Hi! Just got a P3B-F in the mail which had the exact same problem. There was a broken trace near the BIOS eeprom in my case (see photo). After fixing that trace the board was working again. The green led near the AGP slot should definitely be off until the board is fully powered on. The big resistor near the ISA slot should definitely not get warm at all, there should always be zero volts across if the board is on or off.

fool wrote on 2022-02-27, 15:54:

Shorted transistor would conduct 5V straight to Vcore or Vcore to ground. There are also other scenarios but those are the two what would possibly generate short.
I would do similar short check and measure resistance between [Vcore output <-> GND] and [Vcore output <-> 5V input].

Similar for caps. Anything in parallel on a particular segment of circuit which has failed short to ground will result in very low impedance, but you won't be able to differentiate which component has failed without further testing. You can test bipolar transistors and diodes in circuit as noted previously, but MOSFETs won't behave the same, so check the part number when testing transistors. Without fancy TDR gear and such, capacitors, resistors, and other passive components must be lifted on one side to be fully diagnosed electrically.

fool wrote on 2022-02-27, 15:54:

You can also switch multimeter to diode mode and go through those legs of both transistors, all three of them. Thats 6+6 measurements. Diode mode measures the threshold voltage what's needed for current to start flowing in the circuit. I would assume to get readings like over 0.3... to something. Zero reading would be short circuit.

As noted, this works great for bipolar junction transistors. Not so much for MOSFETs.

TheNoOne wrote on 2022-02-28, 17:42:

Hi! Just got a P3B-F in the mail which had the exact same problem. There was a broken trace near the BIOS eeprom in my case (see photo). After fixing that trace the board was working again. The green led near the AGP slot should definitely be off until the board is fully powered on. The big resistor near the ISA slot should definitely not get warm at all, there should always be zero volts across if the board is on or off.

Holy shitballs!

I'm...speechless. This went completely above my head when visually inspecting the card, probably by lack of experience. I won't have time to fix that tonight (starting a new job tomorrow 1st of March heh) but this looks very very good. Well not really, but fixable. That's probably the tiniest repair job I'll ever have done, but I should have everything necessary, experience excepted.

So these P3B-F have a common failure at that very spot? Any potential explanation about why and how to prevent it?

I'm still flabbergasted. I mean...HOW TIMELY CAN YOU BE, YOU WITCH 😁

beefy diode and thin trace, wonder what is being powered down the line and thru the via

Were you using a Voodoo 3 on this board by chance?

Not in a long time, many months, probably more than a year. I was running a FX5950 at FSB133 so AGP 89MHz. Could that be the reason? Overclocking putting too much strain?

edit: I've been running at FSB133 for years on a variety of graphic cards, but the FX5950 has been the one I used for the last few months until the issue

Spooky...

Wow! Sorry for the newbie question, now, but what would be the best way to fix this particular problem?

Just solder a small bridge with e.g. a resistor leg over the broken trace as shown here: download/file.php?id=131578&mode=view

Boohyaka wrote on 2022-02-28, 21:27:

Not in a long time, many months, probably more than a year. I was running a FX5950 at FSB133 so AGP 89MHz. Could that be the reason? Overclocking putting too much strain?

edit: I've been running at FSB133 for years on a variety of graphic cards, but the FX5950 has been the one I used for the last few months until the issue

It could well be. P2B-F/P3B-F are known for being boards that don't have great AGP voltage regulation, some of these boards can't really handle even a Voodoo 3.

appiah4 wrote on 2022-03-01, 06:03:

Boohyaka wrote on 2022-02-28, 21:27:

Not in a long time, many months, probably more than a year. I was running a FX5950 at FSB133 so AGP 89MHz. Could that be the reason? Overclocking putting too much strain?

edit: I've been running at FSB133 for years on a variety of graphic cards, but the FX5950 has been the one I used for the last few months until the issue

It could well be. P2B-F/P3B-F are known for being boards that don't have great AGP voltage regulation, some of these boards can't really handle even a Voodoo 3.

