Amiga 2000 (rev 4.4) - Halting Issue

[GeekNathan]

Hi Everyone

I am fault finding an Amiga 2000 that is presenting with black screen on boot up. Here's what I have been checked and determined so far:

• There is battery damage. It isn't as bad as I have seen in some youtube content, but I've cleaned it up a bit more and checked everything in that area and it appears fine.

• In terms of checking traces, I have toned out all the connections on the CPU, ROM, and Gary. All connections appear to be fine.

• Scoping:

I have checked reset and it is going high after boot as I would expect.
I've checked reset where it terminates on chips around the board. All good.
I have pulled and checker the 4.7k resistor pack that ties in to reset and halt and it measured fine on all pins.
The halt signal goes high at boot but about 12.5 pico seconds later it drops . The reset stays high.
I've chan'd in the kbreset on the 339/gary and the romenable on the rom to watch what happens during that period. It looks fine. kbreset goes high and stays high. Rom enable has a burst of data and stops a couple pico seconds before halt drops.
(I've captured wave forms, so i can post them on request).

Gary appears to be the one asserting halt. I have tried verifying this by booting the CPU with the halt pin out (and watching it not halt) and by booting the system with gary out and the halt also doesn't occur.
From my understanding Gary only asserts a halt based on kbreset. I'm unsure what other circumstance Gary may assert a halt. This is the main hypothesis I am stuck on. Why is Gary asserting a halt?

I've pulled a lot of logic and tested the, in an RCT pro and the ones I have pulled so far have passed, but this was early trouble shooting before I isolated the halting.

I should note this is my first Amiga/16 bit machine troubleshooting. I've done a tonne of 8 bit machines and this has been a fascinating learning experience.

Here's the ROM enable between boot and halt going active low.......

EDIT: Forgot to include that I have a working A500 here and I have verified all the main custom chips as working. So the Gary is among the chips I have tested and seen work on a good A500.

[Shadowfire]

Is BERR being asserted? This could be a double-fault, due to the 68000 not seeing valid ROM data.

[GeekNathan]

Originally Posted by Shadowfire

Is BERR being asserted? This could be a double-fault, due to the 68000 not seeing valid ROM data.

BERR goes high with main power and RST goes high at the usual time and then HLT goes low per the above.

One interesting development from last night. I wanted to see if i could stop Gary from asserting halt and see what would happen. The result is that the CPU eventually (1.5 seconds later, an eternity in a scope lol) asserts halt. See attached.

So in terms of timing, and assuming the circumstance Gary and the CPU would assert halt are the same, I am thinking because Gary senses the issue and asserts halt first, it might be something at fault directly connected to Gary. So I'm going to start pulling the logic around Gary out and testing it in a retro chip tester pro.

[GeekNathan]

Originally Posted by Shadowfire

Is BERR being asserted? This could be a double-fault, due to the 68000 not seeing valid ROM data.

In terms of ROM, I have probed the ROMEN on the kickstart during the RESET to HALT period and I have attached what was seen on the ROMEN line.

[dalek]

Halt being asserted is probably an error on the address or data bus.

Some suggestions:
- sounds like you've already verified Agnus in the A500 - if not do that and also check that each pin is making contact in the PLCC socket
- check if any of the RAM chips are getting warm and that the signals look okay on the scope
- grab diagrom and read the RomStartup.txt to see how far you get

[GeekNathan]

Originally Posted by dalek

Halt being asserted is probably an error on the address or data bus.

Some suggestions:
- sounds like you've already verified Agnus in the A500 - if not do that and also check that each pin is making contact in the PLCC socket
- check if any of the RAM chips are getting warm and that the signals look okay on the scope
- grab diagrom and read the RomStartup.txt to see how far you get

Hi, thanks.

Agnus is good.
None of the ram or the logic near it feels too hot.
I have toned out a lot of the connections on Angus from Gary and the CPU. I've also reflowed the connections on the board.

But what I find difficult to see is if the Agnus chip makes positive contact inside the socket. Even with a hand magnifying glass I'm not seeing the pins touch the plcc socket contacts. I have replacement sockets on the way.

Today I pulled out and tested a lot of logic around Gary and agnus. All passed in an RCT Pro.

[GeekNathan]

Update!

Not fixed yet, but I did find some corroded traces near the battery damage.

This is despite toning out the connections previously and thinking they were fine. There were two small breaks. One was under the CPU socket (E7V signal but doesn't connect to the CPU) which I was able to see once i removed it. The other (connects the CPU to ROM) I was not able to be seen at all until I removed the resistor pack and solder, and then scrub the area with a fibre glass pen and then it was revealed.

A little frustrating, but the lesson learned is that toning things out is reliable, but not infallible.

Tonight I will place in a new CPU socket and re-test everything again.

[GeekNathan]

Update!

After re-socketing the CPU and Agnus and re-testing the halt has gone away

Now I am being presented with a light grey screen and the system resets after about 8-10 seconds. I connected a LED to the led header and it flashes 10 short, 1 long then the reset happens.

I've checked a few references about the system initialisation and found some good ones.....

https://ftp.fau.de/aminet/docs/hard/Amigafix.txt
https://www.lemonamiga.com/forum/viewtopic.php?t=13978

One suggests looking at the CIAs. I have swapped them for known working ones but no joy. The sockets are original though.

Another suggestion is the RAM might not be ok. I've probed around but not sure what to look for if there were faulty RAM in there. None seem too hot to touch.

