Then I try to put all the things together in the case and it never start again. No way, always keyboard error.
I remember many years ago, someone on these forums had a motherboard that worked out of the case, but not when fitted in the case. It turned out that when the motherboard was screwed into the case, part of the motherboard's circuitry was somehow being shorted out to the (metal) case.
Does the symptom only appear when the motherboard is in the case ?
Did the connector get plugged in the wrong orientation ?
I checked measures with other xt motherboard that uses the same 322 and there was different continuity with pins in the 322 and the data pin of the connector.
I wrote earlier, "
PC-class and XT-class computers use variations of that."
Expecting different models of motherboard to use
exactly the same keyboard interface circuitry (and that is more than simply the type of chips used) is invalid.
I'm a little bit lost at this moment...
So, there's the question about whether the symptom only appears when the motherboard is in the case.
Beyond that, it is then a case of using suitable test equipment to look at various places in the keyboard interface circuitry. I have helped people do that diagnosis. In some cases, I had to create small amounts of diagnostic code for the owner to put into a ROM (replacing the existing BIOS ROM).
You do not have a circuit diagram for the motherboard, but
initially, one will not be required. For example, a logic probe is expected to show the POST pulsing the CLOCK line. For example, on the LS322 shift register, a logic probe is expected to show a pulse on pin 12 each time that a key is pressed. Beyond certain basics like that, you would need to work out for me what the circuit is (components, interconnections, etc.)