[SOLVED] PC randomly crashing, restarting, then fails to fully boot. Doesn't even make it to BIOS after self-restart. AMD Ryzen 5 3600X, ATI Radeon RX 580.

[Zavelin]

Hi folks! Mind helping me with some troubleshooting sleuthing? This issue has been going on for over 6 months, and it seems to be worsening.

System:

Operating System
Windows 10 Pro 64-bit

CPU
AMD Ryzen 5 3600X 62 °C (7/2/21: User "Skrybe" noted concerns that this temp is high for an idling computer. Currently using stock CPU cooler. Ordered Noctua CPU cooler on 7/2/21.)

RAM
16.0GB Dual-Channel Unknown @ 1499MHz (16-17-17-35)

Motherboard
ASUSTeK COMPUTER INC. TUF GAMING X570-PLUS (WI-FI) (AM4) 32 °C (7/1/21: User "artk2219" advised checking for motherboard BIOS updates. Found BIOS firmware was almost 2 years outdated. Updated BIOS from version 8/1/2019 to version 6/15/2021.)

Graphics
HP 24yh (1920x1080@60Hz)
ASUS VN248 (1920x1080@60Hz)
8192MB ATI Radeon RX 580 Series (Unknown) 35 °C

PSU
EVGA SuperNOVA 120-GP-0650-X1, 650 G+, 80 Plus Gold 650W

Audio
Realtek High Definition Audio


Issue: My custom-built AMD PC constantly does the following almost daily. Sometimes multiple times in a day.

1. Computer crashes and shuts itself down. No indication of overheating according to temp monitoring software. No BSOD. No error codes.
2. Computer boots back up. Only the computer tower lights up—attached peripherals are unresponsive and do not light up (backlit Razer keyboard, Anker mouse with color-changing lights, etc.)
3. Monitor powers on but only show black. System doesn't even make it to the BIOS after crash+reboot.

Attempted Solutions:

1. Replaced PSU.
2. Updated AMD drivers.
3. Rolled back AMD drivers.
4. Integrity-checked all hard drives with Hard Disk Sentinel. All HDs showing no errors or bad sectors.
5. Reseated RAM.
6. Tried running MemTest86 twice. Computer did the crash+reboot+black screen routine during both tries.

I appreciate any and all help, and will respond to every solution I'm advised! Not just for my sake, but for those that can use and learn from my PC issues to save time, money, and frustration. Topic Creators—don't abandon threads! Share what solutions worked for you. One user's problem thread can be a hundred users' solution thread.

 

[artk2219]

Hi folks! Mind helping me with some troubleshooting sleuthing? This issue has been going on for over 6 months, and it seems to be worsening.

System:

Operating System
Windows 10 Pro 64-bit
CPU
AMD Ryzen 5 3600X 62 °C
Matisse 7nm Technology
RAM
16.0GB Dual-Channel Unknown @ 1499MHz (16-17-17-35)
Motherboard
ASUSTeK COMPUTER INC. TUF GAMING X570-PLUS (WI-FI) (AM4) 32 °C
Graphics
HP 24yh (1920x1080@60Hz)
ASUS VN248 (1920x1080@60Hz)
8192MB ATI Radeon RX 580 Series (Unknown) 35 °C
PSU
EVGA SuperNOVA 120-GP-0650-X1, 650 G+, 80 Plus Gold 650W
Audio
Realtek High Definition Audio


Issue: My custom-built AMD PC constantly does the following almost daily. Sometimes multiple times in a day.

1. Computer crashes and shuts itself down. No indication of overheating according to temp monitoring software. No BSOD. No error codes.
2. Computer boots back up. Only the computer tower lights up—attached peripherals are unresponsive and do not light up (backlit Razer keyboard, Anker mouse with color-changing lights, etc.)
3. Monitor powers on but only show black. System doesn't even make it to the BIOS after crash+reboot.

Attempted Solutions:

1. Replaced PSU.
2. Updated AMD drivers.
3. Rolled back AMD drivers.
4. Integrity-checked all hard drives with Hard Disk Sentinel. All HDs showing no errors or bad sectors.
5. Reseated RAM.
6. Tried running MemTest86. Computer did the crash+reboot+black screen routing again. Tried twice.

