CryoStar Sli and CPU evaporators by Gosmeyer

[runmc]

That is a real champ Eric

[dinos22]

why not have the walls thicker since you are making a much thicker base

still checking out the old thread to see some SS benches & some -100C cascade action loaded

i think Xavier (Before) is building some monster cascade >> wonder if he would make good use of this evap

[gosmeyer]

why not have the walls thicker since you are making a much thicker base

Guess I like to walk before I run Some areas of the walls are if you look close, just not everywhere

[dinos22]

Guess I like to walk before I run Some areas of the walls are if you look close, just not everywhere

you seem to be sprinting already

[jinu117]

No need to make walls any thicker... I'd rather see the space there for other things :P
So... if you use that for SS... does that mean approximately 2 hours of lapping? :P

[andre X_X]

why not add surface area on the bottom of channel? i mean you can drill some hole along the channel road. make it bumpy road. just my 2cents

[Raja@ASUS]

I don't think more surface area is an issue, as it stands the evaps need more protection from the liquid turning to gas too early with high loads, ie cascade @ -100 with quad cores. Hence the revision has been made. If LD ever manages to get the rev1 evaps from customs he will test with a different cap tube size, to see if that helps with high loads, from what I have heard the older revision gets hot in the middle when quad cores are heavily overclocked, while the edges of the evap remain cold, the liquid evaporates before it gets right to the source of heat on the core.


regards
raja

[kayl]

Rev3 going out for testing
This has 6mm base and 11mm channel depth and 170g wt as well as center captube
as opposed to 3mm base and 8mm channel depth and 113g wt and center suction of Rev2 ( on the right )
Same assembly process will apply.
Goal is to get a cascade and a quad @-100 with 1.85+V and 5.5+MHz
Oh well, it's only a goal

nice work , the more copper mass should help with load.
Do you have a quad core to test with?
5.5ghz would be insane and take the WR.
the current WR's for qx6700 is around 5.3x and QX6850 around 5.5x i think.
i did have the WR for QX6850 at one stage not any more

[Fhqwhgads6680]

why not add surface area on the bottom of channel? i mean you can drill some hole along the channel road. make it bumpy road. just my 2cents

I thought about this too, but my friend who is a chemical engineer pointed out that this would create a gas pocket and the liquid would flow right over, better off sandblasting to church the liquid/gas a bit.

Looks great guys! looking forward to seeing your results!

[PPInDaHouse]

nice design on the evap, but I think that the walls are very thin. if they were larger it would help on load because its more cooper to heat and then it takes more time.

[gosmeyer]

nice work , the more copper mass should help with load.
Do you have a quad core to test with?
5.5ghz would be insane and take the WR.
the current WR's for qx6700 is around 5.3x and QX6850 around 5.5x i think.
i did have the WR for QX6850 at one stage not any more

Thanks kayl,
Yes I do, but not the cascade or experience it would take to make a run at that. Going to ship them across the pond to get checked out. I know it's asking allot for that to happen (so many things have to be dead nuts) but in the right hands I think it might fly. Oh well, like I said it's only a goal.

[Pete]

Humm i might have to order one of these. I just really really really hate the holddown on them, look cheap! Sorry i don't mean to be offensive .

[gosmeyer]

Hell Pete, I don't find that offensive. In fact I welcome your criticism

[Dualist]

Erics mounts are really good but ok not pretty but functional.
Would you pay and extra $100 (on top of your $200) for a clear plastic mount.? Sorry but I see it as a waste.
I have a cascade with watercooled other bits and pieces so it's a mass of pipes and insulation so a little pretty bit of led lit plastic stuck in the middle of it all is a waste to me. But on a SS it'll be sweet, granted. Each to there own.

[Pete]

Hell Pete, I don't find that offensive. In fact I welcome your criticism

I'll remember that one to keep it to consrucive criticism then, thanks for being understanding

Erics mounts are really good but ok not pretty but functional.
Would you pay and extra $100 (on top of your $200) for a clear plastic mount.? Sorry but I see it as a waste.
I have a cascade with watercooled other bits and pieces so it's a mass of pipes and insulation so a little pretty bit of led lit plastic stuck in the middle of it all is a waste to me. But on a SS it'll be sweet, granted. Each to there own.

