Thank you for being a valued part of the CNET community. As of December 1, 2020, the forums are in read-only format. In early 2021, CNET Forums will no longer be available. We are grateful for the participation and advice you have provided to one another over the years.

Thanks,

CNET Support

Resolved Question

Help with weird crashing!

Feb 7, 2018 8:41AM PST

Hi!

I have experienced some electric whirring coming from the PSU earlier(and even messaged the xfx support about it, but since it was only occurring sometimes, I let it be.

Now it has become more constant, and I have also experienced constant crashes when playing more demanding games for example. The crashes only cut out the video feed to the monitor, the music continues and the PC is otherwise responsive. A manual restart works most of the time, but sometimes the resolution is ridiculously small and the picture only black&white. Uninstalling Nvidia-drivers solves this.

I have also tested the RAM, which had no problems. The MB has been updated and I have also uninstalled and reinstalled basically all the drivers I can think of numerous times. I have also physically checked the correct alignment of components inside the computer.

The crashes can't be replicated with a graphics-heavy testing program, at least not with furmark or kombustor. I also switched to the onboard graphics and the crashes stopped, but can I still be sure that it is the GTX 970? What if the PSU can now handle the lesser power requirement of the HD 4600 that I'm running now playing the games?

The PSU voltages seem to be fine when tested on programs, but then again I gather that that isn't a good way to measure them.

Can anyone help me, what should I do next?

Windows 10 Pro Version 10.0.16299 Build 16299
16GB (2x 8192MB) HyperX FURY Black DDR3-1866 DIMM CL10-11-10-32 Dual Kit
Intel i5 4690k
4GB MSI GeForce GTX 970 Gaming 4G
XFX XTR 80+ Gold PSU 650W
Crucial SSD 250GB
1000GB WD Blue WD10EZEX 7.200U
MSI H97M-G43 (bios updated to latest version)
No overclocks

Discussion is locked

ECLayla has chosen the best answer to their question. View answer
- Collapse -
Answer
I see the spambot ate your new post.
Jun 16, 2018 8:32AM PDT

The Speccy there shows some of the same troubles (CT250BX100SSD1 error counts) and we know these can lead to very hard to track down crashes.

Some of my other areas seem to still be there as well.
For the Speccy at UpyZLGKE6rpVXuRkXOUNrju

1. The boot drive is still showing errors that could do what you are complaining about. My advice is to get a clean drive in there.

2. The Network area is a mess and non-standard. This is likely due to the VPN but I'm not pointing fingers here. When you have machine troubles like this you head back to stock and see what stands out.

Here we worked a lot of areas but these two remain. If cleaning up these two and all malware scans are clean then it's some hardware issue like an aging motherboard which is only sniffed out by putting in a new one.

Sorry about the forum eating posts with Speccy links.

- Collapse -
Answer
Regarding malware scans. I get Grif's help on that.
Jun 16, 2018 8:32AM PDT
- Collapse -
Answer
Speccy
Jun 16, 2018 4:03PM PDT

Since you have an ssd why do you use hibernate?

This will lead to extra writes to the ssd and wear it out quicker.

Test with hibernate turned off and if acceptable locate the hiberfil file and delete it.

This will free up space on the ssd.

Stock ram speed for that cpu is 1600.

Your running at 1866......why is that?

Look in your power profile and set it to high perf,,,,,,,set everything to never.

- Collapse -
Hibernate
Jun 17, 2018 10:39AM PDT

Hi!

I looked in the settings and it says that hibernate is off already. Anyway just to be sure I turned it off via command prompt now.

As for the ram speed I had no idea that this was the case. So can I just change the speed to 1600 in bios? Or do I need new sticks?

- Collapse -
Good question.
Jun 17, 2018 10:48AM PDT

For the Speccy at UpyZLGKE6rpVXuRkXOUNrju we read:
1. DRAM Frequency: 933.5 MHz
2. From the memory stick itself: Max Bandwidth: PC3-14200 (889 MHz)

3. https://steamcommunity.com/discussions/forum/11/350541595110585473/ kicks this around and I agree with Bob__B that it looks like this stick is being overclocked.

