Honestly sounds like the board is toast. Swapping stuff will not help. Overclocking always has it's risks, specially if you mod a board outside of it's intended specs.
If it has never been re-capped and you overclock you are forcing a much earlier failure than if not OC.
Can you post a good picture of your board.
Am totally against OC anything since about 2005 unless you have an exact spare board and willing to risk totally killing a part (my history with PC's starts in 1992) because have burnt up enough before then to learn 😀

This was posted yesterday and it sounds like a likely explanation for why a Socket 775 P4 motherboard I have suddenly died.
Re: Help with a P4 that turn on just sometimes

"P4 platform (both S478 and LGA775) is notoriously known for losing BGA contact between a PCB and a north bridge or a CPU socket. It's too tight fasteners of the stock CPU heatsink.
Of course it might be not your case. But if your system doesn't start you can do a test.
I suppose the system is taken out of a computer case and built just on a desktop. A CPU cooler is on place but it is unfastened. So try to press carefully onto a CPU or a NB (it's better to put something like a lump of soft fabric under a mainboard into a place of pressing) to temporarily restore losing contact and then press power-on button."

Horun wrote on 2022-09-01, 00:31:

Honestly sounds like the board is toast. Swapping stuff will not help. Overclocking always has it's risks, specially if you mod […]
Show full quote

Honestly sounds like the board is toast. Swapping stuff will not help. Overclocking always has it's risks, specially if you mod a board outside of it's intended specs.
If it has never been re-capped and you overclock you are forcing a much earlier failure than if not OC.
Can you post a good picture of your board.
Am totally against OC anything since about 2005 unless you have an exact spare board and willing to risk totally killing a part (my history with PC's starts in 1992) because have burnt up enough before then to learn 😀

Thanks for the quick reply. I'm not convinced that the board is toast, mostly because I wasn't pushing it hard during these tests. I had the FSB at 250 and voltage at around 1.47V (which is within spec for Prescott).

The board has never been recapped, but it was working perfectly fine before this "bad BIOS checksum" error that popped up during one of my resets. I've never seen a board just suddenly decide to not POST like this.

Repo Man11 wrote on 2022-09-01, 01:08:

This was posted yesterday and it sounds like a likely explanation for why a Socket 775 P4 motherboard I have suddenly died. Re: […]
Show full quote

This was posted yesterday and it sounds like a likely explanation for why a Socket 775 P4 motherboard I have suddenly died.
Re: Help with a P4 that turn on just sometimes

"P4 platform (both S478 and LGA775) is notoriously known for losing BGA contact between a PCB and a north bridge or a CPU socket. It's too tight fasteners of the stock CPU heatsink.
Of course it might be not your case. But if your system doesn't start you can do a test.
I suppose the system is taken out of a computer case and built just on a desktop. A CPU cooler is on place but it is unfastened. So try to press carefully onto a CPU or a NB (it's better to put something like a lump of soft fabric under a mainboard into a place of pressing) to temporarily restore losing contact and then press power-on button."

Repo Man,

So when I had this board on the bench, I had a heavy heatsink (Zalman 9500...the big copper one) sitting on top of the CPU with some thermal paste and no retention bracket. There was no strain on the CPU socket whatsoever.

What you're saying is that it's possible the Northbridge is losing physical connectivity? I can try putting some pressure on it and booting.

Thanks for the link.

My own P4C800 Deluxe that I bought new in 2004 became a victim of the "capacitor-plague" a couple of years later.

A picture of your board would be helpful to decide if this could be the case here!

"First rule out the most common issues, then proceed to look for the hummingbirds."

Just got around to taking a bunch of pictures of the board. More to follow in a second post.

Second batch of photos.

majestyk wrote on 2022-09-01, 08:28:

My own P4C800 Deluxe that I bought new in 2004 became a victim of the "capacitor-plague" a couple of years later.

A picture of your board would be helpful to decide if this could be the case here!

"First rule out the most common issues, then proceed to look for the hummingbirds."

Caps look good before and after the incident with Memtest, at least to my eyes.

Also had a P4P800 Deluxe back in the day. The computer went in auto standby, and then I couldn't wake it up. It never booted again. There were no visible signs of cap plague, overheating, etc.

Did some research, and apparently those P4P800 boards were known to have problems with the CPU VRM MOSFETs. Didn't have the equipment to attempt a component replacement back then, so simply took the opportunity to upgrade, but you might be interested to know.

kaputnik wrote on 2022-09-01, 09:37:

Also had a P4P800 Deluxe back in the day. The computer went in auto standby, and then I couldn't wake it up. It never booted again. There were no visible signs of cap plague, overheating, etc.

Did some research, and apparently those P4P800 boards were known to have problems with the CPU VRM MOSFETs. Didn't have the equipment to attempt a component replacement back then, so simply took the opportunity to upgrade, but you might be interested to know.

