majinga wrote on 2026-06-13, 18:01:

You see any modification to the board?
If I remember correctly, the only way to use a P3 on this board is with a Powerleap NEO s370, or by some mods.

I can't see any modification to the board (nothing done to the CPU pins on the back of the board), which is strange since it was running the P3 BIOS

Got a little more time to work on this. Small update. I should have noticed this before. 3.3v is missing on the PCI power rail. Now I've got to figure out where the 3.3v should be coming from. It might lead me to what's going on with the board.

myne has kindly spent many hours drawing up the 440bx schematics. Going to look at that even though I can't read schematics 🤣.

The thing no one asked for: KICAD 440bx reference schematic

BP6

BX2000

zuldan wrote on 2026-06-15, 07:18:

H3nrik V! wrote on 2026-06-13, 06:21:

FWIV, that card looks exactly like mine, and mine shows nothing anymore - even on a working board 🤷‍♂️

These cards die after a while. I buy a few at a time 😉

Oh, I better get a (couple of) new then 🤣 thanks for the info!

zuldan wrote on 2026-06-15, 07:25:

3.3v is missing on the PCI power rail. Now I've got to figure out where the 3.3v should be coming from. It might lead me to what's going on with the board

That's not good.
Because back then, that +3.3V was causing all sorts of circuit-related chaos.
In Intel’s reference design, the +3.3V supply is fed directly from the ATX connector to the PSI, AGP, chipset (northbridge and southbridge) and memory.
Manufacturers then began to incorporate ACPI support, and a high-power MOSFET was introduced to cut power to everything except the memory and northbridge.
That is, apart from the fact that half of the manufacturers have the AGP powered by its own regulator, and another group have the RAM powered by a separate one.

Insert a spare RAM stick and measure the voltage across the capacitors near the memory chips.
Is there +3.3V there?
And on the AGP pins as well.
And next to the 1084 regulators there are zero-ohm resistors (marked 000). Check their (zero) resistance.

shevalier wrote on 2026-06-15, 09:45:

Because back then, that +3.3V was causing all sorts of circuit-related chaos. In Intel’s reference design, the +3.3V supply is f […]
Show full quote

Because back then, that +3.3V was causing all sorts of circuit-related chaos.
In Intel’s reference design, the +3.3V supply is fed directly from the ATX connector to the PSI, AGP, chipset (northbridge and southbridge) and memory.
Manufacturers then began to incorporate ACPI support, and a high-power MOSFET was introduced to cut power to everything except the memory and northbridge.
That is, apart from the fact that half of the manufacturers have the AGP powered by its own regulator, and another group have the RAM powered by a separate one.

That's interesting information. Thank you!

shevalier wrote on 2026-06-15, 09:45:

Insert a spare RAM stick and measure the voltage across the capacitors near the memory chips. Is there +3.3V there? And on the A […]
Show full quote

Insert a spare RAM stick and measure the voltage across the capacitors near the memory chips.
Is there +3.3V there?
And on the AGP pins as well.
And next to the 1084 regulators there are zero-ohm resistors (marked 000). Check their (zero) resistance.

There is +3.3v on all the memory SMD capacitors and 3.3v on the AGP SMD capacitors. Zero-ohm resistor is 0.2 ohms so that looks fine. There must be something else broken on the board. I've ordered new AGP and PCI chipsets just in case I need to swap them out later on.

If you look at my voltages image further up, one of the IRL3103S chips has a leg (source) is reporting 0v.

zuldan wrote on 2026-06-16, 06:47:

If you look at my voltages image further up, one of the IRL3103S chips has a leg (source) is reporting 0v.

Which one of it? Because I haven’t noticed anything particularly odd in your measurements.
If you’re referring to the paired MOSFETs in the CPU core power supply, the lower one always has its source connected to ground.

shevalier wrote on 2026-06-16, 10:37:

zuldan wrote on 2026-06-16, 06:47:

If you look at my voltages image further up, one of the IRL3103S chips has a leg (source) is reporting 0v.

Which one of it? Because I haven’t noticed anything particularly odd in your measurements.
If you’re referring to the paired MOSFETs in the CPU core power supply, the lower one always has its source connected to ground.

