Tried everything! Random shutdowns and black screens - Need help!

S

SoCalDJ

Limp Gawd
Joined
Jan 7, 2006
Messages
501
Wow it's been forever since I've been on here!


This past year I had both a videocard and some RAM die and I decided to completely redo my entire build.


I decided to go with an AMD chip this time after some reviews and bought a Ryzen 5 2600x. Here is the rest of the specs:


Asus ROG x470-F Mobo

GSkill TridentZ 3200 – 16gb

Gigabyte 1080gtx Windforce

Intel SSDPEKKW 512gb M.2 drive

1300w Seasonic Prime Gold PSU

2x HDD

1x Samsung Evo SSD


Cooling is a custom water loop consisting of:

EK x470 monoblock

EK Gigabyte GTX G1 fullcover block

EK XE 480mm x 60mm radiator

EK XE 240mm x 60mm radiator

EK XRES 140 Revo D5 Pump + Reservoir

Custom D5 Variable Pump + Reservoir

10x 120mm Corsair fans on the Rads (push and pull setup)

4 additional case fans to help with airflow and other components


So here’s my problem. I have random reboots and system freezes. Sometimes it’s while gaming, sometimes it’s just working on a spreadsheet in excel for work. No rhyme or reason, seems completely random.


Everything in the system is brand new except the older HDD/SSDs.


I’ve run Prime95 stress tests on the system for an hour to see if it’s the auto overclocks from the CPU.


I’ve run Furmark for an hour to stress the GPU to see if it’s unstable.


I recently replaced the mobo with a RMA from Asus thinking that it was something there.


I’ve monitored all temps and everything is in check with the watercooling. CPU doesn’t break 65c and vid card doesn’t break 50c. Mobo VRMs also stay pretty cool, much cooler than stock.


I’ve reinstalled Windows 10 twice one as a ‘refresh’ and one from scratch.

For the life of me I can’t figure out why this is happening. I don’t think it’s the CPU because of the stress testing, same for the GPU. New PSU should been perfectly fine on the rails especially one that is that much overkill for a single GPU setup. The only thing I can think of is that perhaps it has to do with the power that is being pulled from the wall being unstable? Or my power strip is dying for some reason?

I never had these issues with my old build (i7 setup with a 780 gtx). So it has to be something with the new one right?

I’m looking for ideas/thoughts/advice from the community here because I’m at a complete loss.
 
Try adding volts to the memory? Might have a bad stick. How about surface Scan of harddrives or removing drive individually.
 
Completely un-build. I mean unplug, unscrew and remove everything. Air blast out the case and all components. Rebuild and re-plug/seat everything.
 
My first suspect is memory if everything is at stock speed (memory running at rated xmp).
Try removing the xmp clock, leave it at 2133MHz and see if it becomes stable.
 
gigaxtreme1 said:
Try adding volts to the memory? Might have a bad stick. How about surface Scan of harddrives or removing drive individually.
Click to expand...

I've run mem tests and Prime SFFand LFF to stress the memory with no issues. I've ran it at the 3200 and manually set timings/voltage and at default for the board which is 2133 but I get the random crashes/reboots with all of them. This was the first thing I suspected too.

funkydmunky said:
Completely un-build. I mean unplug, unscrew and remove everything. Air blast out the case and all components. Rebuild and re-plug/seat everything.
Click to expand...

Just did this when I had to replace the mobo. No luck.

Nenu said:
My first suspect is memory if everything is at stock speed (memory running at rated xmp).
Try removing the xmp clock, leave it at 2133MHz and see if it becomes stable.
Click to expand...

At stock or rated speeds. I will try to increase the voltage but I don't think that's the issue because of the testing I had run prior.
 
1300W PSU damn time to add a few more 1080s! Joking aside I think it's either the PSU or CPU. Since you can't really load up that PSU very much you can still overvolt the CPU a bit and P95 it for a good 24hrs.
 
DDR4 RAM can pass tests most of the time and still be bad. If you have two sticks, try one stick at a time and use the computer normally. I wouldn't be surprised if one of the sticks if bad.
 
Furious_Styles said:
1300W PSU damn time to add a few more 1080s! Joking aside I think it's either the PSU or CPU. Since you can't really load up that PSU very much you can still overvolt the CPU a bit and P95 it for a good 24hrs.
Click to expand...

