[SOLVED] Faulty PCIE slot!? Motherboard on it's way out!?

Status
Not open for further replies.
Sep 29, 2019
38
0
30
Hi there, I have been having two seperate issues that are possibly related.

1. PC restarts whilst gaming (NOT OVERHEATING RELATED) which was narrowed down to a faulty PSU

2. Graphics card fan spinning up to full speed without any load on the GPU.

Number 2 has been solved by swapping the graphics card to another PCIE slot.

Now the fan is operating normally.

This makes me question whether number 1 could be a faulty mobo rather than a faulty PSU.

The mobo is an MSI Z87-G55 which I've had for six years as of November. MSI are supposed to be good...
 
Solution
After 5 minutes of playing the game HATRED, these are my readings

12v

Min 12.144v Max 12.232v avg 12.205v

5v

min 5.080v Max 5.120v avg 5.080v

3.3v

Min 3.360v Max 3.392v avg 3.384v
Nothings looks abnormal by them - but again, that's nowhere near a guarantee that the PSU is functioning normally.
I know you mentioned it previously, what were your MAX CPU and GPU temps?

I'd still be looking at PSU frankly. For random restarts, once software/firmware eliminated, the order of blame tends to go:
PSU > GPU > RAM > MB > CPU
Sep 29, 2019
38
0
30
Right, I've just tried to play Hatred. Restarted after about 15 minutes.

So it would appear that these are two separate issues and the PSU could still be at fault after all.

Not sure why the PCIE slot change allowed the fan on the GPU to work correctly again though.
 

PC Tailor

Illustrious
Ambassador
Well from my recollection, the XFX Pro Core is a more budget orientated PSU by today's standards, and if it is also 6 years old I would be suspecting it too. Effectively when you start or run a game, the GPU draws a sudden amount of power from the PSU, which if the PSU is faulty or has hyper-sensitive protections (like OVP) then it can cause the PSU to shut off.

Once software has been eliminated, if you are encountering random restarts (especially during games) it is usually the GPU or the PSU.

You could monitor rail voltages using HWInfo and see if anything obvious appears (on the 12V, 5V and 3.3V rails) but they aren't overly accurate - it's just a check for anything obvious. Outside of that, the best way to test is to swap out the component and retest.

Overall if you've eliminated software and firmware (BIOS, Drivers etc.) then you would initially suspect PSU based on the age.

Does the PC completely restart or just shut off?
 
Sep 29, 2019
38
0
30
It sounds like a power cut (that's what I thought had happened the first time) then the power comes back on.

12v rail

12.232v

5v rail

5.080v

3.3v rail

3.392v

How do these figures look?

Once I've got the component, it will be too late and the money will be spent, that's why I am trying to get it right the first time.

At the moment it could be the PSU, MoBo or GPU.

Edit: Things tried so far

BIOS update
GPU driver update
Dust cleaned from all components (there was A LOT)
Thermal paste reapplied to GPU core
 

PC Tailor

Illustrious
Ambassador
Rails are normal - are these just current though or MAX / MIN? You also want to check MAX/MIN during load, not just idle, as many times the PSU can faulter when load is induced / power is drawn from the GPU for example.

Once I've got the component, it will be too late and the money will be spent, that's why I am trying to get it right the first time.
I understand your concer, but without swapping and retesting somehow, there is no right first time. The only guaranteed and best way to test a problem is to swap the component and see if the issue persists. Even if this is borrowing components from a friend.

Also if you buy new components you should still be able to return them within a certain timeframe regardless. But that depends on the location.
 
Sep 29, 2019
38
0
30
Silly me, I didn't scroll over where there are more readings

12v rail

Min 12.144 Max 12.232 avg 12.226

5v rail

Min 5.080v Max 5.120v avg 5.080v

3.3v

Min 3.376v Max 3.392v avg 3.391v

Should I now start a game and see if the readings change? Will it be okay to have it minimized in the background or should be be full screen?

Will the title screen be okay or do I have to actually play it?
 

PC Tailor

Illustrious
Ambassador
Should I now start a game and see if the readings change? Will it be okay to have it minimized in the background or should be be full screen?
Play a game for a while.

When you start and play a game the GPU draws a sudden amount of power, and it can be during these stages that you might see a large voltage drop or spike for example. The voltage readings are just a quick idea, they are often inaccurate, but if you see something obvious in the voltage readings, then you know something is likely wrong somewhere.

Basically your tolerances are:

VMaxMin
12V12.611.4
5V5.254.75
3.3V3.4653.135
 

PC Tailor

Illustrious
Ambassador
After 5 minutes of playing the game HATRED, these are my readings

12v

Min 12.144v Max 12.232v avg 12.205v

5v

min 5.080v Max 5.120v avg 5.080v

3.3v

Min 3.360v Max 3.392v avg 3.384v
Nothings looks abnormal by them - but again, that's nowhere near a guarantee that the PSU is functioning normally.
I know you mentioned it previously, what were your MAX CPU and GPU temps?

