Ryzen 7 black screen

[ToothlessFury]

Hello everyone,

 

I have recently purchased a new Ryzen based build.

And I would like to ask for your best guess or advise on what might be wrong.

 

The problem is:

I have been getting random black screen crashes. 

The screen goes black with "no signal". The CPU and case fans keep running, all software sounds stop. 

Holding power button or pressing reset does nothing, I have to turn the PSU off and on.

The system is in stock settings (no OC).

 

When does the problem occur:

• Randomly.
• Most of the times it happens when browsing the internet or at almost "idle" states. 
• I have found that it is more likely to happen after stress testing the system (very rarely it occurred while actually stress testing).
• It also seems that running monitoring software like hwinfo makes it more probable to occur.

I have tried:

• Updated bios to newest version (multiple times).
• Reinstalled windows.
• Replaced the motherboard (with the same model).
• Swapping RAM sticks (have not yet tried a new set of RAM).
• Tried not using any monitoring software (it seems, that this helps a little. I have had only 1 crash per week since not using hwinfo. Opposed to 1 crash every two days when using it. But it might be just a coincidence, because the problem does not occur evenly throughout the week).
• Using windows ryzen power plan.
• Increasing voltage of RAM and CPU manually
• UPDATE: Increasing the RAM voltage (problem still present)
• UPDATE: Increasing the CPU voltage (problem still present)
• UPDATE: Disconnecting and reconnecting all power cables (problem still present)
• UPDATE: Uninstalling AURA and AI suite (problem still present)
• UPDATE: Disable core boosting and SMT (problem still present)
• UPDATE: Setting DOCP on RAM (the lowest possible) (problem still present)

System info (everything stock, no OC):

• CPU: Ryzen 1700X
• MB: ASUS ROG STRIX X370-F
• RAM: Patriot Viper 16GB 3000Mhz CL16 (2 sticks)
• PSU: Corsair RM550X
• GPU: NVIDIA GT710
• SSD: Patriot flare 120GB SSD (System)
• HDD: Multiple WD blue/red HDDs
• CPU Cooler: Arctic freezer 33
• System: Windows 10 64 bit

The only thing I have changed in BIOS is enabling virtualization, because most of the time I have to work in my VM Linux environment.

I have no idea what could it be. My best guess is PSU or RAM incompatibility. But before I will buy a new PSU or RAM sticks I would like to ask for your help.

Thanks everyone for any advise. And sorry for my bad english.

[GOTSpectrum]

3 minutes ago, ToothlessFury said:

Hello everyone,

 

I have recently purchased a new Ryzen based build.

And I would like to ask for your best guess or advise on what might be wrong.

 

The problem is:

I have been getting random black screen crashes. 

The screen goes black with "no signal". The CPU and case fans keep running, all software sounds stop. 

Holding power button or pressing reset does nothing, I have to turn the PSU off and on.

The system is in stock settings (no OC).

 

When does the problem occur:

• Randomly.
• Most of the times it happens when browsing the internet or at almost "idle" states. 
• I have found that it is more likely to happen after stress testing the system (very rarely it occurred while actually stress testing).
• It also seems that running monitoring software like hwinfo makes it more probable to occur.

I have tried:

• Updated bios to newest version (multiple times).
• Reinstalled windows.
• Replaced the motherboard (with the same model).
• Swapping RAM sticks (have not yet tried a new set of RAM).
• Tried not using any monitoring software (it seems, that this helps a little. I have had only 1 crash per week since not using hwinfo. Opposed to 1 crash every two days when using it. But it might be just a coincidence, because the problem does not occur evenly throughout the week).
• Using windows ryzen power plan.

System info (everything stock):

• CPU: Ryzen 1700X
• MB: ASUS ROG STRIX X370-F
• RAM: Patriot Viper 16GB 3000Mhz CL16 (2 sticks)
• PSU: Corsair RM550X
• GPU: NVIDIA GT710
• SSD: Patriot flare 120GB SSD (System)
• HDD: Multiple WD blue/red HDDs
• CPU Cooler: Arctic freezer 33
• System: Windows 10 64 bit

The only thing I have changed in BIOS is enabling virtualization, because most of the time I have to work in my VM Linux environment.