Overclocking can result in very hard to pin down hangs, reboots and oddities.

As to what to do, I'd be sure the BIOS is current ( IT IS! https://www.msi.com/Motherboard/support/Z97M-G43 ) and then check the user manual on how to set the RAM speed down to what the RAM is reported to support.

- Collapse -
Okay, working forward
Jun 17, 2018 10:56AM PDT

Good. I ran the malware scans according to Grif's post. Nothing special found, only ad-cookies.

I'm currently working on cloning the affected drive to the other ssd with AOMEI Backupper.

After this I will see about changing the RAM speed.

I'll post another speccy afterwards.

- Collapse -
Thank you for
Jun 17, 2018 10:59AM PDT

Being patient and sticking to it. Usually when RAM is marked at a certain speed I've taken it as correct but WOW, this one appears to be overclocked and others are noticing.

This is thankfully rare but overclocking is as we know a source of troubles that can be very hard to pin down.

In your case there were a few other items which you cleaned up so it should be interesting if this last item puts a lid on it.

- Collapse -
Never would have gotten this far without the help
Jun 17, 2018 11:04AM PDT

Thank you for still trying with this problem. If these actions won't solve the problem, I'm thinking it might be time for a rebuild with possibly an ATX mobo and otherwise the same basic elements(except the iffy SSD). Of course I would rather not do this but if it is the only way..

- Collapse -
If I was to build today. Here's where I'd start.
Jun 17, 2018 11:14AM PDT
- Collapse -
Well that is better but..
Jun 18, 2018 1:31AM PDT

Ok so I changed the boot drive and the RAM speed. I did both of those at once so I'm not sure which helped but one of them certainly did.

The problem is that it didn't eliminate the problem. Where as the crash would occur about consistently every time in under 25 minutes in Caravan (game), now I could play it for hours with no trouble. Then I tried another game and disappointingly the crash occurred in about 15 minutes,

I do have the suspect SSD still attached but it has been formatted and the game that crashed isn't supposed to have anything to do with it.

What now?

- Collapse -
Better
Jun 18, 2018 1:54AM PDT

Post another speccy and wait for a M to pull it out of the garbage heap.

Disconnect that old ssd.

Connect your new ssd to sata port 0.

- Collapse -
Speccy
Jun 18, 2018 2:03AM PDT
- Collapse -
Sata port
Jun 19, 2018 12:29AM PDT

I have no idea how your mobo labels the ports.

You want your C drive in the lowest number/first port.

The speccy looks better now.

One minor thing in your power profile set your monitor and disk to power off never.

Your getting close to just changing parts.

Can I assume the innards of this machine are nice and clean and that you have given the psu a few shots of compressed air from both directions?

Have you tested this with the side panel off?

Have you tried going back one level on the gpu driver?

- Collapse -
Drive F
Jun 18, 2018 2:05AM PDT

Oh and I don't know what the newly appeared drive F might be. No memory sticks or anything connected, that just appeared after the cloning.

- Collapse -
from the very first question
Jun 17, 2018 1:33PM PDT

I'm surprised nobody has suggested a problem with the GPU. The symptoms you described indicated a power drain, maybe fan caused, making the problem. Video gone, but sound still working, strongly indicates GPU.

- Collapse -
Might be
Jun 18, 2018 1:34AM PDT

Hi!

This might be the case, but I have never gotten the crash in any kind of graphics benchmarks or heavy tasks, other than games. Also the games that have crashed haven't necessarily been very graphically intense at all.

How would one approach this line of thinking next?

- Collapse -
the only way
Jun 18, 2018 4:33AM PDT

Is using the onboard video which may not be good enough for your games, with the discrete video card removed, or using another video card to test it.

- Collapse -
Tried
Jun 18, 2018 11:04PM PDT

I tried that at the very start before posting on the forum. I could not replicate the error, but also wasn't able to play the same games. Maybe I could try that again.

