CryoStar Sli and CPU evaporators by Gosmeyer

[dr.rafi]

Suggestions please -

Compressor - NF5.5CLX - VapoLS - r404/r507 - LBP/MBP
evaporator - gosmeyer
Refrigerant - r402a
Cap tube - 120" (3.048M)

Ambient temp - 24.5C

Idle
Air into front of condenser - 24.5C
Air out of back of condenser - 28.3C

Evap temp @ idle - 43C
Load block temp @ idle - 43C

Low side pressure - 4psi
High side pressure - 210psi

suction line temp 6 inches from compressor - -38.5C

202Watt load

Evap temp - -26.1C
Load block temp - -21.5C

Low side pressure - 13psi
High side pressure - 240psi

suction line temp 6 inches from compressor - -3.5C
gas temp into top of condenser - 48C
liquid temp out of bottom of condenser - 34C

220Watt load

Evaporator temp - -25.1C
Load block temp - -20.2C

suction line temp 6 inches from compressor - +3.4C

low side pressure - 14psi
high side pressure - 245psi

240 watt load

evaporator temp - -24.2C
block temp - -19.2C

low side pressure - 15psi
high side pressure - 250psi

It held 240watts for about 45 min and then pressures dropped to 12psi low side and 240psi high side and evap and block temp rose up to 60F and that's when I turned off load.

I would like to be able to handle 250watt load without crashing I have a BIG charge on it now - no way I can add more r402a.

i suggest to shorten the cap. 15 to 20 cm with larger suction container and tilting more upward towards the compressor that push you to even 270 watt

[krullet]

theese evaps feels like art..cant stop looking at it..

really nice work with them

[wdrzal]

AH ha, floworks!! Which orientation of flow gave the best results in the theoretical model, captube in the middle or suction in the middle. I did wonder why the design looked very mathematical, not human at all

Regards

John.

Does Flo-works model fluids in a transitional state? which would be found in a evap.iirc when it was first added to solid works it didn't, but I haven't followed it.

[metalazzo]

Just one question : Have you considered the fact that the inside of the evaporator must be rough (not smooth :p)? How is the base inside the evap? If it's smooth It won't be the must efficient because a tiny layer vapor can separate both liquid and base which involve bad thermal transfert and less vaporization inside the evaporator.

you rosco parrot :p

[wdrzal]

Clemmaster since this insomnia, there is a lot I can't remember until I hear the name. who was credited with this effect ??????

[gosmeyer]

n00b,
Running a pair inverted thru the atmospheric oven tomorrow with 6" pig tails for suc and cap lines. Completely sealed channels. No leakage from one path to the other. Only thing that will need to be brazed by end user is on to pig tails.
More on this later! 5 days to question no 1

Does Flo-works model fluids in a transitional state?

Walt,
No to directly answer your question, only Incompressible (Liquid) and compressible (gas).

[npk]

n00b,
Running a pair inverted thru the atmospheric oven tomorrow with 6" pig tails for suc and cap lines. Completely sealed channels. No leakage from one path to the other. Only thing that will need to be brazed by end user is on to pig tails.
More on this later! 5 days to question no 1




Walt,
No to directly answer your question, only Incompressible and compressible .

And to answer directly, last version I tried didn´t include phase change models.

[johann]

Hey, is it needed to make 3 posts in a row?

The thermal transfer of copper is fast so it dosn't matter how far it is. The surface area of a stepper or spiral is greater. If you made one that was 6" high it would have more cooling capacity than one that was only 2" high.

LMAO you are funny man! Rather dont post if you are going to make yourself look stupid.

Are you trying to get into the for sale section or something?

[[XC] 2long4u]

I see a difference from the model to the real thing.

Have you thought about a fishbone style channel? Something to privide resistance, or turbulance depending on which way the points are faced.

Quote:
Originally Posted by johann
My feeling tell me this evap design will perform better than any stepper/spiral simply because the surface area is where it needs to be, not away from the CPU core. The evaporation and heat exchange happens where its important, not half an inch up from the CPU.

The thermal transfer of copper is fast so it dosn't matter how far it is. The surface area of a stepper or spiral is greater. If you made one that was 6" high it would have more cooling capacity than one that was only 2" high.

Is that better?????Damn! Do you approve of that Johann?

[johann]

The thermal transfer of copper is fast so it dosn't matter how far it is. The surface area of a stepper or spiral is greater. If you made one that was 6" high it would have more cooling capacity than one that was only 2" high.

This is the biggest rubbish I have ever heard. In fact it was so funny it made my day since I couldnt stop laughing.

1) To start with you dont know that a stepper/spiral has more surface area to start with than this design so you are guessing.

2) Your theory regarding heat transfer is complete crap. According to your "thermal transfer of copper is fast" theory, it would make no difference for instance if an evaporator base is 2mm or 10mm thick?? lol

You have never built a pc phase unit have you? So you have never experimented with different evap designs either?