Thanks for the reply. I had all the MOSFETs heatsinked at the time (took them off for photo) and temperatures were always good on them. I have some replacement ones upstairs that I can swap out if it's determined they're defective but I have no reason to believe they are. Part number is 06N03LA if I remember correctly.

When you plug back the original BIOS ROM chip and do a Clear CMOS. Does it still starts with errors, or is it the same as with the replacement BIOS (nothing happens)?
Are you sure the red wire you have soldered (the one near the CPU socket does not make contact with the nearby capacitor and pad)?
Double checking the modifications made is the first thing to look at.

Datadrainer wrote on 2022-09-01, 10:14:

When you plug back the original BIOS ROM chip. Does it still starts with errors, or is it the same as with the replacement BIOS (nothing happens)?
Are you sure the red wire you have soldered (the one near the CPU socket does not make contact with the nearby capacitor and pad)?
Double checking the modifications made is the first thing to look at.

Same behavior as replacement BIOS (nothing happens, no POST codes on my diagnosis equipment). I did a close inspection of the modification and it does not appear that any part of it is touching the nearby capacitor and pad.

EDIT: I confirmed with a multimeter just now that all the connections for the vdroop mod are proper and that nothing unintentional is touching

The issue can be anywhere then. I know that does not help, but I see no miracle solution here.
Motherboards have multiple copper layers, not just top and bottom with a lot of bridges between them. So can be a broken track that have melted (I have seen that following a shorted dead cap) or can be a defective component. Meaning finding the culprit is a nightmare. I'm saying that because P4 drains a lot a power. The power traces are sized for a max given amperage, and by overclocking you drain more power that was is expected and if the margins are short, that can be a problem. And that is without considering the stress on the power regulation components, not only the MOSFETs.
But you can first check if the ICs are getting abnormally hot by touching them, that would indicate a bad chip. Then check if there is not shorts between power lines and ground from the ATX connector. Normally the PSU would immediately stop or simply not start but who knows. That does not hurt to do it.
Next you can try to check the continuity on 5V and 3.3V lines on the different ICs (datasheets are required to find the VCC pins).
Your caps are visually good but that does not mean they are, especially considering the area. A bad capacitor cannot be excluded but a full cap testing requires an expensive equipment and require to unsolder them. So as they have to be unsoldered its easier to simply replace them. But you can check the resistance with an ohmmeter.
A good clue, is to start from the start. That is a circuit, component are linked through power and data lines. Some are in parallel, but the main ones are in series. Each being activated by the previous on request. Knowing the architecture helps a lot for debugging. You can find info on search engines and ask question for a specific component here. You have some clues from the people who answered before. Its always a good idea to look for know issues at first too.
But you are good for a lot a checking and measurement. Not to discourage you, but solving a "modern" motherboard issue can be very very long. I'm trying to repair a Commodore PC-40 III (i80286 IBM PC compatible, so not considered "modern") from several month now. That was a time where technical documentation was available so I have the schematics, I have understood a lot from the MB design, I have checked the buses signals, most of the ICs. I know what's wrong, but I still don't know why... Despite having A LOT of ICs, the architecture is quite simplier than a P4 MB.
So I which you good luck and hope the issue is easy to find and repair.

cactusjuice wrote on 2022-09-01, 08:22:

Repo Man, […]
Show full quote

Repo Man,

So when I had this board on the bench, I had a heavy heatsink (Zalman 9500...the big copper one) sitting on top of the CPU with some thermal paste and no retention bracket. There was no strain on the CPU socket whatsoever.

What you're saying is that it's possible the Northbridge is losing physical connectivity? I can try putting some pressure on it and booting.

Thanks for the link.

I apologize as it was not a my post , but may I answer here ?
If there was no strain on the CPU socket it's OK. So we can suppose there is no losing contacts of the socket and the northbridge.
I think a debug (POST) card would be helpful. Using a debug card you can see if BIOS functions are running at all.
Also try to measure voltage that voltage regulators produce for chipset (for the bridges or more correctly the hubs : MCH and ICH).
They are mosfets, to measure voltage on 3rd pin of a mosfet relatively to GND rail :
1 - Q191 under AGP slot (expected +1.5v) (for ICH)
2 - Q303 somehow between ends of AGP and DIMM slots (expected +2.5V) (for MCH)

P.S.

cactusjuice wrote on 2022-08-31, 23:31:

***
I'm suspecting that the seller sent me a bad BIOS chip, but I have no way of confirming that because I don't have equipment that can read/flash PLCC BIOS chips.
***

About BIOS.
If you don't have a bios-programmer to flash proper bios-image you can apply a hotswap method using a similar motherboard or just a motherboard that supports the chip. It's Pm49FL004T (both FWH / LPC compatible). So it's almost any mainboard that has FHW or LPC flash in bios-socket.

I wouldn´t trust the blue OST caps beside the CPU socket. If they´re not bulged doesn´t mean they are o.k.

https://www.badcaps.net/forum/showthread.php?t=481

Thanks all for the helpful replies and taking the time to guide me in the right direction. Looks as if I have a lot of testing to do with the multimeter!

