Tower PC freezing on movie playback - graphics card issue?

Hi guys.
I've an old (+7 years) medion tower PC that functions primarily as a server for streaming my plex stuff to my smart TV upstairs.
A few years ago, I got a BSOD, figured it was a graphics card issue and replaced it with a Nvidia cheapy one.
All was fine..
But recently, If I use it to watch any content directly - Plex, media player, streaming services etc, it will intermittently fall over and give me a kernel driver error.
Worst case, it freezes and I have to hard reboot.

I'm running a very clean, recent install of Win 7 and after a lot of research I believe I'm using the most stable driver version, for win7 and my graphics card.

Would replacing the graphics card fix it?
I'm using a 40" sammy, via HDMI as a monitor, would that cause any issues?

Appreciate the help :)

when was the last time you give the pc a clean, a real clean of dust and crap that maybe inside your PC. Maybe its over heating some where due to dust

I had the casing off probably a year ago - gave it a good vacuuming and compressed air combo.
Not a bad point to start at though - thanks Mickey :)

When I was using regular pcs As home servers I had all sorts go. Psu's (caps), gfx card, ram, literally any fucking thing.

Despite the 'load' being low, they just aren't able to hold up compared to a proper server.

Just my 2p


DJ OD

I agree, it's not a bad idea to rule out thermal issues, especially with a machine that's done a few years.

Clean out all heatsinks, change CPU paste (and GPU if you feel up to the job). Also check heatsink mountings, seen a few over the years where heatsink is no longer firmly clamped to the device it is cooling.

Next from there, if possible try to strip out as much as possible (e.g DIMMs, PCI cards, 2nd/3rd/etc hard drives) so you are running your system with the minimal kit you need to work (even disable things like RS232, LPT, additional USB ports in BIOS if possible/relevant)

Ideas being to eliminate stuff you don't actually need o let the PC work to see if that helps. If it does, then reintroduce parts 1 by 1 until issue comes back.

If that doesn't yield results, next I would be using prime95 and memtest on various settings to torture test CPU/RAM/chipset.

If that doesn't work, time to swap parts/drivers until you yield a result.

Or the other way, download the Windows SDK , don't install all, but install debugging tools for Windows, set this up to get symbols from the Microsoft server as required and analyse your crash dumps. I don't treat analysis of 1 crash dump as gospel, but if I have 5-10 crash dumps from a machine and 4 out of 5 seem to suggest the same problem then you probably are onto something.

Or the real lazy way, post your crash dumps and me (or others) can analyse and give you feedback.

As above make sure its clean and not over heating. Failing that I would plug for GPU. Not got a cheap as shit one lying around to test the theory before purchase ???

Also could be worth a wipe / reinstall before spending out.

Also do you have spare RAM to swap out with, never mind how shit it is .. just eliminates another thing.

In truth, a lot of what you guys have said, is way over my head and beyond my abilities.
I guess I was looking for a simple answer I could bung a few quid at to fix.

In terms of using it as a server, It's only direct playback that causes problems.
using it to stream either via file share to my notebook, or through Plex to my smart TV has never caused a problem

Look in C:\windows\minidump and see if you have any *.dmp files there (these are the crash dumps I referred to earlier).

If you have files, post the recent ones. If you don't, post back and I will post instructions so your computer will save crash dumps when it crashes in the future.

Then I can (or others on here can) run them through a tool which will suggest what is causing your issue.

From what you say, my money's on the graphics card, but I don't want to tell you to fork out on a new one if we haven't got more info to solidly point more fingers at the gfx card.

Ok, so I have no minidump folder
did a bit of googling and I'm hopeful I've now enabled a dump folder, but I'm guessing one won't appear until it crashes again?
If so, its a wait and see moment: i'll start playing vids ans see if it generates one.

I'll be back just as soon as the fucker crashes!

thanks guys :)

Yep, if dumps weren't enabled, now we need to wait until it crashes and then we can look at the dumps we hopefully should have.

Once you have 5-10 dumps, post them up and I will happily take a look.

