Help with A600 Green Screen diag

[psoma]

Hi all,

Having a ball getting back into Amigas after 20 years in the wilderness... picked up a couple of non-booting 600s and managed to recap and repair one (cap juice damaged 12V via), the other was in a light green screen reboot loop with 10 power LED flashes. I understand this is Amiga for “RAM issue” but could be due to a faulty Agnus.

Troubleshooting/Diagnostics;

- unplugged everything
- decapped and inspected board - all looks in-tact
- booted without caps - same symptoms
- all chips heat up except the IDE controller (GARY??) and the two RAM chips.
- noticed the corner of the left hand CIA chip is slightly damaged. I don’t think this could cause a green screen though as it’s unrelated to RAM control

From here:

- I’ve ordered a DiagROM 1.2
- I’ve ordered a fresh cap set

Questions:

- My understanding is, a 600 should boot even without the caps as they are primarily for power conditioning. In essence, if I have a perfectly working 600 and I remove all the caps, I *shouldn’t* be getting a green screen at least - is anyone able to confirm as this will eliminate the caps alone being an issue.
- Anything else to try/look at before trying to source replacement RAM or a Fat Agnus?

Many thanks, loving the community!!

Tom
Melbourne, Australia

[solarmon]

A green screen on boot means there is an issue with configuring or setting up of RAM.

This means either the ram chips are faulty, or the memory address and data lines are faulty, or the chips using the memory and address lines are faulty.

Obviously the permutations of one or more of the above can sometimes seem like looking for a needle in a haystack.

I just recently had the same issue with an acquired A600 that was booting to green screen and resetting after 10 Power LED flashes.

When inspecting the motherboard I found corrosion next to U7 memory chip. However the issue was a trace that was broken UNDER the chip. I made a temporary wire jumper and it booted OK.

The above example shows that the issue can be found if you carefully inspect and use the appropriate tools. And some damage is not visible without using the appropriate tools to detect it - for example, a microscope and a multimeter.

The first step is that careful visual inspection. To the untrained and naked eyes things might not be so obvious and look OK. So if that is the case, take some clear and close up pictures and post it here for others to give further advice on.

And yes, the A600 should be able to boot without any caps (although some are used for timing circuits), assuming there is nothing else wrong with it. A recap won't fix physical damage.

[Mark sealey]

I used Amiga PCB explorer on the internet to diagnose my A500+...might help a bit ?

[psoma]

Thanks guys - will take a closer look tonight.

One thought I had though... should the memory chips be getting warm? I’m concerned that both are cold meaning something providing power/signal *to* the chips is failing...

Cheers,

Tom

[solarmon]

Originally Posted by Mark sealey

I used Amiga PCB explorer on the internet to diagnose my A500+...might help a bit ?

Yes, the two best online resources to help with troubleshooting and diagnosis are:

http://www.amigapcb.org/
https://www.amigawiki.org/doku.php?i...ice:schematics

Originally Posted by psoma

One thought I had though... should the memory chips be getting warm? I’m concerned that both are cold meaning something providing power/signal *to* the chips is failing...

You can use a multimeter to check that power are getting to the chips - http://www.amigapcb.org/ will help you identify the supply pins.

However, being cold is not nescessary an indicator that something is wrong - especially at this stage when it hasn't even booted up and it is not doing much.

[psoma]

Originally Posted by solarmon

You can use a multimeter to check that power are getting to the chips - http://www.amigapcb.org/ will help you identify the supply pins.

However, being cold is not nescessary an indicator that something is wrong - especially at this stage when it hasn't even booted up and it is not doing much.

Thanks @solarmon - good call, will consult the schematics and probe for voltage... stay tuned.

Cheers,

Tom

[psoma]

OK, with the help of amigapcb, I traced out all the mem chip pins back to their source ICs successfully and confirmed correct GND and VCC supplies to the chips in the correct locations... It all looks good.

Could it actually be I just have a bad RAM chip - if so, is that the sort of thing DiagROM will let me know?!?

Cheers,

Tom

[solarmon]

Originally Posted by psoma

OK, with the help of amigapcb, I traced out all the mem chip pins back to their source ICs successfully and confirmed correct GND and VCC supplies to the chips in the correct locations... It all looks good.

Could it actually be I just have a bad RAM chip - if so, is that the sort of thing DiagROM will let me know?!?

Cheers,

Tom

Now check them again! Seriously!

Also, there are memory address/data traces going between the CPU and Agnus and Gayle - have you checked those?

And sometimes the traces may buzz out OK with the multi-meter but they may have degraded enough to be too high a resistance for the address and data signals. So this is where inspection of the traces with a microscope is useful to identify any suspect traces, then use the multimeter to check the resistance of those traces.

