How can a PIII run at 1.25 volts?? Please Read

[fireman233]

What's the minimun Vcore needed to run a PIII700? Please read!

Hi guys. My new gigabyte GA-6oxet (i815 Bo) board only goes as high as 1.800 volts. I previously had my PIII700 Oc'ed to 1015 w/ my old Soyo board (on a slocket) doing 2.050 volts. Anyway, I wanted to up the voltage, so I found a web page which references Intel Pentium III spec's from 500 to 1.13ghz (http://developer.intel.com/design/pentiumiii/datashts/245264.htm). Found on page 21 of this spec sheet is Table2, shown here.
http://www-personal.umich.edu/~mikesp/Voltage.gif
So accordingly, if I connect all 5 pins colored below, Vid0, Vid1, Vid2, Vid3, and Vss, I should be able to run 2.050 volts.

http://www.extremeoverclocking.com/articles/howto/images/PIII_Layout_sm.gif http://www-personal.umich.edu/~mikesp/MBM.gif

Well, I did this and plugged the chip in and it's not default 1.65 volts or what I wanted at 2.05 volts.... it's 1.25 volts

It read 1.25 in both the Bios and in MBM5. Weird since Intel's spec sheet doesn't even have a listing for 1.25 volts :confused: . What do you guys think? Did I do something wrong? Does this have something to do with the i815 B0 chipset?

Also, WHY IS THIS THING EVEN RUNNING?? :mad: :confused: :mad: ?? It's at 1.25 volts. The temperature plummeted, so that's good. But I didn't even think this would boot at 1.25 volts??????
I ran Prime95 and Seti overnight and it's still purring w/o a problem in the morning. Any ideas?

[vimal]

Maybe the board simply doesn't do anything higher than a certain voltage.

Since you have them all connected to Vss, try coating some of the pins to get other voltages, (maybe 1.70 - I know you can do it in bios, but see if you can do it through pins as well)

If that works, see if you can get 1.85 by coating Vid2

Though I will admint having a proc humming at 1.25V is pretty impressive.

[fireman233]

Unfortunately, there are no other voltages possible. The reason is that the cB0 SL45Y PIII700 that I have, and all PIII cB0 for that matter, already have Vid0,1,2 grounded to Vss. Default is 1.65 volts and it is internally grounded, so you cannot remove the ground to Vid0,1,2. The only room for play is whether Vid3 is connected. So you either get 1.65 or 2.05. I can live with 2.05, this proc has run it before w/o a problem.

The cC0 have more variety, but alas.

I'm not positive about the internal ground, can someone confirm that?

[Muggy]

I wonder if this applies to the slot versions? I mean I know MBM shows 1.75 but is it somehow different internally?

[vimal]

The pins may be internally grounded, but the socket/motherboard doesn't need to know that. Give it a light coat of glue/nailpolish, or if you're using it in a slocket, tape over the corresponding pin in the socket.

Breaking the connection is read as 1, while a direct connection to ground is read as 0

[Mictlan]

Might be the following:

Some instruments are calibrated to be used in a certain range. If you use them out of these range they give you false readings. How false? It depends on the quality of the instrument or, in the case of digital instruments, it's converter. I really don't know exactly how does the mobo measures voltage, but, as the CPU is running, I would think that you are experimenting a false output from your V monitor.

[vimal]

Originally posted by Mictlan
Might be the following:

Some instruments are calibrated to be used in a certain range. If you use them out of these range they give you false readings. How false? It depends on the quality of the instrument or, in the case of digital instruments, it's converter. I really don't know exactly how does the mobo measures voltage, but, as the CPU is running, I would think that you are experimenting a false output from your V monitor.

That would have been my first guess, except fireman233 did say that temps plummeted. Did they go below temps at 1.65V? If so, you really are running at 1.25V. If not, then the voltage converter is wacky. But I would think that it should be fine up to 2 or even 3V, since i'm pretty sure mobo manufacturers don't want to change the circuitry for each different chip they support (probably have the same circuits as from the older PII as well as Celerons, some of which run in the 2V range.

[fireman233]

Yeah, like what has been said, the temps plummetted. Crunching Seti and Prime95 all night, the temp never exceed 27C while with 1.65 volts, they shot it to around 37-40C. At 1.65 volts, I was able to OC to 884 w/o failure. I tried playin ith OCing this thing at 1.25 volts, but it wouldn't even do 785. These tell me that I really am running a lower voltage.

As for the mods, I have read of people plucking the pins off that they don't want to alleviate the problem of the pre-grounded Vid 0,1,2... I could do that, I just didn't know if that was correct and wanted to find out before I broke this thing permanently. I really want this thing at 2 volts or above (parts are coming in now for the watercooling setup), I could aim at 2.0 volts and remove some pins, but if that doesn't work either, then I'm screwed with a permanent mod that doesn't work.

[vimal]

Don't remove. Just isolate with a thin, non-conductive coating like nail-polish or super-glue/epoxy (very thin)

That way, if it doesn't work, you can scrape it off and be abck where you left it. Pin clipping is recommended if you know exactly waht voltage you want, and never ever want to chagne it.

[fireman233]

I've heard some bad stories about nail polish and pins.... that layers were too thick and it didn't work well because the polish was scraped off when inserted the PGA into the socket. Anybody ever try this, or is it just speculation.

Also, this doesn't solve the original issue of why the Vcore dropped to 1.25.

[vimal]

I used nail-polish on a PIII 750 for a little experimenting and it worked fine. Clogging the socket could be an issue, as could the scraping when you insert it into the socket - but with my socket (on some cheap slot adapter), the processor just drops in (no pressure at all) so there was no scraping.

As to why your mobo defaults to 1.25V, that could just be it saying "I can't give ya 2.05 no matter how much you ask for it" It would be nice to test 1.85V just so you know whether the limit is in BIOS software or mobo hardware.