Dragon Evo

[afireinside]

ROFL that's tiny.....nice craftsmanship/idea for sure.

what is the advantage or disadvantage in terms of performance while comparing tall and short post with DICE/LN2 cooling

No performance change. Just if you use pellets of dry ice the height lets you fill it and forget it.

[Entsafter]

A damn nice pot!
I'm out of my mind when I can hold mine in my hands...

btw: Special thanks to k|ngp|n that he use some extra time to produce 3 of them till the end of week -> overclocking team switzerland has to thank you and we try to achieve some great results.

[dinos22]

No performance change. Just if you use pellets of dry ice the height lets you fill it and forget it.

i understand

what difference in benching time would a standard height pot of KP's last with DICE compared to the small dragon with full pot

[dinos22]

isn't the base designed for LN2

good tip about LN2 benching and pot walls...first time i hear that

[k|ngp|n]

OK...benched DI all day yesterday in SLI on the EVO 2.2 unit. Keep in mind my best conroe b1 stepping isn't the best clocker out there, but on a scale of 1-10 maybe something like an 8. These results are typical of what one could expect using this unit,the best DI benching method, and a decent clocking chip. All tests were done using vga's on air with def gpu clocks and not a whole lot of tweaking. Really just testing/emphasizing max cpu mhz/vcore using a good slushy DI + acetone mix. Now I learned how to bench DI like this from one of the best DI benchers in the world. This is the only way to bench DI IMO.
The results are petty damn good I think for a unit which is really engineered for ln2 benching and is in actuality only a 3inch high copper unit , as the aluminum does nothing really for better performance but really helps with DI functionality.

Conroe b1 6800/evo 2.2 1.76v-1.77v actual vcore
EVGA 680iSLI/air def coolers
Corsair 6400C3
Silvestone 850w zeus

Now a little info on the DI/acetone benching method used. This is how you get your best loaded temps using DI + acetone.
To start, take maybe 5-10lbs of DI put it in a bag and smash it into mostly tiny chunks and powder using a hammer:


Fill your container with the powder DI using a scooper or something maybe 3/4 full. Then pour in about a full shot of acetone AND NO MORE! You want the mix to have the consistency of a thick slush:


Another very important thing with DI benching is under heavy load the bottom part of the mix will evap off quicker and form air pockets near the base of the container(expecially in a unit with a more complex base). You need to constantly work the air pockets out of the mixture while you are benching using some sort of a tamp. I use this steel rod, and I am constantly packing it down, refilling, and stirring out the air pockets mid bench. If you dont do this you will lose you best base temps under load, and wont get the best results possible:


Ok once the mix is set, time to start clocking...
All runs done setting clocks in bios.
Set 407x12 in bios for 4884mhz and around 1.76vcore and the bios/meter temp comparisons are looking pretty good for start.
Approx -70c meter = -61 bios:



Start off with some 1M. I really think I could have managed a 4900mhz run with some things disabled and some more effort, but didnt want to spend to much time running a pretty meaningless benchmark when it comes to load testing a dual core cpu:


Best 8m I could manage....I could make it to last few loops about 40mhz higher, but it would error out. Again som more time spent here would yield some better numbers I think:


Now for 3d testing. I had to raise vcore one bump to get most of the 3d benchmarks to fully pass compared to pi. Pi could pass at lower vcore than shown. The 01 for sure could have gone a little higher I think as I ran 01,03,05 pretty much at the same speed.
Again, all 3d tests run with def gpu clocks.

01:


03:


05:


06:


Absolutely no leaking with the two-piece design and some pretty decent results I would think using DI + acetone. This container does it's best work with ln2, but this testing shows that it can run DI pretty damn good too.
Then when all finished, just power teh rig off, disconnect everything and turn mobo upside down with unit attached to dump out mixture:

[TaPaKaH]

thanks for sharing

[G H Z]

Great post K|ngp|n, I use a wooden back scratcher I got with a drink in Hawaii to tamp the mix

These results give me some real hope for my 6800 and prove the efficiency of the Evo design.

[LittleDevil]

Great and nice pics

Regards

[Misko]

awesome temps, great info.

[massman]

KP: Will you test the tube with ln² as well (as the tube is designed for ln&#178?

