Memtest Deluxe Download

[Brigitta Martini]

Forthe first few weeks after build, there were no issues. Since then I've been having numerous BSOD with many different error codes being listed. On the mobo forum site, there are others with similar board and memory that appear to be having same issues.

The memory is rated for 5-5-5-15 and 2.1 V. At these values, I've had various results when running memtest under different settings in the bios. Only when I turn down the speed to 7-7-7-24 and 1.8 V, will memtest run without error. I stopped it after 20 cycles. memtest has failed from immediately to after 10 cycles at different bios configurations attempted. If I turn off computer for 5 minutes or longer, memory test will run longer before failure.

Tried bios set up last night suggested by another user with same problems. He said these worked for him, though I still got BSOD in less than 12 hrs. I still have yet to revert back to original bios version as he also suggested.

Under all settings/scenarios tested so far, eventually I will get BSOD, usually within 12 hrs. Running memtest immediately after failure will result in many errors. If I power down system for 5 minutes or more, memtest will run for many passes without error. Leads me to believe that the root of the problem is a charge dissipation on the memory. Either the memory is bad or the memory controller on the board is.

Exited utility (did not power down for reboot), started memtest, instantly failed thousands of errors. 2600 before I turned off. Power down for 1 minute and restarted. Got message that overclocking failed. Back into set up utility. Made no changes. Saved and exit. Launched into memtest without issue, turned off after 1 pass.

Upon boot into windows, PC probe application started (turned on to monitor temps and voltages. Got all the same warning windows of win32_ failures and MB temp reporting at 47C, CPU at 39C. Open door to cabinet. MB temp dropped to 44C within 5 minutes.

Reboots, launches into memtest. Got failure in 1st pass, test 3. In test 4, most errors have similar address around 159 mb, all reporting same err-bits at 00010000. total of 1 err in test 3, 17 in test 4, 1 error so far in test 7. Turned off.

Have tested them individually a while ago. Similar results when running separate or as pair. Have noticed that after swapping out one stick for another, the eventual BSOD failure may not happen for a day or two instead of within 12 hrs.

Video issue still present. Manually power cycle. Notice that each time when I press the power button, power comes on for a second or two, then shuts off and turns on again automatically. Got the overclocking failure again. Ugh.

Started testing system using prime95 stress testing. Don't think the memtest is telling me anything anymore. Test fails no matter what sort of configuration. Stress test fails anywhere from immediately to runs for almost 24 hrs. No consistency here either.

Also partitioned the new HD to leave room for Linux OS. Likely going to install Fedora or Ubuntu to test out issue being tied to windows OS. Not expecting it to play out, but I've been looking for a good reason to start migrating for a while now.

Been an avid reader of the forums for a few months now but decided to make an account just now because I have an issue with my P4C800-E Deluxe. I'm hoping that the experts and enthusiasts here can guide me in the right direction because I'm at a total loss for what to do next.

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". This is something I've never seen before. It gave me the opportunity to load a BIOS file from floppy/CD, but I instead reset the machine again. After that, the machine won't POST at all anymore. My heart sank.

I've been using an UltraX PHDPCI2 PCI card to diagnose issues on my various old hardware. During a normal boot, it cycles through various POST codes...but now it says at all zeroes since the Memtest incident. It's acting as if there's no BIOS at all.

My first thought was that the BIOS got corrupted somehow during the resets I was doing when Memtest showed memory errors. So I ordered a replacement BIOS chip on eBay from a reseller called "biosdepot". I dealt with them successfully in the past to resurrect a bad BIOS on my ASUS P4B and that worked wonderfully, so it seemed appropriate to go to them again since it seems to be a BIOS issue. The chip arrived today and upon installing it, I'm experiencing the same issue. It's not POSTing and the PHDPCI2 card is still showing all zeroes. The board powers up completely fine but refuses to do anything whatsoever.

"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."

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.

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.

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.

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.

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.

3a8082e126