100% GPU Fan Problem

Associate
Joined
21 Jan 2004
Posts
161
Location
Planet Krypton
Hey guys. I don't really expect a reply here over christmas, but i'll throw up a thread anyway.

The Scenario:

I've had a problem when powering up my computer at boot, the GPU (Geforce 275 896MB) just powers at 100% fan, and the computer doesn't post on occassion. Although 9/10 it will post after a long period of 100% fan, this feels really wrong. 10-15seconds.

I've included a screenshot of my setup. As you can see i've got a nice overclock that is running stable in Games, with load temps of 50c. I have the Corsair H50.

My Bios settings that i've set manually are: Multiplier 18x, CPU Voltage, 1.400v, NB Voltage 1.20, (It goes red text when 1.30v), DRAM 1.9v, everything else Auto or stock.

Today I switched my computer on and it 100% fan'd for ages, and didnt work. I only fixed it by taking on the GPU, and then resetting my BIOS.

How can I fix this annoying problem? Is it my graphics card? It did the same exact thing on my last board.

Merry Christmas and if your bored, help!!

http://www.macfadyen.org/desktop_settings.jpg
 
Ok are you trying to say... How can i stop my graphics running at 100% ???
If so there should be software inside the box which should once installed let you control the fan speed and do overclocks e.t.c
 
My computer does not post/boot sometimes, and the GPU fan spins at 100%. Sometimes it takes 10-15 seconds at 100% before it boots, sometimes it just spins at 100% for more, and I switch it off.
 
Make sure all of the expansion cards / ram / etc. are 100% in place, the GPU fan doesn't slow until a sucessful POST
 
The thing is, it DOES post, after a expanded amount of time, and i've played lots of games, Crysis on high etc, no problems. It just intermittently stops posting/working and I have to reset CMOS/take card out.

It hasn't been moved or anything.
 
No signs of the GPU temperatures and settings in that screenshot.
Is it hot when playing games?
 
Have you rechecked your OC? Also CPU temperatures are the least of your problems with a CPU only WC setup!! Have you checked the VRM temperatures, etc.? Does the case your using provide adequate cooling? Don't forget that the setpoint of an OC can shift (as components age due increased V) and ambient temperatures will affect this... I would be testing the CPU OC (only) with something serious (like F@H) running on all 4 cores for 24 hours.

Also is it a cold boot problem (when the machine has been off overnight for example) or when the components are still warm?

I do find those PCI BIOS checkers to be useful. It can be instructive to see where the POST holdup is by buying one of these gadgets (a few quid on the 'bay) as the POST sequence code will stop at the problem device. I've tracked down problems with these myself. :cool:

Bob
 
You don't say what psu you're using? it might be an idea to list all your components to see if people with similar hardware may have come across the same issues.
Intermittent faults are always a nightmare to try and diagnose so the more details the better.

USB devices I've found can lock up machines before finishing POST, have you tried disconnecting various USB devices, don't forget things like card readers which are connected to the internal USB headers.
 
I've got a 650W CoolerMaster Silent Pro. I do mean to check the O/c, been have been busy with christmas. It's the same problem without the OC, and also, with a different motherboard/CPU combo. Same RAM, though.

That USB device you mentioned sounds interesting!
 
you've not got extended memory checking on have you? I know if this is on it can really extend the POST, had a few servers a couple of years ago that would take about 25-30 mins to POST
 
That USB device you mentioned sounds interesting!

You mean PCI right!! I thought I was buying a turkey (:D) but these devices are useful. Of course you need to have a PCI slot and need to be able to see "under" the PCI card (I think you can put the output on a flyer lead) - in a standard vertical tower case - not designed by someone too bright me thinks!!

Bob
 
I don't think I have extended memory checking on..

I've tried increasing my NB voltages.

I think I might try a random PCI-E card, see if it's the same.

I can then blame the card?
 
Back
Top Bottom