PC restarting help please (hardware)

[Trackman]

I have a PC that is restarting itself, this is not an OS problem but a hardware one.
What I have found if I isolate it from the mains after a while it will boot then restart again.
I have also checked the Q codes on the motherboard, but it’s showing everything is hunky dory and it does not stick on a particular Q code, just the ones that initialise the computer.
Ive also removed the gfx card and memory, but the post Q codes sticks on a number. It does this on both gfx and memory so I’m not sure what is going on.
Ive swapped the gfx card for a back up which was the main suspect, but it’s doing the same, also replaced memory modules.
The only thing I’ve done in the last couple of weeks is to install a GeForce rtx 2060 so I can play with my trains and planes again. I’ve looked it up and the power drain is about 60W less.
The PSU is a RT1000 but I’ve had for donkey years, motherboard is knocking on a bit too.
Anyone any ideas what is happening? I’m leaning towards a PSU problem.
Thanks in advance.

 

[Ediswan]

For clarity, was the PC in the habit of restarting before you installed the GeForce RTX 2060 ?

 

[Trackman]

For clarity, was the PC in the habit of restarting before you installed the GeForce RTX 2060 ?

No, it wasn’t.

 

[GS250]

No, it wasn’t.

Firmware on the MB? Or The GPU if it's possible?

 

[Ediswan]

No, it wasn’t.

OK ... How consistent is the run time between restarts ? Is it shorter if you are using your trains and planes ? Any other observed or suspected patterns ?

And, as a sanity check, do you definitely meet the power supply requirements, with the supplementary power connected ?

System Power Supply
Minimum 500 W or greater system power supply with one 8-pin PCI Express
supplementary power connectors.

 

[david1212]

I would check all the cables are properly together / pushed onto the motherboard and device sockets plus swap out the PSU to eliminate it. PSU capacitors deteriorate and hence the maximum power the PSU can deliver falls.
If the board has onboard graphics they will be lower power so remove the graphics card then a smaller PSU would suffice for testing.

With the system off check all the fans spin freely. Start the system and check they start up and not running slow. Check the CPU heatsink is properly seated, no clips are loose or broken and that it is clean not full of dust. Check all the other heatsinks on the mainboard devices are properly in place so no gap between the heatsink and the device it is supposed to be taking the heat away form. I have seen the clips pull out of the board.

If still an issue there is the possibility of static damage when you swapped the graphics card.

 

[nlogax]

Check that you’ve not accidentally unseated memory modules while installing the new GPU. Consider swapping slots if you have more than one module. I also second the PSU advice above, if it’s not of sufficient output to cope with the GPU you’re going to have problems.

Once you’ve checked that, disable any cpu or memory overclocking profiles that you may have active in the bios. Reset to stock mobo settings if at all possible, invoke overclocking on one component at a time, crank it up slowly and test each time. Maybe also consider an Nvidia studio driver in place of a game-ready one.

What cooling do you have in place? An RTX is likely going to run hotter than your previous card.

 

[WatcherZero]

Three main causes for sudden restarts:
1. Bad Memory
2. Overheating
3. Insufficent power such that when it warms up/GPU goes to max performance the voltage drops too low.


1. Check in the bios that the voltages are set right for your ram chips, i.e if they are 1.7 volts or 1.9 etc the correct voltage is being supplied. Run a ram verifier program (free ones are available you never need to buy one) and try removing one RAM stick at a time and seeing if it still crashes.
(You say youve done this)

2. On you motherboard/GPU performance app check the CPU & GPU temperatures at rest, are they in the 50's/low 60's? Then run a hardware intensive application like a game for 10 minutes, check back on your App if GPU still in 60's-low 80's, if CPU is in 60's-low 70's thats alright. If they are in 80's/90's youve got an overheating problem. Pull out the computer and check whether all the case fans are spinning correctly (they may stop or slow right down if temps in the 40's/50's).

3. Turn off any overclocking first as this causes extra power drain, a few decimal points drop in power to RAM can be enough for it to crash, If your computer always crashes when your GPU is working at its hardest this is another sign of low voltage. Ive had Power Supplys that have been fine for years then just start gradually dropping in performance causing problems with other components.


Of the 4 computers where I have had serious hardware failures over the last 20 years once it was bad ram (would work fine for hours then randomly crash, or a lot sooner after 20-30m minutes if using a program that used a lot of RAM like a large game or watching Videos), once it was a slowly dying PSU (would crash whenever my GPU was working really hard) and twice it was the GPU burning out (which often means computer wont even boot)

 

[Trackman]

Thank you for all your replies.
I’m getting somewhere now and have stopped the boot loop (so he says lol etc)
I‘ll explain tomorrow or whenever as I am not mauling with it tonight, especially when there is ale inside me! Hopefully I can do a full test tomorrow.

