New Build - Crashing & Corrupt Graphics

Jinxter

Jinxter

Jinxter

Associate
Joined
24 May 2009
Posts
322
Location
Southampton
Hi,

I built a new PC which has been running OK for the last couple of months however over the last couple of days it has intermittently stalled every couple of hours where the screen will either go blank or exhibit black/white vertical lines and once green / red vertical. It has been hotter in the room where it is due to weather so I was wondering whether this might be something to do with it. I have overclocked in past but not nothing too strenuous. Here's my spec...

ASUS M3A78 PRO - 780G onboard graphics (oc'd to 650mhz)
AMD 7750BE 2.7GHZ (oc'd to less than 2800mhz on 1.25vcore) using standard HSF
OCZ REAPER 4GB 1066MHZ (oc'd to 1099mhz)
WD 320GB SATA HD (very recently changed from IDE to AHCI)
DVDRW
Coolermaster 330 case with CM 460w PSU (includes one exhaust 120mm fan at rear)

While I've had a dabble with overclocking I haven't really pushed the hardware to what I would consider dangerous, it actually crashes past 2800mhz most of the time and the onboard graphics I've researched can successfully be pushed to around 800mhz mark but never got that far. I've done some short runs of PRIME 95 and MEM86 and it's not actually crashed or shown problems.

When pushed core temps are around 30degree mark (51degree on the fan control screen of AMD Overdrive)

My initial thoughts are that this is a hardware problem relating to the onboard graphics and might be my chipset overheating as result of the recent hotter weather. The problem has also happened at stock speeds to suggest OCing might not necessarily be the issue here.

I would like to know whether this sort of problem might indicate possible hardware damage or is just purely because it has been hotter weather and it's just safely locked up before damage can be done. I've ordered another 12" fan for the front of the case incase this will help airflow through the case. As I type it has been cooler today and the PC has been on and not crashed for past five hours.

Lastly, can anyone recommend a CPU cooler for my board which will clear the ram slots as the reaper memory has heatsinks making them pretty tall.

Any opinions / suggestions greatly appreciated.
 
First - welocme to the forums.

Second, you list the current spec as clocked -
(oc'd to 650mhz)
Click to expand...
- have you tried running the rig at stock settings and checked to see if you get the same problems?

If you try it and you don't get the same problems then it's probably a heat issue as you suspect.
 
Last edited:
Thanks for the reply Plec - yes have tried at stock settings and same problem did occur, still running stock at mo and everything fine so far today so fingers crossed it is just heat and the extra fan on order might help (hopefully) !
 
Yeah sounds like heat issue, try running with the side panel off so your case gets more air, it does sound like gfx rather than cpu which is overheating too, you could also try a fan pointed in, if you get no problems doing this then its definately heat, you could buy a better case if you arent able to get good enough airflow, or try some spot coolers pointed at the gfx chipset.
This would be decent CPU cooler and should fit
 
Thanks - that's great help and assurance, definitely going to consider the HSF recommended as never happy with bog standard stock ones. Will see how the extra fan for intake goes in bringing temps down when it arrives.

BTW my motherboard temp is around 32 degree mark is this OK or for that matter, accurate?
 
Ok, bought fan for case front and installed couple of weeks ago and things had been fine until today! Getting a crash again with black /white vertical lines. Everything at stock speeds nice and cool. Any further suggestions as to what might be causing this please folks?
 
Do you have any old graphics cards around the house that you could install into your rig and use that to see whether the crashes still happen??
 
Sorry to hear your still having problems...

What were you running at the time of the crash - was it idle or were you stressing the system with an app or game?

EDIT: DavyBoy is right (i was on the phone and delayed submitting my post) - it's my suspicion too which is why i asked the above - if you're running a gfx intensive app or game at the time of the crash it would point towards the onboard gfx - especially with the screen you're getting. If you can get hold of a gfx to install do as DavyBoy suggets and see if it still replicates the problem (remember to disable onboard vga if you do.)
 
Last edited:
I was just general web browsing when it happened - nothing to stress it - processor at 35 degrees and mobo 30degree...
 
Bum!

Then i suggest you start stressing components and hopefully you may get a clearer picture.

Run memtest for a few hours (not in a windowed environment) and if that passes run prime95 (or it's equivalent) and see if it freaks out. If it remains sane run furmark and note what temps and reactions you get from that.

Then get back to us...

Good luck.

EDIT: have a quick look in the BIOS and look at your voltages in hardware monitor.
 
Last edited:
ok thanks for the suggestions - unfortunately don't have a spare PCIE gfx to try at mo although do plan on getting something like a HD 4850 in the near future. By the way I have a low vcore of 1.21v which is just above min at stock speed for 7750BE. Not sure if changing over the AHCI might have something to with it as it started out just after this change and I'm aware it is preferable to clean install when changing (I just tinkered with thre registry so vista recognised drivers after a bios change) I'll give memtest a go and get back to you after the weekend....
 
Just had a thought and might appear a complete idiot - I have a 24pin mobo with only 20pin connected from psu - can this be anything to do with it??
 
Jinxter said:
Just had a thought and might appear a complete idiot - I have a 24pin mobo with only 20pin connected from psu - can this be anything to do with it??
Click to expand...

This may well prove to be your problem as some components may not be getting the power they need. The extra 4 pins give extra voltage to aid stabilty but your system should, theoretically, cope with the 20 pin.

You have a very undemanding rig - and i've seen other rigs running ok with 24pin mobos with only a 20 pin connection attached - but it's still not an ideal set up.

Run the above tests i suggeted before or you could try running prime 95 and furmark at the same time and see what happens - if you have got some power issues your system should buckle quite quickly.

The problem you're faced with now is that even if any of the above tests fail - you can't rule out the 20 pin connection as the problem - even though you may suspect another component is at fault. Unfortunately, the 20 pin connection will always have to be taken into consideration until disproven - thus making trouble shooting all the more difficult. Can you borrow a 24 pin PSU from someone?
 
Last edited:
ok im getting furmark now and will give that a go - im just hoping I havent damaged the onboard as it did run overclocked for about 1 month....
 
You won't have damaged anything if you're running system under voltage - and your stability has been pretty good up until this recent event.

It may not be the fault of the 20 pin connection as your rig doesn't draw much power - but it can't be ruled out - and, as mentioned earlier, it's not ideal.

Is that your total spec hooked up to the PSU, BTW?

Also did you try 95thrifles suggestion of spot cooling the onboard gfx as this may help?
 
Last edited:
Yes that is the total spec which like you say is well under load, even had my old IDE hooked up for a while without issues. I've tried 95 and furmark for a while and it was ok, could try spot coolers but it should in theory be OK without ocing the graphics. Think I might go the route of getting 20-24 pin convertor just to rule this out (and hopefully cure the problem one hopes!)
 
Jinxter said:
Think I might go the route of getting 20-24 pin convertor just to rule this out (and hopefully cure the problem one hopes!)
Click to expand...

It's definitely worth the effort just to rule it out.

Jinxter said:
could try spot coolers but it should in theory be OK without ocing the graphics. -
Click to expand...

True enough, but sometimes some components just require that extra bit of cooling. If the above 24 pin adapter doesn't make a difference you could just leave the side off and blow a normal fan directed at the mobo. If the sytsem then becomes stable then it's probably worth investing in the spot cooler. (probably worth trying this now if you have a fan to hand.)
 
You must log in or register to reply here.
Back
Top Bottom