held active by a faulty IC, but I have not been able to prove this. Looking at the MIDI out, I occasionally see some CCs, but otherwise the note data is nonexistent. Based on my understanding of the system architecture, I am fairly certain that the fault only exists within the controller/programmer half of the mainboard.

I borrowed the mainboard from a known working 8P that seems to work fine when swapped in this chassis, so the power supply, display, and panel boards should all be good. I tested the traces connecting ICs and connectors. I also swapped ROMs, replaced CPU and RAM just to confirm that those were good. I probed all of the pins on the suspect ICs between both boards and the only major difference I noticed was that the CPU was being sent a 32kHz interrupt signal on the bad board.

At this point, I’ve tested every IC in the vicinity of the panel switch driver circuit and CPU/data bus the with my tester. All of them tested OK, although I’m really not sure how thorough my tester is. The only IC I have no way of testing is the Roland gate array (IC2), but otherwise I have run out of ideas and things to test.

Any suggestions? This is not my first synth repair rodeo, but I’m really stumped on this one…

1 Like

It sounds like you’re on the track with that interrupt signal.

Which CPU are you looking at? I see that the HD63B03 has IRQ going to the gate array, but the NMI is running off to the 8051 CPU. There’s an interrupt on the 8051 that seems to be connecting into a counter.


ProfTrudo, thanks for the reply. I assume the NMI signal is active low? It looks like the NMI between the 8051 and the hitachi assigner CPU is always held at ~5v on both boards, so I don’t think the Synth side of the mainboard is causing the problem.

I’m looking at a poor pdf copy of the service manual, so it’s a bit of a chore to figure out what’s going on with this synth. And I don’t have a JX-8P to reference.

I gather that the HD63B03RP is responsible for interpreting the key, MIDI IN, control panels, and cartridge, while the 8051 handles voice assignment and MIDI OUT.

Which line did you find the 32kHz signal?

According to the datasheet for the HD63B03RP, a low on pin 4 will cause the CPU to jump to the NMI routine. And a low on pin 5 will send it to the IRQ routine.

That IRQ line is coming from the gate array pin 28 starting from SN530350. That’s making me wonder about the board you swapped in.

As long as neither pins are held low, I think the processor should be running. Can you verify that the address and data lines are active when the keyboard is locked up? Clock signals?

How is the power supply? If a supply voltage goes low, it might force the cpu into a standby state. Are any of the chips getting too warm?

After many hours of additional troubleshooting, I’ve ultimately decided to throw in the towel on fixing this one. I reached a point where I thought I had narrowed it down to the proprietary Roland gate array through process of elimination (swapping ICs), and then the problematic board completely stopped booting after letting it sit for a few days. All I can get now are a few flashing panel LEDs (no display though). I’m absolutely certain that my power supply and other boards are fine because my other mainboard still works perfectly (different year, same revision). I’m fairly confident that the gate array is faulty, because now all the divided clock signals are completely wrong even though the16MHz crystal still looks good. Given the amount of time I’ve already wasted on this, combined with the scarcity of the gate arrays, and the number of remaining unknowns on this board, I simply no longer have the motivation or desire to continue down this wild goose chase.
Pure hubris, thinking that I could buy 2x broken 8P’s and fix them BOTH. I guess that’s how these things go sometimes. Hopefully this doorstop can bring life back to someone else’s faltering JX-8P.

Not sure if this thread is dead, but I too have a JX-8P on my bench. TL;DR, but I had a perfectly operating unit that I had re-capped the power supply and put in a battery holder. No problems. None. Then I bought the Vecoven EPROM and installed it. Once I booted the synth I noted a very odd issue: The top row of preset buttons did nothing, and the bottom row thought they were the top row. All the other buttons on the board were a bit wonked too. I said, ah, oh well I’ll just swap the old EPROM back in. Did, but to my dismay, the synth STILL behaved the same. Now, I suppose I could have zapped something swapping EPROMs, but I’m completely lost. I’m about to get a scope on it, but like Jerzilla said, I’m not sure I can repair it, or, not sure how deep I want to go. On a hunch looking at the schematic I swapped in new Toshiba TC40H245P logic chips, as one controlled the button matrix. No luck. If anyone is still reading this thread, ANY clues or thoughts? Thanks!

I totally understand. That’s a bummer.

Yes, it’s a bummer. But I think he made one good working one out of it with some spare parts.

Might get ahold of the folks at Kiwi Technics and see if their mod makes for a workaround on the PGA:

You can spend a lot of time repairing something like this, but if it comes down to a custom part that is not available you are just wasting your time.

I spent a week on a Roland Juno 2 trying to disprove that one of the custom chips in it was bad. I finally bit the bullet and replaced it. Works like new. I knew it when I started…

Right Prof. I’m glad to have the service manual, but in reality it leaves a lot of matters in the dark. Well, maybe 40 years ago a lot of it was ‘common knowledge’, but unlike a vintage ARP where I can just hail mary and replace all the opamps and logic ICs, it’s virtually an impasse with the 80s computerized boards. I’ll keep working.

I looked at the KIWI stuff a while ago, just not what I’m looking for in my 8P. Very cool, indeed, but not my cup o`

I looked at the Kiwi for the Juno 2. It adds some nice features, but I’m reluctant to invest more money into the keyboard.