Temperature Problems since HD replacement

[JohnRD]

Hi guys,

You'll have to pardon me, I'm a bit new here. Generally I just call up my dad or Google if I'm having problems along these lines, but I'm having trouble getting good information today.

Recently, my hard drive crapped out. I took it in to Fry's (under warranty) and had the hard drive replaced. Since then, however, I've continued to have problems, first with power and now with heat.

I use an HP Pavilion Elite m9350f. I've been made aware that one of the problems with this system (other than the insides being a total freakin' mess) is that it only comes with a 500W power supply, not really enough to run everything on top of a 460W graphics card (Nvidia Geforce 9800 GT). So I'm looking at replacing my power supply.

However, when I turned on SpeedFan to see what my temperatures were like before I started jacking around with things, I saw some startling numbers. I've never been good at interpreting them before, but they're definitely higher. Every single icon (except Temp2, obv.) has a flame icon next to it, just sitting here idling, with the side of the case off. Taking the side off seemed to push everything up a degree or three from these numbers:

GPU: 80 Celsius
CPU: 72
Local: 67
Remote 2: 72
Temp1: 57
Temp2: -128
Temp3: 55
Core: 57

I regularly blow out my fans, although I have no clue how to do so with the gfx card (which I've read is usable up to 105C, though I really don't want to push that.)

Before going and adding a power supply, I'd like to know what to do about my temperatures. I'm not sure which temp matches up with which meatspace component. I'm covered under my warranty still, but I wasn't having temperatures like this (that I recall) before I brought it in to Fry's. Plus, they took 29 days just to replace the hard drive*, so I'm not eager to jump back into that hole just yet.

By the by, I'm not overclocking or doing anything crazy. I just want to play some games. World of Warcraft crashes everything within three seconds of loading. Team Fortress 2 crashes everything inside of five minutes.

* for the record, they replaced my 750GB Seagate HD with a 750GB 7200RPM Western Digital Caviar Black w/32MB cache. For cryin' out loud, I even crashed one time while playing ROM CHECK FAIL (with its incredible 8-bit graphics.)

 

[guanyu210379]

Everything is overheated?
Wow!
I don't even know for sure where to start.
Try checking the BIOS settings.

 

[JohnRD]

This is probably obvious, but check BIOS for what?

What would you guys recommend to help get my temperature down?

 

[Twoboxer]

OK, there are two issues - cpu temp and gpu/case temps.

Before panicking, please download CPUID's Hardware Monitor and make sure the temps it gives roughly correlate to Speedfan. If so,

You say the inside of the case was a mess. Was it a mess *before* you took it to Fry's? If so, bring it back is an option. Usually, HP/Dell/etc route cables and stuff very well, to minimize cooling requirements and heat issues. Otherwise, try as best you can to get the mess out of the way the air will flow, presumably from the low front to the high rear.

Check the cpu heat/sink cooler. Gently try to lift each corner where its secured to the mobo. If it wobbles in any way, bring it back (or secure it, or remove it, clean off the heat sink and cpu, add new thermal paste, and resecure.)

They should not have touched BIOS. However, what you can do is copy down all the BIOS settings (or photograph). Then clear CMOS (remove power cord from wall, remove CMOS battery, press case power on key 5 times, wait 5 minutes, reinstall CMOS battery. Power up, go immediately into BIOS.

Load whatever "Safe Defaults" exist. Compare the new BIOS settings to your old ones, and change ONLY things you know you MUST change, eg, RAID (which you do not have) would be one of those. Save and exit, boot up, and check cpu core temps again with HWM as suggested.

(They put a nice HD in there - wonder if they were feeling guilty about what they did/how long they took lol.)

 

[JohnRD]

(aside: Yeah, earlier this year I'd looked into replacing the HD myself and grabbed a WD Caviar Green, liked it. I was a little surprised myself to see the WD Caviar Black when I was checking out the guts of the beast before signing it out at the service desk.)

As regards your question: the cords were a mess to begin with. They were a bit worse, though, after the service.

SpeedFan has given me numbers different from those I got last night, so I'll re-list here, since everything doesn't quite match up. I can't quite tell what matches up other than the GPU.

