A4000 Repair Help

[Khyron]

Hi All,
 
I pulled my amiga stuff out of storage this weekend, mainly to pull the batteries out of them, and all but one system I can't get up and running, funny part is.. that it barely had any damage, I have a 3000 that looks like an alien drooled on it, and after is was cleaned it booted right up..
 
The damage on this one was just corrosion on exposed metal surfaces around the battery, The battery was swapped about 5 years ago so It didn't actuall start leaking on to the board yet.. I cleaned it up, and tried booting the system, and nothing but a black screen.
 
Pulled all the ram, still black screen..
Swapped kickstarts and CPU board with ones that I know work, and still black screen.
 
Verified the all the power rails, and they're fine
System comes out of reset properly
 
I found a continiuty issue on the 74f245 located between the ram / battery, I cleaned it up more, and resoldered the contacts, at the same time I pulled the last simm socket, since it was corroded a bit too much and I thought there might be a chance of a short I can't see underneath..
 
after that the symptoms changed,
With (good simm) chipram installed, I still get a black screen
If I pull the chip ram I get a flashing green screen
If I put the wrong type of simm in the chipram slot I get a yellow flashing screen.
 
I'm assuming there's either an issue with the simms sockets, or U891-894, but I'd like to get some more opinions before I go ordering the parts.
Hopefully someone has seen this before?
 
 
Note: All the custom chips on this board are soldered, so there's no "easy" swap, and I don't have any spares in a socket to swap with..

[tokyoracer]

Have you checked the caps?

They wern't of great quality and are almost as prone to leaking asmuch as batteries are.

[Tumbleweed]

A black screen may well inidcate a dodge processor board. Can you check with a known good board? Is an A3640 or A3630 board? have you ckhecked the jumper settings on the main board to make sure that they are set up for the CPU like int or ext for the clock?

Weed

[TjLaZer]

Well the A4000 is prone to fatal failure when the battery even leaks a little bit.  The A2000 is a beast and will survive a major leakage.  I think it's the SIMM contacts that are nearby the clock that are one of the issues.  Good luck getting it up and running.

[Castellen]

I found a continiuty issue on the 74f245 located between the ram / battery, I cleaned it up more, and resoldered the contacts, at the same time I pulled the last simm socket, since it was corroded a bit too much and I thought there might be a chance of a short I can't see underneath..
 
after that the symptoms changed,
With (good simm) chipram installed, I still get a black screen
If I pull the chip ram I get a flashing green screen
If I put the wrong type of simm in the chipram slot I get a yellow flashing screen.

You've done a good job in covering all the common stuff so far, though failed fast memory interface is unlikely to prevent the computer from booting as you describe.

At a guess based on what you've said it's doing something along the lines of:

000,$00,First CPU instruction on powerup
001,$01,Finished checksumming the ROM
005,$05,Returned(or skipped) from Debug BootROM or XIP card
009,$09,About to disable all interrupts etc
010,$0A,Interrupts off, DMA off, screen should be dark grey ($111)
011,$0B,Set bus timeout stuff in Gary (A4000/A3000)
012,$0C,Hardware setup, ROM checksum passed OK
013,$0D,Exception vectors setup OK. (this also means low chipram OK)
015,$0F,Calling 'CPU type' test
016,$10,Returned from CPU test
017,$11,Chip RAM test (tests one word in every 16kB)
018,$12,End of Chip RAM test
023,$17,Some Error/Exception happened.

After the (not very thorough) chip RAM tests, the CPU begins the setup of exec.library in RAM.  If any of this data is corrupted, then you'll have the problem you're seeing.

So this could be caused by a problem with the memory bridge (U250, U121, U123, U211, U216).  Or the chip RAM data bus becoming corrupted by anything that lives on there; i.e U211, U400, U450, etc.  Or of course a PCB problem in any of these circuits due to an open circuit via for example.

In summary, not much more you can do yourself without some SMD tools and a collection of spare parts.

If you're interested in taking it further, Email me for a repair estimate.
http://amiga.serveftp.net

[Khyron]

Thank you all for replying!!
 

Have you checked the caps?
They wern't of great quality and are almost as prone to leaking asmuch as batteries are.

One of the first things I looked for visually before powering it up, and then I powered it up and checked the supplies with an oscilloscope, and felt each of the larger 47uF one's to see if they were getting hot
 

A black screen may well inidcate a dodge processor board. Can you check with a known good board? Is an A3640 or A3630 board? have you ckhecked the jumper settings on the main board to make sure that they are set up for the CPU like int or ext for the clock?
Weed

Yup, I tested it with a 3630/3640/Appollo 4060 which I pulled from the A3K and my other a4K
 
 

Well the A4000 is prone to fatal failure when the battery even leaks a little bit. The A2000 is a beast and will survive a major leakage. I think it's the SIMM contacts that are nearby the clock that are one of the issues. Good luck getting it up and running.

Thanks, I've still got my A2000 to check, it's been sitting the longest, so I'm kinda saving what I think is the worst for last..
 

You've done a good job in covering all the common stuff so far, though failed fast memory interface is unlikely to prevent the computer from booting as you describe.
 
At a guess based on what you've said it's doing something along the lines of:
 
000,$00,First CPU instruction on powerup
001,$01,Finished checksumming the ROM
005,$05,Returned(or skipped) from Debug BootROM or XIP card
009,$09,About to disable all interrupts etc
010,$0A,Interrupts off, DMA off, screen should be dark grey ($111)
011,$0B,Set bus timeout stuff in Gary (A4000/A3000)
012,$0C,Hardware setup, ROM checksum passed OK
013,$0D,Exception vectors setup OK. (this also means low chipram OK)
015,$0F,Calling 'CPU type' test
016,$10,Returned from CPU test
017,$11,Chip RAM test (tests one word in every 16kB)
018,$12,End of Chip RAM test
023,$17,Some Error/Exception happened.
 
After the (not very thorough) chip RAM tests, the CPU begins the setup of exec.library in RAM. If any of this data is corrupted, then you'll have the problem you're seeing.
 
So this could be caused by a problem with the memory bridge (U250, U121, U123, U211, U216). Or the chip RAM data bus becoming corrupted by anything that lives on there; i.e U211, U400, U450, etc. Or of course a PCB problem in any of these circuits due to an open circuit via for example.
 
In summary, not much more you can do yourself without some SMD tools and a collection of spare parts.
 
If you're interested in taking it further, Email me for a repair estimate.
http://amiga.serveftp.net

Thanks! I was more focused on where the battery dammage took place, it didn't pop in my mind that the chip/fast bus were isolated
 
I'll dig down that path a bit more and look for the obvious stuff, it's soo much hassle when you don't have SMT equipment at home, changing $5 worth of glue/buffers to debug which should take minutes becomes a workload.. especially when you don't have any of the parts in your stock..
I'll contact you, just my fear is shipping costs..