Ouch ! A friend once pulled an ISA modem out when computer was still on and it fried part of the south bridge/peripheral interface as it is near impossible to pull a ISA card perfectly vertical out of slot without some tilt.
If you look at ISA slot connections you can see that on the B side it is easy to get any of the AT supply volts being pumped to very important signal lines which can easily fry them.
The ones commonly fried are lines to Reset, IRQ2, DRQ2, or the ALE or OSC lines.

First I would reset the CMOS. Then, reseat the BIOS EPROM. If it still doesn't work, I'd make sure the BIOS rom is intact, and in any case, reprogram it, or even try a different chip.

tauro wrote on 2023-07-04, 01:04:

First I would reset the CMOS. Then, reseat the BIOS EPROM. If it still doesn't work, I'd make sure the BIOS rom is intact, and in any case, reprogram it, or even try a different chip.

That's was exactly my first thought. This motherboard has L27C512U EPROM Bios chip. Now, I was never able to read it for some reason... My T48 programmer simply refuses to work with it. I've tried to dump the rom before my accident, so I know that at least then BIOS was OK. I did ordered few w27C512 chips from China, but it will take some time to get it. I kind of hoped that some programmable chip was the culprit, so BIOS was the first suspect. There is another programmable chip,
PALCE16V8, however I don't know how to test it. It is socketed.

Horun wrote on 2023-07-04, 00:50:

Ouch ! A friend once pulled an ISA modem out when computer was still on and it fried part of the south bridge/peripheral interface as it is near impossible to pull a ISA card perfectly vertical out of slot without some tilt.
If you look at ISA slot connections you can see that on the B side it is easy to get any of the AT supply volts being pumped to very important signal lines which can easily fry them.
The ones commonly fried are lines to Reset, IRQ2, DRQ2, or the ALE or OSC lines.

You are unfortunately right, whenever you try to pull full size ISA card, longer side goes out first and shorter side comes out harder... I think I should be able to maybe try to test Reset line....

Does the Reset LED flash on your Diag card or stay on or never do anything ?

Horun wrote on 2023-07-04, 13:24:

Does the Reset LED flash on your Diag card or stay on or never do anything ?

Actually it flickers if I do resets, also transitions when initially powered on.

I did the same not so long ago... motherboard on a test bench and I put some fingers on the near by ISA slot to get leverage and pulled the ISA VGA card out of the slot. Only later to realize it was powered on 🤣 Actually nothing happened but it didn't want to POST in the slot the ISA card was in. Some weeks later I tested it again and all the ISA slots worked. Strange...

kixs wrote on 2023-07-04, 15:48:

I did the same not so long ago... motherboard on a test bench and I put some fingers on the near by ISA slot to get leverage and pulled the ISA VGA card out of the slot. Only later to realize it was powered on 🤣 Actually nothing happened but it didn't want to POST in the slot the ISA card was in. Some weeks later I tested it again and all the ISA slots worked. Strange...

I confess that I did it once too, and the board survived.

What I've seen with 386 motherboards is that the traces sometimes are very frail, very delicate, and plugging or unplugging cards or even slightly bending the board can be the difference between post and no post. It depends on the quality of the board and the possible damage it has seen over the years.
Sometimes these frail traces are not visible, as they are in an inner layer. These are delicate things...

Still you can probably fix this board, provided you have the knowledge/patience/time/will.

I'd check the CPU on a different board, just in case.
Try plugging some cards on different ISA slots to see if there's any reaction. Plug in a speaker in case the board partially works but the ISA slots don't.

HansZ wrote on 2023-07-04, 02:13:

That's was exactly my first thought. This motherboard has L27C512U EPROM Bios chip. Now, I was never able to read it for some reason... My T48 programmer simply refuses to work with it. I've tried to dump the rom before my accident, so I know that at least then BIOS was OK. I did ordered few w27C512 chips from China, but it will take some time to get it. I kind of hoped that some programmable chip was the culprit, so BIOS was the first suspect.

The first thing is to make sure your BIOS is intact, could you make a dump before the incident?

Here's some bios images that should be compatible
https://theretroweb.com/bios/?chipsetId=735

HansZ wrote on 2023-07-04, 02:13:

There is another programmable chip,
PALCE16V8, however I don't know how to test it. It is socketed.

I wouldn't mess with that one!

Usually the one that comes out first was 16 bit part when card is pulled first is the free end, This means back 8 bit part with bracket pushes at the back panel, causing card to pivot forward in the ISA slot (sliding forward slightly) Causing B7 to juice -12V into next B8 pin. Also any ground pin that is next to signal in ahead will be pulled down hard, blowing the TTL driver.

If you pull from bracket end, it is very hard to do and usually comes out clean.

Cheers,

CPU is surface mounted, I think I can replace it if required. if 14mhz Freq generator is damaged, would that cause motherboard to not to post anything at all?

What does the "CLK" LED on your POST-card say?

I wouldn´t worry about the CPU. The ISA bus is hooked to the peripheral controller "VT82C481" (Venus Chipset).
If there´s any damage that was caused by mishandling an ISA card it´s probably inside this chip.

To verfy this you should probe all the IRQ, DRQ, DACK, data and address lines with a multimeter. Measure the resistance between each line and ground with the red probe grounded and watch out for any aberrations / non-typical values.