Yea I bought it so I'd hit about 50%-60% usage and keep the stress on the PSU low which would hopefully keep the rails tighter and caused less issues. It wasn't much more than a regular 800-900w PSU either. I don't think the PSU is the issue because I was also testing the system on a 1000w Thermaltake (10 years old though) and it was having the same issues. CPU I don't know what would be causing that either since if it was bad at the boost clocks it would fail Prime95 for sure under a full stress load.

ToastyX said:
DDR4 RAM can pass tests most of the time and still be bad. If you have two sticks, try one stick at a time and use the computer normally. I wouldn't be surprised if one of the sticks if bad.
Click to expand...

Agreed, although this has passed multiple tests and has been stress tested via Prime95 too for more memory focused form factors. I'll try running one stick at a time and see where that takes me but I doubt I can RMA them if I have no testing proof that they are bad.
 
SoCalDJ said:
Yea I bought it so I'd hit about 50%-60% usage and keep the stress on the PSU low which would hopefully keep the rails tighter and caused less issues. It wasn't much more than a regular 800-900w PSU either. I don't think the PSU is the issue because I was also testing the system on a 1000w Thermaltake (10 years old though) and it was having the same issues. CPU I don't know what would be causing that either since if it was bad at the boost clocks it would fail Prime95 for sure under a full stress load.



Agreed, although this has passed multiple tests and has been stress tested via Prime95 too for more memory focused form factors. I'll try running one stick at a time and see where that takes me but I doubt I can RMA them if I have no testing proof that they are bad.
Click to expand...

That 1kw TT PSU should only get about 50% utilization so it should be a good test to see if it's the SS. I agree with the memory re-check but honestly once the PSU is ruled out the last thing is the CPU.
 
Furious_Styles said:
That 1kw TT PSU should only get about 50% utilization so it should be a good test to see if it's the SS. I agree with the memory re-check but honestly once the PSU is ruled out the last thing is the CPU.
Click to expand...

Agreed but the SS replaced the TT after the system was built and it was happening with both so the SS can't be the cause unless BOTH PSUs are bad. That's why I was thinking it could potentially be the draw from the wall outlet or something crazy like that.
 
SoCalDJ said:
Agreed but the SS replaced the TT after the system was built and it was happening with both so the SS can't be the cause unless BOTH PSUs are bad. That's why I was thinking it could potentially be the draw from the wall outlet or something crazy like that.
Click to expand...

Wow that would truly be rare. You can try a different outlet if you think that's it I suppose. And there's a low chance you got another bad mobo from Asus but that's doubtful...
 
AP514 said:
My money is on.................. the 1 BAD stick of MEMORY
Click to expand...

DrDoU said:
Agree with this post
Click to expand...

Well I've got the manual timings and voltages setup again, this time back at 3200. Been out of town for work so haven't been able to test this weekend but will try to later this week.

Any suggestions on how to test it outside of Prime95 and memtest?

Chapeau said:
Cables? - I've has loose pins inside connectors without realizing.
Click to expand...

Shouldn't be the case since I just reseated/replaced the mobo and all the cables had to be reconnected.
 
Run Thaiphoon burner find out if the chips are hynix chips.Some hynix based chips get finicky. Latest BIOS? Some of the latest BIOS Ironed out memory problems.
 
Update: Last night I had the system locked up again under no load. I pulled out one of the mem sticks with the intent to test today if one was possibly bad.

This morning while just browsing the web the system rebooted for no reason (same issue as last night). So I swapped out the memory stick and will be hoping for the best today!

Both of these happened under little to no stress on the system.

gigaxtreme1 said:
Run Thaiphoon burner find out if the chips are hynix chips.Some hynix based chips get finicky. Latest BIOS? Some of the latest BIOS Ironed out memory problems.
Click to expand...

They are Hynix chips it appears. Hynix - H5AN8G8NAFR-TFC

I just replaced the board with one ASUS sent me but it looks like the bios is version 4602 which is the most recent on their website.

It's odd that the timestamp on it though is 3/7/2019 which on the download center correlates to version 4406 and 3/25/2019 correlates to the 4602 version.

Not sure if I should just try to flash and see if it updates the timestamp?

AP514 said:
Memtest86 has always rooted out my bad sticks....
Click to expand...

