Question Random Crashes / No Boot

[TallMan21]

My workstation PC, which I built about a year ago, is crashing seemingly randomly. I haven't been able to lock down a specific cause, or repeat the issue consistently.
It has crashed running heavier CAD programs, but also crashed just reading email. I checked cpu (i7 7820x) temps, which seemed normal both at idle and under load (getting no higher than 70C during a 20min stress test), and cleaned all dust out of the case. I tried checking out fan speeds in the BIOS, and the PC crashed when I set all fans to max rpm. This lead me to believe that it was the PSU (Corsair cx650 bronze), rather than an overheating issue. However, sometimes I am able to set fans to max speed for several minutes with no crash. Sometimes the computer won't boot all the way, fans will spin and hard drive light is on, but I get no display. Sometimes, the PC locks up and wont respond until I perform a hard shut down. From the research I have done, most of these issues could be related to a bad PSU, but I wanted to see if anybody had other thoughts before I spring for a newer, bigger power supply.

There have been 0 other symptoms besides the crashing and boot. No odd sounds, change in performance, etc.

For those who might want all the system specs.
Mobo: MSI X299 Raider
CPU: i7 7820x
Ram: Corsair vengeance 32gb 3000
GPU: NVIDIA quadro P4000
PSU: Corsair cx650m

 

[natcha12]

If its overheating, the pc will probably shutdown
If its the power supply, the pc will probably shutdown
But if its the RAM, the PC will DEFINATELY crash and/or shutdown

What exactly is happening? Crash with a frozen screen, crash with blue screen, instant restarting/powering off/restart loop?

With bad RAM, your pc can act fine until the bad sectors are used. This could be after a time doing low power stuff like emails or browsing, or within seconds gaming or using your CAD software.

PS. It could also be bad HDD/SSD storage for your OS , or even display issues:

1. If you use a wifi card/adaptor and it has a light indicator, and it eventually starts flashing while the display is blank, this means your PC has booted wifi drivers and connected to the wifi but doesn't output video, its a video issue.
2. If its a storage issue, unplugging the drives should make an "error no drive" message after turning on and POSTing the PC. That would mean its POSTing properly, but nothing good is happening after it tries getting the OS off the drive.

Hope this helps, let me know how it goes.

 

[TallMan21]

If its overheating, the pc will probably shutdown
If its the power supply, the pc will probably shutdown
But if its the RAM, the PC will DEFINATELY crash and/or shutdown

What exactly is happening? Crash with a frozen screen, crash with blue screen, instant restarting/powering off/restart loop?

With bad RAM, your pc can act fine until the bad sectors are used. This could be after a time doing low power stuff like emails or browsing, or within seconds gaming or using your CAD software.

PS. It could also be bad HDD/SSD storage for your OS , or even display issues:

1. If you use a wifi card/adaptor and it has a light indicator, and it eventually starts flashing while the display is blank, this means your PC has booted wifi drivers and connected to the wifi but doesn't output video, its a video issue.
2. If its a storage issue, unplugging the drives should make an "error no drive" message after turning on and POSTing the PC. That would mean its POSTing properly, but nothing good is happening after it tries getting the OS off the drive.

Hope this helps, let me know how it goes.

A couple of different crashes occured. Sometimes there would be a frozen screen, unresponsive until I performed a hard shut down. Sometimes the screen would freeze and start flashing, and would continue to flash until a hard shut down. Sometimes it would instant restart (sometimes it would boot up fine when it restarts this way, other times it would just power on and leave a black screen). Never had a blue screen or any sort of crash message screen.

Connected via ethernet, no wifi. I will check the storage.
Is there any way I might test the RAM?

 

[natcha12]

For ethernet you could still login to your router and see if your pc is on the connected list, if its there then its like the wifi example, "working" but with no visuals.

If you are able to boot up the pc, run Windows Memory Diagnostics, just search the start page. This is fully automatic and will show you the all clear/errors reported the next time you login.
If that fails/cannot finish, give it an hour if it seems stuck before thinking its failed, there's a good chance there is a very badly dead RAM onboard. You would then need to diagnose which RAM and/or motherboard slot is failing by trial and error. Again, Windows Memory Diagnostics is your friend here. Seeing as you get only infrequent crashes, that tool is more likely to find errors than just leaving your CAD on for an hour

Using only 1 RAM stick at a time, plug it in, try diagnostics, if it works, onto the next stick/slot. Rinse and repeat till you find the evil one

 

[Zizo007]

Test each RAM alone using 1 pass of memtest86+.

Is your RAM in the memory compatibility list of your motherboard?

 

[Calvin7]

i7 7820x supports DDR4-2666 Quad Channel memory.

What "32gb 3000" is, I wouldn't know. Maybe your the kind of guy who buys one stick now and further down the road picks up another stick. With no part numbers or link to it, it's anyones guess what your using.

 

[TallMan21]

i7 7820x supports DDR4-2666 Quad Channel memory.

What "32gb 3000" is, I wouldn't know. Maybe your the kind of guy who buys one stick now and further down the road picks up another stick. With no part numbers or link to it, it's anyones guess what your using.

I was mistaken. I had it in my mind that I had purchased DDR4-3000, but I did end up buying DDR4-2666 when I found it was the highest RAM frequency it supported.

 

[TallMan21]

I have run both windows memory diagnostics and memtest86+ on each individual ram stick/slot and ran memtest86+ for 6 full passes (22 hours) using all 4 sticks/slots at once. No errors detected. Storage tests went well, didn’t find any issues there.