Question New system is bricking RAM sticks

[jkrm]

Hello, recently I built this setup for a client:

I3-10100F
ASRock H510M-HVS R2.0
Corsair Vengeance LPG 2X8 16GB 2666MHZ CMK16GX4M2A2666C16
Zotac GTX 1650
EVGA ATX 550W 80 Plus White

The client was playing CSGO and suddenly the PC rebooted (without any freezing or bsod), and was power cycling, but never booted again, so he removed one RAM stick and the PC booted again. We tested the two slots of the mobo and both worked. We tested the other RAM stick and apparently it died. Then, today, the PC did the same thing and the second RAM stick died too. We tried everything already, clearing cmos, etc etc.. Why is this happening?

Fresh win 11 install, all drivers updated, latest bios, etc

 

[Ralston18]

Where was the PSU purchased - source?

Were any cables from any other PSU used for the build?

Where was the RAM purchased?

Was the 2X8 RAM configuration from a matched kit for dual channel use?

 

[jkrm]

Where was the PSU purchased - source?

Were any cables from any other PSU used for the build?

Where was the RAM purchased?

Was the 2X8 RAM configuration from a matched kit for dual channel use?

All the parts from this build was purchased on Kabum, the biggest store here in Brazil. Everything brand new. I only used this PSU for this build, and this PSU isn't modular, so I only used the cables that comes with it (and there's no cable extensions). The RAM came in a kit 2x8 (one box with 2 sticks), both exact same model.

 

[Ralston18]

At the next sucessful boot look in Reliability History and Event Viewer for error codes, warnings, and even informational events that were captured just before or at the time of reboots.

Look for the date and time around when the client was playing CSGO or the date and times other reboots occurred.

For the most part, my thought is that there is a loose connection somewhare.

No one wants to force a connection at any time especially with a new build. So something that seems connected and in place is really not fully in place. Then heat (expansion/contraction) and vibrations cause intermittent loss of connectivity and the build reboots/crashes.

Power down the build, unplug, open the case.

Carefully and methodically plug and unplug everthing a few times. No rush - be methodical.

Inspect plugs, slots, pins, etc. for signs of damage, debris, bare conductor showing, damage of any sort. Remember to check case connections as well. Verify that all the connectors there are correctly installed and securely in place.

Pay close attention to the I/O panel. Use a bright flashlight to carefully inspect everything everywhere in the case.

What might happen is that suddenly some connection will seat itself very easy and smoothly. You may well feel the difference.

RAM would be a good starting point. Some motherboards require that the first physically install RAM stick be placed in a specific slot.

As for clearing CMOS refer to the Motherboard's User Guide/Manual for the correct procedure.

 

[jkrm]

At the next sucessful boot look in Reliability History and Event Viewer for error codes, warnings, and even informational events that were captured just before or at the time of reboots.

Look for the date and time around when the client was playing CSGO or the date and times other reboots occurred.

For the most part, my thought is that there is a loose connection somewhare.

No one wants to force a connection at any time especially with a new build. So something that seems connected and in place is really not fully in place. Then heat (expansion/contraction) and vibrations cause intermittent loss of connectivity and the build reboots/crashes.

Power down the build, unplug, open the case.

Carefully and methodically plug and unplug everthing a few times. No rush - be methodical.

Inspect plugs, slots, pins, etc. for signs of damage, debris, bare conductor showing, damage of any sort. Remember to check case connections as well. Verify that all the connectors there are correctly installed and securely in place.

Pay close attention to the I/O panel. Use a bright flashlight to carefully inspect everything everywhere in the case.

What might happen is that suddenly some connection will seat itself very easy and smoothly. You may well feel the difference.

RAM would be a good starting point. Some motherboards require that the first physically install RAM stick be placed in a specific slot.

As for clearing CMOS refer to the Motherboard's User Guide/Manual for the correct procedure.

