G’day from Sydney Australia, Synth wizards!
Hoping someone can help me with a main board problem on my Roland Jx-3p!
It appeared to be quite dead but after checking voltages and signals on the main board with my scope I can see quite a bit of activity. None of the LEDs are lit and it doesn’t respond to buttons or keypresses. Power supply is working the main board has +5v, +15v and -15v as expected. The CPU reset, RAM reset, CPU clock and ALE all working and the address bus is very busy. What I notice is that the logic used to scan the keyboard & panel or drive the LEDs requires the CPU to toggle the WR and RD lines (pins 16 and 17) but they both stay high. I haven’t seen either line switch LOW. I can see some of the other logic working; for example the I/O decoder chip (IC41) is cycling through Y1-9, so it appears to be attempting to read the panel switches & keys but IC40 needs the CPU RD line to be set low to gate the signal. CPU RD doesn’t change at all. In the hope that the CPU RD & WR lines were being pulled high by something external to the CPU I replaced IC 30, 40 and 53 but no change.
The only other chips that use the RD / WR lines are the static RAM (TC5517APL and the 8031 itself). I’ve ordered both chips so my next step will be to replace one or both…but…I’m hoping there might be something I’ve missed…?
Could there be another reason why the 8031 would hold the READ and WRITE lines both high…? - it seems inconsistent to have them both at the same logic level as it can’t do both at the same time?
1 Like
G’day
Having worked on several of these, the most common suspect are the Toshiba TC series logic chips but I’ve had a couple where the RAM was responsible so that’s quite likely here too. It would partially boot strangely but would also stay frozen. The CPU’s do fail also, sometimes only partially. I’m in Sydney also
Thanks for chiming in Ivan…!
Nice to hear from you mate.
I had to order some parts for the Jx-3p, so I used the opportunity to study the Intel MCS-51 series of microcontroller chips a bit more. The critical signal missing was the PSEN, which should have been pulsing low whenever the 8031 tried to read from the external EPROM. I ordered a replacement 8031 microcontroller and installed a 40 pin socket when I replaced the chip today. Happy to report that the new MPU has restored normal function. I upgraded the EPROM to version 4.00 software and have spent the afternoon in poly synth heaven…! The Jx-3p sounds great.
So to answer a few of my own questions;
- The RD and WR lines float high unless pulled low by the 8031 when it tries to read from or write to an external component.
- The 8031 pulls PSER low to enable reading from EPROM. Normally this pulses low at about 2MHz, along with most ALE pulses that latch an address to read. A picture of the waveform on pin 29 of the 8031 is attached.
- In my case the 8031 couldn’t assert PSEN so the EPROM could not read the address and respond with any data. So, with no instructions coming in, the MPC just kept scrolls through the program counter - but doing nothing. This explains why the signals on the address bus looked reasonable.
Anyway, I have a few cosmetic things to attend to but otherwise it’s a happy end to this repair story. A few snapshots to follow.
It’s easily one of my favorite synths of all time. I’ve had mine for a while now. Let me recommend an upgrade for you. It was originally the Organix midi upgrade that’s been reissued with permission from the original owner and it’s local here in Sydney! It upgrades the MIDI spec of the JX3P and enables you to use MIDI and the PG200 at the same time
