Archon

[gdesmo]

Physics at work here, as temperature rises so does the resistance, The molecules are a little further apart due to expansion. To get the same power it requires a slightly higher voltage to push it's way through. Conversely when chilled the voltage needs are lessened as everything is a little closer together.

[xmanrigger]

Physics at work here, as temperature rises so does the resistance, The molecules are a little further apart due to expansion. To get the same power it requires a slightly higher voltage to push it's way through. Conversely when chilled the voltage needs are lessened as everything is a little closer together.

Thanks, I will take your word for it, as I am not a electronics engineer. Just sounded kinda fuzzy to me and have never really experienced that. Even with sub-zero cooling.

[Cyclops]

Thanks, I will take your word for it, as I am not a electronics engineer. Just sounded kinda fuzzy to me and have never really experienced that. Even with sub-zero cooling.

It's true. At a constant voltage, a chip might be able to do 4 GHz at 50C. The same chip can probably do 5 GHz at -50C using the same voltage. That's why all world overclocking records are set using liquid nitrogen/helium.

[xmanrigger]

It's true. At a constant voltage, a chip might be able to do 4 GHz at 50C. The same chip can probably do 5 GHz at -50C using the same voltage. That's why all world overclocking records are set using liquid nitrogen/helium.

True, a lot of records set with LN2 and helium. But look at the voltages they are pushing! The voltages they are applying would instantly fry a cpu if cooled with anything less. It is the voltage that is causing the high temperature more so than the high clock speed. And to attain those clock speeds, you need a retarded amount of voltage which requires extreme sub-zero cooling to achieve.

I am not trying to argue gdesmo's theory, but go have a look at the voltages those records were set with. Those voltages are unattainable without ln2 or helium.

I can run my 3930K at 5.0ghz (1.63v) with only ambient room temp water cooling, all day long. I have had it as low or lower than -10C on the die, and still couldnt hit 5.2ghz without jacking the voltage to =>1.7v, using sub-zero liquid and air. That kinda throws a f**k into the argument doesnt it?

Please, educate me. I am not saying all of you are wrong and I am right, but this goes against everything I have learned and done with overclocking. A few years ago, I consistently set the highest VGA clock speeds in EVGA's forum on air or water. Only clocks higher than mine were LN2. Actually, my highest VGA clock speed in that forum was done with sub-zero air cooling. Never fried one card.



EDIT:Sorry Grinder. Can a mod please move this discussion to a new thread? I had no intention of hijacking this thread, but would like to continue the discussion in the forum.

[Grinder]

No worries xman, I suppose I got this ball rolling when I asked 10e how much delidding dropped voltage.

And the reason you don't hear too much about this is that people who use extreme cooling aren't looking to merely lower voltage at the same frequency, but to max the frequency. The chip may have a limit that it just can't get past. But who wants to waste dry ice, liquid nitrogen or liquid helium to keep the speed the same and just run at a slightly lower voltage? Maybe if you were using single stage or peltier, but both of those cost you electricity so what's the point, unless you were worried about frying your chip with too much volts.

Once I set up water on my other rig, with the rad in the garage at around 1C, I'll see if I observe much difference in the power required. Or you could try your 3930k at 5Ghz with -10 on the die, and see if you need less than 1.63v.

*Edit* Also, if I end up delidding my 3770k I'll see if it lowers the volts required for 4.7Ghz.

[xmanrigger]

Cool, I am actually trying to behave myself here now.

Unfortunetly, Iam working outn of town likely until spring time. When I get home, roads could be dry, and it will likely be bike season, so not sure how much time I will put into it.

[Grinder]

Received a small order from Daz with some stuff for my Blood and Guts build, plus the 4 wide angle white LEDs for the Archon fan mod. Will try to post some pics later today, since that's the last cosmetic item.

[Grinder]

As promised, some pics with the LEDs installed on the fan mod, sleeved extensions and ram fans installed. Also tidied up some cables.

*edit* Ugh, forgot to move the ram fan cables...

[Hooded]

Looking good bro.

[nass]

ya looks nice...

[10e]

Looks great. Very nice B&W build, Grinder!

You put all the LEDs in perfect spots.

[Grinder]

Thanks guys! I guess unless I decide to delid, I'm done for now.

Finally I'll return my attention to my Blood and Guts build.

[bungwirez]

I really like this one Grinder. congrats.

the back lighting makes it all spectacular.

[Grinder]

Thanks bungz. Not including the lights on the cpu/gpus/fanmod, it was $17.99 well spent NZXT 2m LEDs was the perfect length to do the entire perimeter of the case, plus has on/off and low/med/high switches. First time I've had a window this large so had to show it off. Hard to believe the side of the case is attached in these shots.

[Grinder]

I've removed the mobo, cpu, and both 290's from this build in order to use them in my H220 build. I found a kijiji deal whereby I traded my 990FXA-UD3 and 2x4GB of 1600 C7, along with some cash, for a Crosshair V Formula-Z and 4x4GB of 1866 C9. The 1866 ram is also going to the H220, and Archon gets the C5F-Z along with the 8350 and 290X from the H220, and another 290X and 140 rad purchased from vrdublu.

I'll return to this once the parts have arrived and the dust has settled on the H220!