Yes his works and yes it's simple and form over function is not needed all the time. However with such a pimp bit of kit you kinda want it to be. Something like Chilly1's 2007 hold down or a mount that Duniek makes even Kayl would be better looking. An @Inatour would be the top dog but there we go.

It just personal prefrance and i'd like to get one but i have to have it with encoloser sadly

[gosmeyer]

i'd like to get one but i have to have it with encoloser sadly

Why is that can't you just mod one of the others?

[Pete]

Why is that can't you just mod one of the others?

Heard from manny others you have to have both from the shop not evap only! In all honesty i was just looking at a way of being a cheap skate no offence as i have an encloser i wana use

[gosmeyer]

See what we can do for evap only. To be perfectly honest I would love never to machine another encloser. It contaminates my coolant too quickly.

[Pete]

See what we can do for evap only. To be perfectly honest I would love never to machine another encloser. It contaminates my coolant too quickly.

Tell Ron that then! I'm sure i know one man that will be more than willing to do an encloser for you

[killermiller]

You guys should offer "step up" program

[jinu117]

Well testing just started Finally got evac down and getting to it...
This is same evap Ron was playing around with almost same length of cap tubing i believe. Might need to adjust it later so could take few days to get to final verdict. After all compressor used here is NEK2134GK much different animal than NF5.5clx.
Should be interesting
The rubber mallet on right side is there in case I get really frustrated (nah... just to knock compressor around when evac'ing) Man I need cheap heat lamp :P

[PhilippF]

Man I need cheap heat lamp :P

Heat lamp should just be a fraction of what the manifold gauges did cost

[TopherTony]

(nah... just to knock compressor around when evac'ing) :P

Why do you need to do that?

[Pete]

Digital manafold gauges, Jinu you pimp

[jinu117]

Actually I've been using this toy for some time now... And hate using regular manifold gauge if I can help it. It gives me everything I need in short glance more accurately than analog ever will. As for heat lamp.... I got the gauge when I had some $$$... Right now, I need to pinch save on anything I do to make sure my family is fed/have housing, etc. E-bay might be route to look at. It really isn't easy time for me right now but I will get through it
(I reckon push comes to shove I could sell all my tools and live off of it for at least few months.. :p nah)

Back to the topic...

I am quite delighted at result of this evap.It does have some very good property that I have been looking at evaporator.
Well, let the data show you first.
I have forgone evap temp on this one as it has no bearing to me for figuring out few very important thing in high load handling SS... (namely superheat/subcooling/evap's outlet temp to suction, etc, etc).
Suction line temp around -33.5c, load block temp around -18.4c. Some other block I've used which has been heavily modded would have had same suction temp and load block temp around -8.5c. Oh in case I forgot it is the damned 250w load... This probably roughly translates to about 3-5c better CPU temp under that kind load condition. However, since I didn't have quad to test, I can't say for sure how it would have been in actual CPU granted heat pattern might be different than load block.
Few things I have noticed.
#1. Block is exceptionally finicky... very fast reacting to change in load, etc. I think this can have negative impact on stable overclock situation or extreme end of benching. We have already talked about this before and to be honest, rather than making base 6mm, I would have preferred to have seen 4mm base with taller wall to compensate for mass distribution. If all we needed was bottom thickness, we could as easily have put cold plate and tested it for that :P I believe anything beyond 4mm is going to be detrimental on some good characteristics of this evaporator especially regarding cooling done by surface area on side as heat won't travel much to there.
#2. I think keeping the liquid flow in from outside to inside idea is for the keeps for single stage at the moment. This will give it best chance to boil off most of refrigerant before it hits the suction port of compressor from where I see it. Starting from middle might not aid in full evaporation and combined with blocks fast reaction, could cause floodback while fully operating.
#3. Cap tubing for my compressor probably is much shorter from what I see. I saw very smooth transition of charge vs. capacity till I hit 200w. After that I had to put a LOT more on to get to 250w. Will have to see static pressure after turning it off on testing. I doubt I will get better temp than what I have but it should help on safety & balance of system.
#4. Evap is quite large on diameter. It probably is needed space in current state to be able to have enough space for both liquid and gaseous refrigerant. With depth increase though I am thinking maybe it can be changed...
#5. Once flood back test passes (I get the feeling there will be some after amount of gas I put in), I will retune the unit for shorter cap to see what it holds.
That's all I can think for now... 20-30 more minutes of measuring and time to hit the sack.