[psoma]

OK, have checked again, and beeped out all from-CPU traces. 100%. The board is in great condition in general, so my gut feel says it's not a trace problem. I think I have a dead something - be it Agnus or RAM chip(s).

Does the RAM piggyback trick work on A600 RAM?

Cheers,

Tom

[solarmon]

Since the RAM in the A600 are surface mount, not through hole, I'm not sure whether this will work - I haven't tried it.

Have you checked that the 5V and 12V rails are good and getting to the chips?

It might not be a broken trace, but it could be a short to ground, VCC or something else - so try to inspect for that too. Use a USB microscope as things like solder balls underneath and between component legs are hard to spot with the naked eye.

If something is shorting to ground, VCC, etc it could get hotter than normal - so do a heat (and smell!) test and see if anything is unusually hot.

I would also recommend getting a logic probe to make it easy to check whether there is any activity on the traces.

[psoma]

Thanks again solarmon - checked VCC as I was going along to ensure there were no shorts to ground... looked good. Was getting 4.8V on the 5V feeds which I presume is within tolerance.

Nothing is getting hot, but one thing I did notice is that Gayle is cold... it's getting its 5V feed, but it's not warming up at all. Not sure if it should, but I know it plays a part in the memory addressing process so... maybe?!?

Cheers,

Tom

[psoma]

As mentioned earlier though, this is my one area of concern - the corner of the U7 CIA seems damaged:



I can't see how this could cause a green screen though?!?

[solarmon]

4.8V is a bit low but within tolerance.

Yes, a part of Gayles function is for memory address decoding, and other functions too:

https://www.amigawiki.org/lib/exe/fe...cification.pdf

If it is cold then I suppose that is not a good sign. I can check and compare against my A600 (rev 2B) - I can just undo the temporary jumper and it will give me a green screen, LED flashes and reboot - so similar to yours.

[psoma]

That would be awesome man, thanks!

[solarmon]

Originally Posted by psoma

As mentioned earlier though, this is my one area of concern - the corner of the U7 CIA seems damaged:



I can't see how this could cause a green screen though?!?

Yeah, that doesn't look good. That is quite a lot of force to get it like that. I'm surprised that the motherboard is not damaged, so it might have gone in like that from factory?

There are actually some memory address/data signals that goes to the CIA's.

However, I believe, with the A500 anyways, that the Amiga should be able to boot without the CIA's. Unless the CIA's themselves are dragging something down.

[hooverphonique]

Originally Posted by psoma

As mentioned earlier though, this is my one area of concern - the corner of the U7 CIA seems damaged:

I don't think that makes any difference overall - the die is deeply embedded in the package, and no bond wires should be present in the corner.
Unless it's no longer air tight and moisture has been absorbed, then it's a different matter.

[solarmon]

Also, try pressing down on a chip, or parts of a chip, whilst it is boot looping to see if that makes a difference. There might be a dry joint somewhere.

[solarmon]

Below is the corrosion and fix on my A600:





There is a trace under U17 that is broken so a temproary wire jumper was put in.

With the wire jumper removed. I can confirm that both U5 (Gayle) and the two U16 and U17 stay cold even after about 5 minutes of boot looping. Eveything else was relatively warm.

Even when OK (with wire jumper put back in), on the insert disk boot screen, it was the same symptons - cold U5 (Gayle) and the two U16 and U17 and everything else was relatively warm.

[psoma]

Originally Posted by solarmon

With the wire jumper removed. I can confirm that both U5 (Gayle) and the two U16 and U17 stay cold even after about 5 minutes of boot looping. Eveything else was relatively warm.

Even when OK (with wire jumper put back in), on the insert disk boot screen, it was the same symptons - cold U5 (Gayle) and the two U16 and U17 and everything else was relatively warm.

Thanks man, that's awesome... saves me chasing after red herrings. Definitely looking pristine around my RAM though - no hint or suggestion of corrosion (will take pics) so I *think* it's either dead RAM or something further up the chain... I'll eat my hat if it's a bad trace.

[psoma]

Originally Posted by hooverphonique

I don't think that makes any difference overall - the die is deeply embedded in the package, and no bond wires should be present in the corner.
Unless it's no longer air tight and moisture has been absorbed, then it's a different matter.

Many thanks for confirming what I was hoping was the case (I'm not fully across how ICs are built, but figured tolerances wouldn't be THAT particular). Either way, I can't see how the CIA could cause a green screen as it isn't involved in the POST memory check sequence.

Cheers,

Tom