Question RAM causing BSODs then suddenly started working - confused.

Jun 25, 2021
5
1
15
0
Hi All,

New to this forum and PC building in general. 12 months ago I built a PC which I have used as a workstation (architectural design). This week I began getting BSODs and automatic restarts. Eventually they were so frequent I was unable to even log in. On occasions after powering on nothing happened - no screen, no BSODs, nothing.

After searching some of the codes it appeared to be a memory problem. I don't have another PC or other components on hand to swap stuff out (I live in fairly remote spot in Costa Rica - to get PC parts is a 2 day planned trip).

On a hunch (I don't why I did this) I removed 1 of my 2 ram sticks and everything started working again. This was a relief as I was working towards a deadline and needed to crack on. Now the deadline has been met I am trying to figure out what has happened and what I need to look at replacing or fixing. So the PC was running fine on one stick in one slot. To see if it was the ram stick or the slot that was broken I swapped the 2 sticks over (still only 1 in) and the PC works fine - I thought ah ok it must be the slot and probably would require a repair or a new MB. Out of curiosity I put 2 sticks back in to try and generate the BSOD codes again. Now the PC has the same 2 sticks of ram in the same 2 slots and it is working fine. This is the same configuration it had when it repeatedly crashed all day earlier in the week.

This is great that it is working again however I am nervous that it will only be a matter of time before I get BSODs again. I am wondering if anyone has any suggestions as to what may have caused this, what I can do to pin point the problem and how I can prevent this happening again?

My components are as follows (everything is at stock settings I have never overclocked anything):

intel i5-10400 (stock cooler)
2x T-Force Delta RGB 8gb DDR4 2666
ASUS Prime X490P
ASRock Challenger AMD Radeon RX 5600 XT
Corsair CV650
2xADATA Swordfish M.2 2280PCIe Gen3x4 SSD 500GB
NZXT H510 Case

All drivers and BIOS are up to date

Windows 10 Pro

27inch Samsung Monitor

Thank you for taking the time to read and any comments are much appreciated.

Tom
 
Last edited:

keith12

Illustrious
Hey there,

It defo points to ram issues, and possibly a PSU issue.

Test the ram with memtest86+ outside of windows. You will need a usb stick, and set up as a bootable device. Instructions on memtest website how to do this. Set it to boot from your bios, and run memtest and test your ram for 4 passes. Test each stick individually. If there is a problem with the ram, memtest will find it. Then you can narrow it down to which DIMM is faulty (that's what I suspect is happening.) If there are no erros, then it's likely a different issue.

On the PSU side, the CV isn't a great choice for a modern system. It can cause random restarts, and more serious problems down the line if it's on it's way out. I'd be looking at getting a replacement for that one either way.

Start with the ram first, and let us know how you get on testing it.
 
Reactions: TomG1
Jun 25, 2021
5
1
15
0
Hi Keith,

Thank you for the time and reply. I have run the memtest86 this afternoon on your recommendation and in my eyes it is a pretty conclusive result that it is indeed one particular stick of ram. I will summarize the results as I am no expert and may be jumping to the wrong conclusions.

(all with 4 passes)

ran the test with both sticks (stick a and stick b) in - failed on test 6

ran the test with only stick a in - passed no errors

ran the test with only stick a in a different slot - passed no errors

ran the test with only stick b in - failed on test 6

ran the test with only stick b in a different slot - failed on test 6

To me this indicates stick b is the problem/broken component. As of writing I am using the PC with no problems with only stick a in. It seems an obvious conclusion and I will plan to pick up a new set of ram - do you concur?

Do you know why this might of happened? just a bad component? ....sadly it is 2 weeks out of warranty, typical.

I was not aware of the CV being a potential problem but as I am making the trip I will upgrade this also as I plan to upgrade the CPU, cooler and MB in the not too distant future.

Thanks again for taking the time to reply, much appreciated.
 

Jason H.

Distinguished
Hi Keith,

Thank you for the time and reply. I have run the memtest86 this afternoon on your recommendation and in my eyes it is a pretty conclusive result that it is indeed one particular stick of ram. I will summarize the results as I am no expert and may be jumping to the wrong conclusions.