If it's not saving dumps, let us know how you enabled dumps and I can see if I can give any ideas on why it's not saving dumps.

First step. Buy can of air !!! Maplin, Staples, PC world etc. About a fiver, probably cheaper online. Open PC and blast the shit out of it with can of air then hoover up the fallout.

DO NOT hoover up inside the PC. have fixed so many PC's when bellends have stuck the hoover in the case and fucked up tiny components !!

You say most of this is above you but you refit a GPU at some point so don't sell yourself short. Not a lot more complicated than that other than new MOBO.

Just be gentle and don't forget to talk to it all the while.

Quote:

---

Originally Posted by Over Carl

Yep, if dumps weren't enabled, now we need to wait until it crashes and then we can look at the dumps we hopefully should have.

Once you have 5-10 dumps, post them up and I will happily take a look.

If it's not saving dumps, let us know how you enabled dumps and I can see if I can give any ideas on why it's not saving dumps.

---

I enabled dumps by doing this:

To ensure minidumps are enabled:
Go to Start, in the Search Box type: sysdm.cpl, press Enter.
Under the Advanced tab, click on the Startup and Recovery Settings... button.
Ensure that Automatically restart is unchecked.
Under the Write Debugging Information header select Small memory dump (256 kB) in the dropdown box (the 256kb varies).
Ensure that the Small Dump Directory is listed as %systemroot%\Minidump.
OK your way out.
Reboot if changes have been made.

The .dmp files are located at C:\Windows\Minidump. Until a .dmp file is generated, the Minidump folder may not exist.

I've rebooted, and enabled the show hidden files and folders option, so on the premise that the above works, it's a wait and see moment :)

I wouldn't select small memory dump as they contain limited info, if you are expecting a BSOD and want to diagnose it, temporarily set it as full/kernel memory dump (and change it back after the issue is resolved). If you're lucky then a free program called WhoCrashed will tell you exactly what is causing the dumps.

okay, so it's been doing it's crashy thing for a couple of days.
Still no dump folder, but having a mooch I've got a C:/Windows/LiveKernelReports/WATCHDOG folder with loads of dumps, which looks to correlate against the times of crashing so:

A. Are these relevant?
B. Can I upload the dmp files directly here as attachments, or
C. Do I need to find a program to read them, (recommendations??) then cut/paste the info into a reply?

Thanks chaps :)

My guess is the PSU has gone out of tolerance. As its a clean install, and thermally all good & within temps, i would look at the PSU.

Sorry looks like my skills are showing their age, seems like Win8/Win10 may save the dumps there if it's not a full BSOD crash and the systems keeps running, while BSOD minidumps still go to c:\windows\minidump.

If you can post the files I can take a look, but you may need to load them onto something like dropbox.

If you have a whole mess of files and aren't sure which are relevant just post a screenshot of the lot.

Quote:

---

Originally Posted by I Black Belt

okay, so it's been doing it's crashy thing for a couple of days.
Still no dump folder, but having a mooch I've got a C:/Windows/LiveKernelReports/WATCHDOG folder with loads of dumps, which looks to correlate against the times of crashing so:

A. Are these relevant?
B. Can I upload the dmp files directly here as attachments, or
C. Do I need to find a program to read them, (recommendations??) then cut/paste the info into a reply?

Thanks chaps :)

---

That's interesting, can't say I've noticed that before but when I expect there to be a dump but can't find it I tend to go to %windir% and search *.dmp in the top right search box but this never brought that location up.

Download WhoCrashed for free and try to open the directory into it and see if it helps, its not a full debug but occasionally flags up a specific driver or helps work out the issue.

Check out your PSU. Some PSUs do not give a very clean +12v on the rail which can cause kernel panics. I used to get intermittent crashes on an old PC and it turned out that my gfx card (ATI) was picky about the voltage and needed AT LEAST 12v. I opted for a new PSU in the end that all the reviews stated gave a solid voltage that didn't fluctuate. Problem went away :)

So it's worth checking the graphics card requirements (wattage and any known issues with PSUs).