AM4 B350/X370 black screen crash thread

[shalafi]

I am an utter and complete idiot.
The reason for my blackscreens was that the fan curve for the Wraith was too optimistic. Basically the cooler was running at 50% during benchmarks and temps were slowly creeping up until some thermal mojo locked up. Now the fun part - Tctl and CPU temp in HWinfo peaked at some non-scary value and didn't increase further. The only hint were some of the Temperature 3/4/5/6 readings in HWinfo. Depending on how I tinkered with case fans and whatnot, one or two of these temps would never peak and continue to increase until a blackscreen crash. Now that the Wraith is running full tilt at load, Temperature 6 peaks at around 62 in Prime95 Blend and the system is stable at 3800MHz, whereas before I couldn't even get 3650 to not crash.

 

[eyetrip]

So black screen crashes can also come down to vrm temps? Good to know.

Don't think that was an issue on mine. They don't seem to even have reached 60 from what I saw.

Interestingly, when I was having a quick play with a dirty OC, I had a black screen crash - but for the first time ever, I eventually restarted. Normally it would just sit there until I rebooted.

Wonder if Windows Update had any bearing on that...

 

[shalafi]

VRM along with DRAM and all the other temps in HWinfo were stable at some point into the benchmarks. The only one that was creeping up until it crashed was "Temperature 6". The thing is these temps are not even reported in any of the basic ASUS / AMD tools, not even the BIOS. It seems to be a sensor specific to the Crosshair VI Hero, for now we don't know what it relates to (the author of HWinfo is reaching out to ASUS).

 

[eyetrip]

Ah, right. Be interesting to know what it is.

 

[Wah007]

Anyone know how to find the VRM temperature on HWinfo for the Asus Prime X370 Pro?

I really wanted a motherboard with good VRM cooling as my previous AM3+ motherboard VRMs would start to go over a 100C even on stock clocks and undervolting. I really wanted a decent mobo this time I hope I don't have the same issue again, I did choose the Taichi and Gigabyte K7 at first but went with the Prime as they had stock.

 

[eyetrip]

Anyone know how to find the VRM temperature on HWinfo for the Asus Prime X370 Pro?

I think its those "temperature 3/4/5/6", etc temps. But not completely sure.

Dont think I have even seen those temps on mine go near 60 yet... which is why I am a bit skeptical they are VRM temps.

 

[Wah007]

I think its those "temperature 3/4/5/6", etc temps. But not completely sure.

Dont think I have even seen those temps on mine go near 60 yet... which is why I am a bit skeptical they are VRM temps.

Thanks.

I saw a post on Reddit I think where someone posted a link to a website where a person checked the VRM temperature of this board on his infrared heat thing, the temp was on 85C at 3.8ghz. Nothing was mentioned about the voltage, room temperature cooler etc.

 

[Xyz]

I know this thread hasn't been actively recently, but is there any solution for solving this problem?

1800x, C6H, latest bios
Custom monoblock cooling
17D-16GTZR RAM
HD7950 graphics

Stock settings at 3.6ghz and 2133 ram black screen almost immediately. Longest Aida stress time is 40 minutes by reducing 1.8vPLL. Can't remember the last time cinebench ran without crashing.

Running fresh instal of windows 10, chipset, amd drivers. Reinstall to no effect.

 

[gavinh87]

I know this thread hasn't been actively recently, but is there any solution for solving this problem?

1800x, C6H, latest bios
Custom monoblock cooling
17D-16GTZR RAM
HD7950 graphics

Stock settings at 3.6ghz and 2133 ram black screen almost immediately. Longest Aida stress time is 40 minutes by reducing 1.8vPLL. Can't remember the last time cinebench ran without crashing.

Running fresh instal of windows 10, chipset, amd drivers. Reinstall to no effect.

What code is the motherboard displaying when the crash happens?

 

[Xyz]

What code is the motherboard displaying when the crash happens?

"8"

 

[gavinh87]

Have you seen the problem with the ek gaskets? These were causing problems I wonder if your block has uses the same material....
Do you have any other coolers you can try?

 

[eyetrip]

I get these still myself.

When stress testing an oc under load, screen will go blank - system still keeps running fine. Can unplug hdmi and plug back in and get desktop.

That's using occt linpack avx for testing. So not likely same driver issue from before.

 

[gavinh87]

I get these still myself.

When stress testing an oc under load, screen will go blank - system still keeps running fine. Can unplug hdmi and plug back in and get desktop.

That's using occt linpack avx for testing. So not likely same driver issue from before.

This is what I've been calling an overclock crash, dialling the frequency back solves it for me.

 

[Xyz]

Have you seen the problem with the ek gaskets? These were causing problems I wonder if your block has uses the same material....
Do you have any other coolers you can try?

I've read about them for the EK Supremacy/EVO blocks? Haven't heard anything about the full monoblocks though. I'll look into it a bit more. The back is a blueish paper like material between the PCB and the back-plate.

I get these still myself.

When stress testing an oc under load, screen will go blank - system still keeps running fine. Can unplug hdmi and plug back in and get desktop.

That's using occt linpack avx for testing. So not likely same driver issue from before.

Hmm, no error code on your board though? Will definitely try a display unplug/replug when it happens again. Curious I never thought to do that before.

This is what I've been calling an overclock crash, dialing the frequency back solves it for me.

Normally that would make sense, but I can't run an 1800x at 3.6..which is stock. Would rather RMA it than run it at 3.5 (which does alleviate the issue).

