Increasing voltage of p-750 above 1.65

Hi everyone,
I am trying to overclock p-750 Processor. I would appreciate if anyone could explain some things that i don't understand.
Unfor I have Intel board which is "anti-overcloer" - i don't have any kind of control over cpu voltage/multiplier/fsb. I already know from the sticky that covering bsel (aj31) pin will force mb to detect 133fsb instead of 100. And, as mul. is built-in in cpu, cpu "may" work at 1000 mhz. Once i did this trick with slot-1 cpu and whenever i raised fsb from 66->100 (celeron 266->400) that was not enough, i had to cover VID pins on cpu cartridge to raise cpu from 2.0 to 2.2 V. Than everything worked OK and stable.
So i guess i will have to do same way here. Disconnecting pins is not a problem - i only need to put on the specific pin 4-5 layers of superglue - it is not conductive and very strong.
This is Vid specification from Intel's manual of Socket 370.
The default voltage of P3-750 is 1.6v(i think) So to receive any entry Above 1.6 v - i need somehow to CONNECT vid3(aj37) pin. As i understand, this pin is of course presen(cose all of the do) but "disabled" some-hon in the cpu itself.
So is there anything i could to to encrease a voltage?
Any help is really appreciated.

VID3 / AJ37 VID2 / AL37 VID1 / AM36 VID0 / AL35 voltage
1 1 1 1 no
1 1 1 0 1.35
1 1 0 1 1.40
1 1 0 0 1.45
1 0 1 1 1.50
1 0 1 0 1.55
1 0 0 1 1.60
1 0 0 0 1.65
0 1 1 1 1.70
0 1 1 0 1.75
0 1 0 1 1.80
0 1 0 0 1.85
0 0 1 1 1.90
0 0 1 0 1.95
0 0 0 1 2.00
0 0 0 0 2.05

By the way, is there any difference in VID pins between [fc-pga] and [ppga] cpu's?

FCPGA and PPGA use the same VID pins.... only the newer Tualatin FCPGA2 CPU's VID pins are (slightly) different as intel changed 1 VSS to a VID so the tualatin could have a .05V....

as for what you can VID you CPU to... we need to know what your defualt is. So when you are 100% sure its 1.6V, i'll help you out... best way to know what the defualt is, is the sspec or stepping of the CPU. If you are not 100% sure, see what the monitor in the BIOS (if you even have one) says

The CPU could be one of 2 voltages:

1.65v (cB0 stepping)
1.70v (cC0 stepping)

Its most likely a cB0 which is 1.65v. This however leaves you with a very large problem.

On a 1.65v CPU, VID0, VID1 and VID2 are ALL connected internally to the VSS pin in the CPU. This means the only possible voltage you can get is 2.05v by connecting VID3 to VSS. This however is a very high and probably lethal voltage for a Coppermine CPU.

if you are lucky enough to have a cC0 stepping CPU, only VID3 is connected internally to VSS. Meaning you can select any voltage you wish from 1.70v to 2.05v by choosing which pins you wrap. A good one would be 1.85v, by connecting VID0 and VID1 to VSS.

Here is a good article from HardOCP which explains it and has a picture. (Celeron II is the coppermine core so it applies here)

http://www.hardocp.com/article.html?art=MjIx

There was never any 1.6v... like Reef says, it's probably a cB0 which is 1.65v. That's the worst for trying to do VID pinning, but if you have success in isolating the BSEL pin with a coating or something, then there is hope, you might be able to isolate the VID2 pin and wire the VID3 pin to the VSS pin for 1.85v. You need to be sure what default voltage you have first though. It'll say right on the CPU.

Thank you guys for the vital info, i'll try that out.

I just checked the cpu - it is written on it
750/256/100/1.65v
l021a489-0793- sl462
So, as i understand, as far as vid0,vid1,vid2 are already internally connected to vsss - the only thing i could to - is to connect vid3 - which will give me 2.05v, as Reef said.
Do you think it is too risky, even applying very strong cooling.?
This computer is opened so, i connected it's own heatsink to 2 additionals big heatsinks(with some white mass- don't know what it is called in english, it just passes though heat very well.) And on all 3 there are coolers, I tried this combined cooler on c266->400mhz and it almost the same temperature as room, just 2-3c higher, wich is much lower thatn it's default hetsink+cooler in it's nativ 266mhz.
So, please i need your opinion, Do you think that cpu can just burn out(not from heat)?
Have you ever heard of success of that before (1.65v->2.05=ok?)

I ran at 2.00v on a Celeron 600 for around 6 months without any problems. In Intels specs they say 2.1v is dangerous for the Coppermine Core, 2.05v (give or take a little for fluctuation) is just below this so in theory should be ok . But it is still very risky and could kill your CPU. I did try 2.05v on the Celeron 600 I had and it simply refused to post, seems to be very random results at voltages this high. Very good cooling is a MUST if you try it though, they heat up a lot after 1.9v.

Thanks for the info Reef, i'll try that.

No offense, but that heatsink combination you cobbled together probably ain't gonna to be enough to risk 2.05v. That's a lot of voltage even if you had a nice watercooling system. If you run 2.05v, it'll work at least for a while, but you run the risk of frying the CPU with long term usage. I think you're getting the cart before the horse here. You don't even know if this CPU will work at 133 FSB yet and now you're worrying about extra voltage. Like I said in my other post, if you are able to isolate the pin to make it run at 133 FSB, then why not isolate VID2 and then wire VID3 to VSS for 1.85V, I think that's your best option here.