@ Pallmall: acetone is certainly the most used liquid when benching with dry ice. Basicly you just need an alcohol-based liquid.

[k|ngp|n]

Hey king what difference would it make if you used pure methanol?

Last time i used pure methanol,but i was wondering if the acetone is really the best out there?

Never used it. Nasty fumes?

[k|ngp|n]

nope.

I always used this because the best overclocker in my country said it was better then acetone. But it has to be the purest of the purest, straight from the factory.

Ofcourse as i haven't tested it myself to see the difference i cannot tell, but i was hoping you also tested it

Acetone hasn't let me down yet. I just try to find some with the highest percenatge purity. A slushier mix definetly works better with this unit than a chunkier/snowier one does. I tried it alot of diff ways and slushy/packed down hard gave the best base temps. I think same hold true for any unit and benching DI this way. You really have to work the container while benching making sure there are no air pockets forming down there. Even just idling in windows after bench no load at these speeds. Better temps with it packed down tightly.

[ozzimark]

You really have to work the container while benching making sure there are no air pockets forming down there. Even just idling in windows after bench no load at these speeds. Better temps with it packed down tightly.

how about making a wire mesh thing to simply push down on the mixture to force the bubbles up?

[k|ngp|n]

how about making a wire mesh thing to simply push down on the mixture to force the bubbles up?

I like that idea alot! Is that what you use man? Sort of like a french fry maker that you would push down on a potato? Only in place of the sharp wires on french fry maker, have a stainless mesh.
I could machine a piece of aluminum (rod with flat circle on one end) to maybe a hundreth under the inside diameter of the unit, and put some reliefs into the tamping end for air to pass through. That would work great. Make it as heavy as possible so you you could leave it in the container while benching to keep the pressure and compacted DI on the base. Not sure how much I would really use it though

[Misko]

Maybe better some sort of mixer, wire on a el. motor so one man could do the bench by himself, Kingpin how short is period between air pockets starts to show up?

[afireinside]

I think same hold true for any unit and benching DI this way. You really have to work the container while benching making sure there are no air pockets forming down there. Even just idling in windows after bench no load at these speeds. Better temps with it packed down tightly.

Yep! Exactly why I don't fill past the acetone, makes it easier to pack it hard. I use a wooden spoon to do it.

Thanks for testing Vince, the results are great! Now I just need a conroe and some money... BTW where did you get the nvidia logo steelpad Qck?

[NickS]

Aw, I want an nVidia logo on my Steelpad S&S

[Ad Rock]

I like that idea alot! Is that what you use man? Sort of like a french fry maker that you would push down on a potato? Only in place of the sharp wires on french fry maker, have a stainless mesh.
I could machine a piece of aluminum (rod with flat circle on one end) to maybe a hundreth under the inside diameter of the unit, and put some reliefs into the tamping end for air to pass through. That would work great. Make it as heavy as possible so you you could leave it in the container while benching to keep the pressure and compacted DI on the base. Not sure how much I would really use it though

You mean something like this with a finer wire mesh on it? I think its a great idea, and would not be that hard to make either.

[T_M]

So long as you can pull it up/out without sucking up your mixture

[ozzimark]

You mean something like this with a finer wire mesh on it? I think its a great idea, and would not be that hard to make either.

yup, that's almost exactly what I was thinking of

T_M, you have a good point, I didn't think of that. I guess hands on testing will tell the story

[k|ngp|n]

any updates on the other pot vince?the one you could'nt find a name for

Still testing it. It works well so far. Similar peformance to the evo with ln2. Definitely designed with ln2 in mind and not really DI on that one.

Some units going out tommorrow . One is going to be used at the Abit event this weekend to be tortured with LN2. Among the other units is one going to another overseas ripper who was a huge influence to me in overclocking...and I am so honored to have this dude use it. I really hope he likes it and can remind everyone he is still around with his results

[T_M]

Opp?

[Stapler]

KP, what do you use for your mounting hardware?

[afireinside]

Among the other units is one going to another overseas ripper who was a huge influence to me in overclocking...and I am so honored to have this dude use it. I really hope he likes it and can remind everyone he is still around with his results

Would this dude happen to reside in Finland?

[blind_ripper]

that woeld be macci