I started maxing my fan/pump curve (100% by 57C) and can get through Cinebench now (three runs back to back), and at least 20 minute stress of Aida (cpu, fpu, cache, memory). Started to get through single runs of different pc benchmarks in Strada lite as well. Will have to load up realbench and occt tonight.

Maybe it is an OC issue. My question would be why though. Under water and using HWinfo (latest beta) max temps I've been able to see are: cpu 57, temp6 33, vrm 37, pch 44. No cause for alarm anywhere.

 

[gavinh87]

I've read about them for the EK Supremacy/EVO blocks? Haven't heard anything about the full monoblocks though. I'll look into it a bit more. The back is a blueish paper like material between the PCB and the back-plate.



Hmm, no error code on your board though? Will definitely try a display unplug/replug when it happens again. Curious I never thought to do that before.



Normally that would make sense, but I can't run an 1800x at 3.6..which is stock. Would rather RMA it than run it at 3.5 (which does alleviate the issue).

I started maxing my fan/pump curve (100% by 57C) and can get through Cinebench now (three runs back to back), and at least 20 minute stress of Aida (cpu, fpu, cache, memory). Started to get through single runs of different pc benchmarks in Strada lite as well. Will have to load up realbench and occt tonight.

Maybe it is an OC issue. My question would be why though. Under water and using HWinfo (latest beta) max temps I've been able to see are: cpu 57, temp6 33, vrm 37, pch 44. No cause for alarm anywhere.

Try taking the vcore voltage up a notch. It's possible you may have a bad sample and the board is undervolting or the leakage is high with your chip. If it works with a slight voltage bump then you are into the right track.

 

[Xyz]

Try taking the vcore voltage up a notch. It's possible you may have a bad sample and the board is undervolting or the leakage is high with your chip. If it works with a slight voltage bump then you are into the right track.

Should also note that I built this last week and have been messing with this for 3 solid days now. Ran with BIOS 1201 on Friday night until beta 9943 was released. Have been running with 9943 since then.

Everything on [Auto] in bios cranks the vcore up to 1.45/1.5 sometimes. When I manually set the multi at 36 I think vcore still went up to about 1.22 or so. I almost think the c6h throws way to much voltage everywhere. Initially when troubleshooting I had vcore 1.4 and SoC 1.15/1.2. Messed around with LLC and current settings a bit, but had nearly instant crashes. 'Out of the box' just 'never' worked on any bench/stress.

I think when i finished up last night i had 27.25 multi, vcore 1.2, soc .95, and PLL 1.81, but also only running 20 minutes in aida. Ran through one pass of CPU only y-cruncher as well.

Typically I'm just used to stressing for at least an hour and I'm only kind of worried what it means if I can't reach that. I feel like I haven't done my due diligence when others can manage hours of stressing just fine.

 

[gavinh87]

Should also note that I built this last week and have been messing with this for 3 solid days now. Ran with BIOS 1201 on Friday night until beta 9943 was released. Have been running with 9943 since then.

Everything on [Auto] in bios cranks the vcore up to 1.45/1.5 sometimes. When I manually set the multi at 36 I think vcore still went up to about 1.22 or so. I almost think the c6h throws way to much voltage everywhere. Initially when troubleshooting I had vcore 1.4 and SoC 1.15/1.2. Messed around with LLC and current settings a bit, but had nearly instant crashes. 'Out of the box' just 'never' worked on any bench/stress.

I think when i finished up last night i had 27.25 multi, vcore 1.2, soc .95, and PLL 1.81, but also only running 20 minutes in aida. Ran through one pass of CPU only y-cruncher as well.

Typically I'm just used to stressing for at least an hour and I'm only kind of worried what it means if I can't reach that. I feel like I haven't done my due diligence when others can manage hours of stressing just fine.

In that case something is certainly wrong. I'd be looking at RMA for some components. Code 8 could also be related to memory. Try running memory tests for a good few hours and try 1 stick at a time also. If they come back as good and still can't pass at stock I'd return the CPU but I think it's unlikely to be that. They either work or don't in my experience.

 

[Xyz]

In that case something is certainly wrong. I'd be looking at RMA for some components. Code 8 could also be related to memory. Try running memory tests for a good few hours and try 1 stick at a time also. If they come back as good and still can't pass at stock I'd return the CPU but I think it's unlikely to be that. They either work or don't in my experience.

I've run Memtest64 to 200% with no errors, though maybe I should let it run a lot longer. Checked with Thaiphoon burner and SPD checks out (OK).

Edit: Meminfo: G.Skill 17D-16GTZR (rgb) running at default 2133 and default dram voltage (1.25 I believe).

I think it was you who was considering a replacement CPU? Or was that someone else.
Edit: was Eyetrip. Did that replacement ever happen to see if CPU replace fixes the issue?

 

[gavinh87]

I've run Memtest64 to 200% with no errors, though maybe I should let it run a lot longer. Checked with Thaiphoon burner and SPD checks out (OK).

Edit: Meminfo: G.Skill 17D-16GTZR (rgb) running at default 2133 and default dram voltage (1.25 I believe).

I think it was you who was considering a replacement CPU? Or was that someone else.
Edit: was Eyetrip. Did that replacement ever happen to see if CPU replace fixes the issue?

I would be looking into sending that CPU back tbh.

 

[haszek]

I get these still myself.

When stress testing an oc under load, screen will go blank - system still keeps running fine. Can unplug hdmi and plug back in and get desktop.

That's using occt linpack avx for testing. So not likely same driver issue from before.

Yeah, I have them as well. It's strange, why would PC be loosing video signal while system is still running fine? Is it really reason of OC stability sign?

I have a lot of them while testing with new CH6 bioses. Mostly when testing memory HCI or RB