I have no idea what could it be. My best guess is PSU or RAM incompatibility. But before I will buy a new PSU or RAM sticks I would like to ask for your help.

Thanks everyone for any advise. And sorry for my bad english.

Have you updated the BIOS on the Motherboard?

 

Have you tried a fresh install of windows? 

 

What build of windows are you currently running? 

[ToothlessFury]

21 minutes ago, Ben Quigley said:

Have you updated the BIOS on the Motherboard?

 

Have you tried a fresh install of windows? 

 

What build of windows are you currently running? 

Really?  

There is whole "section" dedicated to "i have tried" where i stated, that i have tried reinstalling the windows and updated the bios (multiple times when a new update was released, to the newest version).

 

Windows 10 build is 16299.192

[GOTSpectrum]

2 minutes ago, ToothlessFury said:

Really?  

There is whole "section" dedicated to "i have tried" where i stated, that i have tried reinstalling the windows and updated the bios (multiple times when a new update was released, to the newest version).

 

Windows 10 build is 16299.192

Shhh, I haven't slept in two days, Sorry about that mate. 

[GOTSpectrum]

6 minutes ago, ToothlessFury said:

Really?  

There is whole "section" dedicated to "i have tried" where i stated, that i have tried reinstalling the windows and updated the bios (multiple times when a new update was released, to the newest version).

 

Windows 10 build is 16299.192

I am thinking this could very much be a RAM issue. With my 1700 my first set up did this I replaced the RAM and motherboard at the same time. 

 

When exactly did you build the system? And were the components new or second hand? 

[MasterChromatica]

So the "Black Screen with everything else still running" is a failed OC. It just means that the OC set was not stable, try turning down the GHz or turning up the Voltage.

[GOTSpectrum]

2 minutes ago, StepinCurry said:

So the "Black Screen with everything else still running" is a failed OC. It just means that the OC set was not stable, try turning down the GHz or turning up the Voltage.

Hes not OCing anything in his system.

[ToothlessFury]

7 minutes ago, Ben Quigley said:

I am thinking this could very much be a RAM issue. With my 1700 my first set up did this I replaced the RAM and motherboard at the same time. 

 

When exactly did you build the system? And were the components new or second hand? 

Everything was new (no second hand) and the system is cca 2 months old.

 

My first guess was also RAM and second guess PSU. I will have a perfectly working PSU at my hands in a few days, so i will try it with that and then the next step is probably different RAM then.

[MasterChromatica]

So what I would then recommend is a 3.7GHz OC at 1.37500 volts, run a stress test. If it works great just leave it there. The black screen from Ryzen is how it handles bad OC, I know because I get that when I OC my R5 1600 to 4.0 it will crash midway through the Cinebench R15 test and the fans and RGB of the case will stay on, but the system is still unresponsive.

[GOTSpectrum]

1 minute ago, StepinCurry said:

So what I would then recommend is a 3.7GHz OC at 1.37500 volts, run a stress test. If it works great just leave it there. The black screen from Ryzen is how it handles bad OC, I know because I get that when I OC my R5 1600 to 4.0 it will crash midway through the Cinebench R15 test and the fans and RGB of the case will stay on, but the system is still unresponsive.

There Is No Overclock

[MasterChromatica]

Just now, Ben Quigley said:

There Is No Overclock

If you read what I wrote, I said "I would recommend a OC of 3.7GHz @ 1.375GHz"  That way we could tell if it is a Ram issue or just AMD setting weird voltage levels when set on "AUTO".

[GOTSpectrum]

Just now, StepinCurry said:

If you read what I wrote, I said "I would recommend a OC of 3.7GHz @ 1.375GHz"  That way we could tell if it is a Ram issue or just AMD setting weird voltage levels when set on "AUTO".

If you have stability issues with a system, generally, the last thing you want to be doing is adding in more factors that could add instability. 

 

@ToothlessFury One thing is, what CPU voltage are you running at in CPU-Z?

[MasterChromatica]

Just now, Ben Quigley said:

If you have stability issues with a system, generally, the last thing you want to be doing is adding in more factors that could add instability. 

 

@ToothlessFury One thing is, what CPU voltage are you running at in CPU-Z?

Well, what would you suggest Mr I changed my MB and Ram at the same time?

