DFI Ultra-D random overheating???

[macgamesrule]

In the last few weeks my CPU has started to overheat on startup . Sometimes when I start up it's normal, ~30 C etc., but once in a while, it'll be ~55-58 C as soon as I start up- no programs running, nothing! But as soon as I restart, temps go back to normal I checked the fans, etc., they're running fine... sometimes I take the side off the case and it still overheats. I ran Prime95 and some games on one of the restarts where it was behaving itself, and it never went over 40 C at full load. Below are some screens, taken immediately after 2 successive restarts:

I assume the lower voltage is the mobo's attempt to bring the temps down ... thx in advance

 

[FragTek]

Looks like a problem only a DFI could have, rofl.

 

[FarFarAway]

What PSU do you have mate? Those are some huge voltage fluctuations in the pics. I would say that thats why your getting overheating

 

[JN]

The +12v rail dipping that low is just nasty. Looks like it might be time to get a new PSU.

Edit: Just noticed you've got an OCZ - it might be worth testing those rails with a voltmeter to ensure you are getting correct readouts, and then RMA the PSU if your rails are indeed dropping that low.

 

[FragTek]

You wont be hassled by OCZ for the RMA either, they are the best of the best. They'll ship a brand spankin new one out to u no questions asked ASAP.

 

[NickS]

What PSU do you have mate? Those are some huge voltage fluctuations in the pics. I would say that thats why your getting overheating

Smartguardian's +12v rail is anywhere from .20-.30 off comapred to a DMM.

Nick

 

[FarFarAway]

Yes but the voltages are fluctuating by quite a lot=bad

And don't even talk to me about Core Centre - I just tested a mates MSI K8N (or whatever) and his core centre reading were 100% wrong missing by as much as .3-5v

 

[NickS]

LOL. Mines accurate .

Must have been a Neo2 board. Oh well.

All I know is when I had my DFI, the readings drove me mad.

Nick

 

[macgamesrule]

So the consensus is it's a bad PSU? meh so much for $130 for a top brand... I'll check the volts and they're really that far off, I'll have a chat with OCZ

Interesting though that the voltage going down would cause overheating... I would think the voltage would have to go up for overheating to occur.

 

[Raven]

well if a device doesnt get the V it needs it compensates by using more A.

 

[macgamesrule]

The overheating problem occured again and this time I was able to test the voltage rails as OCZ wished... the good news is, the voltages were almost exactly their correct values. Bad news, something else besides the PSU must be causing the voltage on the mobo to be too low, causing the CPU to overheat... voltage regulators went bad, maybe? It appears my next step is to contact DFI . I'm rather hoping to get this resolved and everything working well in about 2 weeks time when I get out of school for the summer

 

[Raven]

checked if the caps are round and fine ?

EDIT:

you do know many temp sensors on motherboards are a joke ?

amd have measured up over 30 celsius wrong on an asus board and all other boards they had tested was more or less off.

is the heatsink hot ?

reapplied AS5 ?

 

[macgamesrule]

I looked around at the caps, none appear to be blown as far as I can see, and I doubt the temp sensors are off by almost 30C. Also I'm not using AS5, just whatever came with the stock AMD HS.

Although the problem may have occurred at system startup a few times in the past, I haven't been able to get it to occur at startup when I've tried. After doing a bit more testing, I have found another way to reliably duplicate the problem. If I put the computer on standby, then try to bring it back up, the overheating problem occurs every time. Further evidence that this is a hardware issue, and not just SmartGuardian reporting incorrect values, is that the first few times I attempted to bring up the system from standby, the whole computer shut down. When I went into the BIOS and upped the CPU shutdown temperature from 60C to 65C, the system shutdown ceased occurring and I was able to bring the computer back from standby into Windows, where SmartGuardian reported temps of ~59C. Since the shutdown is controlled by the BIOS, not SmartGuardian, it appears that the BIOS is seeing the high temps too. And so the plot thickens...

 

[FarFarAway]

Have you got cool and quiet disabled in BIOS? And CPU thermal throttling turned off?

You don't have wake-on enabled or anything like that do you?

 

[Raven]

If I put the computer on standby, then try to bring it back up, the overheating problem occurs every time. Further evidence that this is a hardware issue, and not just SmartGuardian reporting incorrect values, is that the first few times I attempted to bring up the system from standby,

seen many strange things happen when using standby mode so i still believe its a software issue.

i know you dont have a msi motherboard but if you go to their forum and make a search on standby you will se many different problems.

monitors not starting mice and keyboard doesnt work

cd's stop working networkproblems... the list can be made very long.

 

[macgamesrule]

Cool and Quiet is disabled, as are the power on for mouse and keyboard. CPU thermal throttling is enabled at 50%... I thought this is for overheat prevention? Come to think of it, thermal throttling is probably what my mobo is doing, to prevent further overheating. If my temps hit 60C with it on, how high would they go without it?? I don't really want to know.

It *could* be a software problem relating to standby, but still, standby SHOULD work. It is preferred over shutting down the comp all the time to save on the power, wear and tear that the system goes through on startup, etc. If for some reason my system is incapable of using standby correctly, then somebody screwed up the design somewhere.

Also btw I have 2 Windows installations on this system, both of which exhibit the exact same overheat behavior on startup. Therefore it is at least not a probem with a particular installation of Windows.

 

[macgamesrule]

From Angry_Games on DFI Street:

actually if you check the real temps, I think you'll see that only the temp sensors are reporting incorrectly resuming from standby...the actual temps are not getting that high.

This is a known issue (resuming from standby and having temps show about 20C too high).

There is no fix for this.

There is nothing wrong with your motherboard.

The DFI Lanparty series motherboards are built purely for overclocking, and because of such a design, certain features on 'vanilla' boards don't always work correctly anymore (resume from standby and hibernate are two features that usually take a hit along with Cool & Quiet).

You are free to RMA your board as many times as you want, but you will find that each and every board will do the same thing upon resume from Standby.

The only real answer is just shut the computer down when not using it, or just have the monitor turn off after 15 minutes or whatever and leave the rig running.

So, it is a software reporting issue after all. Guess I'm good to go on overclocking, as long as I stay away from standby. As I said, not a good design, but meh maybe I can live with it as long as it's not actually damaging the CPU.

Thanks for the help everyone :worship: :worship:

 

[Raven]

why not get a third party temp sensor for the cpu instead of using the builtin ?

third party is often much more accurate.

 

[macgamesrule]

That is a possibility... could you recommend a good model?