Same, but I guess it's best to rule everything out!
 
Well another update. After 2 days of just working on the PC (Excel, Chrome, etc) I was able to test both sticks one at a time in the system. Both eventually locked up or crashed.

One other odd thing I notice when this all happens is that the keyboard and mouse lose functionality, i.e. my Logitech G910 keyboard goes dark when it should be lit up. I also can't use the reset or on/off button on the mobo front panel to reboot or turn off the system.

Only two things remaining are CPU and wall outlet....guess I'll get an extension cord and test a different wall outlet to see if that's it....
 
SoCalDJ said:
I also can't use the reset or on/off button on the mobo front panel to reboot or turn off the system.
Click to expand...
The only things that can impact this are the motherboard or PSU.
My first suspect would be motherboard.

Make sure the CMOS battery is in good order, but I have never seen this affect the reset or power buttons.
They are unaffected by anything, a failure is extreme and rare.
Reset should be a pure hardware function that is unimpeachable.

Its possible mobo mfr(s) are doing things differently now such that a system crash can affect the mobo this badly but I have not heard the like of this ever.
I very much doubt this is the case.

edit
Hmm, I wonder if this is a low power state issue.
For many years now motherboards/cpus have been able to invoke a lower power state that older PSUs were unable to handle.
Older PSUs could be made to function by disabling the lower power states in the CMOS setup.
But I have not done this myself so it would be best if someone else could cover this with you.

However, I wouldnt expect this to affect the reset button unless the crash somehow made the PSU stop responding.
After all, there are now micro circuits in some PSUs looking after them.
I hope PSU micromanagement is not becoming part of the weird scheme of PC nuttiness.

ps its normal to lose keyboard and mouse control with a full system crash.
I always check if my caps lock still works if I think the system has crashed.
 
Last edited:
Nenu said:
The only things that can impact this are the motherboard or PSU.
My first suspect would be motherboard.

Make sure the CMOS battery is in good order, but I have never seen this affect the reset or power buttons.
They are unaffected by anything, a failure is extreme and rare.
Reset should be a pure hardware function that is unimpeachable.

Its possible mobo mfr(s) are doing things differently now such that a system crash can affect the mobo this badly but I have not heard the like of this ever.
I very much doubt this is the case.

edit
Hmm, I wonder if this is a low power state issue.
For many years now motherboards/cpus have been able to invoke a lower power state that older PSUs were unable to handle.
Older PSUs could be made to function by disabling the lower power states in the CMOS setup.
But I have not done this myself so it would be best if someone else could cover this with you.

However, I wouldnt expect this to affect the reset button unless the crash somehow made the PSU stop responding.
After all, there are now micro circuits in some PSUs looking after them.
I hope PSU micromanagement is not becoming part of the weird scheme of PC nuttiness.

ps its normal to lose keyboard and mouse control with a full system crash.
I always check if my caps lock still works if I think the system has crashed.
Click to expand...

Yea the PSU was just replaced so I don't think that's it. I'll double check the next time it locks up but I'm 99% sure the Caps Lock button doesn't work since all the power to the keyboard cuts off since the LEDs go out. Not sure I could even check that since there's no indicator. Mobo was just replaced too....I'm at my wits end as to why this brand new system is still crashing....
 
SoCalDJ said:
Yea the PSU was just replaced so I don't think that's it. I'll double check the next time it locks up but I'm 99% sure the Caps Lock button doesn't work since all the power to the keyboard cuts off since the LEDs go out. Not sure I could even check that since there's no indicator. Mobo was just replaced too....I'm at my wits end as to why this brand new system is still crashing....
Click to expand...
If you have a multimeter, when the system locks up again, check the 5V and 12V rails to the motherboard.
There are 2x 4pin plugs carrying +12V and the large motherboard connector which carries both +12V and +5V.
Check them while the system is still running, do not unplug the connectors.
We are looking to see that there is voltage within 5% tolerance.

Black = gnd/0V
Yellow = +12V
Red = +5V
Connect the meters black lead to gnd and the other to either red or yellow.
You are looking for 4.75V to 5.25V (red) and 11.4V to 12.6V (yellow).
And crucially the voltages should remain stable, ie they should not change/fluctuate at all.
This will only explain why reset doesnt work if the voltages are way way out of spec.
A fluctuating voltage wouldnt stop reset working but it will show a problem PSU regardless.