That's the first time I hear of that.

doogie wrote on 2022-03-01, 02:02:

Wow! Sorry for the newbie question, now, but what would be the best way to fix this particular problem?

Best would be tracing what is being powered by that diode and running thicker wire between two points directly.

rasz_pl wrote on 2022-03-01, 12:16:

doogie wrote on 2022-03-01, 02:02:

Wow! Sorry for the newbie question, now, but what would be the best way to fix this particular problem?

Best would be tracing what is being powered by that diode and running thicker wire between two points directly.

Yes. Use the diode through hole on the right side of the picture as a solder point. Find a through hole or good sized pad on the other side of the broken trace for the other side of the wire.

gdjacobs wrote on 2022-03-01, 21:20:

Yes. Use the diode through hole on the right side of the picture as a solder point. Find a through hole or good sized pad on the other side of the broken trace for the other side of the wire.

Right, that should work and be easy enough shouldn't it?
Would a strand of Kynar wire be too fine for the potential load? Maybe a resistor leg as suggested before would be better? I may have slightly thicker cable laying around too

further, the trace has same width behind the via. Find what is being powered.

rasz_pl wrote on 2022-03-02, 05:12:

further, the trace has same width behind the via. Find what is being powered.

I have no idea how to tell, it disappears under the ISA slot

I probably wouldn't use a via as they're usually not as robust for soldering on. I think the two pin jumper pad by the corner mounting hole would work. 24 ga wire looks to be about right. If so, you can run a jumper wire along the board edge and tape it down.

Finally had the time to do it and guess who's back, alive and kicking? Good ol' P3B-F!

TheNoOne wrote on 2022-03-01, 05:48:

Just solder a small bridge with e.g. a resistor leg over the broken trace as shown here: download/file.php?id=131578&mode=view

Man, thank you SO much. I'm still not over the fact you casually passed by with a couple of posts on this board and shown a picture of the exact issue my board was having, 1 day after I posted. As someone else mentioned: spooky 😁
I can confirm the resistor is not hot anymore, and the LED doesn't light up as soon as PSU is turned on neither.

gdjacobs wrote on 2022-03-03, 01:11:

I probably wouldn't use a via as they're usually not as robust for soldering on. I think the two pin jumper pad by the corner mounting hole would work. 24 ga wire looks to be about right. If so, you can run a jumper wire along the board edge and tape it down.

Good call, that was easy enough and I did exactly that!

For the record and anybody in the future getting to this thread with the same issue: I have been running this board for months at FSB133, with AGP cards on the bus running at 89MHz. I absolutely can't say for sure (by lack of electronic knowledge) this is the reason of the burnt trace, but seeing two cards with such an identical problem this looks like textbook common mode failure, and running things slightly out of specs is as good an explanation as any.

Thanks to anyone else that chimed in along the way, much appreciated!

Sweeeet! I am thrilled to see success here. I will keep this in mind on my board if/when it dies.

Separately, I'm happy to report that the recap with the Nichicon polys was a success, and the system is stable with my SL4KL 1GHz/100MHz. The caps look futuristic and hilariously out of place, but they won't likely be the cause of failure any decade soon. 😀

Ah, I've just tried to run my P3B-F rev. 1.04 after ~ 1 year of being turned off and I has not started. Similar symptoms - does not boot, without CPU it beeps, but without RAM or Graphics - no beep.
I've cleaned all pads (I had this many times when old computer did not start) - but no change - also tried to search for burned path, but on my motherboard it looks fine.

I'll try to dig further - maybe even put out BIOS chip and reprogram it - but I doubt this is the cause.

Reading through this topic and some scraps of information it seems that this board is prone to fail for no apparent reason. However this topic proves it can be fixed sometimes.

Any suggestions? Maybe I also have a burned path - even the same - but somewhere else?