Everything feels right. We already checked for loose cables and stuff. Clearing CMOS is a standard procedure, there's no need to check for the mobo manual.. Now we are waiting for the new RAM to arrive. If this one bricks too, we are gonna to get a new model to test it. And, if the new model bricks too, we are gonna to change the mobo I think. I don't think the PSU are causing this, because I've never seen a reboot/crash bricking a RAM.

When the new RAM arrives, we are gonna downgrade the BIOS to the 1.50 version, because now it's on the 1.40G Beta version, maybe it's because it's on beta bios.

 

[DRagor]

The question: was the sticks that died installed in the same mobo slot in both cases?

 

[jkrm]

The question: was the sticks that died installed in the same mobo slot in both cases?

We don't remember for sure, but last time the stick was in the #2 slot. First time I think it was in the #1 slot. We don't remeber for sure.

I think it's a bad RAM kit. I hope it is.

If it was bricking just in one slot, what could it be? CPU?

And what do you guys think about this PSU? I know it's a low end psu, but for this system, I think it's ok.

 

[DRagor]

If it was bricking just in one slot, what could it be?

In that case it would be strong possibility of bad motherboard.

 

[Ralston18]

This motherboard?

https://download.asrock.com/Manual/H510M-HVS R2.0.pdf

[Do verify that I found the correct User Manual.]

Agree that clearing CMOS is a standard procedure.

However, details matter:

Did you follow the procedure per physically numbered Page 19 of the User Manual?

Section 2.5 - Jumpers Setup.

"CLRMOS1 allows you to clear the data in CMOS. To clear and reset the system parameters to default setup, please turn off the computer and unplug the power cord from the power supply. After waiting for 15 seconds, use a jumper cap to short the pins on CLRMOS1 for 5 seconds. However, please do not clear the CMOS right after you update the BIOS. If you need to clear the CMOS when you just finish updating the BIOS, you must boot up the system first, and then shut it down before you do the clear-CMOS action. Please be noted that the password, date, time, and user default profile will be cleared only if the CMOS battery is removed. Please remember toremove the jumper cap after clearing the CMOS "

If you did not follow that procedure then something may still be astray.

Also there is a footnote on the page:

"If you clear the CMOS, the case open may be detected. Please adjust the BIOS option “Clear Status” to clear the record of previous chassis intrusion status. "

Yes, a bad motherboard is quite possible.

However, if the documented procedures are not followed then what was done (or perhaps not done) can be called into question......

Leave no room for doubt.

 

[jkrm]

This motherboard?

https://download.asrock.com/Manual/H510M-HVS R2.0.pdf

[Do verify that I found the correct User Manual.]

Agree that clearing CMOS is a standard procedure.

However, details matter:

Did you follow the procedure per physically numbered Page 19 of the User Manual?

Section 2.5 - Jumpers Setup.

"CLRMOS1 allows you to clear the data in CMOS. To clear and reset the system parameters to default setup, please turn off the computer and unplug the power cord from the power supply. After waiting for 15 seconds, use a jumper cap to short the pins on CLRMOS1 for 5 seconds. However, please do not clear the CMOS right after you update the BIOS. If you need to clear the CMOS when you just finish updating the BIOS, you must boot up the system first, and then shut it down before you do the clear-CMOS action. Please be noted that the password, date, time, and user default profile will be cleared only if the CMOS battery is removed. Please remember toremove the jumper cap after clearing the CMOS "

If you did not follow that procedure then something may still be astray.

Also there is a footnote on the page:

"If you clear the CMOS, the case open may be detected. Please adjust the BIOS option “Clear Status” to clear the record of previous chassis intrusion status. "

Yes, a bad motherboard is quite possible.

However, if the documented procedures are not followed then what was done (or perhaps not done) can be called into question......

Leave no room for doubt.

Yes, I made all the right steps for clearing the CMOS, including removing the battery for 5 min.

Last time, the pc crashed on idle, he wasn't even gaming... Just on teamspeak with his friends. I think this mobo is very poorly built, idk.. Surely it can handle a 10100f, but the quality just sucks. Idle temps was 35c last time I saw, stress testing like 70 75c, but I think this mobo is sendind a lot more vcore than the cpu needs.. For the RAM too.