I appreciate any and all help, and will respond to every solution I'm advised! Not just for my sake, but for those that can use and learn from my PC issues to save time, money, and frustration. Topic Creators—don't abandon threads! Share what solutions worked for you. One user's problem thread can be a hundred users' solution thread.

Potenitally the motherboard, a slight ram compatibility problem, a buggy gpu, or least likely, a problem with your cpu. If it crashed with memtest, im thinking its a memory compatibility problem most likely, but it still doesnt rule out mobo, cpu, and ram. You can rule out graphics by switching to a known good gpu that doesnt require external power if you have like a gt 710 or r5 240 or something laying around. Is everything running at stock clocks? Have you done any BIOS updates on the motherboard?

 

[Zavelin]

Potenitally the motherboard, a slight ram compatibility problem, a buggy gpu, or least likely, a problem with your cpu. Is everything running at stock clocks? Have you done any BIOS updates on the motherboard?

Hey, thanks for stopping by. Everything running stock, I don't dabble with overclocking and such. BIOS updates, not in a long time. Let me check and see if I have the latest.

 

[artk2219]

Hey, thanks for stopping by. Everything running stock, I don't dabble with overclocking and such. BIOS updates, not in a long time. Let me check and see if I have the latest.

Excellent and to simplify things i would pull out your gpu and go with a simple graphics card if you have something laying around, then check the sticks one at a time after youve updated the bios.

 

[Zavelin]

Potenitally the motherboard, a slight ram compatibility problem, a buggy gpu, or least likely, a problem with your cpu. If it crashed with memtest, im thinking its a memory compatibility problem most likely, but it still doesnt rule out mobo, cpu, and ram. You can rule out graphics by switching to a known good gpu that doesnt require external power if you have like a gt 710 or r5 240 or something laying around. Is everything running at stock clocks? Have you done any BIOS updates on the motherboard?

UPDATE: I'm on ASUS's driver page for my X570-PLUS motherboard. A whole load of BIOS and chipset updates that I've never touched. Nice call! Since the crash+reboot problem happens as often as multiple times a day, how about I install the new drivers, wait three days, and report back?

 

[artk2219]

UPDATE: I'm on ASUS's driver page for my X570-PLUS motherboard. A whole load of BIOS and chipset updates that I've never touched. Nice call! Since the crash+reboot problem happens as often as multiple times a day, how about I install the new drivers, wait three days, and report back?

Sounds like a plan, definitely update that BIOS, i know that TUF X570 was a bit flaky when it came out and Asus has done some good work on the BIOS for it, there have been like 6 BIOS updates for it since the end of last year.

 

[Zavelin]

Update #2: Turns out there's multiple versions of the X570-PLUS. A "WIFI" version and a non-WiFi. I was wondering why the ASUS EZ Flash utility was saying the BIOS update I downloaded was not compatible. Turns out I was trying to put a non-WIFI BIOS on my WIFI-enabled motherboard. Whoops!

Also, my current BIOS version date is 8/1/2019. Latest BIOS update:? 6/15/2021. Woooooow.

 

[Zavelin]

Update #3: The BIOS update went through fine. After restarting, my PC automatically started updating/reinstalling drivers for my mouse and keyboard. Then boom, the crash+reboot+black screen combo again. Should I give things more time settle?

 

[Skrybe]

I know you said it's not overheating but that CPU temp bothers me. Is it 62 degrees C at idle? If it is then you should probably look at removing the CPU cooler, replacing the CPU paste and seeing if you can get the temps down. If it's the stock AMD cooler you could consider getting a better aftermarket one too. I had random crashes with a 2700x using the stock cooler. Switched to a Noctua and the crashes disappeared.

Guessing from what you've said in the last post the crashes happen in Windows, not just games? So you could rule out a specific app/game misbehaving.

You said you tried a different PSU, what size/make was it?

 

[Zavelin]

I know you said it's not overheating but that CPU temp bothers me. Is it 62 degrees C at idle? If it is then you should probably look at removing the CPU cooler, replacing the CPU paste and seeing if you can get the temps down. If it's the stock AMD cooler you could consider getting a better aftermarket one too. I had random crashes with a 2700x using the stock cooler. Switched to a Noctua and the crashes disappeared.

