My PC keeps crashing when under load. Help!

Kirth Gersen · 23 Aug 2018 at 10:32

Gosh, it's been no less than 8 years since I last posted on here. I used to be a keen overclocker/system builder, but eventually was happy with my rig, stopped playing Flight Simulator, stopped overclocking and lost interest. Since then, I have upgraded a few parts: the power supply, RAM etc to keep it working and relatively current.

So I've started getting crashes under load (games or using Photoshop). Not BSODs, but the screen freezes and the speakers let out a terrible wailing sound. This has happened once when not under load too.

I've stripped the PC down totally, including taking out the mobo, cleaned up every component, replaced the battery and put it back together... twice Still bad.

I'm running Coretemp, and there is nothing scary. Temps under load don't go above 50C, usually mid 40s, and I put the fans on full blast when gaming etc. Mid 30s when not under load. Occasionally RAM gets to 75% (12Gb) when in photoshop, and CPU hits 100%. the PSU is never challenged. It's times like that when it crashes.

My Rig:
Intel Core i5 2500K
Mobo: Asus P8Z68-V LX
RAM: 16Mb G-Skill RipJaws PC3-12800
GPU: Radeon HD5770 - yeah its 8 or 9 years old
PSU: 600Watt Corsair
2 Hard drives and Boot drive is a 128Mb SSD

I'm certain there is a component failing, but what? I'm on a budget, so don't want to do a total rebuild, just identify the culprit and replace. I've checked RAM using mdsched.exe and that seems fine. My best guess is the GPU. How do I check that?

Any help would be appreciated.

Thanks

Donnie Fisher · 24 Aug 2018 at 00:50

Test the ram by running only 1 stick at a time.

Gpu in a different slot?

Jamin280672 · 24 Aug 2018 at 06:44

Sorry wrong thread

Kirth Gersen · 24 Aug 2018 at 08:42

Donnie Fisher said:
Test the ram by running only 1 stick at a time.

Gpu in a different slot?

Will give it a shot and report back.

LuckyBenski · 24 Aug 2018 at 08:51

Is the CPU still running an overclock? If so, what frequency/voltage? Chips do age and become unstable, though only to a noticeable degree if abused I think.

Jamin280672 · 24 Aug 2018 at 09:31

I know you say you've removed the overclock and it's still doing it how is the ram set? Have you loaded xmp settings if so, make sure the ram is getting the correct voltage, and also bump up the uncore voltage a little, that's for the memory controller in your chip, usually freezing without a bsod is ram related, crashes with a bsod or instant reboot is cpu.

This time, the right thread lol

EsaT · 24 Aug 2018 at 15:15

Kirth Gersen said:
My best guess is the GPU.

If crashes happen also in Photoshop GPU isn't likely cause.
CPU-motherboard-RAM is more likely.

What's the particular PSU model?
Not all Corsairs are quality PSUs.

And doesn't Event viewer show any more serious errors with time stamp close to crashes?

Kirth Gersen · 24 Aug 2018 at 17:07

LuckyBenski said:
Is the CPU still running an overclock? If so, what frequency/voltage? Chips do age and become unstable, though only to a noticeable degree if abused I think.

Not overclocked in years.

Kirth Gersen · 24 Aug 2018 at 17:08

Jamin280672 said:
I know you say you've removed the overclock and it's still doing it how is the ram set? Have you loaded xmp settings if so, make sure the ram is getting the correct voltage, and also bump up the uncore voltage a little, that's for the memory controller in your chip, usually freezing without a bsod is ram related, crashes with a bsod or instant reboot is cpu.

This time, the right thread lol

Thanks - will check RAM voltage etc.

Kirth Gersen · 24 Aug 2018 at 17:15

EsaT said:
If crashes happen also in Photoshop GPU isn't likely cause.
CPU-motherboard-RAM is more likely.

What's the particular PSU model?
Not all Corsairs are quality PSUs.

And doesn't Event viewer show any more serious errors with time stamp close to crashes?

It's a Corsair HX620W. Checking my records, that must have been in there about 8-9 years also...