GPU: (SF)77-78; (HWM)77-78

HWM says
TMPIN0: 56-58
TMPIN1: 55-57
Core #0: 56-59
Core #1: 56-59
Core #2: 56-59
Core #3: 56-59

SpeedFan says
CPU: 70
Local: 65
Remote 2: 70
Temp 1: 57
Temp 2: -128
Temp 3: 56
Core: 58

The Hardware Monitor numbers look a bit nicer, but I don't see CPU listed there explicitly (or Remote 2, whatever that is.) Also, SpeedFan is still listing everything with a flame icon.

I'm wondering if the insufficient power supply is part of the reason things are so hot. I'm considering replacing it (and in the process, cleaning up the ridiculous cord mess - gonna steal some cable ties from work tomorrow) and hoping that will fix everything.

 

[Twoboxer]

Now you may understand why we ask for HWM numbers lol.

Judging by its label (500W you told us), the power supply isn't insufficent. The 9800 only draws 108W, and only requires a typical 400W psu. Here's a review showing total system power needed with a 9800gtx at 267W, at the wall - power from the psu is 10%-20% less:

http://www.hardwarecanucks.com/foru...0-gt-roundup-evga-asus-gigabyte-palit-19.html

So let's get back to basics -

Mid-50s cpu at idle is high. I don't normally do AMD stuff, but a quick scan says it shouldn't be running this far over ambient temps.

are all the fans running freely?

is the cpu heat sink/cooler properly attached? Gently grasp each corner where its attached to the mobo, and try to lift/rock. You should feel no movement, other than perhaps the motherboard flexing as you pull.

is the heat sink clean of dust? not just the fan, but between the "radiator coils". If air doesn't move past/through there, heat is trapped.

If that doesn't work, we go back to resetting CMOS and loading BIOS defaults.

(also, you have the cover off - but also lay the case gently on its side - heat rises, lets keep it as cool as we can for now.)

 

[JohnRD]

- I did some more reading, it's actually a 460W PSU I've got now, not 500W.
- Is that survey done for the 9800GT? Looks like 9800GTX (I think there's a difference.)
- The fans seem to be running freely. I checked them pretty thoroughly when searching for the source of a small buzzing noise (just a vibrating bay door, but damn it, Fry's).
- I can only access 2/4 corners of the heat sink, due to PCI slot placement. Those two corners felt very stable and would not rock.
- I can see no dust, but have taken some pressurized air to it anyway.

That said, I spoke with my dad today. He's been doing IT stuff for the school district for decades (mostly networking lately, but he knows his way around a computer.) He told me it was very likely that a CMOS reset would stop one or more of my components from functioning.

Thing is, when I got my hard drive replaced by Fry's, they had difficulty installing the operating system (as I did when I tried the same tack with the aforementioned WD Caviar Green.) They had to install something else (beyond my ken) to get it to work. Whenever I start up my computer, instead of listing the Western Digital hard drive, it just lists "NVIDIA SPAN". Because of this, he's betting a CMOS reset would very likely break something.

I'm not sure what else can be done about the heat. As you might have inferred, my proficiency at this sort of thing is armchair-level at best. I've been digging around inside computers for eight years or so, but only as needed. I'm gonna go to town with some cable ties, triple-check the heat sink, and see if I can't air out the gfx card fan (pretty hard to get to.)

 

[guanyu210379]

CMOS reset would stop one or more hardware from functioning?
???
Is he really an IT guy? Which I doubt!
I reset my BIOS perhaps millions of time during the first OC steps, everytime I got a really unstable setting.

Just reset the d*mn thing !

 

[Twoboxer]

A CMOS reset would require you to re-establish your RAID array, eg, if you have one, if that's what nVidia Span is. But you can ensure you get it right by recording BIOS settings or using a digital camera to photo them before the reset.

I can't think of anything that could be affected by a CMOS reset that can't be restored by choosing the right BIOS options.

 

[guanyu210379]

I believe the thread poster is not using RAID configuration.

 

[Twoboxer]

Yeah, I don't think so either, and I think nVidia Span is a video thing, but . . . the only issue he could have with a CMOS reset is a RAID re-establish.