Faulty Amiga 2000

[salocinx]

Hi - I am working on a faulty A2000. When I received the computer, diodes D400 and D401 (on +12V and -12V lines) were completly burned! I replaced both diodes and now I get a green boot screen when I have Kickstart 1.2 mounted and a red boot screen when Kickstart 2.04 is mounted.

Any ideas? How would you proceed?

I want this Amiga working again !

[jkp]

This page: https://en.wikipedia.org/wiki/Kickstart_(Amiga) has a list of colour codes for the boot screen. Also, the keyboard led blinks to indicate state. Start by looking at that.

[Arnie]

If you haven't done so already, then I would check the lines going to/from D400/401 for continuity.

[salocinx]

Thanks for your suggestions. I already checked the meanings of the different diagnostic screen colours:

- Kickstart 1.2/1.3: Green Screen, which indicates a problem with Chip-Mem
- Kickstart 2.04: Red Screen and blinking power LED 9x times short, 1x long indicates Chip-Mem problem

From this website, I got the following information:

A2000 (see also A500 & A1000 & A2500 for hints)
1) Red screen; power LED blinks 9 short 1 long continuously
a) bad ram with open

So it looks like I've got a problem with Chip-Mem... But on this A2000 rev. 4.1 there are 32 RAM chips soldered on the motherboard which doesn't make it easy to replace.

@Arnie: I also checked the continuity of the +12V/-12V lines connected to the diodes (D400/D401). They seem to be fine.

Any ideas how to find the bad RAM chips without de-soldering and replacing ?

[dalek]

If you have a known good ram chip you can piggy back it on top of each chip on the motherboard in turn to see if it fixes the issue. Also, feel the top of each chip and see if any are hotter than the others.

But first, what I would do is get a proper PLCC extractor and remove the AGNUS chip. Most of the green screens I've come up against have been poor seating/contact of the AGNUS chip.

So try removing it with the proper tool, cleaning the pins if both the socket and the chip to remove any of the grey corrosion and clean the legs and pin socket with isopropyl alcohol and cotton tips.

I have had success using a safety pin around both the socket and the chip to pull outward the pins so that they make tighter contact - be careful though, you don't want to pull them out too far as they could break. To make sure, you could do a continuity test from the top and bottom of the board to make sure each pin is making good contact.

If that fails, you will need to go through the ram chips one by one.

[salocinx]

Thank you very much for your suggestions! I meanwhile checked all AGNUS pins on continuity by fitting a needle to my multimeter, checking continuity directly from the pin on the IC itself to the corresponding solder points on the backplane of the motherboard. These seem all to be fine. I also replaced the AGNUS IC with another (known working) one.

Still the same problem.

Most of the custom chips (Paula, Denise, etc.) get warm after some minutes running the machine, also the 68k CPU. But the RAM ICs keep cool, there's no feel-able difference between them. Should they get warm when the Amiga is staying in the "red-screen" state ?

[dalek]

I have a rev 4.3 (same 32 ram chips) here on the bench and I ran it for a few minutes at the workbench screen. Only Paula and Denise were warm, and the CPU a tiny bit. All the ram chips were pretty cool. So that seems similar to what you are seeing.

The only other thing I could suggest is to remove the CPU and kickstart rom chips and check underneath for any damage to the traces. See if you can pin-out with your multimeter all the traces from the CPU to the right hand side of the board (ram area) - especially the address and data pins.

[cpiac64]

ignore color, verify connection from kick and cpu, next verify chip cia, paula, denise and 1488/1489

[salocinx]

Originally Posted by dalek

The only other thing I could suggest is to remove the CPU and kickstart rom chips and check underneath for any damage to the traces. See if you can pin-out with your multimeter all the traces from the CPU to the right hand side of the board (ram area) - especially the address and data pins.

Okay I will have a check on these traces and report back asap. Thanks.

Originally Posted by cpiac64

ignore color, verify connection from kick and cpu, next verify chip cia, paula, denise and 1488/1489

Do you mean U300/U301, the CIA chips. 1488/1489 are U304/305.

[cpiac64]

12 and -12 are used for serial port, video out and audio amplifier

[salocinx]

Originally Posted by cpiac64

12 and -12 are used for serial port, video out and audio amplifier

Ok, I will check those too...

[salocinx]

WOW :-D I found the solution to the problem thanks to you guys! I printed all the schematics for the A2000 and checked the address bus (A[1:23]) and data bus (D[0:15]) traces from the 68k CPU going outwards. The address line A15 was broken and not connected to the CPU. This line leads from the CPU to the Kickstart ROM and from there to the Fat Lady (aka AGNUS). So afterwards, everything makes sense and once more, I learned a lot new things about the Amiga :-) !

So many thanks for your help !!!

[dalek]

well done! another one saved

[cpiac64]

another damage caused from battery

[salocinx]

Originally Posted by dalek

well done! another one saved

Originally Posted by cpiac64

another damage caused from battery

yes, probably. the pre-owner replaced the 68k socket due to battery leakage, so I didn't thought that there was an open circuit underneath...