The top one (in front of the ATX connector). Yes, it looks like the leg is connected to ground but the bottom one has a leg also sitting on ground (I’ll need to check that) and that’s reading 2v. If you think it’s normal I’ll ignore it 😀 my electronics knowledge is very basic 😗

Have you got any suggestions of anything else I should be looking at?

zuldan wrote on 2026-06-16, 10:46:

Have you got any suggestions of anything else I should be looking at?

Re: Abit BP6 - Restoration Journey
Check the motherboard’s response to the RESET button.
Monitor the corresponding LED on the POST board.

zuldan wrote on 2026-06-16, 06:47:

If you look at my voltages image further up, one of the IRL3103S chips has a leg (source) is reporting 0v.

According to your readings it's completely right, the one FET (low side) connects to ground while the other one (high side) connects to the 5V rail:

So that's totally fine.

shevalier wrote on 2026-06-16, 11:05:

Re: Abit BP6 - Restoration Journey Check the motherboard’s response to the RESET button. Monitor the corresponding LED on the PO […]
Show full quote

zuldan wrote on 2026-06-16, 10:46:

Have you got any suggestions of anything else I should be looking at?

Re: Abit BP6 - Restoration Journey
Check the motherboard’s response to the RESET button.
Monitor the corresponding LED on the POST board.

When I push the reset button the reset light shows up on the post card so that's good. If I hold down the power button for a few seconds, the motherboard turns off. Also a good sign.

When I first powered up the motherboard tonight, C3 appeared on the post card but I think that was just a blip in the system. I couldn't get it to that again.

tehsiggi wrote on 2026-06-16, 11:15:

According to your readings it's completely right, the one FET (low side) connects to ground while the other one (high side) conn […]
Show full quote

zuldan wrote on 2026-06-16, 06:47:

If you look at my voltages image further up, one of the IRL3103S chips has a leg (source) is reporting 0v.

According to your readings it's completely right, the one FET (low side) connects to ground while the other one (high side) connects to the 5V rail:

The attachment Screenshot 2026-06-16 at 13.12.20.png is no longer available

So that's totally fine.

Thanks for the confirmation!

My (not particularly beloved) CUBX-L, which had been stored in the loft for about seven years, but it wasn’t in this condition.
Could it be that the motherboard was damaged by fire or water?

shevalier wrote on 2026-06-17, 09:51:

My (not particularly beloved) CUBX-L, which had been stored in the loft for about seven years, but it wasn’t in this condition.
Could it be that the motherboard was damaged by fire or water?

I believe the motherboard was in a fire but not directly in the heat. The board was covered in black soot which probably gave it a bit of protection over the years. None of the plastic is melted and no signs of heat damage. The BIOS chip reads perfectly fine.

zuldan wrote on 2026-06-17, 09:57:

shevalier wrote on 2026-06-17, 09:51:

My (not particularly beloved) CUBX-L, which had been stored in the loft for about seven years, but it wasn’t in this condition.
Could it be that the motherboard was damaged by fire or water?

I believe the motherboard was in a fire but not directly in the heat. The board was covered in black soot which probably gave it a bit of protection over the years. None of the plastic is melted and no signs of heat damage. The BIOS chip reads perfectly fine.

It’s probably a goner – the circuit board tracks have rotted away.

I’d try to find out why there’s no 3.3V on the PSI bus.
1. It might not be there at all, but that was the case on erlier Socket 7 platforms.
2. The vias may have broken somewhere.
Check the connections between all the PSI slots on the +3.3V bus.
And everything else. The southbridge is powered by +3.3V; they may share a common trace route.

shevalier wrote on 2026-06-17, 11:19:

It’s probably a goner – the circuit board tracks have rotted away. […]
Show full quote

It’s probably a goner – the circuit board tracks have rotted away.

I’d try to find out why there’s no 3.3V on the PSI bus.
1. It might not be there at all, but that was the case on erlier Socket 7 platforms.
2. The vias may have broken somewhere.
Check the connections between all the PSI slots on the +3.3V bus.
And everything else. The southbridge is powered by +3.3V; they may share a common trace route.

I hope not, I’ll never be able to afford a working one 🤣. Besides the missing 3.3v, I guess it’s possible that soot landed on the CPU socket pins. I’ll take the socket cover off for one of the sockets and see what the pins look like.