One thing that I think it's strange it was the VRM sensors and stuff on HWMONITOR... All of them was like 110-120c on idle.. Stressing was the same... I don't know if it was a bug or something..

 

[Ralston18]

In the meantime as a matter of elimination try testing the PSU.

If you have a multimeter or know someone who does then test the PSU's output voltages.

Not a full test because the PSU is not under load. However, any voltages out of tolerance make the PSU suspect.

FYI:

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

 

[jkrm]

In the meantime as a matter of elimination try testing the PSU.

If you have a multimeter or know someone who does then test the PSU's output voltages.

Not a full test because the PSU is not under load. However, any voltages out of tolerance make the PSU suspect.

FYI:

https://www.lifewire.com/how-to-manually-test-a-power-supply-with-a-multimeter-2626158

Do you think it could be just a bad RAM kit? And we are breaking ours head over nothing

I think I don't have the skills to test a PSU like this

 

[Roland Of Gilead]

Hey there,

All the recommended checks above are exactly the way to go.

Are you sure the PSU is 550w? I can only find the 500w version.

With that said, the PSU in that build is pretty weak. I'd put on par with a Corsair VS. Not suitable for gaming systems. It's cheap for a reason, and covered with almost standard warranty of 3 years.

The PSU is the heart of each system. You don't want to start off with a bad heart that causes issues. I'm not saying this is the whole issue, but being underpowered, and made with old topology, it's not a good PSU at all.

 

[jkrm]

Hey there,

All the recommended checks above are exactly the way to go.

Are you sure the PSU is 550w? I can only find the 500w version.

With that said, the PSU in that build is pretty weak. I'd put on par with a Corsair VS. Not suitable for gaming systems. It's cheap for a reason, and covered with almost standard warranty of 3 years.

The PSU is the heart of each system. You don't want to start off with a bad heart that causes issues. I'm not saying this is the whole issue, but being underpowered, and made with old topology, it's not a good PSU at all.

https://imgur.com/U9rG0J5

View: https://imgur.com/U9rG0J5


Here's the PSU

Seeing more closely, I think it's not even 80P at all.. But this is a very low power draw system, I tought that PSU would handle this fine.

And, even if it is the PSU, why a crash/reboot is bricking a RAM stick? Over voltage to the vRAM?

Man, I love EVGA, I never tought they would make garbage PSU's

 

[Roland Of Gilead]

https://imgur.com/U9rG0J5

View: https://imgur.com/U9rG0J5


Here's the PSU

Seeing more closely, I think it's not even 80P at all.. But this is a very low power draw system, I tought that PSU would handle this fine.

And, even if it is the PSU, why a crash/reboot is bricking a RAM stick? Over voltage to the vRAM?

Yeah, it's not great. Although the symptoms seem different, if the PSU wasn't proving enough power, then you might experience some degradation in performance from the system.

Run memtest86+ on the mem, one at a time, then both together. If you have any errors after 4 passes, your ram could be faulty. This, based on the system, is indicative of ram issues.

Ruling everything out one by one is all we can help you do. Not having access to the host PC, it's possibly an easy fix, or not! In this case more investigation is needed to rule everything out, and be left with the lowest common denominator.

 

[jkrm]

Yeah, it's not great. Although the symptoms seem different, if the PSU wasn't proving enough power, then you might experience some degradation in performance from the system.

Run memtest96+ on the mem, one at a time, then both together. If you have any errors after 4 passes, your ram could be faulty. This, based on the system, is indicative of ram issues.

Ruling everything out one by one is all we can help you do. Not having access to the host PC, it's possibly an easy fix, or not! In this case more investigation is needed to rule everything out, and be left with the lowest common denominator.

