You don't need to be an 'investor' to invest in Singletrack: 6 days left: 95% of target - Find out more
Following on from my earlier thread, after a lot of good advice and information form the folks on here, I bought a new SSD and upgraded to Win 10 Pro from Win 7 Pro.
Everything appeared good, really quick to boot up and all working fine.
However, I have started to have quite regular system crashes. It normally starts with an error in Chrome of being to open a page, the "Aw Snap" thing. Then I might get some insufficient memory errors from Windows or Display Driver failure and my 2nd screen going down.
I have updated all my drivers and Windows.
I did a bit of Google and have tried the System File checker thing but it found no errors and the crashing has continued.
Could it be a hardware issue? The new drive is a 1TB SSD, it seems to run hotter than the previous 250GB SSD. Is there some way to cool it? Paste or a heat sink?
Any ideas?
If it had been a mechanical hard drive I'd be recommending running a diagnostic on that, but generally SSD's work or they fail completely, I'd think about running memtest+ or something similar to test the RAM.
I’d think about running memtest+ or something similar to test the RAM.
As would I, that's highly symptomatic of RAM being on the wonk. As you've been rummaging around inside I'd double-check the DIMMs are firmly seated also (I'd expect a DIMM half-hanging out to either work or not boot at all, but it can't hurt to check).
If you don't need it any longer, you could try refitting the 250 and doing a clean W10 install onto that. That'd indicate whether it was a faulty drive or a faulty something else.
Is there some way to cool it?
Assuming a desktop, you can (or at least, used to be able to) get adapters which mount drives in a 5.25" bay with additional fans. It's not something I've seen in decades though and they were kinda gimmicky to start with.
It does rather beg the question as to why you should though. A new drive shouldn't be getting so hot that it melts its own brains out unless you've mounted it by stuffing it in with cotton wool padding. That's not normal, it's faulty.
Thermal issues usually present themselves when the system has been on for a while and get progressively worse until it's allowed to cool down again. I've never come across this issue with an SSD before (and I've no idea how long it takes for one to heat up) but I'd be looking for that sort of pattern in the crashes.
Random thought - you did do a clean install, yes? It's not a clone from the old SSD?
I’ve put ssds in some terrible places (I’m a roboticist) and they tend to fail completely as above, rather than go dodgy. Exception being dodgy cables or cables fitting.
Sounds like ram or failing that a bad data connection. I’d be reseating everything then memtest overnight and a long disk test too.
I’ve put ssds in some terrible places
For a brief but horrifying moment there I wondered where the hell you were going with that...
I'm not surprised they are unstable if you have 10 widows and they have been 'upgraded' for a new one ? 🙂
It was a clean install.
I ran memtest last night and it found no errors.
I have already checked the physical connection of the drive and it appears fine.
I will have a look at the RAM later today and see if that's the issues.
Memtest will have tested the RAM, while it's a good test it doesn't fully stress RAM like a Windows environment can do, if there are two sticks of RAM try running one at a time and swapping over. That said it's perhaps more likely that it is a dodgy SSD especially if it's running very hot.
Very unstable Widows
Perhaps bereavement counselling?
It does sound like a hardware error. If you're memory is good might it be gfx card related (as that seems to be what's choking)? Perhaps try re-seating the card.
Does it always start with a Chrome error? Occasionally software errors can lead to hardware errors. Chrome has a hardware acceleration toggle in its setting and you could try running it toggled off to see if that helps (equally the new Chromium based Microsoft Edge browser is available and seems very good).
Would there be any indicators in the Event Viewer logs?
This could be a bug in the graphics driver. I had similar problems with an early AMD driver on my Ryzen laptop.
I know you said you updated all drivers, but as a first step I would first try uninstalling the graphics driver and if you were using the official amd/intel/nvidia ones, let it use the ones from Windows Update (they will appear in the Device Manager window as being published by Microsoft on the Driver tab). If however you previously let Windows Update install them, I would get the driver directly from the manufacturer's site.
Could it be a hardware issue? The new drive is a 1TB SSD, it seems to run hotter than the previous 250GB SSD
Seems to? I've a couple of SSDs, a Samsung and a SanDisk. Both came with health monitoring software that reads the temperature, have you got anything like that? Mine have been running trouble free for years, typically at 30-35C - are yours significantly hotter?
why don't you try not using Chrome for a while, and see if it is more stable?
why don’t you try not using Chrome for a while, and see if it is more stable?
I did. Started using edge but it still crashed.
Mine have been running trouble free for years, typically at 30-35C – are yours significantly hotter?
I have installed Open hardware monitor and they are 30C. I will keep an eye on them.
Would there be any indicators in the Event Viewer logs?
This is what I have found from a crash this morning. Although I restarted the PC as soon as Chrome played up as it was always just before the display driver fails.
Doesn't mean anything to me?
chrome.exe
79.0.3945.130
5e1f9839
KERNELBASE.dll
10.0.18362.535
50cc8d5a
e0000008
000000000003a839
2ad8
01d5dacaad8a2e15
C:\Program Files (x86)\Google\Chrome\Application\chrome.exe
C:\Windows\System32\KERNELBASE.dll
0eb7ea35-e847-4fd8-9512-346a7af2b21d
Thats a memory leak or something to do with addressing from storage.
50cc8d5a
e0000008
000000000003a839
2ad8
01d5dacaad8a2e15
These are memory/storage block allocations.
0eb7ea35-e847-4fd8-9512-346a7af2b21d
This is a registry key perhaps? not totally sure.
Looks like something to do with memory and/or storage.
Easy option is to remove a stick of RAM and give it a go for a bit. If it crashes try the other stick. If still crashes remove your GPU and run off the integrated graphics. If it still crashed your SSD is the answer.
Although, as above - SSD's tend to work or don't work. It could have been a dodgy install or corrupt driver install that is causing the memory leak - a format and reinstall of everything may fix it. You could try just removing the GPU driver first (there are cleaning tools out there also that'll fully remove it) and reinstalling that if it seems to be related to visual output.
A bit more digging and I found this one yesterday.
EventData
Library C:\Windows\System32\perfts.dll
Service LSM
BufferSize 27896
RequiredSize 33224
This led me to this page Answers MS
So I ran the tools on the page through the command prompt and I will see if it fixes it.