• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Uber ghetto dice pot!

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

ivanlabrie

Member
Joined
Jul 29, 2011
Location
Buenos Aires, Argentina
Well, I have two 775 chips and some d9's...no mobo yet, but that's in the works.
I was wondering if I could make a crappy pot with a stock 775 cooler and some plastic pipe/duct tape surrounding it to put the dice in. :rofl:

I've seen it done and a guy had -40c load on a core 2 duo at 4800mhz running spi32m and Vantage.

Any thoughts?
I have no money, and a stock cooler...:popcorn:
 
Try. Necessity is the mother of invention. I would imagine you would want a real good seal if you plan on using acetone.
 
Not sure on the acetone part...perhaps 96% alcohol would be better.
I figure it would make for a pretty good NB pot, but I was thinking the mass is pretty low, copper wise for a cpu, right?
A guy sells pretty good copper pots for 100 locally, but I don't have money to burn atm.
 
the ducktape bond is gonna get rather fragile reguardless of if you use isoprophyl or acetone i would think. I think you would be better off just getting a large copper pipe cap, and either a copper tube or some sort of metal tube, and either solder/braze them or possibly use something like JB weld and then leak test them before use.

A large amount of leaked coolant across a board if its dice or ln2 isn't a great thing, It some cases it in result in cracking of the traces or damage to the solder joints on some of the small stuff. Had this happen first hand once, not so much fun.
 
Yeah that's why I intended this as an experiment with cheap or free components xD
As for the whole brazing thing, that's impossible for now.
It was in my plans but with the price of copper I'll have to do the tube part with aluminum but not sure how that will impact performance...
 
If you do decide to build/test one please oh please stream it!!!!!!! I've been tossing around ideas in my head on how to make a wicked cheap dice pot and would love some inspiration!
 
I will most certainly try my first idea with the stock 775 cooler which will probably end up being an nb pot, and I'll begin to save for a copper base plate with good mass and an aluminum tube to hold the coolant.

Feedback? Not sure how to make the mounting mechanism, I was thinking wood and some long washers...

Also, I'll keep scouting for a cheap or free 775 board to bench with.
 
Last edited:
Yeah that's why I intended this as an experiment with cheap or free components xD
As for the whole brazing thing, that's impossible for now.
It was in my plans but with the price of copper I'll have to do the tube part with aluminum but not sure how that will impact performance...

Aluminum tube, copper end cap would be best, maybe take the copper slug out of the HS for added mass.
 
There are three things that a successful DIce pot needs, to some extent or another.

1) Surface to surface contact with dry ice.
2) Mass, to resist temperature changes.
3) Tight mounting to the CPU.

#1 Dry ice does not like to touch solid surfaces that are warmer than the DIce, at all. When it does it generates a gas layer that insulates it. This is what the acetone is for. It has nearly perfect contact with the dry ice as well as nearly perfect contact with the pot. #1 really comes down to the ability to hold acetone inside the pot in sufficient volume to touch a lot of DIce and a lot of pot surface area. CPU heatsinks are terrible for this.

#2 is fairly simple, a sudden change in CPU load can be soaked up to an extent by the mass of the pot. This is crucial if the pot lacks internal surface area. With enough mass you can spend 5m pulling the temps down and then the pot mass can sustain a load for a bit, even though there isn't enough surface area to hold the temp down. Think of it as a reservoir of cold. CPU heatsinks are terrible for this, too.

#3 is fairly obvious, if you aren't touching the CPU you aren't transfering cold. To actually touch the CPU in a meaningful way you need very tight mounting. Plastic frame heatsinks need not apply, you have to take the mounts off to attempt to use them as a CPU cooler. Metal mount CPU heatsinks are pretty good this way.


As an example, if you compare an F1EE and a Koolance V2 on Dry Ice on a low heat processor, say a core2duo, they're nearly identical. Both have plenty of surface area to keep the pot cold.
If you compare them on a 980x however, the koolance warms up a bit (fairly quickly) under load and then stays there at that temperature, as long as you keep it fed with dry ice. The F1EE does the same thing, but warms up a lot more, as it needs a larger temperature delta between the acetone and the pot to transfer enough heat due to the lower surface area.
For short benches, like WP32 (mighty short on a 980x) the F1EE can actually be better, that heatload for 3 seconds can't heat 5lbs of copper up nearly as much as it can heat the 2lbs of the Koolance (weights are estimates). Over the longer haul the F1EE warms up more than the Koolance, hence the WP1024 result.

A CPU heatsink will work to an extent on a low heat processor, and I mean really low heat. It won't get to full DIce temps due to a lack of contact.

As far as alcohol goes, you won't get as cold with it. The details of that are lounge material.
 
Back