It would be helpful to check the same voltages while the system is running ok at idle (ie no programs running after a fresh boot).

If this shows a problem the PSU is at fault.
If it doesnt show a problem the PSU could still be at fault in other ways so cant be ruled out but being new this is less likely.
Which would mean the motherboard is most likely the problem.
However you must go through your mobo mfrs support, be guided by them to troubleshoot for an RMA.
***Although***
If they determine it is reasonable to send the mobo back, its better to send it back to the shop and get a new board from them.
Sending a mobo back to the mfr can result in getting the same mobo back (they find no fault) or getting another refurb that was previously sent back by someone else.

ps being new doesnt rule out any product when there is a serious problem.
 
Last edited:
Update - Tried running an industrial grade extension cord to another plug in another room (One that was on a different circuit in the house).....system still locked up.

Nenu said:
If you have a multimeter, when the system locks up again, check the 5V and 12V rails to the motherboard.
There are 2x 4pin plugs carrying +12V and the large motherboard connector which carries both +12V and +5V.
Check them while the system is still running, do not unplug the connectors.
We are looking to see that there is voltage within 5% tolerance.

Black = gnd/0V
Yellow = +12V
Red = +5V
Connect the meters black lead to gnd and the other to either red or yellow.
You are looking for 4.75V to 5.25V (red) and 11.4V to 12.6V (yellow).
And crucially the voltages should remain stable, ie they should not change/fluctuate at all.
This will only explain why reset doesnt work if the voltages are way way out of spec.
A fluctuating voltage wouldnt stop reset working but it will show a problem PSU regardless.

It would be helpful to check the same voltages while the system is running ok at idle (ie no programs running after a fresh boot).

If this shows a problem the PSU is at fault.
If it doesnt show a problem the PSU could still be at fault in other ways so cant be ruled out but being new this is less likely.
Which would mean the motherboard is most likely the problem.
However you must go through your mobo mfrs support, be guided by them to troubleshoot for an RMA.
***Although***
If they determine it is reasonable to send the mobo back, its better to send it back to the shop and get a new board from them.
Sending a mobo back to the mfr can result in getting the same mobo back (they find no fault) or getting another refurb that was previously sent back by someone else.

ps being new doesnt rule out any product when there is a serious problem.
Click to expand...

How would I read the voltages of those plugs if the system is still running? There's no place to stick a multi-meter into the backside of the cable and get a reading.

If the PSU is the cause, I find it strange that BOTH the old 1000w PSU and the new 1300w PSU have the exact same issues.

As for the motherboard I just replaced it last weekend. I had ASUS cross ship a new board (Though it could have been refurbished/repaired as that's what their policy says) but it's not the original board I had. So unless that board was also bad in the exact same way then I doubt that's it it either.

I'm seriously pulling my hair out here as I'm running out of options....:mad::rage::punch:
 
You can stick the probe in the top of the connectors unless the probes are too fat.
If the probe tip has a bobble end you can wrap a small length of wire round it to stick in the connector.

I feel your pain with the problem.
Having replaced everything and the problem remains perhaps its down to how the motherboard is mounted.
ie there could be a short circuit. It certainly fits with the type of problem you are having.
Check to see if the same happens with the motherboard outside the case, mounted on some books for example.
 
Nenu said:
You can stick the probe in the top of the connectors unless the probes are too fat.
If the probe tip has a bobble end you can wrap a small length of wire round it to stick in the connector.

I feel your pain with the problem.
Having replaced everything and the problem remains perhaps its down to how the motherboard is mounted.
ie there could be a short circuit. It certainly fits with the type of problem you are having.
Check to see if the same happens with the motherboard outside the case, mounted on some books for example.
Click to expand...
Well today I bought a new case, 2 DIFFERENT types of RAM (3200 same as I currently have, and 3000 same brand and listed as compatible with the mobo), and a new Ryzen 2600x. I plan to try and figure it out before Disneyland on Sunday :)

On a side note I was just trying to stress the system and try to trigger a crash with Prime95 while having a conversation with the wife. I put in the new 3200 sticks and ran a blend test. Cores 3 AND 4 failed after 7 mins. Going to try a small and large FFT and see what happens.
 