Guessing from what you've said in the last post the crashes happen in Windows, not just games? So you could rule out a specific app/game misbehaving.

I'm going to take a cue from your advice and personal experience. I have a Noctua NH-D15 on the way now. I do not play PC games, but I do notice a higher rate of crashing when using more resource-intensive programs. The best example I currently have is Any Video Converter, a new video conversion program I just started using. The two or three times that I've tried to convert episodes of a show, the programs gets through a few episodes before my PC crashes.

The computer also refuses to fully boot up again until I've given it at least 5 to 10 minutes to cool off since its last crash. It'll power on, but nothing more than a black screen. Would that be a clue to overheating being an issue?

You said you tried a different PSU, what size/make was it?

I replaced it with the same one: EVGA SuperNOVA 120-GP-0650-X1, 650 G+, 80 Plus Gold 650W. Since it's apparently considered a good quality PSU, I RMA'd the original on the chance that I had gotten a faulty one.

 

[Skrybe]

I couldn't guarantee it, but that does sound like a heat problem to me. At least that's the first thing I'd be checking. Especially when you have to wait a few minutes before you can boot again.

If you haven't got quality thermal paste I'd get some of the Noctua NT-H1 paste. It's relatively cheap and works quite well. Lots of people recommend Thermal Grizzly Kryonaut but it feels like spreading melted cheese on your CPU. It works well, but it's super thick and stringy and hard to work with. If you've got decent thermal paste now you could even try redoing your CPU with that while you wait for the new cooler to show up.

As for the PSU; fair enough. The 650W from EVGA should be enough for that setup. Was just checking to make sure the other PSU wasn't something like a 350W cheapo PSU

 

[artk2219]

I couldn't guarantee it, but that does sound like a heat problem to me. At least that's the first thing I'd be checking. Especially when you have to wait a few minutes before you can boot again.

If you haven't got quality thermal paste I'd get some of the Noctua NT-H1 paste. It's relatively cheap and works quite well. Lots of people recommend Thermal Grizzly Kryonaut but it feels like spreading melted cheese on your CPU. It works well, but it's super thick and stringy and hard to work with. If you've got decent thermal paste now you could even try redoing your CPU with that while you wait for the new cooler to show up.

As for the PSU; fair enough. The 650W from EVGA should be enough for that setup. Was just checking to make sure the other PSU wasn't something like a 350W cheapo PSU

Yep Skrybe is right, it could be a heat problem if your cooler isnt mounted correctly, you could also try repasting and remounting your cpu cooler, also, have you run a memtest yet?

 

[Zavelin]

I couldn't guarantee it, but that does sound like a heat problem to me. At least that's the first thing I'd be checking. Especially when you have to wait a few minutes before you can boot again.

If you haven't got quality thermal paste I'd get some of the Noctua NT-H1 paste. It's relatively cheap and works quite well. Lots of people recommend Thermal Grizzly Kryonaut but it feels like spreading melted cheese on your CPU. It works well, but it's super thick and stringy and hard to work with. If you've got decent thermal paste now you could even try redoing your CPU with that while you wait for the new cooler to show up.

As for the PSU; fair enough. The 650W from EVGA should be enough for that setup. Was just checking to make sure the other PSU wasn't something like a 350W cheapo PSU

Funny thing: I had already planned to replace the thermal paste, and I ordered the NT-H1 paste about a month ago. Just hadn't gotten around to disassembling my build and redoing the paste application. It looks like my procrastination paid off—having to disassemble AGAIN to install the new cooler I ordered would have been annoying! I can't even remember what my current thermal paste is, probably something from the Arctic Silver series.

The cooler should be here in a few days. Thanks so much for your help so far.

 

[Zavelin]

Yep Skrybe is right, it could be a head problem if your cooler isnt mounted correctly, you could also try repasting and remounting your cpu cooler, also, have you run a memtest yet?

Yep, I'll go ahead and replace the stock cooler entirely. I would have thought that for a build that I don't even use for gaming, the stock CPU cooler would have been adequate. Phbbbbt.

