More Voltage Not Always the Best Idea when Overclocking...

[dasickninja]

Taken from the Tech Repository. All credit goes to those masters of the art. Its an intresting read for all owners of Core 2 Duo's that engage in overclocking...

Linkage Maaaan.

 

[ajfink]

Typically when I OC I only go as high as I can on stock volts or with very minor voltage increases. Voltage is bad for chips. I undervolt CPUs I plan to have at stock.

 

[dasickninja]

Same here, we typically test how high we can overclock from stock voltages as a true metric of the potential of a chip. So if a 3060 can clock higher than a E6600 equivalent on stock voltages, we know that the 3060 is truly a higher binned chip. But anyway, the article is still a pretty nice read for all technically minded, and for new overclockers hat crank up the volts just to get to clock.

 

[m25]

I always like to be a stock-voltage overclocker; have to use RMClock to get the Vcore to 1.35V because this stupid board only supplies the highest allowed voltage for a CPU abd sent my 4200+ to 1.40V by default and with the 3000+ @ 2.0V I could unervolt it to 1.25V.

 

[apt403]

Thats a bunch of BS, i have my E6300 running with a 5.7v vcore. Its doing fine!

 

[dasickninja]

What the hell voltage mod would you need to do to supply that much to the Core?

 

[gOJDO]

What the hell voltage mod would you need to do to supply that much to the Core?

4 PSU's in a serial combo

 

[apt403]

A car battery and a pair of jumper cables.

 

[tool_462]

Very cool article. I've read some "technical spec sheets" and they are way to confusing for me. Never had any schooling in stuff like that.

If I understood the FSB thing, that somewhat explains having "FSB walls" like I do on my ABIT NI8. I can run 175 x 20 and 195 x 20 but anything in between is either unstable or wont even boot. I've heard of the DS3 having a wall somewhere around 330 too.

 

[dasickninja]

Very cool article. I've read some "technical spec sheets" and they are way to confusing for me. Never had any schooling in stuff like that.

If I understood the FSB thing, that somewhat explains having "FSB walls" like I do on my ABIT NI8. I can run 175 x 20 and 195 x 20 but anything in between is either unstable or wont even boot. I've heard of the DS3 having a wall somewhere around 330 too.

Yep. The tech specs can get a bit woobly woo.. about those walls.. what about adjusting the RAM to high latency to squeeze some stability?

@apt403 & gODJO
Fockin wow.

 

[will14]

Yeah @ 5.7 vcore I think you'd need phase change or some crazy cooling.

 

[apt403]

Yeah @ 5.7 vcore I think you'd need phase change or some crazy cooling.

I dont think a processor would even function with a 5.7v vcore, no matter what kind of cooler you have. Its like a toddler trying to snort a fist full of uncut coke, its just dead.

 

[joefriday]

I think a Pentium 100 could handle it.

 

[ajfink]

I upped the volts a little on my overclock. I am kinda wary of adding all that much though. Went from the 1.35 to 1.4. Still a good deal less than what my 3000+ ran at stock. It ran 1.5 stock and 1.525 on overclock.

I think your OC is fine. Five hundredths of a volt increase for that dramatic of a clock gain seems...fair.

 

[kwalker]

This is a good article and points out the fact that most of us know from experience that the termination voltage at higher levels does little to the stabilization of the FSB frequencies in fact the north bridge can heat up to the point of degraded performance.
This is the point of the article that is of importance.
You can have a much higher FSB with less than 1.4 Volts Termination with less Vdroop or worse (extreme overshoot).
It is better on some boards to set this manually because auto configuration can increase the termination voltage above 1.4v causing an overclock failure.

 

[dragonsprayer]

I second that for the most part - 1.4v p4's run good at 1.45v usually 1.43v real

965 1.3v max runs good at 1.33v

core 2 duo 1.4-1.45v

 

[Eurasianman]

Good article.

Too techy and wordy for me to read :tongue:

Lamen terms?

 

[dasickninja]

That a high vCore can actually make a overclock unsuccessful and unstable. Contrary to popular belief, keeping the vCore a few points lower, such as at 1.35v instead of 1.41v can contribute to a more stable, and possibly higher clock. This also means that voltages can be lowered significantly at stock speed to further the lifespan of your chip. Now applying that to the high end, you can see a higher clock head room for Conroe/Allendale/Kentsfield chips.

Just imagine how high it will be when Penryn comes out.. ()...

 

[Eurasianman]

So, are you saying that if you want to find a happy medium overclock and not harm the lifespan of your CPU, basically, only overclock your CPU until it requires an upped voltage.

Or, lower the voltage and see when it can run fully stable at stock speeds...

Sorry... I haven't had enough caffeine yet...

If so, why in the world didn't Intel do that in the first place with their Core 2 Duo? Wouldn't this show even more power consumption efficiency?

 

[dasickninja]

The chip has much higher tolerances when it comes to v/clock than anyone anticipated. This is freaking awesome...

 

[kwalker]

Actually the article is talking about VTT (termination voltage) not Vcore.
Termination voltage increases the current flow through a given circuit.
Raising the Vcore to stabilize your overclock is acceptable as long as it stays within the tolerances values at 1.55 absolute maximum.
Raising the termination voltage can be the harmful aspect of the overclock setting the fluctuation of voltages in a given circuit too high or too low.
It all depends on the quality of the chip you’re trying to OC as to the amount of Vcore needed to run at an optimum FSB.
The article explains the effects of Termination circuitry not Vcore.

 

[dasickninja]

Well fock me.. I've been schooled once again. But what I say is true.. every last word... except for the 50th word.. that was a lie.. ^_^
Thanks for the correction Kwalker.

 

[kwalker]

8)

 

[Major_Spittle]

I started reading it, but quit after I realized there were no pictures of monkeys.

I like monkeys.

I took electronics 8 years ago, when electrons always flowed counter clockwise.

 

[flasher702]

Taken from the Tech Repository. All credit goes to those masters of the art. Its an intresting read for all owners of Core 2 Duo's that engage in overclocking...

Linkage Maaaan.

holy crapzors they are much l33ter than I...