U34 8253 failure - replace or new motherboard?

[Divad]

Hi all

I'd hoped that after replacing the caps, RAM bank 0, and the U33 chip this board might finally be ready to work, but sadly it has now thrown up a failure of all U34 8253 timer channels

(This is a 16-64K 5150 motherboard)

How practical is it to replace this chip, and/or how easy is it to find a replacement? Given the problems I've had with this board would it be easier to source another one and use this as a parts donor? On the other hand I've put a fair bit of time and effort into this one already...

Thanks

David

 

[3pcedev]

Not hard to find a replacement. All the clones etc have one. Some clones even have socketed 8253's which make life very easy.

As for physically replacing it; it depends on your soldering skills. Be patient, use a good quality desoldering braid and it's really not too bad.

 

[H-A-L-9000]

The 82C54 can still be bought new. The 8253 is available at ebay.

To replace, as you probably don't have a rework station, you can carefully cut the pins at the chip then un-solder them one by one.

But be aware that this failiure can also be caused by a broken clock input signal, or addressing problem.

 

[Divad]

Thanks, I was just concerned this might be an indication of a more general failure, or that a compatible chip might be vanishingly rare.

Anyway, I found a replacement 8253-5 on ebay. Hopefully it will get this working at last...

Thanks again for your advice

David

 

[Chuck(G)]

Before you haul out the soldering iron, do check the input and output of the 8253--they rarely (but have been known to) go bad. It could be one of the associated ICs.

 

[Divad]

Checked with a logic probe, all 3 CLK pins (input?) are pulsing, and OUT 0 and OUT 2 are also pulsing. OUT 1 is low. This is with the standard U33 IBM BIOS installed.

I assume this means it's getting a signal from the clock crystal at least?

Full chip check:

1. TO 8. HIGH
9. Pulsing
10. Pulsing
11. HIGH
12. LOW
13. LOW
14. HIGH
15. PULSING
16. OPEN
17. PULSING
18. PULSING
19. TO 24. HIGH

Before you haul out the soldering iron, do check the input and output of the 8253--they rarely (but have been known to) go bad. It could be one of the associated ICs.

 

[Chuck(G)]

OUT1 is the DMA request for memory refresh, so it's entirely possible that the 8253 has gone to the angels. It's unusual, however. But I'd still use a diagnostic ROM (Supersoft or Landmark) to pin the issue down before you start warming up the iron.

 

[Divad]

Sorry, I thought I'd included this in my first post, but checking back I didn't. Anyway - I've used a Supersoft/Landmark ROM on this board and it indicated a failure of all 3 8253 timer channels...

OUT1 is the DMA request for memory refresh, so it's entirely possible that the 8253 has gone to the angels. It's unusual, however. But I'd still use a diagnostic ROM (Supersoft or Landmark) to pin the issue down before you start warming up the iron.

 

[Chuck(G)]

Hmmm. So, during the whole boot-up process you get no pulses at all on pin 21 (CS/) of the 8253? In that case, I'd start by replacing the LS138 at U66 that supplies this signal. If that does nothing, then I'd attach the 8253.

Does that make sense?

 

[Divad]

Just checked with the probe again there, for the first couple of seconds after power on, pin 21 does pulse, then goes high (and stays high).

I also checked pins 9-16 on the LS138, pins 9, 12, & 13 pulse for the first couple of seconds then go and stay high. Pins 10, 11, 14, 15 & 16 pulse twice then go and stay high if that's any help?

Hmmm. So, during the whole boot-up process you get no pulses at all on pin 21 (CS/) of the 8253? In that case, I'd start by replacing the LS138 at U66 that supplies this signal. If that does nothing, then I'd attach the 8253.

Does that make sense?

 

[Chuck(G)]

I wonder if there's any way you can get the diagnostics to keep hitting the 8253 chip-select. If that creates pulses, then yup, you might as well replace the 8253 as planned.