Problems about repairing a PC/XT 5160 motherboard

[modem7]

As I guess, there is a problem with Base 16K RAM.

Maybe, but of course, to successfully read/write RAM, the entire RAM subsystem needs to be working. That involves many components, some common to other subsystems. One example of a component in the RAM subsystem is the delay line.

But it can't be just only the RAM subsystem that is faulty. TEST6083 fails, and that has nothing to do with RAM. (TEST6083 runs from ROM, does not use RAM, and only manipulates a pin on the 8255 chip.)

 

[OctSL]

Do you have a means to independently prove that the memory chips you’re using for swaps are good?

I used one of these to prove bad 4116 and 4164 memory in my 5150 and to validate the replacement stock I bought from EBay also was working correctly.

I have many 4164 chips and they may not all be bad, I only used 9 chips in Bank 0.
How to simply test them?

 

[OctSL]

Maybe, but of course, to successfully read/write RAM, the entire RAM subsystem needs to be working. That involves many components, some common to other subsystems. One example of a component in the RAM subsystem is the delay line.

But it can't be just only the RAM subsystem that is faulty. TEST6083 fails, and that has nothing to do with RAM. (TEST6083 runs from ROM, does not use RAM, and only manipulates a pin on the 8255 chip.)

Is TEST6083 showed the 8255 chip is bad?
When testing, the voltage of 8255 Pin 3 dropped only twice, and the rest were 5V.

 

[modem7]

Is TEST6083 showed the 8255 chip is bad?

Failure of TEST6083 could be because of a bad 8255, but there are other possible causes.

Examples:
- Maybe some of the code in the TEST6083 ROM is corrupted on its way to the CPU.
- Maybe I/O writes are not working.
- Maybe the chip-select logic is faulty in way that results in the 8255 not getting selected.
- Maybe 8255 commands (some of the TEST6083 code) from the CPU are corrupted on their way to the 8255.

A logic analyser (and the knowledge of how to use it), together with reference to the TEST6083 code and 8255 data sheet would show what is and what is not happening.

As can be seen from the TEST6083 code source (targeting the IBM 5160) at [here], TEST6083 is not doing much. (Yes, I know - the code is not optimised.)

When testing, the voltage of 8255 Pin 3 dropped only twice, and the rest were 5V.

Twice is odd.

 

[OctSL]

Failure of TEST6083 could be because of a bad 8255, but there are other possible causes.

Examples:
- Maybe some of the code in the TEST6083 ROM is corrupted on its way to the CPU.
- Maybe I/O writes are not working.
- Maybe the chip-select logic is faulty in way that results in the 8255 not getting selected.
- Maybe 8255 commands (some of the TEST6083 code) from the CPU are corrupted on their way to the 8255.

A logic analyser (and the knowledge of how to use it), together with reference to the TEST6083 code and 8255 data sheet would show what is and what is not happening.

As can be seen from the TEST6083 code source (targeting the IBM 5160) at [here], TEST6083 is not doing much. (Yes, I know - the code is not optimised.)


Twice is odd.

Thanks for your help!

 

[OctSL]

Maybe, but of course, to successfully read/write RAM, the entire RAM subsystem needs to be working. That involves many components, some common to other subsystems. One example of a component in the RAM subsystem is the delay line.

But it can't be just only the RAM subsystem that is faulty. TEST6083 fails, and that has nothing to do with RAM. (TEST6083 runs from ROM, does not use RAM, and only manipulates a pin on the 8255 chip.)

After testing, 36 RAM chips are all good, it verified my prediction that the circuit controlling RAM was damaged.
However, there was a problem with the 8255 chip, which was unexpected. I will test it when the MDA card and monitor arrive in a few days.

 

[OctSL]

Per [here], if you step back through the received codes, you expect to see {00}{01}{02}{03}{04}{05}{06} before the {07}.


Not at this point. The code in TEST6083 is small, and I was curious to see if it worked for you.