[[XC] 2long4u]

Ok Johann what takes longer to melt a 50lb block of ice or a 1lb block? Same thing. How long would it take to heat up a 1oz block of copper compared to a 16oz block of copper. YOU obviously don't understand the theory behind that!

[n00b 0f l337]

They take the same time if you apply 50x the amount of energy to the ice, and 16x to the copper. But for a single amount of heaet, you are wrong 2long4u.

[[XC] 2long4u]

Please call me 2long for short. And what you said makes no sense to me. Could you explain? What I'm talking about is mass. The mass acts like a buffer. A greater mass won't fluctuate like a evaporator with less mass. Granted it won't matter at full load for 4 hours.

[Unknown_road]

mass acts as a buffer, this is true but what is not true is:

The thermal transfer of copper is fast so it dosn't matter how far it is.

although copper conducts heat very well it isn't ideal, a thick base will give you higher core temps.

If you guys want to read a good book read : thermal physics

it's not very easy, bachelor university level, but it's a very good book.

[SoddemFX]

If you guys want to read a good book read : thermal physics

it's not very easy, bachelor university level, but it's a very good book.

I have that book, a first edition but it's a very good book

Is that one of your course books Jurgen?

Tom

[liloaty]

The thermal transfer of copper is fast so it dosn't matter how far it is. The surface area of a stepper or spiral is greater. If you made one that was 6" high it would have more cooling capacity than one that was only 2" high.

This part is pretty incorrect. If the evap was 6" it really wouldnt provide any real cooling advantage. Although the thermal transfer rate is "fast" as you put it, or its thermal transfer coefficient is relatively high, like the poster above me put it, its not ideal. You will still get a temperature gradient in the copper, and when you get to say, 6" up, its not going to provide much resistance to the bottom of the block heating up. If I get a chance tonight I'll run some quick and dirty numbers and post them.

[Unknown_road]

I have that book, a first edition but it's a very good book

Is that one of your course books Jurgen?

Tom

yes it is.

[[XC] 2long4u]

mass acts as a buffer, this is true but what is not true is:



although copper conducts heat very well it isn't ideal, a thick base will give you higher core temps.

If you guys want to read a good book read : thermal physics

it's not very easy, bachelor university level, but it's a very good book.

I didn't say anything about base thickness. But thats ok. The height is by far more important than the diameter of the evap. The more volumn of refrigerent it can hold the higher heat load it should be able to handle. Or am I wrong?

[n00b 0f l337]

Higher and higher doesnt help as much at some point as all things in life. Overkill is overkill at some point. But I agree though, height can be very helpful if you can use it to increase the surface area to mass ratio.

[[XC] 2long4u]

I wonder if a 3" tall evaporator would be able to put a smack down on the quad?

[liloaty]

You cannot argue height vs diameter, it is about the SURFACE AREA that contacts the CPU. That value is constrained to how large the die/IHS is. Making an evap 3" I don't feel would help the quads all that much, because it is so far away from the CPU itself. If it was that simple, systems that could handle quads very well would already have been made. Its a matter of getting enough surface area on those dies, which, with quad cores is very tough because the IHS remained the same size, if those dies were spread out proportionally to their dual core brothers, it wouldnt be much of an issue.

[runmc]

To dramatically improve performance and ease of assembly the evaporators are now run thru an atmospheric oven (thank you Eric ) at a temp of 1465F. Initial purge is argon then once in the oven it is pure nitrogen and a little hydrogen. The evap is run thru the oven upside down, (stub lines facing down). The cap joint /channel tops / and lines are all flowed it once. When the cap and the base fit together there is .015” clearance between top of channels and bottom of cap. 15% silver 79% copper 6% phop. rod is cut into pices with a total of 2.5gm wt. This is known by surface area of joints . There is zero oxidation. There is NO oxygen at all in the oven. Electric is the heat source. The evaporators are set on a material called fiberfrax good to 3000F. Gosmeyer ran two yesterday on initial trial and it went very well. Eric cut them in half on a band saw to view the joints and as you can see in the pictures (please enlarge image) the brazing alloy flowed very evenly throughout the joints and channels.. The best thing about this process is that it allows an evaporator to truly be as efficient as it was originally designed to be. The refrigerant is forced to travel the complete designed path. The refrigerant can not slip from one channel to another leading to a significant decrease in performance. NO Cross Flow!! Have you ever wondered how much refrigerant slipped between chambers of your evaporators??

No evaporator assembly will be required by end user. Think about how much time that will save in assembly. On top of that, the end user will not have to worry about trashing any evaporators. I know for myself I have destroyed at least 5 evaporator during assembly.

Evenflow

Slick

Atmosphere Oven

[n00b 0f l337]

Damn now thats a beautiful evap. I'm glad your in USA

[sacha35]

Great evap guys well done, would like to test one of these evaps when they are ready.

[Ssilencer]

Have you ever wondered how much refrigerant slipped between chambers of your evaporators??

Yes, that's why we are thermal shielding our evaps since 2005...