I don't understand , Batboy. Of course i would set 1.8-1.85 v only if i could. If i understand it correctly, in the table given in the first post - the same table from the link to 3rd reply - I have Vid0, vid1,vid2 - already internally connected to vss. So the only increase in voltage i could get - is to isolate vid3 - which will give me 2.05 volts.(the last row in the table).
Do i have any other choice?

Yes, the default 1.65v means that VID0, VID1, and VID2 are internally connected (closed) to VSS. VID3 is open (not connected). To get 2.05v, you need to wire VID3 to VSS. Ok, if you connect VID3 to VSS and also isolate VID2, then you'll have 1.85v. Look at the chart, if VID3, VID1, and VID0 are connected and if VID2 is open, then that's 1.85v.

Finally i tried to do as Bathboy offered. And it "almost" worked. I mean, i Connected vid3 to vss(that was easy) and isolated vid2 which gave me 1.85v. Next i isolated bsel(aj31) to make mb detect fsb=133. It "almost" worked, cos i was able to start windows(2000) without problem, but that it just hung up. And this is after 1-1.5 mins after sturtup, when heatsink was still cold.
Then i tried 1.9v and 1.95 (by scratching and applying superglue to pins) and this time pc hung up much earlier - after post or in the middle windows boot. So i had to give up o'cing and live it default 750. Very strange. i thought giving the cpu more voltage will make it more stable (or, better say less unstable) when overclocked. But for some reason this is vise versa.
So thank you all, especially Batboy.
P.s. Overclocking this cpu is VERY important to me. It is not just saying to yourself "i did it". Most of the time i compressing video using tv-card(bt878 chip) When dealing with _realtime_ compression - if cpu is too slow - you will have dropped frame and you capture will be spoiled. With it's default 750mhz I can compress only with 480X360, if i tried 640x480 - i had 100% cpu load and i had 10-15% dropped frames. So o'cing 750->1000 would probably solve this problem.

The more voltage you use, the higher the load CPU temp. Those Coppermine P-III CPUs don't like getting above 45 degrees when overclocked. It sounds like a temp problem to me. Maybe try improving cooling. I'm having trouble invisioning how you connected three heatsiks together, but I suspect you need more air flow across the fins. Also, increasing case ventilation might help. Maybe cut a hole in the side of the case to mount a fan that will blow fresh air right at the CPU heatsink.

Being in Israel I bet computer prices are expensive, but for video rendering you really need the greater bandwidth that DDR RAM and the P-4 has to offer.

I connected huge heatsink to 750's native heatsink(i connected them with white mass, that passes through heat well) And on that heatsink i put big fan (bigger than in the power block) My comp is open - it has no cover so, no matter how hard i overclock the cpu, heatsink's teperature are hardly more than the room's one.(my bios doesn't have any monitoring tools so i don't know cpu's temperature) But maybe you are right about the heat. Cos i have to consider the "heat resistence" between heatsink and cpu itself. For example maybe the cpu is 50 gradus more than heatsink. And there is notheing i could do. I notice that when i unplug the cpu just right after such stress test - i can feel that the cpu itsef is very hot - much more that heatsink. This problem is because the p3-copermine core is very small. (i think 1/2-1/3cm*cm). For example i have celeron 266 - it has some metal on the cpu - like p4 - very big metal , and that's why it has much lower heat resistence than p3.

I'm assuming this white gunk you're talking about is thermal paste? Anyway, if you've got that between the heatsinks, then you are cutting off air flow to the lower fins. You need to start over with the cooling. Maybe try to figure out how to mount that bigger heatsink by itself. I really think your combining heatsinks is not working. Do you have a thin layer of thermal paste between the CPU core and the heatsink? Lapping the heatsink has been proven to lower temps by a few degrees along with using a good thermal paste. Don't give up, but you need to make some more cooling improvements.

No, i mena that the bigger heatsink is not "avobe" the one 750 has. it is "near - adjucent to one side (which has sharp "cutting" edge) between them is thermal paste for better thermal connection. Each one has it's own cooler. So Second(bigger) one helps to spread the heat to the main(smaller). I can even figure it out from that fact that the same side of main(small) heatsink which is adjacent to second heatsink IS colder, while other side is wormer while work. I found this as a good(costless) solution and made this changes to many my friends' pc's. In most cases you cannot put bigger heatsink INSTEAD default one cos there rarely is free space from all 4 sides, But in most (or at least many) cases there is free space at least at ONE side of default heatsink, where you can attach even just a small heatsink. I know that this method cannot encreas heat spred too much, maybe 15-30% depending on many things. My father has alredy 4 year a comp that i made for him - it is Super Socket 7 mobo, o o'clocked p1-166mmx(66*[email protected]) to 250mmx(100*[email protected]). Leaving it's default cooler with 250 would lead to freezing of pc just 2-3 min after start. Applying add. heatsink using above method soleved this problem permanently (he is still using this pc). Of course Bying some brand cooler (with copper heatsink) would be better (maybe) but , as you can understand this cooler would cost more that that whole ancient p166@250
But this is just my opinion. I personnaly think that one should try to make as much as he can by his hands and buy additional equipment only when he hasen't got choice