Given the information that you have presented so far, I think there may be a problem related to the address and/or data buses. Do all 13 steps at [here] produce the expected results? That is not a complete test of the address and data buses, but if you do not see the expected results, the result can lead the way to identifying the problem cause.

I now suspect that the 8259 is broken because the keyboard does not work, diagnostic program often gets stuck and cga card often does not work normally.
Is there a diagnostic ROM for 8259?

 

[modem7]

I now suspect that the 8259 is broken because the keyboard does not work, diagnostic program often gets stuck and cga card often does not work normally.
Is there a diagnostic ROM for 8259?

I am only aware of the Supersoft/Landmark Diagnostic ROM and Ruud's Diagnostic ROM.

 

[OctSL]

I am only aware of the Supersoft/Landmark Diagnostic ROM and Ruud's Diagnostic ROM.

After inserting the CGA or mda card, it can start most of the time, but it cannot display normally and the ROM cannot recognize the video card.
No open circuit on ISA slots.

 

[OctSL]

It is worse that the CGA cannot work and the Supersoft ROM cannot find the card and beeped.

 

[modem7]

Back to low level.
Do all 13 steps at [here] produce the expected results ?

 

[OctSL]

Back to low level.
Do all 13 steps at [here] produce the expected results ?

Fortunately, yes.
But you said " No error beeps from the Supersoft ROM",actually the board beeped.

 

[OctSL]

What a pity that this forum cannot upload the full video of testing the board by SuperSoft ROM.

 

[OctSL]

 

[modem7]

Back to low level.
Do all 13 steps at [here] produce the expected results ?

Fortunately, yes.

Sometimes the Supersoft ROM is displaying something to you, other times is it is not not. But maybe the Supersoft ROM is the problem (e.g. not created properly).

Question #1

Does the I/O CH RDY procedure (uses original IBM ROM) always work, or does it sometimes fail? If it sometimes fails, then we know that the Supersoft ROM is good, and that it is the motherboard that is intermittent. And if it sometimes fails, is it always the same step (of 13) that fails, or is it a different step each time?

Question #2

If I look back through the posts, I see that TEST6083 failed. Does that now work, or does it just work some times?

I hate intermittent symptoms.

Some things to try

* In case of poor connection between the 8088 CPU and its IC socket, re-seat the CPU a few times.

* Remove all RAM chips, in case of some form of interference. They are not needed for the Supersoft ROM to display something.

* If an 8087 NPU is fitted, remove it, in case of some form of interference. It is not needed.

 

[OctSL]

New progress!
The board worked after I use the 82’ IBM BIOS and fixed the RAM controlling circuits.
But the 8255 maybe died because the keyboard is unavailable (Error code 601) and the board will often be force reset after memory testing.
This is a picture of started displays that was not reset (sadly I did not take the picture of it displayed 601 error):

And the 8087 died.

 

[OctSL]

Sometimes the Supersoft ROM is displaying something to you, other times is it is not not. But maybe the Supersoft ROM is the problem (e.g. not created properly).

Question #1

Does the I/O CH RDY procedure (uses original IBM ROM) always work, or does it sometimes fail? If it sometimes fails, then we know that the Supersoft ROM is good, and that it is the motherboard that is intermittent. And if it sometimes fails, is it always the same step (of 13) that fails, or is it a different step each time?

Question #2

If I look back through the posts, I see that TEST6083 failed. Does that now work, or does it just work some times?

I hate intermittent symptoms.

Some things to try

* In case of poor connection between the 8088 CPU and its IC socket, re-seat the CPU a few times.

* Remove all RAM chips, in case of some form of interference. They are not needed for the Supersoft ROM to display something.

* If an 8087 NPU is fitted, remove it, in case of some form of interference. It is not needed.

According to the things after the board is turned on,I think the 8255 died and unstable causes these problems.

 

[chjmartin2]

Or is there a converter project that can convert cga to vga?

MCE2VGA is the best one IMHO.

 

[OctSL]

MCE2VGA is the best one IMHO.

Thanks a lot!

 

[OctSL]

Oh no