Sorry, I wasn't clear in my initial post: I tried running memtest twice and the system crashed both times. Is running that program considered a high load activity? I've had consistent crashes when using other intensive programs, like the program I use for mass conversion of video files into other formats.

Thanks so much for your help, I'm looking forward to seeing how better cooling and 2 years worth of BIOS updates turn out. I'll keep everyone up to date.

 

[artk2219]

Yep, I'll go ahead and replace the stock cooler entirely. I would have thought that for a build that I don't even use for gaming, the stock CPU cooler would have been adequate. Phbbbbt.

Sorry, I wasn't clear in my initial post: I tried running memtest twice and the system crashed both times. Is running that program considered a high load activity? I've had consistent crashes when using other intensive programs, like the program I use mass conversion of video files into other formats.

Thanks so much for your help, I'm looking forward to seeing how better cooling and 2 years worth of BIOS updates turn out. I'll keep everyone up to date.

Sounds good, and if it was arctic silver 5, keep in mind that that stuff dries out, which in itself isn't the worst thing, but if it dries out and you bump the computer and break the seal that it made, then that can lead to bad contact and can cause a heating issue. Good luck though, let us know what you find!

 

[ohio_buckeye]

As far as Arctic Silver 5, I've been using that for years, and it may dry, but it doesn't seem to dry nearly like the white past that comes in prebuilt systems. Also, you should not be hitting your cooler or anything after you get it set.

 

[Zavelin]

I double-checked my purchase history. It was "ARCTIC MX-4," purchased in 2019.

 

[Danra]

ARCTIC MX-4 is a good thermal paste, not the very best but still good. Regarding AS 5, I worked in the computer hardware field for over 30 years and tested Arctic Silver 5, over time components would separate and cracks would form causing hot spots.

EVGA has not been making the best power supplies for the past few years. The EVGA SuperNOVA 120-GP-0650-X1 is known to have poor transient response and and inaccurate power good signal, still it should be good enough for mid range computer system like yours, I would not use it in a high end gaming rig though.

65 degrees C is much to hot for idle CPU temps and first on my list would be proper cooler placement, your thermal paste should be fine. NOTE: store TIM with the tip facing down, it contains an emulsifier that can separate and you don't want a lot of that on any surface. Regarding Skrybe's comment indicating Thermal Grizzly Kryonaut being more difficult to use is incorrect, in any event if you ordered a Noctua cooler it will come with a tube of good TIM, just remember to store it with the tip pointed down.

 

[ohio_buckeye]

I haven't experienced that with Silver 5, but I always did the method of manually spreading the paste over the chip. I know some preferred to apply and put the cooler down to squeeze the paste. But interesting.

 

[Skrybe]

@Danra maybe the Kryonaut I got was a bad batch but it was atrocious. It does it's job ok, but in my experience it's super hard to apply properly. Like I said it was literally like trying to smear melted cheese on the CPU. And yes, I verified that it was a genuine product using the code that came with it. If you google "Kryonaut tacky" you'll find a number of people having similar problems. On a different note, google "Kryonaut scratching" it's also not pleasant, though I haven't had that problem.

I'm sticking with the Noctua paste from now on.

@Zavelin I think the standard cooler is fine for basic computing that doesn't really stress the CPU, email, spreadsheet, browsing and so on. But if you're doing video encoding that's going to strain the CPU. Even when using the GPU encoders handbrake smashes the CPU for example. Memtest is high load, while it is focused more on the memory the CPU is part of the testing. From memtest themselves:
"Please be aware that not all errors reported by MemTest86 are due to bad memory. The test implicitly tests the CPU, L1 and L2 caches as well as the motherboard. "
https://www.memtest86.com/troubleshooting.htm

 

[Zavelin]

@Danra @artk2219 @Skrybe @ohio_buckeye

The Noctua cooler is here—CHRIST this thing is a monster. I hope it can fit in the case. I have the paste and cooler ready and unopened. Could someone recommend a good method for applying the paste and setting the cooler? Thanks!

 

[ohio_buckeye]

Place the paste on the cpu only, then attach the cooler.

 

[dorsai]

@Danra @artk2219 @Skrybe @ohio_buckeye

The Noctua cooler is here—CHRIST this thing is a monster. I hope it can fit in the case. I have the paste and cooler ready and unopened. Could someone recommend a good method for applying the paste and setting the cooler? Thanks!

