CryoStar Sli and CPU evaporators by Gosmeyer

[Dualist]

Jin, are you load testing the evaporator with a flat plate load simulator?

[jinu117]

Huh? Flat plate load sim? I just checked static pressure and it seems to be overcharged. However, behaviour under low load, etc was quite good. Could be that suction accumulator forcing it. (which I prefer not to have if possible). Static pressure sits at around 95 psi right now. Time to recover and rework cap tubing a little to see what diff it can make.

[n00b 0f l337]

He means is your load tester flat, which I beleive it is, vs something milled to be the size of a cpu IHS.

[jinu117]

It is flat but not size of heat spreader :P Probably tad larger. I need to get my mill back in action to do that but waiting for few parts at the moment... or money to get that :P (It has to do with health condition inside shop and I don't want to be knocked back out again).

You will see when I get the setup fully ready to rock as there are quite a few NC code I need to run through.
Now, back to topic on evap.... More I think of it, more I think total flooding is out of question for heat capacity unless we want to risk the liquid slug back. Problem is, flooded evap by nature are much more efficient than mist wetted evap... at risk of flood back of course. There probably need to be some measure of controlling this flood back without using excessive sized accumulator... Also, larger sized spun type accumulator can possibly contain enough oil to have oil slug back happen (if you have enough superheat but there is slugging, chances are, it is oil slug backing)... which is just as bad as liquid refrigerant slug back.
Now, as for foggy evap style which this particular evap is designed more likely for it from what I see, it seems to handle the load just as well as flooded type (probably due to lack of space in flooded type to correctly control behaviour of flash gas)... I remember some EU builders using AEV at one time with good success which makes me think we still have room with dry evap idea. (AEV never floods... it's always mist or foggy)
The more I think about it, the more variables keep coming into question without solid thermodynamic calculation to back things up... Maybe one of these days, I will go back to school in that side of skill sets. Thermodynamics engineer on top of Computer Science and Engineering, HVAC/R, than thermodynamics... there is so much to learn in this world.
Enough babbling for now... cap tubing has been resized and pulling 3rd evac with heatlamp (nah... just pulled out 500w halogen lamp next to compressor... I hate idea of using 500w for such long time considering electricity but hoping it is counter balanced by shorter evac time by pump...)

[jinu117]

Quick update: After swithching the cap length a little...
Idle: -44c load block.
50w: -42c load block.
100w: -36c load block.
150w: -30c load block
200w: -23c load block
250w: -18c load block.
One thing is extremely clear from the pattern... it is handling load in very linear fashion. Which is what I would like to see.
No flood back on this situation as well as first test. Superheat is tad lower now that it is behaving more like it should. I am wondering about effect of over suction heat exchange at the moment as it is definitely putting more strain on compressor and condenser than I think it should. Around 250w is the max I would tune my system for using the current condenser (which is incredible to begin with but proper tuning dictates that I need even bigger condenser beyond this point). Desuperheater is viable option for another 20-30w but I rather not risk few undesirable effect... besides, can't imagine benefitting more than 250w on actual CPU with temp drop vs voltage increase (just like I used to tune for best temp around 180w in time of X2s, and 200w if requested).
Cascades, I can see the need for 300w capacity which can't really easily be reproduced from single stage as gas utilized is vastly different... maybe one of them is in target for my next cascade build which is planned up.
One thing I would like to really see in this evap actually is the size reduction if possible on diameter while increase in height with chamber depth. Total inner volume greater than what it is now, total mass being greater as well without adding too much height on bottom plate part of it. Instead of keep increasing bottom plate, I believe thermal inertia can be obtained by higher wall height to add in the mass. Another thing I would like to see has already been personally mentioned to Ron and Goss, so that could be viable option if it IS practical for production process in brazing.
I wonder how newer revision (believe I am playing with Rev.1) works like...
I might have to change my production with NEK on this evap while leaving other line of NF series with previous evap. Evap works wonderfully with some characteristics of NEK compressor I didn't quite like too much and makes strong combination that NF compressor probably won't be able to touch.

[Fhqwhgads6680]

Evap works wonderfully with some characteristics of NEK compressor I didn't quite like too much and makes strong combination that NF compressor probably won't be able to touch.

so wait, you are saying that the NEK's work better with this evap? Just curious and slightly off topic but do you feel these are much better than the Nf's? I am trying to decide what compressors I want to go with, and haven't seen many using the NEK's.

That new cascade rev. looks nice guys! keep it up!

[jinu117]

so wait, you are saying that the NEK's work better with this evap? Just curious and slightly off topic but do you feel these are much better than the Nf's? I am trying to decide what compressors I want to go with, and haven't seen many using the NEK's.

That new cascade rev. looks nice guys! keep it up!

It will work great with NF too but this evap design REALLY helps showing true nature of NEK. I was a little miffed before by NEK vs NF9FX considering it is newer, etc... on paper, it should have about 20-30% more capacity but that wasn't the case. Probably about 5-10% was what I was able to garner. And add to the fact that I can't use it on vapochill LS modification, having to have 2 inventories (drives up cost in many ways), I was almost ready to move back to NF9FX/NF10FX... With the new evap though, I am getting the about 15% of it :P
Now my limitation has shifted from compressor to condenser... and I think I am stuck on condenser side of story due to size issue. Pretty much as efficient as it can possibly get condenser design... (rifled tubing inside, optimized fin density, optimized tubing size, optimized flow pattern). And I am paying a lot for custom making these to begin with -_-;
One option is to add desuperheater, and another option is to include faster fan..
Basically, by increasing airflow, I was able to get load block temp another 2c. (about 30 cfm more than the panaflo m1a that was in there). I would estimate desuperheater can probably do same too.
- Yes... the 250w testing was done with Panflo M1a which is as slow as it gets in phase change cooling ahaha :P I think I forgot to mention one important thing.

[DagoDuck]

Today my CryoStar evaporator from the USA arrived. Thanks Eric and Ron! I will show pictures and results, as soon as the unit is finished. Waiting for refrigerant (R507).



Best regards,
Chronis

[runmc]

Today my CryoStar evaporator from the USA arrived. Thanks Eric an Ron! I will show pictures and results, as soon as the unit is finished. Waiting for refrigerant (R507).



Best regards,
Chronis

Thank you for your patience - Did you also get the filter/driers?

[DagoDuck]

Everything is fine, THANKS! I received the filters/dryers. The package arrived very fast, about 7 days. And NO problems with the customs administration. I did not have to pay anything.

[jinu117]

Okay, quick question.
How are evaps cleaned after sand blasted?