PDA

View Full Version : Celeron 900 @ 1008 Problem


apachean
01-22-02, 10:56 AM
Right here's the problem.

I've overclocked my Celeron 900 to 1008 by changing the FSB to 112Mhz, it boots up fine and runs for a while(about 5 mins then
my screens go blank. It doesn't reboot it just hangs.

This also happens at any FSB that isn't 100MHz

I'm using the heatsink that came with the boxed Celeron 900.... it's looks like it should be prity adaquite, it better be after the nightmare I had trying to fit it (nearly lost a finger!!)

I'm a bit of a newbie. so some of the following information maybe alittle odd.

System info:
Celron 900
196 Mb of PC100 Ram
ECS D6VAA Mobo with VIA686B chipset
SuperGrace Radeon VE, Dual Monitor edition.

Overclocking information:
FSB 112MHz
Tempratures(i'm unsure of what the sensors relate to but using MBM5 it shows me 3 temprature sensors):
Sensor 1: 28C
Sensor 2:32C
Sensor 3:31C

And the core voltage is 1.83.

I don't know much about core voltages but I think I can get hold of a bios for my mother board that allows me to change it.

Can anyone help

apachean
01-22-02, 11:00 AM
ooops probably should have said I'm running XP pro

Pinky
01-22-02, 11:22 AM
Initially I was thinking you weren't giving the CPU enough juice/voltage to run at anything above stock/default. I'd be sure that you are increasing the voltage, and if you're not sure be sure (get that bios update if need be).

temps look fine.

apachean
01-22-02, 11:26 AM
I just got the chip and I don't want to fry it, what kind of voltage am I looking at?

apachean
01-22-02, 12:16 PM
Thanks for the help, it's all fine at 1.98, temps all the same as before and nothings has dived ...yet

takiwa
01-22-02, 12:48 PM
Well, how good is your memory? typically , it should run alright at 112 if it is good, but there is some crappy stuff out there...

I didn't have to increase my voltage to run my 900 at 112, but I did to run it at 124...you should bump it to 2.05 (max on that celly with a wire trick) and go for 133FSB to get your PCI/AGP back in spec...

At 112 your PCI is WAY outta spec...HDD corruption, maybe? Good reason to lock-up...

Also, temps can be misleading for you...without knowing anything about your sensor chip. You may need to compensate for those sensors...

I just got the chip and I don't want to fry it, what kind of voltage am I looking at?

Also, a burn-in wouldn't hurt...especially with that chip...

kraanul
01-22-02, 12:59 PM
takiwa is right.
Put some stress on and then go directly to bios and check temps there.

batboy
01-22-02, 01:18 PM
That voltage is way too high for just using a retail heatsink and fan. I'm guessing those temps you are quoting are idle temps. They sure can't be full load CPU temps, no way. You need to resolve your cooling problems FIRST before trying to increase the voltage. A better CPU cooler is in order before attempting any significant overclocking. However, 112 FSB is well within the range of even a factory retail cooler, but you need to remove it, clean off that crappy thermal pad, and reinstall it with a good thermal paste. That CPU won't last long at those high voltages without much better cooling. In my opinion, 112 FSB is a fairly mild overclock that most systems should be able to handle without too much problem using default voltage or at most, only slightly higher voltage. Also, sometimes PC-100 RAM starts getting flakey at 112 to 115 FSB (especially if it's generic). You might think about getting better memory too.

apachean
01-22-02, 03:04 PM
I put the voltage back on auto to see what would happen and everything seemed fine, running at 112. untill all of my screen fonts went all wonky (hard to explain it look like some middle eastern script) which made me obviously asume that I never had a problem in the first place. It was just my card, when trying to reboot to find out if that would make my problem go away it crashed and when windows restarted it gave me the error log which basicaly put it down to the card/drivers.

SO now i've got another question, my card is a supergrace (ati based) Radeon VE Dual monitor (intrestingly the font scrambling only happened on the primary display). Would UnderClocking the cards proccesor help? to compencate for the AGP bus running faster... or am I WAY of the mark... fell free to call me stupid.

batboy
01-22-02, 03:24 PM
It's not uncommon to corrupt the Windows registry or various Windows drivers when overclocking. As was mentioned earlier, some harddrives don't like running very far out of spec and you can also get data corruption on the harddrive.

If you think it might be video drivers messed up, try removing them and then reinstalling them. To remove the vid drivers, first install the standard VGA graphic drivers and reboot, then reinstall the video drivers that it is supposed to have.

If you have access to Norton Utilities program, run Disk Doctor and Win Doctor. If not, at least run the Windows Scandisk and then download a registry cleaner program.

The other thing that sometimes works is to try and reinstall Windows. If all else fails, you might have to reformat the harddrive and do a totally clean reinstall of Windows.

Most overclockers at one time or another have had to do this. If you do a lot of overclocking, tweaking, and playing with your system, you'll soon learn the value of backing stuff up.

apachean
01-22-02, 03:30 PM
The thing is it's fine when I put the FSB back to 100. without re-installing the drivers, it just goes funny when I put it above that.

The drivers work fine again now.

batboy
01-22-02, 04:15 PM
Possibly one of your components don't like to be overclocked then. The vid card is a possible suspect. I doubt underclocking the vid card will help and besides you'll lower your graphics benchmarks, but hey... what do you have to lose if you try? Do you have a network card (ethernet) because often those don't like being overclocked. It might even be the RAM giving you fits, after all, it is running out of spec.

apachean
01-22-02, 04:42 PM
Well, I'm not really bothered about the graphics, I only use the card cos it's dual header, usefull for audio production.

So I've underclocked the card by a few Mhz (in think it may have been over clocked in the first place, which may have been the problem) and everything seems ok so far...........if it all goes tits up again then I'll guess I'll have to stick with 900mHz...which is a massive improment on the 350 before it!(I'm just being gready for that 108MHz)

Thanks Pinky and Batboy for the help, much apreciated and hopefully I won't need to bother as many people next time I'm doing something I really don't understand

cheers:beer:

apachean
01-22-02, 04:44 PM
oh sorry should have thanks all the other that helped too, didn't mean to offend anyone