(all with 4 passes)

ran the test with both sticks (stick a and stick b) in - failed on test 6

ran the test with only stick a in - passed no errors

ran the test with only stick a in a different slot - passed no errors

ran the test with only stick b in - failed on test 6

ran the test with only stick b in a different slot - failed on test 6

To me this indicates stick b is the problem/broken component. As of writing I am using the PC with no problems with only stick a in. It seems an obvious conclusion and I will plan to pick up a new set of ram - do you concur?

Do you know why this might of happened? just a bad component? ....sadly it is 2 weeks out of warranty, typical.

I was not aware of the CV being a potential problem but as I am making the trip I will upgrade this also as I plan to upgrade the CPU, cooler and MB in the not too distant future.

Thanks again for taking the time to reply, much appreciated.
Yea just need new ram.

As to why it happened? Not too sure, but hardware sometimes just fails and it could be as simple as that and Im sure thats probably what it was considering you didnt do anything to the ram in the bios, etc.

So yea, just buy a new whole kit of ram, and you should be good. You possibly, maybe, might be able to sell your 1 working chip to recoup a small amount of the cost. Or just keep it around as a backup in case anything happens with your new kit in the future :)
 
Reactions: keith12

keith12

Illustrious
Hey Tom,

Sorry for the late reply. Yes, as the other posters have mentioned, getting new ram is the way to go. I'd suggest rather than getting a similar single DIMM, you buy a matched kit. This will ensure they work together rather than two mixed DIMMS (even tough might be very close in all specs), which could cause compatibility issues. You can then sell the single working DIMM you have to offset the cost.

Do you know why this might of happened? just a bad component? ....sadly it is 2 weeks out of warranty, typical.
More often than not, it's just down to a bad minor component on the DIMM. It does happen, but not really that often.

Yes, the CV although good enough for a basic enough system, it's not recommended for modern builds, primarily because CPU's/GPU's have become so fast, the frequency swicthes/power draw from low to high load (which happen in microseconds) trip the overprotection's on the CV, and can cause random restarts.

If you are going to do a new build soon, then that might be the time to get a new decent gold quality unit, which will last you from your new build and on into the next one. If you need any suggestions, feel free to PM me.
 
Last edited:
Jun 25, 2021
5
1
15
0
Yea just need new ram.

As to why it happened? Not too sure, but hardware sometimes just fails and it could be as simple as that and Im sure thats probably what it was considering you didnt do anything to the ram in the bios, etc.

So yea, just buy a new whole kit of ram, and you should be good. You possibly, maybe, might be able to sell your 1 working chip to recoup a small amount of the cost. Or just keep it around as a backup in case anything happens with your new kit in the future :)
Hi Jason, thank you for your reply. I will get a new kit ordered (I was planning to upgrade the RAM about now anyway) and just keep the one good stick in case of future emergencies.
 
Reactions: Jason H.
Jun 25, 2021
5
1
15
0
Hey Tom,

Sorry for the late reply. Yes, as the other posters have mentioned, getting new ram is the way to go. I'd suggest rather than getting a similar single DIMM, you buy a matched kit. This will ensure they work together rather than two mixed DIMMS (even tough might be very close in all specs), which could cause compatibility issues. You can then sell the single working DIMM you have to offset the cost.



More often than not, it's just down to a bad minor component on the DIMM. It does happen, but not really that often.

Yes, the CV although good enough for a basic enough system, it's not recommended for modern builds, primarily because CPU's/GPU's have become so fast, the frequency swicthes/power draw from low to high load (which happen in microseconds) trip the overprotection's on the CV, and can cause random restarts.

If you are going to do a new build soon, then that might be the time to get a new decent gold quality unit, which will last you from your new build and on into the next one. If you need any suggestions, feel free to PM me.
Hi Keith, no problem and thank you I appreciate your advice. I will order a new kit and just hold onto the working stick for any future needs.

That's good to know about the CV, I think I will slowly be upgrading everything to the point it is basically a new build (I just learned there is no update available for my MB and the 11th gen CPUs) so will definitely start with a gold unit to support a more powerful future build.

One final question if you (or anyone else) have any thoughts I would be interested to hear - when starting the PC from cold I am 9 times out of 10 getting no response the first press. The components start up up but nothing comes up on the screen, I have to power off with a long press on the power then start up again - it always starts and runs fine on the second press. This started happening following the RAM drama earlier this week. Any idea what this might be, it's not a major issue but slightly odd?
 

ASK THE COMMUNITY