This is a solution that will isolate the CPU in the equation given. It will be a constant voltage instead of a fluctuating one. It will also have a constant GHz clock as well, That IMO would remove two factors instead of introduce new ones.

[ToothlessFury]

1 minute ago, Ben Quigley said:

If you have stability issues with a system, generally, the last thing you want to be doing is adding in more factors that could add instability. 

 

@ToothlessFury One thing is, what CPU voltage are you running at in CPU-Z?

Jumping from 1.243 to 1.439. Quite steady 1.22V when stress testing.

[GOTSpectrum]

2 minutes ago, StepinCurry said:

Well, what would you suggest Mr I changed my MB and Ram at the same time?

This is a solution that will isolate the CPU in the equation given. It will be a constant voltage instead of a fluctuating one. It will also have a constant GHz clock as well, That IMO would remove two factors instead of introduce new ones.

Rather than setting an OC, why not just set a voltage?

 

I didn't replace those to fix my issue, I did it to match the ascetic and it just happened to fix the issue.

[GOTSpectrum]

1 minute ago, ToothlessFury said:

Jumping from 1.243 to 1.439. Quite steady 1.22V when stress testing.

1.22 should we'll be stable, try 1.3 as that will definitely be stable.

[ToothlessFury]

6 minutes ago, Ben Quigley said:

1.22 should we'll be stable, try 1.3 as that will definitely be stable.

Ok i will try that later this day, when i have a little bit more free time. But the weirdest thing is, that it very very rarely happens while stress testing. It has happend only once while stress testing and it was after i was playing with the voltages. As i said in the original post, it mostly happens at "idle" or after stress testing. So i think the problem is somewhere else. But i will certainly try it. 

[GOTSpectrum]

2 minutes ago, ToothlessFury said:

Ok i will try that later this day, when i have a little bit more free time. But the weirdest thing is, that it very very rarely happens while stress testing. It has happend only once while stress testing and it was after i was playing with the voltages. As i said in the original post, it mostly happens at "idle" or after stress testing. So i think the problem is somewhere else. But i will certainly try it. 

Maybe the Auto setting is causing it to drop the voltage too low while you are idle? 

[ToothlessFury]

15 minutes ago, Ben Quigley said:

Maybe the Auto setting is causing it to drop the voltage too low while you are idle? 

That is possible, but it would have to be a CPU related problem then. Because two separate motherboards did the same thing.

i will try setting a stable 1.3 volts and disable core boosting and will see if it makes any difference.

[GOTSpectrum]

1 minute ago, ToothlessFury said:

That is possible, but it would have to be a CPU related problem then. Because two separate motherboards did the same thing.

i will try setting a stable 1.3 volts and disable core boosting and will see if it makes any difference.

Yes, in this situation it would be a CPU problem and i would RMA the chip personally. 

[MasterChromatica]

Soooo dont OC and set the Voltage to the aforementioned 1.3 By @Ben Quigley You will be fine.

[MasterChromatica]

7 minutes ago, ToothlessFury said:

That is possible, but it would have to be a CPU related problem then. Because two separate motherboards did the same thing.

i will try setting a stable 1.3 volts and disable core boosting and will see if it makes any difference.

Dont turn off Core boosting you will loose performance that way.

 

[ToothlessFury]

35 minutes ago, StepinCurry said:

Dont turn off Core boosting you will loose performance that way.

 

The thing is, that i want to isolate the problem to the CPU. I have already sent a MB to RMA and they straight up gave me a new one. So i want to be dead certain that it is the CPU before i go storming into the shop with the supposably "bad" CPU. 

 

And after i have a stable non-OC build with no faulty components I will want to OC the hell out of it.

[MasterChromatica]

Ok.

[ToothlessFury]

So i have tried setting the core ratio to 34 an Vcore to 1.3V (manual, no offset) and disabled core boosting. And the pc crashed into black screen while simply browsing the web. It passed 10 minute stress test a few minutes before that.

 

I have rebooted, run a 7 minute test. After stoping the test i went to browse the net, and it crashed within 2 minutes of browsing.

I am lost, is it a faulty PSU or CPU?

 

Tommorow i will send the CPU for RMA and will let you know.