I use a credit card cut down to the same size as the cpu die to spread out a thin even layer across the entire surface. With Ryzen employing separate chips under the heat spreader using the old pea sized drop method is no longer good enough in my opinion...others I'm sure will have their own opinon.

After reading through the thread I would look at system ram being the most likely culprit for the crashing. You've already swapped out the PSU if I followed along correctly, as well as updated your BIOS. The "unknown" ram is a giant question mark in my opinion and it's worth swapping out to name brand modules from Corsair or GSkill...some 3600mhz cas 16 ram would pair with the 3600x nicely.

At this point I feel it's also important to test the GPU using some demanding benchmark runs to see if it's stable. The RX 580 is getting long in the tooth and could also be failing.

 

[Skrybe]

Like dorsai said, small amount of paste on the CPU die and spread it smoothly in a thin layer. A lot of guides seem to just do the "pea" thing but I've always used the thinly spread layer method and it works.

From memory the Noctua's came with a little instruction booklet on how to install them. Noctua have setup videos on their site too: https://noctua.at/en/nh-d15/video The stock cooler uses a different mounting method so unfortunately you'll have to remove the mobo from the case and replace the CPU backplate with the noctua one.

Mine is set up with two fans, one at the front of the cooler pulling air into it and another in the gap in the centre of the cooler pulling the air through. And I have a case fan that sits right near the back of the cooler pulling air out of the case.

The fans themselves have a little directional arrow on them showing which way they move the air. Make sure they're blowing air through the heatsink and out of the case.

Let us know how that goes, it should drop your temps at idle down to high 30s or low 40s if you've done it right. Once you're seeing that it's time to run through your stability tests again and see if the problem is gone.

 

[Zavelin]

Like dorsai said, small amount of paste on the CPU die and spread it smoothly in a thin layer. A lot of guides seem to just do the "pea" thing but I've always used the thinly spread layer method and it works.

From memory the Noctua's came with a little instruction booklet on how to install them. Noctua have setup videos on their site too: https://noctua.at/en/nh-d15/video The stock cooler uses a different mounting method so unfortunately you'll have to remove the mobo from the case and replace the CPU backplate with the noctua one.

Mine is set up with two fans, one at the front of the cooler pulling air into it and another in the gap in the centre of the cooler pulling the air through. And I have a case fan that sits right near the back of the cooler pulling air out of the case.

The fans themselves have a little directional arrow on them showing which way they move the air. Make sure they're blowing air through the heatsink and out of the case.

Let us know how that goes, it should drop your temps at idle down to high 30s or low 40s if you've done it right. Once you're seeing that it's time to run through your stability tests again and see if the problem is gone.

Installed the Noctua cooler. New temps, I let the PC run overnight:


Operating System
Windows 10 Pro 64-bit
CPU
AMD Ryzen 5 3600X 34 °C
RAM
16.0GB Dual-Channel Unknown @ 1066MHz (15-15-15-36)
Motherboard
ASUSTeK COMPUTER INC. TUF GAMING X570-PLUS (WI-FI) (AM4) 31 °C
Graphics
8192MB ATI Radeon RX 580 Series (Unknown) 34 °C
Storage
1863GB Hitachi HUA723020ALA641 (SATA ) 24 °C
931GB Western Digital WDC WDS100T2B0A-00SM50 (SATA (SSD)) 27 °C
3726GB Western Digital WDC WD40EZRZ-22GXCB0 (SATA ) 27 °C
5589GB Western Digital WDC WD60EZRZ-00GZ5B1 (SATA ) 25 °C
1863GB ASMT ASM1156-PM SCSI Disk Device (USB (SATA) ) 27 °C
11176GB Western Digital WD My Book 25EE USB Device (USB (SATA) ) 48 °C
4657GB Western Digital WD Elements 2620 USB Device (USB (SATA) (SSD)) 30 °C
1862GB Western Digital WD My Passport 25E1 USB Device (USB (SATA) (SSD)) 30 °C
Optical Drives
HL-DT-ST BD-RE WH16NS40
Audio
CORSAIR VOID PRO Wireless Gaming Headset