- Collapse -
Answer
Done these steps, I assume sata port 0 is labeled C by defau
Jun 18, 2018 8:20AM PDT

"Done these steps, I assume sata port 0 is labeled C by default?"

Drive letters are assigned when the OS boots. There are articles about this but for this discussion I think you are nearing the end with many items covered but that overclocked RAM then we get to back up and think about other parts that have aged and now are giving up troubles.

- Collapse -
Frustrating
Jun 18, 2018 11:11PM PDT

So the crashes have been reduced but are still occurring quite regularly unfortunately.

Are there any other things that might give clues as to what is the source?

It feels wasteful (and expensive) to start a new build at this point, but then again I really don't know what else to do.

Again also the rig is only 3 years old, I wouldn't consider that a long period of time and certainly expected the parts that I really tried to research at the time to be of better quality. Then again, you never know of course and there just might be one defective part in the whole system.

- Collapse -
Test the RAM
Jun 19, 2018 1:45AM PDT

memtest86 can be used. Maybe a stick of RAM is going bad.

- Collapse -
This has been explored
Jun 19, 2018 2:47AM PDT

I already did this the first time this was happening unfortunately. Nothing wrong with the ram then.

- Collapse -
As long as the clocking to the RAM exceeds it's specs.
Jun 19, 2018 9:02AM PDT

All bets are off and I expect crashes from time to time. Why it worked for years is easy. Motherboards have caps that age and over time the noise on the voltages increases and then it's too much and you get crashes on a formerly fine system with not one change in settings.

I've shown this to folk where we have "identical" motherboards and one crashes and the other does not. I can't show this often due to cost and parts availability but aging is going to decimate some owners since they can't accept it.

So here we are, we have a very good reason for it to crash and history on why.

- Collapse -
I guess so
Jun 19, 2018 9:23AM PDT

It's hard to comprehend how this could be the case as the crashes are so predictable and still I have not gotten even one outside a game. But then again I don't really know or understand enough so that's that.

So maybe I should replace the motherboard and the RAM? I guess I could do another build basically around the GPU and the CPU. Then again probably for the best to update the mobo to a newer model so then I have to get a new CPU as well.

Or could I sell the whole system with big caveats and warnings maybe?

I appreciate all the help I've gotten here so so much!

- Collapse -
Here I would
Jun 19, 2018 9:30AM PDT

Be setting the BIOS to slow the RAM to what the RAM is reporting it's speed to be. But that's me with having to deal with too many PCs over the years.

We don't like changing motherboards since it's cost of the motherboard plus 150 labor and maybe more for an OS reinstall and if the OS is OEM or W10 you have to deal with the license.

Let me share that Microsoft has issues new KEYS if you know what to say. Which is "My motherboard died." This is why it's so much work that for me I'm changing memory speed in the BIOS because the work to change the motherboard is a PITW (pain in the wallet). Yes you could try it with one stick ram as a test too to see if a RAM change may work.

- Collapse -
Clarification
Jun 19, 2018 9:53AM PDT

The change of RAM speed that you mention is the same thing I've done isn't it? Well that did not help unfortunately. (As a non-native English speaker I'm not sure of your meaning concerning the ram speed Happy )

I could try only the one RAM stick though.

- Collapse -
Removing ram
Jun 19, 2018 9:54AM PDT

Do I need to do anything else than physically removing the stick? Change something in bios for example?

- Collapse -
Nope.
Jun 19, 2018 10:52AM PDT

Just unplug (power off and power unplugged of course) but no SPECCY REPORT has shown the RAM not being overclocked. So I have no proof you accomplished this change.

- Collapse -
Speccy
Jun 19, 2018 10:57AM PDT

RAM
16.0GB Dual-Channel DDR3 @ 799MHz (9-10-9-26)


That's from the latest speccy I posted, doesn't this mean that it isn't overclocked now?

- Collapse -
You dropped the MHz
Jun 19, 2018 11:09AM PDT

But the wait cycles were reduced so you may be back to the start.
16.0GB Dual-Channel DDR3 @ 933MHz (10-11-10-30)

Try 799MHz with 10-11-10-30. And with one stick RAM too.