Well, right now my client's pc is on the other side of the country and the 2 sticks that he have are dead, so we are waiting for the new sticks to arrive, so I think I will start with Prime95 and let the test run for 24hrs, if nothing happens, we good, if the system crashes and bricks the RAM again, we are gonna swap to a better mobo, because this one is trash, for real. (and a better psu too)

I feel sorry for him, he saved money for 2 years for this pc and this sh1t happens, my man just want to play csgo in peace

 

[punkncat]

When installing the RAM, are you getting a good and solid click into the slot and can verify that the "pivot" (there may be one or two each slot) that locks is in the closed position?

It sounds much like the RAM stick is not going fully into the slot and causing issue.

 

[Roland Of Gilead]

Well, right now my client's pc is on the other side of the country and the 2 sticks that he has are dead, so we are waiting for the new sticks to arrive, so I think I will start with Prime95 and let the test run for 24hrs, if nothing happens, we good, if the system crashes and bricks the RAM again, we are gonna swap to a better mobo, because this one is trash, for real. (and a better psu too)

I feel sorry for him, he saved money for 2 years for this pc and this sh1t happens, my man just want to play csgo in peace

Ah, okay. Yes, having spent some, we expect (or hope) it will work flawlessly.

In this case, you could use Prime95 for testing, but for ram, doing the blend test. I don't particularly use Prime95 for testing mem, as I think something outside of the windows environment is more reliable. Prime is certainly useful (on small ffts) for testing CPU reliability by loading only the cores and cache on the CPU. Note: This can seriously make you take a gasp at how high the temps can go! So be prepared if you've not seen it in action yet.

 

[jkrm]

When installing the RAM, are you getting a good and solid click into the slot and can verify that the "pivot" (there may be one or two each slot) that locks is in the closed position?

It sounds much like the RAM stick is not going fully into the slot and causing issue.

Yes, 100% solid click and locked in position. Maybe the slot has some type of dust on it, but I don't think so. The mobo was brand new.

 

[jkrm]

Ah, okay. Yes, having spent some, we expect (or hope) it will work flawlessly.

In this case, you could use Prime95 for testing, but for ram, doing the blend test. I don't particularly use Prime95 for testing mem, as I think something outside of the windows environment is more reliable. Prime is certainly useful (on small ffts) for testing CPU reliability by loading only the cores and cache on the CPU. Note: This can seriously make you take a gasp at how high the temps can go! So be prepared if you've not seen it in action yet.

Yeah, I've done a few P95 passes in some systems, idk wth this thing do, but like in furmark stress test my 5800x with pbo enabled with +200mhz override don't pass 75c with an capellix 280mm, and on P95 goes like 85 90c+ lol

 

[punkncat]

I might suggest that if you can keep the system running long enough to download HW Info/monitor and see if voltages being reported look correct.

Those White rated EVGA units are junk. They are cheap junk, and I have used them in lower power builds, but am not going to say it hasn't bitten me in the hiney before.

 

[DRagor]

Man, I love EVGA, I never tought they would make garbage PSU's

The point is, EVGA does not make PSUs, they just brand units made by other companies. So they have some good models but also some absolutely crappy models. Yours is closer to the latter end of scale.

 

[jkrm]

So, should I run memtest86+, p95 or another stress test when the new kit arrives to see if the system is stable?

In my view, the p95 would be the more suitable, because it will stress not only the RAM, but the whole system. But I could be wrong.

 

[Roland Of Gilead]

So, should I run memtest86+, p95 or another stress test when the new kit arrives to see if the system is stable?

In my view, the p95 would be the more suitable, because it will stress not only the RAM, but the whole system. But I could be wrong.

For purely testing ram, to ensure no instability, Memtest is the way to go. Once you know they are good (an d as a result the issue is solved) or not. If that's the case and you have the same issues with new ram, it maybe the mobo.

The blend test in P95 is useful for a more everyday usage type. Small ffts, purely for testing CPU stability. But in no way do the small ffts test is comparable to everyday use. It's really to find the most minuscule of instability.

 

[jkrm]

All right. Thank you guys, I'll update you all as I find something.