Kirth Gersen · 24 Aug 2018 at 17:16

EsaT said:
And doesn't Event viewer show any more serious errors with time stamp close to crashes?

I will push it and check event viewer next time. Didn't think of that before.

EsaT · 24 Aug 2018 at 18:27

Kirth Gersen said:
I will push it and check event viewer next time. Didn't think of that before.

I think logs should show it if something was recorded last time.
Unless it was long time ago...

Kirth Gersen · 24 Aug 2018 at 21:49

Didn't take long while playing an undemanding game. kernel-power 41 - bugcheckcode 0 - task 63. I've had 107 of these in the past 3 months.

- System
- Provider
[ Name] Microsoft-Windows-Kernel-Power
[ Guid] {331C3B3A-2005-44C2-AC5E-77220C37D6B4}
EventID 41
Version 6
Level 1
Task 63
Opcode 0
Keywords 0x8000400000000002
- TimeCreated
[ SystemTime] 2018-08-24T20:03:05.075079800Z
EventRecordID 30340
Correlation
- Execution
[ ProcessID] 4
[ ThreadID] 8
Channel System
Computer Richard-PC
- Security
[ UserID] S-1-5-18
- EventData
BugcheckCode 0
BugcheckParameter1 0x0
BugcheckParameter2 0x0
BugcheckParameter3 0x0
BugcheckParameter4 0x0
SleepInProgress 0
PowerButtonTimestamp 0
BootAppStatus 0
Checkpoint 0
ConnectedStandbyInProgress false
SystemSleepTransitionsToOn 1
CsEntryScenarioInstanceId 0
BugcheckInfoFromEFI false
CheckpointStatus 0

I've done a bit of googling and the conclusions are wide and varied - ie Power Plan, PSU, CPU, RAM and Drivers in that order of frequency of issue resolution. My power plan has not changed for years, and I have checked and updated all the drivers. So I conclude it is a piece of hardware, but which one? We know both the PSU and CPU are old. The RAM was upgraded a few years ago to get photoshop to run better.

Running out of steam today, but I have a multimeter hidden away somewhere, so may try this https://www.wikihow.com/Check-a-Power-Supply as a next step. Any other ideas?

t31os · 27 Aug 2018 at 03:52

Given the age of the PSU that would be my guess, check your voltage rails(in the bios and/or with a motherboard app) and see if any are off by any significant amount.

Kirth Gersen · 31 Aug 2018 at 19:18

Quick update. I don't believe the PSU is at fault. All the voltages are bang on. Which leaves the CPU. I'm beginning to think I am going to have to dip into savings as a new CPU means a new mobo, means probably new RAM (mine are DDR3)... means, well pretty much a total upgrade. Suppose 8 years isn't bad in terms of depreciation.

Away for 2 weeks now. Time to mull things over.

EsaT · 31 Aug 2018 at 20:58

Kirth Gersen said:
Quick update. I don't believe the PSU is at fault. All the voltages are bang on. Which leaves the CPU.

While DC average voltage is correct there's still that AC component called ripple.
Only multimeters which might give some hint of its level cost same as couple good PSUs.
And for truly measuring ripple you need (oscillo)scope costing even more.
But while really old design now seriously outdated design HX620 has high quality capacitors, which are primary component's behind PSU's aging/wearing.

Again that motherboard was rather basic mothebaord so its CPU VRM might be tiring.
CPU itself is extremely reliable part if temperature and voltage are kept reasonable.

Kirth Gersen · 18 Sep 2018 at 13:29

Further update. I decided to bite the bullet and upgrade the Mobo, CPU and RAM to modern day standards. My work in Lightroom will benefit with the boost in CPU. Given it a bit of a stress test in LR and gaming and it seems solid as a rock (fingers crossed), and surprisingly runs a little bit cooler.

Mobo - MSi Z370-A pro
CPU - Intel i7 8700k
RAM - 16 Gb Corsair Vengeance DDR4 3000Mhz

Thanks for your help. In November, I will travel to Spain and test the CPU and RAM on my mother's PC. She should benefit from the part that works!