Pot vs direct contact with Dice

[Disparaitre]

I'm considering a possible Dice run this coming week and was wondering if it'd be feasible to cut the Dice to about the diameter of the cpu ihs and make a rig to keep pressure on the Dice forcing it down onto the. Cpu or do I really need a pot?

[Drdeath]

The answer is simply don't do it. You need a pot. If you need advise on how to line your board let me know. I discover an easy easy way to prep the board.

 

Edited May 16, 2011 by Drdeath

[Locutus]

Yeah, you don't want to run DICE without a pot.

[Drdeath]

Hey Ward, I live in Schaumburg, Il. If you wanna dice let me know. I have a pot and can come by if you wanna try it. You cannot keep it though

[Nyt]

Yeah definately use a pot ... Im also going to go sub-zero later this year ...

[Disparaitre]

Well, why I was even wondering is, the CPU core has a layer of tim, then theres the IHS, then another layer of tim, then the pot with the shaved Dice resting in it. Thats 4 layers where the conduction of heat is becoming less efficient as you go. I figured making an unsulated square copper tube open on both ends, loaded with a square cut length of dice which would be forced onto the CPU via a plunger.

 

Honestly, I understand everyone is used to using a pot for Dice and it's kinda a necessity for LN2 but not for Dice. I just don't see any risk cutting out the middle man for direct contact which tbh, still isn't direct as there is still the thermal paste between the IHS then the IHS itself.

 

So far I think the whole pot thing is overrated for Dice if the Dice can be cut to just under the dimensions of the IHS and forced directly onto the IHS instead of being shaved, and letting gravity do it's thing this cannot in any transfer the temperatures from the Dice to the put like Dice being forced onto the IHS like I am proposing.

 

Can someone backup the actual pressing need for a pot despite my proposal? I see absolutely no downside of using my method outside of a limited ammount of time to bench which I can make tweaks under multiple runs as I immagine a 6 - 8" bar of Dice should give me at least 5 - 10 minutes of benching unless I can make a reloadable system that will allow me to keep adding Dice rods which I am already trying to work out.

 

I do understand that with my method the Dice will not last as long as it will come closer to the heat source than if I used a pot but the IHS should reach lower temps with direct contact than they would with a pot setup.

Edited May 16, 2011 by WARDOZERX

[Disparaitre]

I just had a knucle headed idea. This would turn my Q6600 into a work of OverClocking art as well I can solder the copper tube directly to the IHS Make a Dice slurry and voila. Refillable, closer to the heat source and It's basically a pot, only one that is right on the source cutting out the layer of tim between the ish and the pot and of course the pot itself.

 

If I make a diethyl ether Dice slurry it can get almost -15* C cooler than Dice alone.

 

The only downside I see if while I can and would have to solder the tube to the IHS with the CPU clamped down, removal of the tube once soldered will be near impossible without destroying the retention mechanism.

Edited May 16, 2011 by WARDOZERX

[ARandomOWL]

The key thing with a pot is that it provides a lot of mass for the CPU to dump heat into whilst it's load is changing. The CPU's heat output varies greatly with load. With DICE directly on the IHS there will be very little mass to keep the temperature swings of the CPU under control.

 

In conclusion, I don't think your idea would do much better than having no heatsink at all. Although I have not tried this, nor know of anyone who has. Nothing like testing it for yourself

Edited May 16, 2011 by ARandomOWl

[Disparaitre]

The key thing with a pot is that it provides a lot of mass for the CPU to dump heat into whilst it's load is changing. The CPU's heat output varies greatly with load. With DICE directly on the IHS there will be very little mass to keep the temperature swings of the CPU under control.

 

In conclusion, I don't think your idea would do much better than having no heatsink at all. Although I have not tried this, nor know of anyone who has. Nothing like testing it for yourself

 

 

I think I actually will test the soldered tube thing. I'll try it on an AM2 setup first as with an AM2 chip I have a larger IHS to work with and with the AM2 locking mechanism I have a good chance of being able to remove a CPU.

 

I'ma start getting stuff together for this because I would indeed like to try this though I doubt even with good success many people will be willing to solder mini pot's to their CPU's for short benching runs.

 

And using a Dice slurry the method I'm proposing will have the exact same effect as using a pot for Dice on a lidless CPU. No difference at all aside from the diameter of the pot obviously.

 

. . . . . . I'll be damned, I was right all along, there is no downside to direct contact with the IHS, soldering a tube to the IHS because it literally IS the exact same thing as using a pot ona lidless CPU, the IHS becomes the bottom of the pot is all so damn myself for not seeing that sooner but my proposed method has in a way already been done with everyone who ever used a Dice pot on a lidless CPU

 

Kinda funny how that worked out eh? I guess now it comes down to, is there really any difference in what the diameter of the pot is if a slurry is used in place of shaved dice?

 

Almost makes it not worth going thru the hassle with all the soldering seeing as in the end it's no more original than deliding a CPU then using a Dice pot I was kinda hoping to do something with more originality :'(

Edited May 16, 2011 by WARDOZERX

[ARandomOWL]

. . . . . . I'll be damned, I was right all along, there is no downside to direct contact with the IHS, soldering a tube to the IHS because it literally IS the exact same thing as using a pot ona lidless CPU, the IHS becomes the bottom of the pot is all so damn myself for not seeing that sooner but my proposed method has in a way already been done with everyone who ever used a Dice pot on a lidless CPU

 

As I said, I think mass will be the problem. A "light" pot has more than 1KG of copper to absorb the heat load. A decent pot that can handle the load of a Q6600 at 1.8v has 2 or more kilograms of mass. You might be able to cool the CPU down to a low temperature at first, but it won't stay there for long when you put any load on it (even just idling at high voltage).

[Nyt]

Yeah a heavier pot is always better . A ln2 buddy of mine can sell me a 2.65kg pot for only $150

[Disparaitre]

I would make my own. I have done it before a long time ago and it's part of the fun. A good frind of mine's father has a somewhat basic metal lathe and other equipment that he usually uses to make simple replacement parts for his tractors and he has other equipment like nice drill presses and such so though I don't have access to a machine shop I do have access to people good at welding soft metals, basic machining equipment and I can get the materials easily enough and save myself almost $50 - $100 vs buying a pre built pot.

 

I was more curious than anything if it'd be worth it or the difference and was hoping to do something relatively original but using a direct contact method isn't much different than all the people that used pot's on lidless chips

 

Thanks for the input all