Core voltages

Red Dawn · Post by **Red Dawn** » Wed Oct 09, 2002 2:17 pm

I know that overvolting may be dangerous to your chip - shorten it's life span etc., but how does undervolting affect it?

I noticed that when running my t-bird 1.8V at 1.3V one of my hd's didn't show up during the boot sequence, and after I cranked up the voltage a bit, it magically appeared again. Hmmm.

Right now I'm at 1.45 with no instabilities, I can probably go lower though.

ChiefWeasel · Post by **ChiefWeasel** » Wed Oct 09, 2002 2:39 pm

I think it could only cause instability, which shouldnt cause any physical damage, but just had a thought that it could cause data corruption on the HD maybe?

Red Dawn · Post by **Red Dawn** » Wed Oct 09, 2002 2:44 pm

Yes, I had the same thought... but as I'm not tampering with my bus, how could it have that affect on it?

It seems kind of odd...

ChiefWeasel · Post by **ChiefWeasel** » Thu Oct 10, 2002 12:31 am

Not sure, it might not, but i was just thinking that if at 1.3V it wont show up, then maybe at 1.45V it might show up but still not be 100% stable?

You werent tampering with the bus when it was @ 1.3V were you? but the HD didnt show...

BTW being stable at 1.45V is v. impressive for a 1.8V chip and your not underclocking it are you?

Red Dawn · Post by **Red Dawn** » Thu Oct 10, 2002 4:18 am

I'm not sure about this either, I'm pretty new to this undervolting business... I learn as I go.

Nope, I didn't touch any of the bus settings at all.

No underclocking, everything is left in "auto", ie. no manual override.
I might have another go at all this, the hd not showing up at 1.3 might be unrelated to all this... on the other hand, I don't want my hd to die on me.

Also, what's the "track record" for TwinMos memory? Is it supposed to be a good overclocking memory, or otherwise "resistant" to extreme working conditions?

ChiefWeasel · Post by **ChiefWeasel** » Thu Oct 10, 2002 7:21 am

I think the problems you encountered are the same problems people encounter when overclocking, i.e. sometimes you need to increase the voltage to maintain stability. Since 1.3V is to low for your chip, you need to increase it to maintain stability (1.45V). I wouldnt think anything permanent would happen to your HD.

Just bear in mind that when overclocking it is necessary to run stability checks to ensure your system isnt creating any errors. The same applys when undervolting. Think of it as checking a certain voltage and speed is stable, irrespective of the stock settings. Have you tried your system with Prime95 yet? Sometimes it will find errors after hours of stable 100% load number crunching.

I was only thinking that errors created in the CPU could be written to disk when saving data... If Prime95 runs ok for 24hrs, then i think your system would be fine under-volted.

I've not heard of TwinMos memory before , but just searched the [H] forum and the PC3200 seems to be held in high respect!

Red Dawn · Post by **Red Dawn** » Thu Oct 10, 2002 8:47 am

Yes, 1.45V seems the way to go, at least for now. My chip doesn't get much cooler because of the case temperature versus a relatively high ambient temperature at the moment, there simply isn't much differance so that the cooling effect will be dramatic I'spose.

While I have not run Prime95 24 hours or something like that, simply because I like to shut my computer off while I go to bed, I did run it a few hours last night, no problems arose, so I should say it's 'stable enough'. It's not like I normally (when web browsing and listening to music) use my cpu to 100%.

I use PC-133 memory, but I hope they're of the same quality as of those PC-3200 ones!

PS. 1 GB of ram proved to be a massive overkill, my highest count memory usage has been around 33-35% of total memory so far.

DS.