What's new

xodaraP

Legendary
Joined
Sep 21, 2016
Messages
4,750
Reaction score
3,793
Location
Gold Coast, Australia
Hi everyone,
I've been working on a TMNT board for a while and I'm at the point where if I can't get it working now, it's going to be scrapped. I can see why it won't boot but not how to resolve it so hoping someone can help :(

The board initially had quite a lot of track damage due to heavy corrosion on H16 and H17, most likely from mouse pee. I've repaired a dozen traces and all pins now been out to where they should.

The board is rebooting constantly on a white screen, but without the clicking you usually get from the watchdog which would suggest the watchdog is only kicking in after it attempts to boot. The sound system is good (can get crackling from speakers if I run my fingers on the amp, and previously had watchdog clicking)

There are no floating pins that shouldn't be disconnected. I've been over it multiple times to confirm wiring is correct.

I've been following and testing the reset lines and noticed that 2 LS138 multiplexers are related to the CPU ROM output enable and CPU RAM chip select lines. These are both jammed on HI, but the datasheet says this is because one of the inputs is high (which it is)

G2A on H25 is locked on HI, and this needs to be low for the enable lines to be activated. This is listed as A20, which comes from the 68k

G2B on G24 is also locked on HI, as a result of the output from H25.

I am at a complete loss as to why the address signal coming from the 68K is HI instead of LO or changing between states and this seems to be the final hurdle to get the board to fire up.

Hoping someone might be able to help me out here, am I even looking in the right place?

Things done:
- Repaired traces
- ROMs verified
- Crystal replaced in sound section
- Repaired more traces
- Checked for floating and stuck pins and verified against schematics and datasheets
- Repaired more traces

The sound pot is missing, but that shouldn't prevent it from booting. The sound section is alive with data on the sound ROM.
 
Last edited:
I had one similar white screen symptoms and it ended up I needed to replace the CPU. I had no way to test that until it was replaced, so it's a bit of a gamble.
 
Hi everyone,


...

I am at a complete loss as to why the address signal coming from the 68K is HI instead of LO or changing between states and this seems to be the final hurdle to get the board to fire up.
Hi,
I'd say it's because CPU doesn't reach the command that would change that address line state.
Faulty work RAM?
 
I was actually considering that last night, the work RAM is Fujitsu 6264 which is usually pretty reliable so I initially discounted it but the more I think about it the more it seems the main CPU RAM is the likely culprit.

I would've thought I'd get the test screen at least but if it's very dead I guess not.

I might try piggybacking it and see if I get anything.

If not, time to consider desoldering the CPU
 
Last edited:
I was actually considering that last night, the work RAM is Fujitsu 6264 which is usually pretty reliable so I initially discounted it but the more I think about it the more it seems the main CPU RAM is the likely culprit.

I would've thought I'd get the test screen at least but if it's very dead I guess not.

I might try piggybacking it and see if I get anything.
The thing is RAM chips are probably the most sensitive parts on our beloved PCBs. And the first thing people do when a game doesn't boot is raising +5V... :cursing:
 
I've been beeping out all connections from the CPU, RAM and ROMs, nothing so far but I'm getting there. There's hundreds of traces, if everything checks out - RAM replacement

Where I'm getting stuck is that the RAM and ROM /OE lines are all high, so inactive, connected to the LS138 which is activated by A20 on the CPU

So is the CPU waiting for something before it turns on RAM and ROMs? Or are you saying because the RAMs are bad the CPU is getting stuck and not going any further?
 
Last edited:
Sometimes you have activity for a fraction of second on buses (too short to be visible with a logic probe or analogue oscilloscope) before CPU goes in high impedance. If pull-up resistors are used then you got everything stuck high.
 
Ok so I'm revisiting this board after I put it on the too hard pile for a while.

I completely redid the previous trace repair to do it more nicely where possible. Socketed everything I removed and replaced 3 chips that even though they tested ok off board they just weren't doing what they should be doing.

I was missing main CPU clock signal, this is now fixed after replacing H17 and H18

I'm now getting watchdog but it's not white screen watchdog, it's a rug/garbage pattern or a solid blue screen (I've tried 2 different sets of RAM).

Work RAM has been removed and tested off board and is ok, ROMs have all tested good.

I feel like I'm close here but I'm missing something simple. I have found another dead trace going to the sound RAM so I'm going to repair that (address line) but I don't feel like this would prevent it from booting.

I think there is one other dead trace so I'm going to see where this is going and patch it if it's bad. I think it is sound system related so again shouldn't be preventing the board from booting.

I am missing 12v on the bench right now (need to redo my harness) but again that shouldn't prevent the board from booting just without sound.
 
Update:

I'm not getting a SNDON signal which appears to be a bad LS138, will replace that to confirm. Also noticed that the 051550 custom that handles reset has both +5 and +12 connected to it. I'm guessing this means with no +12 it won't boot so I'll try it in a cab

Edit, nope, same thing. It's attempting to boot but crashing so not sure what's going on here
 
Last edited:
Check trace integrity and the TTL logic chips on the bus for the address, data and control lines between the CPU, program ROMs, and work RAM.

Also, you can disable the watchdog reset by connecting JP1-check and see if it gets a little further into the diagnostic boot process before the watchdog resets it.
 
Last edited:
This is where I've been looking at it looks ok but there is a LOT of Fujitsu. My next step is going to be checking the traces for address/data bus
 
Back
Top