I'd still be looking at PSU frankly. For random restarts, once software/firmware eliminated, the order of blame tends to go:
PSU > GPU > RAM > MB > CPU
 
Solution
Sep 29, 2019
38
0
30
Right, I've just been playing the game HATRED for 1/2 an hour.

I have enabled the graphs from HWinfo and they are consistant. No voltage spikes, just a fairly smooth graph.

Very small spikes on the 3.3v rail, but the 12v and 5v are smooth.

If only there was an option to upload an image directly to the site but there isn't.
 
Sep 29, 2019
38
0
30
Max GPU temp

80 degrees (and that's with new thermal paste!) but it only gets this hot playing HATRED.

I've played Crysis often where it only gets to 55.

From what I've read a hot GPU will enter thermal throttling rather than cutting the power.

Max CPU temp 43 degrees.
 

PC Tailor

Illustrious
Ambassador
From what I've read a hot GPU will enter thermal throttling rather than cutting the power.
Technically that will apply to everything. And is more of a "definitely should" - modern components should throttle to prevent overheating damage.

Again just the MAX and MIN would suffice. Even a single spike above 3.465 is a cause for concern.
Either way, my suspicion would be PSU due to it's age (if it is as old as the rest of the rig).
 
Sep 29, 2019
38
0
30
Well I've just finished a 30 minute session on HATRED.

By the way, here is a review of the PSU, so it was SUPPOSED to be good quality

 
Sep 29, 2019
38
0
30
Technically that will apply to everything. And is more of a "definitely should" - modern components should throttle to prevent overheating damage.

Again just the MAX and MIN would suffice. Even a single spike above 3.465 is a cause for concern.
Either way, my suspicion would be PSU due to it's age (if it is as old as the rest of the rig).

Max-80

Min-31

Edit: You meant the max and min of the power rails?

In my post above.
 
Last edited:
Sep 29, 2019
38
0
30
Okay I'm just browsing and the GPU fan has gone wild. PSU, MoBo or GPU fault?

I am running the HWinfo graph and 5v is flat.

12v has frequent minor fluctuations and 3.3v has infrequent very minor fluctations.

I am going to restart my system now to try to get the GPU fan to go back to idle speed as it is loud.
 

PC Tailor

Illustrious
Ambassador
so it was SUPPOSED to be good quality
WAS is one of the key words though. Not stipulating it is BAD but those kind of PSUs will certainly show their age after a little while. And I was assisting someone not long back who had an issue with the exact same PSU after 5/6 years.

Do you notice if the fan only ramps up when it hits 80 degrees or is it random?
Have you tried manually setting a fan curve?

Following that, it's difficult to identify what the faulty component is, best bet would be to see if your GPU works normally in another system, as the only sure fire way of identifynig faulty hardware is by swapping and retesting.
 
Sep 29, 2019
38
0
30
WAS is one of the key words though. Not stipulating it is BAD but those kind of PSUs will certainly show their age after a little while. And I was assisting someone not long back who had an issue with the exact same PSU after 5/6 years.

Do you notice if the fan only ramps up when it hits 80 degrees or is it random?
Have you tried manually setting a fan curve?

Following that, it's difficult to identify what the faulty component is, best bet would be to see if your GPU works normally in another system, as the only sure fire way of identifynig faulty hardware is by swapping and retesting.

That's a useless suggestion as I don't have another system.

"
WAS is one of the key words though. Not stipulating it is BAD but those kind of PSUs will certainly show their age after a little while. And I was assisting someone not long back who had an issue with the exact same PSU after 5/6 years."

<Mod Edit>, there were threads on THIS FORUM recommending it at the time.

Look at the best answer on this thread and that's just one example



What more evidence do you WANT!?
 
Last edited by a moderator:

PC Tailor

Illustrious
Ambassador
Woah where did that come from? I'm trying to help. Apologies if it came across badly, but it certainly wasn't my intention.

Unfortunately the first link is from 2013 which proves my point, it is a good PSU, but when it's been used for 6 years, it deteriorates. The XFX was made by Seasonic if I am not mistaken, and at that time is good quality. Never said it was bad quality, just it's aged..

I wont warrant the second post with a response because it's very incorrect unfortunately.

We're all volunteers here with different views. There will always be conflicting answers on any forum.

I equally never said it IS your PSU, and I never said it was bad quality, just that after 6 years, it certainly COULD be the issue, thus my other questions. And apologies if it is a useless suggestion, but that's your option to identify correctly what is causing it, try another system or leave it with a shop who can identify it for you.

Again apologies if anything came across badly but certainly wasn't my intention. I'm ultimately a voluteer trying to help.
 
Last edited:
Status
Not open for further replies.