Questions

Why would the BIOS return a memory error on a random DIMM slot?

+
0 Votes
Locked

Why would the BIOS return a memory error on a random DIMM slot?

jtjenkins213
On a Dell Precision desktop T3500 with Windows XP Professional 64-bit installed, the operating system was consistently receiving BSOD screens (actual BSOD, as in a hard stop and not a general page fault). When I ran the Dell hardware tests, the memory came back as bad. It currently has 4-1GB DDR3 1333 DIMMs installed.
I researched the type of memory that the system would support, and purchased 3-4GB DDR3 1333 ECC memory as an upgrade. When I installed the memory, I received a beep code of 1-3-2. According to the Dell website, that is a memory error. Thinking I may have for some reason ordered the wrong memory, I put the old memory back in. Once again, I heard the beep code, same error.
I contacted Dell Tech Support, and they had me run through the motions of troubleshooting (take all memory out but one, try each slot, insert one at a time). Eventually they determined that it was not the memory but the systemboard. So according to our service agreement, I had them send a tech out to replace it on site.
Next day, the tech came, replaced the system board. Tried the new memory, received the same error, 1-3-2 beep code. He called tech support, went through the same thing I did. Mysteriously, however, when he put a single 1GB DIMM in, he received the beep code. He then hit the power button accidentally twice (double-tap) which shut power off and turn back on-It then worked. It went through POST and gave the message of system memory size has changed. So he added a second 1 GB DIMM, and same thing, beep code, power off, power on, and it boots but it gives an error stating an error on DIMM2. So he takes that DIMM out and replaces it with the third 1GB DIMM. Same issue, error on DIMM2. He then ignored that, added a third, same thing, beep code, double-tap, it works but returns an error stating that there is a memory problem on DIMM3. So he powers off, powers on and the memory error disappears. He then powers off, puts in the fourth 1GB stick, and same thing, beep code, power off, power on, error on DIMM3, power off, power on, error disappears. It now works as it did before, with the random BSOD.
It's a brand new system board, same CPU, the old memory (obviously, since the new memory always returns the beep code), same video card, same power supply. Nothing else has changed other than memory, and all tests were done with minimal hardware attached (monitor, kb/mouse, no HDD)
Anyone have any clue what might cause this? In 4 years of professional IT support, and 15 years as a hobby, I have never seen anything like this. Any help or advice would be most appreciated.

(edited to clarify subject)
  • +
    0 Votes

    Just a guess, maybe something on the case has warped, bent, etc. and is shorting
    across the DIMM controller or something.

    +
    0 Votes
    jtjenkins213

    There is nothing obvious that would cause this, and considering that the system board was replaced, a bad capacitor or transistor is not very likely either... and none appear popped on the new board...

    +
    0 Votes
    robo_dev

    if you've basically swapped memory and system board, then something else is wrong.

    Did the Tech swap the processor?

    The other variable is the power supply.

    Maybe a wild guess, but I know that Dell uses what are called 'active PFC' power supplies on some models. These PSUs do not play well with a UPS, for example...you need a 'real' true-sine-wave UPS to run one of these units. While some new technology is really great, these new power supplies can die a horrible death if they are fed a pseudo-sinusoidal input power signal (the stepped approximation of a sine wave provided by many normal UPS units).

    +
    0 Votes
    jtjenkins213

    There is no UPS on this system, it's used as a client machine with a basic surge protector on it. I did try plugging into a different surge protector to see if it was a power issue, but that made no difference.
    The tech did not swap the processor, and that was something I had mentioned to Dell. Dell tech support dismissed that as a possibility out of hand, which is documented in the chat I had with them (online chat) so chances of me convincing them to send a replacement CPU when it was already dismissed as a source of the issue is slim at best.

    +
    0 Votes
    Merlin the Wiz

    This problem sounds like a mis-match between the BIOS and the embedded pre-CPU controller firmware. I check for and update the BIOS. After reading all of the notes available on the changes made to the BIOS. With todays motherboard designs, I doubt if the CPU has not initialized when you get this error. Particularly since pressing the power on/off button causes the error to step from one memory slot to another. The motherboard manufacturer (not Dell) may have a flash firmware fix for this problem, but finding the manufacturer may cost more in time than the whole computer is worth. I ahva seen similar problems due to defective surface mount capacitors. Good Luck!

    +
    0 Votes
    jtjenkins213

    This is something I had not thought about. I do want to attempt to install fresh memory sent from Dell first before I try something drastic as flashing the BIOS. I've had bad experiences with Dells when it comes to upgrading the BIOS, so I think of that as a last ditch effort only before just buying a new system.
    As I mentioned before, there is no obvious burned capacitors anywhere. As this has been mentioned twice, I will take a closer look.

    Thank you all for your replies.

    +
    0 Votes
    IcebergTitanic

    I don't know about that particular system but some newer systems are using the Processor as the memory controller. On servers at least, when you have issues like that you swap out the processor and see if the problems go away.

  • +
    0 Votes

    Just a guess, maybe something on the case has warped, bent, etc. and is shorting
    across the DIMM controller or something.

    +
    0 Votes
    jtjenkins213

    There is nothing obvious that would cause this, and considering that the system board was replaced, a bad capacitor or transistor is not very likely either... and none appear popped on the new board...

    +
    0 Votes
    robo_dev

    if you've basically swapped memory and system board, then something else is wrong.

    Did the Tech swap the processor?

    The other variable is the power supply.

    Maybe a wild guess, but I know that Dell uses what are called 'active PFC' power supplies on some models. These PSUs do not play well with a UPS, for example...you need a 'real' true-sine-wave UPS to run one of these units. While some new technology is really great, these new power supplies can die a horrible death if they are fed a pseudo-sinusoidal input power signal (the stepped approximation of a sine wave provided by many normal UPS units).

    +
    0 Votes
    jtjenkins213

    There is no UPS on this system, it's used as a client machine with a basic surge protector on it. I did try plugging into a different surge protector to see if it was a power issue, but that made no difference.
    The tech did not swap the processor, and that was something I had mentioned to Dell. Dell tech support dismissed that as a possibility out of hand, which is documented in the chat I had with them (online chat) so chances of me convincing them to send a replacement CPU when it was already dismissed as a source of the issue is slim at best.

    +
    0 Votes
    Merlin the Wiz

    This problem sounds like a mis-match between the BIOS and the embedded pre-CPU controller firmware. I check for and update the BIOS. After reading all of the notes available on the changes made to the BIOS. With todays motherboard designs, I doubt if the CPU has not initialized when you get this error. Particularly since pressing the power on/off button causes the error to step from one memory slot to another. The motherboard manufacturer (not Dell) may have a flash firmware fix for this problem, but finding the manufacturer may cost more in time than the whole computer is worth. I ahva seen similar problems due to defective surface mount capacitors. Good Luck!

    +
    0 Votes
    jtjenkins213

    This is something I had not thought about. I do want to attempt to install fresh memory sent from Dell first before I try something drastic as flashing the BIOS. I've had bad experiences with Dells when it comes to upgrading the BIOS, so I think of that as a last ditch effort only before just buying a new system.
    As I mentioned before, there is no obvious burned capacitors anywhere. As this has been mentioned twice, I will take a closer look.

    Thank you all for your replies.

    +
    0 Votes
    IcebergTitanic

    I don't know about that particular system but some newer systems are using the Processor as the memory controller. On servers at least, when you have issues like that you swap out the processor and see if the problems go away.