Ok I've continued testing....down-clocked the new memory sticks and running LFF. I'm assuming this is going to pass. If it does then that would mean the it's something to do with the memory, but having changed out 2 sets of sticks I'm starting to lean on it potentially being the on-board CPU memory controller? Anyone ever had this happen before?
 
So it passed for 30+ mins @ ~2128 mhz or 'auto' on the MB under LFF. It's got to be memory related. I'm doubting it's the actual sticks though....
 
Are you running any RGB lighting software? Try disabling it. I've heard some MBs have trouble with some brands of RGB software such as Corsair.
 
Update - Just finished rebuilding into a new case (Moiught as well if I'ms tripping everything out again because of CPU testing.

I put a new Ryzen 2600x in there and I'm hoping there will be some resolution this week with some additional memory testing with this new CPU. Here's to hoping it's the IMC! :D

Azrak said:
Are you running any RGB lighting software? Try disabling it. I've heard some MBs have trouble with some brands of RGB software such as Corsair.
Click to expand...

Just the Asus software from my Asus board. I can stop it but not sure it would really be the culprit.
 
So I've been continuing to troubleshoot this issue. I've now tested 3 sets of memory all SK Hynix ICs.

I was unable to get any of them to remain stable even after changing the CPU.

Trident Z RBGs 3200 and 3000 CAS 16. I even tried pushing 1.4v on the memory. All of them either black screen crashed or rebooted.

I'm starting to doubt it's the memory OR the RAM now....but I have some Corsair memory that should arrive tonight for additional testing, including what I believe is Samsug B-Die ICs.

While I had downtime during all of these tests I stumbled onto another possible cause relating to there being multiple applications accessing the sensors.....

https://rog.asus.com/forum/showthre...-AI-Suite-3-s-IOMap64-sys&p=734622#post734622
https://rog.asus.com/forum/showthre...and-Ryzen-2700x-Random-shutdown-s-anyone-else
https://rog.asus.com/forum/showthre...m-and-weird-black-screen&highlight=x470+b-die

And while some of these are different configurations and setups, and some are BSOD I'm going to uninstall AURA and HWInfo and leave AI Suite for testing. I'll try removing AI Suite next and see if that fixes it....
 
Throw some cash at it and get a low end board and see if that's the problem. Good to have backups in case you have to RMA either the processor or the board. Some times you get a bad board.
 
gigaxtreme1 said:
Throw some cash at it and get a low end board and see if that's the problem. Good to have backups in case you have to RMA either the processor or the board. Some times you get a bad board.
Click to expand...

Already replaced the board via RMA with Asus. Same issues from their new board and my original.
 
gigaxtreme1 said:
Then you may have a bad IMC. RMA the processor.
Click to expand...

Already swapped it for a new one.

Update: Looks like I may have finally figured it out. I uninstalled ASUS AI Suite III, HWInfo, and ASUS Aura Sync. Started testing with Prime95 again last night, 1 hour of LFF passed, then another 2! Later in the evening I started a blend test in P95 and it ran overnight with no errors for over 12 hours!

While I'm not going to hold my breath, I can't believe this may be a poor motherboard design. Apparently it is a known issue with AMD ASUS boards (Though most of the cases were from the beginning of 2018 and thus why I didn't find the issue since I assumed it was fixed at this point) that when there are two different programs trying to access the on-board temp sensors it causes random system lock ups and crashes. Inf fact many people described my EXACT same issue.

I guess we'll see what happens!

I'm incredibly surprised that when I called ASUS to RMA the board they didn't have someone who would have known about this issues and suggested a fix....
 
DrDoU said:
Maybe ASUS should pay you.
Click to expand...

I wish! Man that process was a total PITA. It took nearly 3 months to get them to finally not only HAVE a board to advance ship but also convince them to do it in the first place....

Been a full day and no black screen yet! :nailbiting:
 
Final update - After multiple tests and another overnight P95 run it appears that this has been resolved. I have now tested it at 3200 mhz on the ram timings and 3000 mhz to be safe but both pass. The last thing is for me to lose out AI Suite and see if it will adjust the fans still based on the settings that were input earlier so that I can run HWInfo. It's at least nice to know that this hell is finally over!
 
You must log in or register to reply here.
Back
Top