Looking at the POST codes in binary, 63h is 1100011, and F7h is 1110111. In both codes, bit 3 (counting from 0) set logic low (0). Also, bit 3 is logic low in that weird post code B6h, and it's naturally zero in C1h too... not sure if that means anything, but if it does it might indicate that data bit 3 isn't getting through to the ISA bus.

Causes of that could be:

• Dirty ISA slot pins
• A lifted pin on the 85C471
• A broken trace or cracked joint someplace
• One of the 74F245 or other TTLs is shot (quite common unfortunately, these things get *thrashed* in regular usage)
• A broken 85C471 (thankfully less likely than the other potential causes)

But before we go down that rabbit hole, have you tried putting the VGA card and POST card in different slots?

Vipersan wrote on 2021-09-09, 16:06:

No speaker atm ...but got a couple of those mini piezos on order. As to the card currently in there ... A basic Trident TVGA 900 […]
Show full quote

Eep386 wrote on 2021-09-09, 16:00:

Do you have a speaker hooked up?
If there's a problem with the VGA, the board typically will give a beep pattern regarding the error.

What VGA cards are you trying? It could be that your VGA cards have issues or need some 'minor' reconfiguring as well.

No speaker atm ...but got a couple of those mini piezos on order.
As to the card currently in there ...
A basic Trident TVGA 9000 B ...which I have tested in another machine ..and works ok
How about the post card codes Eep....do they look ok to you ?
rgds
VS

Your POST board looks like it has a speaker and should have came with leads to connect to the mobo

drosse1meyer wrote on 2021-09-09, 16:34:

Vipersan wrote on 2021-09-09, 16:06:

No speaker atm ...but got a couple of those mini piezos on order. As to the card currently in there ... A basic Trident TVGA 900 […]
Show full quote

Eep386 wrote on 2021-09-09, 16:00:

Do you have a speaker hooked up?
If there's a problem with the VGA, the board typically will give a beep pattern regarding the error.

What VGA cards are you trying? It could be that your VGA cards have issues or need some 'minor' reconfiguring as well.

No speaker atm ...but got a couple of those mini piezos on order.
As to the card currently in there ...
A basic Trident TVGA 9000 B ...which I have tested in another machine ..and works ok
How about the post card codes Eep....do they look ok to you ?
rgds
VS

Your POST board looks like it has a speaker and should have came with leads to connect to the mobo

Sadly came without the leads or indeed a manual...
all for cheapness I guess..

Close inspection of the 85C471 shows no shorts or lifted pins.
I can get 10x new old stock 74LS245N here in the UK for about £5 ..so going to place an order.
You can never have too much retro stuff in stock.
So even if this doesn't help the situation I can discount them.
There are 3 on the mobo ...so will swap all 3 out and take it from there.

...and yes I tried all the ISA slots with the vga card.
rgds
VS

Decided to see how easy or difficult removing one of the 74LS245 chips was going to be..
particularly inaccessible between the ISA slots so you only have the underside of the mobo to work on.
Tried wick ..ineffective..
tried solder suckers ..both electrical and mechanical..
ineffective..
The main problems of course are internal ground planes which wick away the heat.
So what to do..?
Chipquik is the answer ...
yes it is not cheap ...but works by mixing itself with the leaded solder ..and reduces the melting point in the process..remaining 'fluid' and retaining heat for longer.
This allows you time to get a long screwdriver blade under one edge of the chip and apply upward pressure whilst retaining the solders fluidity.
I managed to remove one of the chips without damaging the PCB at all.
Cant say the same for the chip ...which ended up with bent pins ..but intact.
Then used wick to clean out the holes..you can see a light shining through all 20 through holes ...and how clean it came out without pcb damage in the photos.
I shall be fitting turnpin sockets before fitting the new chips.
Anyhoo ..
a couple of photos to show the result
rgds
VS

2 down 1 to go..
I think it can wait 'till tomorrow when my eyes are less tired.

Push through the blah and blur, screw up, ride the panic adrenaline rush until 3AM trying to fix it, pass out on the workbench, that's how computer projects are done! ... just kidding, do not do that. Source, I did that before. 🤣

Vipersan wrote on 2021-09-09, 16:46:

Sadly came without the leads or indeed a manual...

I have the same POST reader, it indeed has a piezo speaker, simply connect its JP2 to the motherboard speaker connector pins 1,4 using a general jumper wire.

