Disk problems….in theory….

[zequikups]

Hello

First of all, I apologize if I'm posting this thread in the wrong place, since I'm not sure where the problem comes from. In any case, I'm clear about where it manifests.

I have a somewhat old computer, a quad core Q6600 at 2.4 GHz on an Asus P5K-VM, with 8 GB of RAM and normally overclocked to 2.8 GHz (400x7) with 8 gb of RAM at 800 MHz stable for years.

A few months ago, suddenly and during certain different periods of time, the PC slowed down until it froze. When you restarted, it gave a disk error. When you restarted again, everything started up ok (without having done anything) and worked OK until it slowed down again.

I tried replacing the hard drive with a new Samsung 840 SSD of 500 Gb, but the problem didn't disappear. I ended up buying a new computer, since it was already time anyway...

Now I've decided to repair the one in question, and I started with a new hard drive (a new Seagate 2 TB at 7200 rpm from the store). The problem is exactly the same.

From here, I'm not able to deduce where the problem comes from. I have suspicions:

• Power supply: that it doesn't supply the constant voltage or amperage and generates errors in the disk.
• Motherboard: that the problem comes from there
• Wiring of the hard drive power or the SATA cable?

Anyway, I would need some pro from the forum to help me and help me locate where the problem may come from.

Thanks in advance

[Sylver]

Hello,

It smells like a motherboard issue, but let's not despair and let's take it step by step.

First, change the power supply for another one you have lying around (that delivers decent power) to test, and also take the opportunity to reset the BIOS and thus remove the OC, to lower consumption and completely stabilize the system.

The wiring seems to me more difficult to be the culprit, unless it is visibly deteriorated, it is very rare that it affects, but if you have another data cable around, try it. I would rule out the power cable if it does not have physical damage in plain sight, the current would arrive anyway, although you can always make sure of it using a multimeter, and thus help to dispel the suspicions of the previous point about the power supply (while you're at it, check all the lines of the same).

If after this the problem persists, another thing I would try would be to remove RAM modules and test them separately, in case any of them is damaged, which should not have a relation with the problem if the error you get is from the disk, but just to be safe, stranger things have been seen. Equally, a memtest would also be good to run, with all or part of the RAM installed, or both.

Keep trying this for the moment and tell us the results, and if there is no news we will continue investigating.

Best regards

[whoololon]

+1
To add something, before you start, check that the CPU and GPU temperatures remain within reasonable limits at rest and under load and that the software environment is stable and properly configured.

It is very important that you run memtest with the modules separately and let it run for the necessary time, it is not worth cheating by leaving it on for an hour and then closing it.

Remember that memtest is left running until it finds an error. The moment it fails, you stop the program, turn off, unplug, discharge static, and replace the module. You turn it back on and run memtest again, and so on with all the modules.

[zequikups]

Good evening

First of all, thank you for your quick response.

I have tried to remove the OC but it keeps happening the same thing, perhaps less often, but it ends up falling the same way.

I have visually checked the power and signal cables of the HD and they seem to be OK.

I have not been able to test the power supply because I do not have one at hand.

Last night I put a Memtest 86 v4.2 for the 4 RAM modules at the same time. We are talking about 4 x 2 gb GSkill DDR2 OCables up to 1000 MHz. In my case, I have them OCed at 400 MHz and voltage 2.1 V, so I have the FSB:RAM ratio at 1:1.
The issue is that after leaving the memtest running for 24 hours, it has thrown several errors (attached link with screenshot), so I understand that I have at least one of the 4 RAM modules damaged.
I do not have experience with Memtest, so I would appreciate your interpretation. The errors seem very close to each other, so I think I can deduce that it could be failing a single module.

http://imageshack.com/a/img540/8292/SHUm3e.jpg

Thanks in advance

zequikups

[whoololon]

@whoololon:

[…] It is very important that you run memtest with the modules separately and let it run for the necessary time […]

Remember that memtest should be left running until it finds an error. The moment it fails, you stop the program, turn off, unplug, discharge static, and replace the module. You turn it back on and run memtest again, and do this with all the modules.

The reason for doing it with the modules separately is because memtest does not distinguish between modules individually, it examines the total installed memory. If it throws an error, you don't know which module is affected.

But anyway, we can already be sure that it's not a disk problem, but a RAM problem.

[zequikups]

Hello

I have tested the 4 modules separately (in the same slot of the board) and it doesn't give me any error. I don't understand anything!

I am thinking that it is not one of the RAM slots on the board that is broken. Should I try testing the modules separately but in different slots on the board?

Thank you

[whoololon]

Man for power can be…

• That one of the connectors is actually faulty (it is quite unlikely, but dirt may have accumulated. Try one of the modules that you are sure works well and test it in each of the banks)
• That one of the modules has been moved by a blow (reconnect them and run the memtest)

...and something more that escapes me now.

If the computer has been overclocked (OC) as you have explained, double-check that the parameters in the UEFI are correct for the operation of the CPU and RAM.
To be completely sure, download and run the Memtest86+ instead of Memtest86.

[Sylver]

And before all that, take a thick brush (but with not very hard bristles) and clean the RAM slots on the motherboard, you will be surprised by the effectiveness of that maneuver.

Best regards

[zequikups]

Hello

I'll go step by step. I'm using Memtest86+ v4.2. I've gone over the slots with an old toothbrush to clean it, although it wasn't dirty.
I've already tested the 4 DIMMs separately and it didn't give any errors. I put the 4 DIMMs back in and the errors come back, apparently in 2 DIMMs I think.
Now I've tested in pairs, that is, first testing 2 DIMMs and then the other pair, using a pair of slots or the other and it hasn't given any errors.
It only gives the error when all 4 DIMMs are connected.
Should I check that the power supply is guaranteeing the correct voltage and amperage? Theoretical values?
Or is it better for me to change the motherboard and stop pulling my hair out?

Anyway, thanks again...

Regards

[Sylver]

I would try another source, that someone lent you or whatever, but it seems to me a necessary test.

On the other hand, a toothbrush is not the best for cleaning the RAM slots, that's why I suggested a brush, because of the reach and the texture, it does the job better. In any case, reconfigure the memories from BIOS and remove the OC you have set, try them at the base frequency and voltage. By the way, have you tried to work with only two RAM modules and the disk? It could narrow down the search if it works without anomalies.

Returning to other possible causes, it is also possible that the disk controller on the motherboard is dying (a teacher diagnosed this problem in my old K8V-X, in the long run he was right). However, I leave you here some guidelines that are recommended on the ASUS website for problems of this type, maybe one of them is the key: ASUSTeK Computer Inc. -Support-

Regards