Weird problem with Vcore settings

[Underwater Mike]

My kids' machine is as follows:

DFI NF2 Ultra Infinity (BIOS 18DEC04) + AMD Athlon XP-M 2500+ @ 230FSB x 10, 1.825v
(2) 512MB Corsair VS512MB400 (DDR460, 2.5-4-4-9 @ 3.1v)
Zalman CNPS7000-Cu
Plextor PX-504A
WD800JB on IDE1
128MB Sapphire Radeon 9600 Pro
Antec SmartBlue350 + Antec Lanboy
Logitech Cordless Elite Duo
XP Home SP2

I recently started raising Vcore from 1.825v to 1.9v to increase the OC. Temps are fine, with load around 49C. When I boot from a cold start with Vcore at 1.825 or below, no problem. From a cold start at 1.85v or above, tho, and I get the high/low BIOS alarm for a misseated or damaged CPU. Lower the voltage back down, and everything's fine.

BUT, if I get that alarm on the higher voltage and then perform a warm restart via the reset switch -- or a warm restart from XP -- I do not get the alarm. That is, always a problem with the higher Vcore from a cold start, but never a problem with the higher Vcore from a warm start.

Anyone know what the cause might be? If the higher voltage was whacking the CPU, why would it stop being a problem on a warm boot vs. a cold start?

[Deer Slayer]

Every voltage regulator has a certain amount of time between the time it's fed power,and it hits it's desired output voltage. This is because all of the capacitors and stuff are completely uncharged, and none of the components are warmud up, so their resistances aren't quite what they should be. When the VRM for the CPU is doing a warm restart, the caps all have a partial charge, and all the components are up to temp, so they are @ the proper resistance. The higher you set the output voltage, the longer it takes to rise to the desired output. There is no real way to fix it, the only thing that might work is to provide the VRM with some more voltage. Which means you'd have to give the 12v rail a kick in the butt. It might be the rails on your PSU are taking a little while to come up to spec. So I would ignore it, and just let it complain. Maybe try checking all of the connections to the mobo in there, and pinching the main atx connector so it grabs the pins better. I don't really think there's much you can do about it. One last thing is to disable the quick post option, so it takes longer to fire up. That might help. Good luck, I hope you get it figured out. :o

[Underwater Mike]

Thanks for the good info! Makes a lot of sense.

Connections are all fine; checking those was my first thought. I tried disabling the quick POST, but that didn't help any -- I guess it's because the problem occurs before the mobo even gets that far.

Given that it's the kids' machine, I'm not inclined to venture into any PSU tweaking. or even springing for a new one. Maybe I'll just tell them to goose the reset switch if they hear the noise. It doesn't seem to have any ill effects, since the machine hasn't begun loading XP or anything.

[Bofinn]

Originally Posted by Underwater Mike

Maybe I'll just tell them to goose the reset switch if they hear the noise.

"Hey kids, if you hear the computer thingy making a noise, push this doohdah over here..."

[Underwater Mike]

Originally Posted by Bofinn

"Hey kids, if you hear the computer thingy making a noise, push this doohdah over here..."

LOL, pretty much. Or, just use my grandpa's approach and whack it a couple times with a hammer and then cover the dent with duct tape.

[StarTraveller]

Originally Posted by Underwater Mike

Thanks for the good info! Makes a lot of sense.

I'm gonna have to 2nd that!

I was looking at this thread when Mike first posted it and I thought it looked similar to issues that I've experienced in the past and never found the cause for (or a solution to).

Deer Slayer's explanation is making perfect sense and explains how seemingly identical scenarios actually do have differences!