It does seem that there are a number of suspicious vias near the keyboard controller. You can see lot of corrosion on chip pins and also ISA slot pins aswell.

Tiido wrote:

It does seem that there are a number of suspicious vias near the keyboard controller. You can see lot of corrosion on chip pins and also ISA slot pins aswell.

Yes, but it does not look that bad. The point is that everything seems to be working, so I dont know where to look for the error...

Made a "Software repair" for that board now. I dumped the BIOS and started PCem with it. Then I used a debugger to search for the code where the "A0" is put out on the post card. Quite easy as there is only one occurence of
mov al,A0
out 80,al
I NOPed everything out till the next port 80 output (mov al,A1 out 80,al).
Did the same with the BIOS checksum test (code 06). Burned back on a EEPROM.

Board is running fine now, keyboard also working 😎 😎 😎

Unfortunately I cant use the original EPROM anymore, but doesnt matter.

I have almost same board and this is interesting, this board isn't supposed to work with 256KB cache and POST never completed (same error) when I installed some. You seem to have 256KB of cache there and with this modification it'll work. Does it show the right amount of cache and is it enabled with this modification ?

My board has a Jumper to select "32 K" or "64K/128K". In the current configuration 64K are detected and used.

But as written, I think my patch fixed something with the keyboard, not the cache. But I am not sure about this, didnt check the code I deleted 😀

At which POST-code does your board hang?

Ok, interesting. Your board has 256KB installed but only sees 64KB of it. I tried my board again and now it actually starts with 256KB and works, with different cache chips than what I used last time. It seems I may be having a bad cache chip at hand instead...

Yes, I am also not sure about my chips. Just removed the 4 Motorolas in Bank 0 and it still detects 64kb. Maybe mine are also bad/not compatible. But I dont have replacements here at the moment.

My board seems to support 128kb maximum anyway.

Mine says 32/64 on one bank and 64/128 at other bank, but then there's a row of jumpers with 32K on one side and 256K on other... Past tests didn't get things going with 256KB but now it does work fine...
I remembered wrong too, my POST code was A1 not A0 when things froze.

Oha, seems to be similar.

I tried to remove the W2465AK in U16. Now it starts with 128kb Cache...

There seems to be something wrong with the configuration. As I got the board there was no cache installed at all.

I never understood the 386/486 cache configuration requirements with TAGs and different sizes and so on, so this was only trial+error...