Today the 74LS245N chips arrived ...
fitted 3 new ones in the sockets I prepared and ...no change..
So I guess I must conclude the original bus line transceivers were ok..
A lot of work for nothing I guess ..but at least these are eliminated.
I just hope the 85C471 is ok.
I guess it's time to break out the logic probe...if I can find it.
It must be years since I last saw it.
Still ..if I can't ..they are cheap enough.
rgds
VS

Still waiting for a replacement logic probe to arrive...
In the meantime I hooked up a piezo speaker to check for beep/error codes.
No actual code as such ...just a string of single long beeps that continue until the mobo is powered down.
I'm no expert here ..but from what I have read on the subject ..it could be that there is a memory issue ?
I have tried quite a few combinations and types of simm with the same result.
I guess there could be a broken/ corroded trace under the simm socket(s)
??

Vipersan wrote on 2021-09-19, 18:46:

just a string of single long beeps that continue until the mobo is powered down.

I've seen some motherboards from the respective era which reacted that way to single-sided EDO RAM modules installed. Installing a FPM module did help.
Do you get the same beep pattern without any RAM installed at all? What about the POST code reached?

Sadly still waiting on my logic probe..so no further progress with this motherboard.
Just thought I'd post a photo of it in my latest creation.
Been busy converting an PC case internals into a test bench of sorts.
As you can see the POST card still stops at F7..but I am now using an ATX power supply with a converter cable for AT.
I guess power supply problems can be crossed off the list.
I also have a proper speaker in the test rig ...
The mobo powers up to 4 beeps ..one long and 3 short ...then silent thereafter.
The only time I get continuous beeps is if I dont use 4 x simms
Then the post card issues C1 ..
I currently have the Trident ISA VGA card installed but I also tried a VLB card just for curiosity.
The post card stops at 05 if I then boot with the VLB video card.

I do like my test rig creation ...
So many mount points for different mobos ...for 486 AT ...through ATX...
amd simply removing the adapter allows the psu to power them all
Also mounted a floppy drive ..not that I can use it on this motherboard.
I just wish I could fix this board.
I haven't given up yet but it's not looking great.

Have you tried an 8 bit MDA card ?

or tape off the 16bit part of the ISA connector on the trident card.
or just shove the trident card in the 8bit slot :-p

Check out Adrian Black on youtube, his channel is focused on doing board repair on c64/apple ii/mac/pcs/etc. Just watching a lot of his videos has helped me get a feel for diagnosing and repair, put your mind at ease on whether you're on the right track and avoiding some missteps.

Eep386 wrote on 2021-09-09, 16:27:

Looking at the POST codes in binary, 63h is 1100011, and F7h is 1110111. In both codes, bit 3 (counting from 0) set logic low (0). Also, bit 3 is logic low in that weird post code B6h, and it's naturally zero in C1h too... not sure if that means anything, but if it does it might indicate that data bit 3 isn't getting through to the ISA bus.

Such a weird problem but I think that's exactly right. At 0x7f2 the BIOS outputs code BE, and at 0x800 it outputs C1. So BE would become B6. With that situation it's hopeless to keep swapping VGA cards in hopes of it working...

Been and borrowed an oscilloscope (damned frustrating my logic probe is taking so long to arrive)...
got a printout of the 8 bit port of the ISA socket and checked side A pins 1 through 9..
Sure enough there appears to be nothing substantial on pin 6 (data pin 3) just noise spikes...all others o,1,2,4,5,6,7 show activity...
So it does look like data bit 3 isn't making it to the ISA slots.
Next question is why ?
would the ISA be driven directly or buffered ?
If only I had a schematic.

Ever been had ????
I certainly have.
The mobo posts just fine now....
I checked with my meter in continuity mode....each data pin to ground ....and data 3 was grounded.
I was just about to pull the mobo out of the test rig ...when a thought ocurred...what if the brand new post card was itself grounding pin 3 ?
So I pulled it and checked again with the meter.
The short to ground was gone.
Popped in the trident and tried booting....IT POSTED !!!
Moral of the story ..dont put your faith in shyte from China...even if it is brand new.
Grrrr

Dang! So the POST card was just faulty? What a weird failure mode.

At least the board's working now!

Yes indeed Eep ...Looks like the repairs I had done to fix the corroded traces were just fine...along with the mod I did to fit a coin cell battery...
If I had not tried a post card at all ...I would have saved myself a lot of wasted time and effort.
You live and learn.
How often do you buy a tool only to find it makes a bad situation worse.
Thanks for your help though.
You diagnosed the problem spot on without being there.
rgds
VS

Looks like I may still have a problem but not as serious though I would like to fix it.
As mentioned already I have fitted a coin new cell battery holder and diode to replace the leaky varta so it is delivering 3.2 v to the motherboard.
Now that I can see the bios I made a few changes last night including time and date.
Finally powered down and went to bed...
This morning however those changes had been forgotten...so the battery voltage is not getting to the CMOS to preserve them.
In the old-er days this would be easily identifyable as a seperate chip or RTC.
So just where are these changes stored and battery backed on motherboards like this one ?
rgds
VS