Ha channelmaniac, 6 slot problem

You quoted in a previous post you can fix the shit out of 4 and 6 slots. I personally never failed to fix 2 and 1 slotters, but I currently am having problems with a six slot.

Any tricks to isolate a watch dog problem

Yup...

I find the 4 & 6 slot boards MUCH easier to fix than the one and two slot boards. I personally think the 2 slot boards are a POS. The one slot boards are a mixed bag depending on the problem.

1. bad CE*, OE*, or WE* traces on the Backup RAM
2. bad Work RAM
3. bad CPU
4. bad BIOS Chip
5. bad BIOS socket
6. cut traces on the bottom of the board
7. shorted 74(LS or F)245 chips connecting the top & bottom boards

Dunno why but if the Backup RAM's control signals aren't getting to the chips it will get stuck in watchdog. This is by far the most common problem.

The 74xx245 chips are 8 bit latches on the data bus. If they are shorted then guess what, the data bus is dead. OOPS. Instant trip to watchdog land.

You can verify either case with a simple logic probe.

The bad Work RAM can be verified with a modified UniBIOS. I PMed Razoola about modifying one to re-write the test routine and got this answer:

Razoola said:

Its possible to do though not so important because you can simply alter the SP of the original Bios to place the stack into palette ram for example just to see if the workram passes.

Click to expand...

I don't have a 16-bit EPROM programmer yet but when I do I'll look more into that as I think having a test BIOS that can verify bad Work RAM better would make it easier to repair the boards.

When fixing the boards I use small Kynar wire-wrap wire.

RJ

Thanks

I got a 6 slotter which battery exploded and went ape shit all on the bottom of the main board. I fixed and confirmed all the traces in that section but still get the watch dog. I will check the latches tonight and replace the work ram chips

How did you confirm the traces?

Instead of confirming traces you should confirm signals on the pins of the chips with the logic probe. That will tell you if you are missing a signal which is better than checking a trace.

It is also the quickest way to find bad traces or a bad chip. If you are missing the signal then check the specific trace then check the specific chips.

RJ

channelmaniac said:

How did you confirm the traces?

Instead of confirming traces you should confirm signals on the pins of the chips with the logic probe. That will tell you if you are missing a signal which is better than checking a trace.

It is also the quickest way to find bad traces or a bad chip. If you are missing the signal then check the specific trace then check the specific chips.

RJ

Click to expand...

I have other working 6 slots, thats how I am more so putting traces back together more so then confirming if they are the right signals, the boards had numerous corroded traces.

Checking for correct signals is stage 2 trouble shooting which is where I am at now. Stage 1 checking traces is the easiest as I do not need to power up both 6 slot and play compare the logic on two powered up boards. Basically I don't want to be bothered with powering up 2 boiards.

Using a logic probe right off the bat is only useful if you know the correct logic levels ahead of time.

Ah, but that my young Padawan is what the datasheets are for!