So progress has been made. But the information on the light grey screen plus 10+1 blinking led is not perfectly clear.

[LIV2]

Your best bet now is to use DiagROM to test the system. If the chip mem is bad it will tell you which address & bits are faulty which will indicate which chip it is.

[GeekNathan]

Originally Posted by LIV2

Your best bet now is to use DiagROM to test the system. If the chip mem is bad it will tell you which address & bits are faulty which will indicate which chip it is.

I agree. I have some 4Mbit ROMs coming so I can burn a diagrom. I also pulled out and tested the buffers on the ram but they passed.

I have seen at least two exinations for the LED blinking. One says to check Agnus, their other suggests RAM. All appears in forums or from old newsgroups. Anyone seen these codes in Amiga service manuals?

[GeekNathan]

Ok.

I managed to find a bad dram, swap it out, and now I am not getting the 10 short/1 long blinking.

I still need to get colour video out because I can't tell what colour the screen is but since fixing the RAM it seems to change colour twice (total 3 colours) before it stops. Maybe someone know what the "3rd" colour is during the start up?

I tried jerry-rigging a patch from the digital rgb/h-v-sync lines (under the video port) to an RGBtoHDMI but I think there is more to getting colour video out of these things.

[mech]

Nice job troubleshooting!
This link might help:

https://retrocomputerverzamelaar.nl/...miga-problems/

The colors should be dark grey,light grey,white

Bad Agnus sockets and chipram problems are not uncommon. Trace damage under the 68000 and kickstart sockets happens too.

When you get it up and running,you may want to check out the upgrades for Rev4.4 boards to bring them up to spec,they were sent to dealers as demo machines and never intended to be sold if my memory is right. There are also brand new reproduction A2000 Pc Boards you can swap parts to.

[GeekNathan]

Originally Posted by mech

The colors should be dark grey,light grey,white

Ha! That's the colours I am getting off the composite output. I wasn't sure if that was just because it was B&W. It stops on white and doesn't appear to proceed further. I have two 2.04 kickstart rom ICs here but the result is the same on both.

I have replaced the 68k and rom socket and I checked all the traces in that area (i repaired two traces near where the battery was). I think I got them all (but i thought that at the start until i found a flaky partially broken connection).

[mech]

Originally Posted by GeekNathan

Ha! That's the colours I am getting off the composite output. I wasn't sure if that was just because it was B&W. It stops on white and doesn't appear to proceed further. I have two 2.04 kickstart rom ICs here but the result is the same on both.

I have replaced the 68k and rom socket and I checked all the traces in that area (i repaired two traces near where the battery was). I think I got them all (but i thought that at the start until i found a flaky partially broken connection).

I have had the same exact experience repairing them over the years.. Its not hard to miss one or have a intermittent connection that just happens to work when you check it,but not under load.


Good luck,i am sure you will get it going!

[GeekNathan]

when you get it up and running,you may want to check out the upgrades for rev4.4 boards to bring them up to spec,they were sent to dealers as demo machines and never intended to be sold if my memory is right. There are also brand new reproduction a2000 pc boards you can swap parts to.

I think I have seen a field upgrade guide to bring them up to 4.5.......

With regards to the troubleshooting guide in that link I have discounted all but one item in the "white screen" row, that being the EMI 1301 1 ohm Resistor. On this A200, the only 1ohm resistor I can find is between Gary and Paula (R200 on this board). Is that the resistor being referred to as EMI 1301?

[Thomas Richter]

Contrary to what some people believe,the Amiga does *not* use blink codes to communicate errors. The blinking is always the same, no matter what is wrong. The type of the error is communicated through the system alert number that appears on the screen. If no alert number appears on the screen, the system is in such a bad shape that it cannot create the display. Only a diag ROM will help you then.

[GeekNathan]

Originally Posted by Thomas Richter

Contrary to what some people believe,the Amiga does *not* use blink codes to communicate errors. The blinking is always the same, no matter what is wrong. The type of the error is communicated through the system alert number that appears on the screen. If no alert number appears on the screen, the system is in such a bad shape that it cannot create the display. Only a diag ROM will help you then.

Interesting! This is part of the reason I was curious about the blinking codes being mentioned in a proper service manual. I could only find these codes mentioned in various forums.

I'm definitely in diagrom territory. I should have eproms in hand in a week or two.

While waiting for those I have been trying to tap the digital rgb (and sync) connections and run them through my rgbtohdmi. I just soldered some right-angle jumpers on the pins under the video connector and connected them to the digital hat for the rgbtohdmi. No joy getting it to work so far.

[GeekNathan]

OOooooookay

It appears to be running again

I know exactly what I have done since the last test, but I'm not clear how it was stopping the system from starting properly.

I have been, and continue to, wait for eproms so i can burn a diagrom and see if it could identify the issue.

While I've been waiting, I stripped the power supply, cleaned it a bit more, and recapped it. The capacitors I pulled out of the power supply had not drifted much, but the power output after i replaced them was much closer to the voltage they should be.

I also replaced some capacitors on the board (see picture). They had drifted quite a lot.

I usually replace capacitors with the view it is preventative maintenance. I have read (possibly in these forums) that some passives can degrade and cause instability. Perhaps this was what I resolved. I'm not sure.

But it appears to be booting properly . Next will be run some software (once I get some on disk).

Any thoughts from more experienced Amiga repairers?