Will trace 3.3v tomorrow.

zuldan wrote on 2026-06-17, 11:31:

shevalier wrote on 2026-06-17, 11:19:

It’s probably a goner – the circuit board tracks have rotted away. […]
Show full quote

It’s probably a goner – the circuit board tracks have rotted away.

I’d try to find out why there’s no 3.3V on the PSI bus.
1. It might not be there at all, but that was the case on erlier Socket 7 platforms.
2. The vias may have broken somewhere.
Check the connections between all the PSI slots on the +3.3V bus.
And everything else. The southbridge is powered by +3.3V; they may share a common trace route.

I hope not, I’ll never be able to afford a working one 🤣. Besides the missing 3.3v, I guess it’s possible that soot landed on the CPU socket pins. I’ll take the socket cover off for one of the sockets and see what the pins look like.

Will trace 3.3v tomorrow.

Honestly, there's a risk. Soot damages, especially in combination with moist can lead to very corrosive stuff, that makes electronics fail "early" as in years sooner than else. I would however suspect that there would be visible damages or residues, which didn't clean out, so there may still be hope 🙂

Your 3.3V issue might be a red-herring:
https://grizlyk1.narod.ru/other/abit2p3.htm

I did some research and it appears, there is no 3.3V power routed to the PCI slots on the BP6.

PCI addon card power 3.3V […]
Show full quote

PCI addon card power 3.3V

Most cheap modern PCI cards (USB 2.0,1G ethernet,SATA1) require 3.3V from motherboard, but BP6 has no the power voltage connected to PCI slots, the PCI pins are NC.

It is interesting, that some of the cards can work even with 486 CPU mobo and oldest PCI versions only by 3.3V power connected.

I was too lazy to sloder all power pins of all PCI slots, so for each used PCI addon card i select appropriate PCI slot, place mark to card with the slot number and for the cards, that require 3.3V, i check used by card 3.3V pins and connect power to PCI slot only to the pins, in fact in my case it was only one wire for each PCI slot, very easy. Of course, i am sure, that one PCI power line is enough for each conrecte card.

tehsiggi wrote on 2026-06-18, 07:18:

Your 3.3V issue might be a red-herring: https://grizlyk1.narod.ru/other/abit2p3.htm […]
Show full quote

Your 3.3V issue might be a red-herring:
https://grizlyk1.narod.ru/other/abit2p3.htm

I did some research and it appears, there is no 3.3V power routed to the PCI slots on the BP6.

PCI addon card power 3.3V […]
Show full quote

PCI addon card power 3.3V

Most cheap modern PCI cards (USB 2.0,1G ethernet,SATA1) require 3.3V from motherboard, but BP6 has no the power voltage connected to PCI slots, the PCI pins are NC.

It is interesting, that some of the cards can work even with 486 CPU mobo and oldest PCI versions only by 3.3V power connected.

I was too lazy to sloder all power pins of all PCI slots, so for each used PCI addon card i select appropriate PCI slot, place mark to card with the slot number and for the cards, that require 3.3V, i check used by card 3.3V pins and connect power to PCI slot only to the pins, in fact in my case it was only one wire for each PCI slot, very easy. Of course, i am sure, that one PCI power line is enough for each conrecte card.

That explains why I'm struggling to trace a 3.3v line to the ATX connector.

I had a close look at the socket pins and they do not look happy. I could replace one of the sockets with a brand new socket 370 (I keep them in stock) but it's A LOT of my time to replace it. Starting to wonder if I should just put this in the "donar" box and move onto the next project like beating your 3dmark score 🤣

If you're planning to move toward the socket, I'd remove the covers, take a look, clean it, and put the covers back on.

I’d insert and remove the CPU about 10 times.
Then I’d insert the Mendocino and test the connection from the top of the CPU package to the socket pins.
Not on the “beeper” setting, but at the 200-ohm range on the multimeter.
From 1 to 2*370

It’s a tedious task.

Small update, I attempted to remove the cover off one of the sockets but I’m worried it’s too brittle and I’m going to snap something so plan B is to give the motherboard a ultrasonic bath. Going to give each side 5 min at 50c. I’ve desoldered all the crystal oscillators on the board and will reinstall them after the ultrasonic bath. If this doesn’t help the board then at least it will be super clean and will be a great display board on the wall.