== Doublepost prevention - post automatically merged: 8 Oct 2022 ==

Update 2: (in bold as post might merge)
Yes, I'm typing on my PC! ..
It seems to be a memory issue, well sort of.
I did some googling yesterday and found a lot of people have had the same problem.
At first I removed each major component and see what would happen, I took the memory and reseated it but was still in a boot loop and reseated it again.
From what I had read, it was a case of swapping the memory around and see what would happen. I downloaded the manual on my iPad to find out the config so put them in relevant slots, it was a quad kit but I gave it a try.
Remember I've done this before using 2 mem modules..
Then Voila! Q code A0, everything hunky dory and boots! .. might have been a bit of dust or something.

Problem is now when I reboot it goes into Q code 32 again, but if I swap them into different banks or take one out it boots. I've tried all 4 modules so there is not a bad apple so to speak.
Someone else is stuck with this same problem, there is a RAM reset work around but not a solution.
My BIOS is up to date and my BIOS settings are not overclocked, going to read more about it, well at least I'm up and running!

 

[Cowley]

Update 2: (in bold as post might merge)
Yes, I'm typing on my PC! ..
It seems to be a memory issue, well sort of.
I did some googling yesterday and found a lot of people have had the same problem.
At first I removed each major component and see what would happen, I took the memory and reseated it but was still in a boot loop and reseated it again.
From what I had read, it was a case of swapping the memory around and see what would happen. I downloaded the manual on my iPad to find out the config so put them in relevant slots, it was a quad kit but I gave it a try.
Remember I've done this before using 2 mem modules..
Then Voila! Q code A0, everything hunky dory and boots! .. might have been a bit of dust or something.

Problem is now when I reboot it goes into Q code 32 again, but if I swap them into different banks or take one out it boots. I've tried all 4 modules so there is not a bad apple so to speak.
Someone else is stuck with this same problem, there is a RAM reset work around but not a solution.
My BIOS is up to date and my BIOS settings are not overclocked, going to read more about it, well at least I'm up and running!

Looks like you’ve got somewhere with this @Trackman?

 

[david1212]

@Trackman

In the BIOS can you set the memory to a lower clock rate as a test?

Another test would be using just two modules at a time with the aim of determining if one out of the four is the problem, perhaps failing once warm / hot.

 

[Trackman]

@Trackman

In the BIOS can you set the memory to a lower clock rate as a test?

Another test would be using just two modules at a time with the aim of determining if one out of the four is the problem, perhaps failing once warm / hot.

I've gone for max stability in the memory settings, in fact all settings as well.
I've tested all 4 modules as well and ran the windows 10 memory boot test on them in different configurations.

Totally baffling, I'm using a workaround I've found in some forums, basically don't power down but go into sleep mode.
Thing is, it will restart- as soon as I power down the Q code for memory error appears. Then I have to remove a module, boot then boot again, replace it and all is fine and dandy.
As I said before totally baffling.

Also checked voltages / temps etc and the BIOS itself is holding the settings.

 

[Pete_uk]

I've had quite a few pc problems in the past that taking out and putting back on the graphics card has sorted.

 

[Trackman]

I've had quite a few pc problems in the past that taking out and putting back on the graphics card has sorted.

This was a straight forward swap and I was careful in what I did, I even put the old gfx card back to try and resolve.

 

[Puffing Devil]

Have you tried https://www.memtest86.com/ ? It's worked for me before spotting a duff stick of RAM.

 

[Trackman]

Well, I've solved it, not sure how unless I disturbed something when changing the gfx card.
I've been through the BIOS settings, I disabled memory fast boot.
What I've found out is when a component fails during the motherboard boot it resets the performance settings in the BIOS.

 

[D365]

Well, I've solved it, not sure how unless I disturbed something when changing the gfx card.
I've been through the BIOS settings, I disabled memory fast boot.
What I've found out is when a component fails during the motherboard boot it resets the performance settings in the BIOS.

That's poor design. Defaulting to a "fast boot" mode is never ideal.

 

[Trackman]

That's poor design. Defaulting to a "fast boot" mode is never ideal.

My thoughts exactly.

I've been through so many forums (mainly the gigabyte ones) with this problem I saw fast boot memory mentioned once and made a mental note of it.
Anyway gave it a try and hey presto!

Thanks to all for your contributions.

If it goes bang again I will know what the problem is - memory.