I have some P4C800s and P4C800-E's around (this is the only E-Deluxe) so I'll see if any of those BIOS chips make the board show signs of life

Roman555 wrote on 2022-09-01, 12:07:

I apologize as it was not a my post , but may I answer here ? If there was no strain on the CPU socket it's OK. So we can suppos […]
Show full quote

I apologize as it was not a my post , but may I answer here ?
If there was no strain on the CPU socket it's OK. So we can suppose there is no losing contacts of the socket and the northbridge.
I think a debug (POST) card would be helpful. Using a debug card you can see if BIOS functions are running at all.
Also try to measure voltage that voltage regulators produce for chipset (for the bridges or more correctly the hubs : MCH and ICH).
They are mosfets, to measure voltage on 3rd pin of a mosfet relatively to GND rail :
1 - Q191 under AGP slot (expected +1.5v) (for ICH)
2 - Q303 somehow between ends of AGP and DIMM slots (expected +2.5V) (for MCH)

Roman,

I did measurements with my multimeter and the AGP voltage, memory voltage, and CPU voltage are all good. It appears as if all components are receiving proper power, but my bios debug card still shows four zeroes.

I poked and prodded all of the other MOSFETs on the board. Several of them seemed to correspond to 3.3v and 12v, but others had no voltage at all on one leg and some had arbitrary values such as 9.7v. Do you or anyone else know of a reference showing expected voltages on each MOSFET?

The next thing I'll be doing it taking a look at the condition of the caps (I'm starting to practice removing caps and checking them).

It's definitely not the BIOS chip, as I've tried two replacements at this point.

Hey mate,
Id recommend you to use Board view programm . There is alot board view files for p4c800 any revision . ill try to attach those files here. This programm will show every component and where it goes and to what connected

ps rev 2.0 is much weaker coz has less components left so it migh be less stable in overclocking . Pre 2.0 revision has more caps at cpu power line.
pss programm itself you can easy find in internet
Psss i think that your bios simply corrupted

cactusjuice wrote on 2022-09-05, 16:46:

Roman, […]
Show full quote

Roman555 wrote on 2022-09-01, 12:07:

I apologize as it was not a my post , but may I answer here ? If there was no strain on the CPU socket it's OK. So we can suppos […]
Show full quote

I apologize as it was not a my post , but may I answer here ?
If there was no strain on the CPU socket it's OK. So we can suppose there is no losing contacts of the socket and the northbridge.
I think a debug (POST) card would be helpful. Using a debug card you can see if BIOS functions are running at all.
Also try to measure voltage that voltage regulators produce for chipset (for the bridges or more correctly the hubs : MCH and ICH).
They are mosfets, to measure voltage on 3rd pin of a mosfet relatively to GND rail :
1 - Q191 under AGP slot (expected +1.5v) (for ICH)
2 - Q303 somehow between ends of AGP and DIMM slots (expected +2.5V) (for MCH)

Roman,

I did measurements with my multimeter and the AGP voltage, memory voltage, and CPU voltage are all good. It appears as if all components are receiving proper power, but my bios debug card still shows four zeroes.

I poked and prodded all of the other MOSFETs on the board. Several of them seemed to correspond to 3.3v and 12v, but others had no voltage at all on one leg and some had arbitrary values such as 9.7v. Do you or anyone else know of a reference showing expected voltages on each MOSFET?

The next thing I'll be doing it taking a look at the condition of the caps (I'm starting to practice removing caps and checking them).

It's definitely not the BIOS chip, as I've tried two replacements at this point.

Hi,
Expected voltages on each MOSFET (though still not on each) are shown on a restricted manufacturer's documents called "repair guide". Sometimes they can be found in Internet. It looks like this (you can download it as an example if you'd like).

cactusjuice wrote on 2022-08-31, 23:31:

I have been playing around with overclocking on a P4c800-E Deluxe and was doing a run with Memtest86 to see if the memory was throwing errors. During one of the memtests, a lot of errors got thrown so I reset the machine (it's on a test bench so I jumped the reset pins) and then I'm brought to a screen that says the "BIOS checksum is bad"

So memory errors could be because of malfunction of MCH.
If you've measured right and mosfet Q303 delivers power +2.5v into MCH (North Bridge i875) that the problem may be exactly the faulty MCH.
Or also bad soldering joints of the chip. Anyway, if I were you I would press the chip by a thumb to exclude this type of defect.

P.S. Also the post above provides helpful schematics. You can download datasheets to better understand purpose of components on the boardviews.

Good luck!

@ciornyi
Thank you for your reply. I didn't know BOM files of motherboard existed... I just downloaded OpenBoardview (great tool) and opened one a the BOM file you have linked.
Everything is here, the link between components, the voltages, the components references. That's really impressive.
Please, can you tell us if there is a place where to get good quality BOM files of MB and add-on cards freely and legally?