Is there a generic XT motherboard troubleshooting guide?

[dhau]

Hi folks, did anyone put together an XT motherboard troubleshooting guide?

I bought a new old stock XT clone motherboard from eBay.
I checked power rails before connecting power and found a shorted capacitor on -12v rail. I removed shorted capacitor.
I tested each 4164 chip individually with Arduino tester, all are fine.
I used a known good 8088 CPU, MDA card and PSU.
I next configured the switches for MDA video output and connected motherboard to power.
Nothing is blowing up, but nothing is seen on MDA output. Also no beeps on pc speaker.

At this point, I want to try landmark or Ruud diagnostic bios. I ran out of eeprom chips, so need to order a wait a bit.
If this doesn't help identify the issue, I will try to follow power good input from power supply and reset signals to all the key chips.
I don't have an oscilloscope, but I may buy one if I have to. I do have an original basic saleae logic probe, but I never used it. Will that help me see whats up with chips?

Any other ideas/advise is highly appreciated!

 

[1ST1]

Do you have a POST diagnostics card for ISA bus or for parallel port? Some BIOS output codes to port 80h on sich an ISA card, some other BIOS output diagnostics code on parallel port (like Olivetti early XT / AT / 386 / 486).

You then need the BIOS specific post code list to see which number means what.

 

[Chuck(G)]

My first impulse would be to use the Landmark BIOS diagnostics. You do need a CGA or MDA/MGA card to see the results however--a VGA or EGA card won't do.
Depending on POST codes requires a good properly functioning BIOS, which may not be the case in a dead system.

 

[dhau]

My first impulse would be to use the Landmark BIOS diagnostics.

I tried Landmark bios on ST M27C256 chip, but didn't get anything. I do use MDA adapter that works fine in other XTs.

I tested Landmark bios on a known good XT clone motherboard, and it works great with the MDA card.

 

[Chuck(G)]

At this point, I think it's time to see if the CPU is even wiggling its address and data lines. Got a logic probe or ' scope? The initial tests done in the Landmark diagnostic don't even use RAM.

 

[Ruud]

One thing you can do is negate the IOCHRDY signal on the ISA bus. Using a simple logic probe the address bus should show FFFF0 and the data bus most likely EA. If this does not happen, then something more seriously is wrong.

Good luck!

 

[Timo W.]

Do you have a POST diagnostics card for ISA bus or for parallel port? Some BIOS output codes to port 80h on sich an ISA card, some other BIOS output diagnostics code on parallel port (like Olivetti early XT / AT / 386 / 486).

You then need the BIOS specific post code list to see which number means what.

Apart from a very few exceptions, XT class machines do not generate POST codes at all. And those that do, are not using 80h, so a standard POST diag card won't work anyway.

Since you have a logic probe: check the data and address lines for activity.

 

[jafir]

I don’t think it will help in this case, since you’ve already got something similar, but I recently became aware of another option for ROM POST diags:

http://bitsavers.org/bits/WindsorTechnologies/WindsorPOST/

 

[dhau]

I don’t think it will help in this case, since you’ve already got something similar, but I recently became aware of another option for ROM POST diags:

http://bitsavers.org/bits/WindsorTec...s/WindsorPOST/

Very neat! Now we have 3 options, Landmark, Ruud's and Windsor! I guess the "unswizzled" is the one to burn in eeprom. And WindsorPOST.gar is some sort of Java program to "unswizzle"?

Google points to an ad of WindsorPOST: https://worldradiohistory.com/hd2/ID...-Page-0027.pdf

 

[jafir]

Not mine…. I couldn’t code my way out of a wet paper sack.

I guess this rom was sold as a pcb with some blinking lights and the “copy protected” the ROM by swapping a couple of pins. See here: https://twitter.com/tubetimeus/statu...486086660?s=21

I think unswizzled is the version that tube time fixed so that he could dissemble it.

Based on the photo, it looks like the PCB also allowed the thing to work in either a PC or XT, so that’s pretty cool. I’m guessing ROM were way more expensive back then?