HansZ wrote on 2023-07-03, 22:41:

I don't have an oscilloscope to test the ISA

A scope would be very useful here. It's not worth buying one just for this mobo but if you were planning to it's a good opportunity. Otherwise - if you really want to try and repair this board, or you are just looking for a "project" - maybe consider some cheap 20Mhz kit with an LCD screen, I've seen those rather cheap. These are not exactly great, especially for a first scope, but will do.

Could be the mobo is somewhat functional but the "south bridge" chip is damaged so ISA isn't working properly. Though depending on the exent of the damage it might prevent the mobo from booting at all. Things to check - is there any activity on BIOS ROM? On CPU data and address lines? On ISA data and address lines? With some luck maybe it wasn't power to I/O short but some driving IC got busted (like a '245 that drives low 8 data bits).

There's quite a lot that can be tested before you start doing any soldering - because otherwise you'd be blindly trying to swap parts, I mean that's a "project" as well but not very fun I think. So consider a scope, if not now then perhaps it would be better to put this mobo into a box and return to it when you have better tools at your disposal.

majestyk wrote on 2023-07-05, 16:03:

What does the "CLK" LED on your POST-card say? […]
Show full quote

What does the "CLK" LED on your POST-card say?

I wouldn´t worry about the CPU. The ISA bus is hooked to the peripheral controller "VT82C481" (Venus Chipset).
If there´s any damage that was caused by mishandling an ISA card it´s probably inside this chip.

To verfy this you should probe all the IRQ, DRQ, DACK, data and address lines with a multimeter. Measure the resistance between each line and ground with the red probe grounded and watch out for any aberrations / non-typical values.

Results from the top ISA slot from which the card was pulled out:
GND to IRQ10 to 14 13k Ohms
DACK 0 12.2M (Mega)
DRQ 0 4.7k
DACK 5 12.6M
DRQ 54.7k
DACK 6 12.3M
DRQ 6 4.7K
DACK 7 13M
DRQ 7 4.7k

All the DACK lines offer resistance in the mega ranges....

Deunan wrote on 2023-07-05, 23:27:

A scope would be very useful here. It's not worth buying one just for this mobo but if you were planning to it's a good opportun […]
Show full quote

HansZ wrote on 2023-07-03, 22:41:

I don't have an oscilloscope to test the ISA

A scope would be very useful here. It's not worth buying one just for this mobo but if you were planning to it's a good opportunity. Otherwise - if you really want to try and repair this board, or you are just looking for a "project" - maybe consider some cheap 20Mhz kit with an LCD screen, I've seen those rather cheap. These are not exactly great, especially for a first scope, but will do.

Could be the mobo is somewhat functional but the "south bridge" chip is damaged so ISA isn't working properly. Though depending on the exent of the damage it might prevent the mobo from booting at all. Things to check - is there any activity on BIOS ROM? On CPU data and address lines? On ISA data and address lines? With some luck maybe it wasn't power to I/O short but some driving IC got busted (like a '245 that drives low 8 data bits).

There's quite a lot that can be tested before you start doing any soldering - because otherwise you'd be blindly trying to swap parts, I mean that's a "project" as well but not very fun I think. So consider a scope, if not now then perhaps it would be better to put this mobo into a box and return to it when you have better tools at your disposal.

Bought this one:

https://www.aliexpress.com/item/33023239239.h … .c5ff1802B3linG

What about DRQ 1, 2, 3, 4 and 5 / DACK 1, 2, 3, 4?

Are the DRQ and DACK lines connested directly to the peripheral controller or to any of the DIL circuits between / right of the ISA slots (244 / 245 buffers)?

HansZ wrote on 2023-07-06, 15:02:

Bought this one:

https://www.aliexpress.com/item/33023239239.h … .c5ff1802B3linG

Yeah, I don't quite belive it has 100MHz BW, or 500Ms/s for that matter (or if it does it's a very small memory buffer), but it should do nicely. Does it come with 10:1 probes? If not get yourself some cheap ones, these days you can actually buy 100MHz 10:1 input probes that cost next to nothing and are pretty decent.

You want to inspect ISA bus, address, data, primary control lines (read, write). You don't need to test every single pin - first try a few bits of each byte (both data and address), see what, if anything, is toggling. If there is no activity at all you need to move to mobo proper - CPU and BIOS are your next targets.

majestyk wrote on 2023-07-06, 18:58:

What about DRQ 1, 2, 3, 4 and 5 / DACK 1, 2, 3, 4?

Are the DRQ and DACK lines connected directly to the peripheral controller or to any of the DIL circuits between / right of the ISA slots (244 / 245 buffers)?

GND to IRQ 2 = 7.6 k
GND to DRQ 2 = 4.7k
GND to DACK 3 = 4M
GND to DRQ 3= 4.7k
GND to DACK 1 = 1.45M
GND to DRQ 1= 4.7k
GND to IRQ 7= 12.9k
GND to IRQ 6=12.9k
GND to IRQ 5 =12.9k
GND to IRQ 4=12.9k
GND to IRQ 3=12.9k
GND to IRQ 2=12.9k
GND to DACK 2 = 8.5M

DACK lines offer unstable resistance measured in MEGA values. Tested DRQ and DACK for connectivity and didn't got any beeps at all.