• Please review our updated Terms and Rules here

Xi8088 won't boot

T

TO.

Member
Joined
Feb 8, 2021
Messages
16
Location
Wisconsin
Having a problem with my Xi8088. The machine will POST (through code 00), I can see RAM count, ROM code is seen and executed if existing.

two issues:

1. The keyboard does not work
2. The machine POSTs all the way to "Booting OS", but then won't go further.

Any thoughts on where to go from here?
 
My first thought: do you have the correct keyboard controller? I had various keyboard controllers laying around and none worked. At the end Sergey had to send me one. And it is possible that a bad controller is the cause of your second problem.
 
It is very possible that I have a bad keyboard controller, but I don't have a spare to test with. I used a VIA VT82c42N as suggested.
 
What crystal are you using right now? Depending on parts you used, it may not work well above ~4.77MHz, so I'd start at 14.31 xtal. If it works, then more to may be 24, then 30 and finally to 36.
 
I'm using a 30MHz Crystal and a V20 CPU. The BIOS POST screen is great, just won't "do the next thing" - POST code is 00

This board has some LS parts, some F parts, I double checked that all parts are in the right place.

I don't know other than "boot the OS" what the next part in the cycle is, so not sure what to troubleshoot on that front.

Yesterday I tested RAM and I swapped the DMA controllers with no improvement.

I've also reflashed the ROM a few times and it gets to the same spot each time.

Today I'm trying to trace out the connections between the VT82c42N and the PS/2 port to make sure the CLK and DATA signals are making it.
 
Following up on this, I figured out why newer ROMs wouldn't flash - MacOS uses a different default language and the command that was `tr "\000" "\377" didn't produce `FF` hex like it should have. After that fix and some additional tracing/testing/troubleshooting, I'm now getting an "08" POST Code that continues to lead me down the "bad VT82c42" route. I've ordered a 10pk and we'll see how that works out. I'm fairly certain I won't wind up with 11 bad keyboard controllers, but you never know.
 
New keyboard controllers arrived, no joy. Although I don't think its the backplane causing grief, I've ordered a 2nd backplane and we'll see if that is the problem. If not, I'm also looking for a 486 SBC
 
Closing out this thread for anybody that comes back to search.

The second back plane came in and didn't offer any relief.

I ordered a TL866 from Ben Eater, the last one I got was a fake and got returned.

From there I took all of the 74 series logic chips off the board and tested them, they all tested fine, except for one that didn't have a test profile... there were a lot of them to test and this was definite progress.

Next up I looked for any other chips I could test and found the Interrupt controllers had a test profile and tested them out. BOTH were bad. I removed an 8259 from another motherboard, tested it clear and then stuck it on the board - the machine now BOOTS. Time to order some 8259s.

Lessons learned:

Fake parts suck and its super hard to tell you're getting/gotten fake parts
POST CODES display the last known success, not the current failure. If I knew this, I could have zeroed in on the 8259s right away. It displayed 08 = keyboard... but that meant "keyboard good"

Thanks for all the help!
 
TO. said:
POST CODES display the last known success, not the current failure.
Click to expand...
Putting aside this particular motherboard BIOS, readers should be aware that some 'POST code' information sources are wrong/misleading. For example, the "IBM AT BIOS:" table at the bioscentral.com web site. Let's pick a line in that table, e.g. POST code 08. 08 is described there as, "DMA page register tested". That is wrong. A look in IBM's published source code for IBM AT's motherboard BIOS reveals the 08 test as:

Step 1: Output the POST code of 08.
Step 2: Test the DMA page register.
Step 3: If the test fails, halt the CPU.

So, on my IBM AT (with IBM BIOS ROM's), if 08 was the last POST code shown on my POST card, the DMA page register test failed. bioscentral.com's "DMA page register tested" would very much mislead me. Instead it should be something like, "Starting test of DMA page register."

Also, bioscentral.com does not show that during the lifetime of the IBM AT, IBM changed the order of two of the tests. (See [here].)

Short descriptions like "RAM refresh" can be misleading. The IBM BIOS for the IBM AT tests only part of RAM refresh.
 
You must log in or register to reply here.
Back
Top