MV1A watchdog repair

[kissedin]

I'm trying to fix a MV1A.

The system is watchdogging with a garbage screen and click (around 3-4 clicks/second). I am using diagrom. 5V is 5.0V.

When probing the 68k cpu w. oscilloscope I see:

* Datalines is going high then low with each click. Except D4 which is stuck high.

* Adresslines A1-A16 and A22-A23 are stuck low all the time. A17-A21 behaving like the datalines; turning high then low with each click.

* Halt and reset are behaving like the datalines; turning high then low with each click.

Are these the signs of a dead cpu, or is it something I've missed?

If it is a dead cpu, are there any typical culprits I should check before I power up with a new one?

 

[Xian Xi]

Trace D4.

 

[kissedin]

Thank you for the advice! I'll trace D4 more closely. But with no schematics or a working board to compare with I'm not sure how to do it thoroughly.

Edit: I suppose half of the ram chips shall have a D4 line.

Cpu D4 pin 1 goes to:
SP1 ROM pin 15
RAM 1 pin 16
RAM 3 pin 16
?????? Pin ??

Edit: D4 has same resistance to ground as the other data lines (can't remember exact value)

One of the color RAMs is not populated but i can't see that this would prevent the board to boot.

I'll try to lift some D4 pins and see what change it brings.

 

[kissedin]

So here's the story:

Traced D4 and removed:
Ram 1 (pin 16)
Ram 3 (pin 16)

Neo buf at ga11 (pin 2)
Neo buf at ga4 (pin 2)
Neo f0 at ga5 (pin 11)

Tested after each removed component. When Neo f0 at ga5 was desoldered, sound turned in to a slower clicking, every 3rd second. Also: activity at D4 and at all the adress- and data lines. (This was no bad 68k after all.)

Put back Neo buf at ga11. Then Neo buf at ga4. But after this the error came back. Desoldered ga4 once again - and D4 went pulsing.

Decided to move on and repopulate rest of the board. Put back RAM1. D4 once again stuck high and rapid clicking. Hmm... Wiggling the board, especially around the Work/backup-RAM-area made the error come and go.

(Strange, as I tried this before with no result. But maybe there were multiple bad joints that were reduced to just one when I removed Neo f0)

Couldn't see anything obvious. Reflowed RAM2, RAM4 and the D4 pin on bios socket. Also cleaned jamma connector (but for the record, I had 5,0 volt at the cpu all the time).

After this I couldn't make the error appear again.

Repopulated rest of the board. Testing after each new component soldered into place. When the final chip (Neo f0) was back the board booted to this:




Put in stock bios and cart. Got z80 error. Cleaned socket. Now boots and plays flawless.

I've had this running now for a couple of hours.

Xian Xi: Thank you again for your laconic advice. I guess a stuck dataline will corrupt all the machines instructions, making it watchdogging.