Evaporator Performance

[mEsUsah1]

I'm trying to design my own evaporators for CPU and GPU cooling. But have no experience in how to do a cuple of things. I'm a capable CNC programmer, and know my way around a workshop. When you guys design yours, what "rules" do you follow conserning:

1) Wall thickness. I guess it's something about the area and pressure...

2) Overall flow design. What's the best way of doing it?

3) The suction flexible-pipe thingie you use. What's it called? (what do you ask for at the hardware store)

4) How do you test it? Do you insert it into a system and test how low and how fast it goes down in temp at a spesific condition or what?

Thanks in advance :)

[Sgrios]

I'm going to quote Soddemmfx's guide because he has said it better than I could. I can no longer find a working link of the guide, but I'll copy and paste my copy.


The evaporator is such a major component of a direct-die system that I felt it warranted its own section in this guide. As mentioned before, the purpose of the evaporator is to allow the transfer of thermal energy from the hot CPU into the cold refrigerant vapor. As I see it, these are the main performance factors which you should hope to achieve in your evaporator:

* Low resistance conduction path(s) from your heat source to the areas in contact with refrigerant vapor. The amount of power which can be transferred per unit temperature difference in a conductor depends on the thermal resistance of the material, the width of the conducting channel and the length of the conducting channel. Whilst there are complex mathematical packages available to simulate this I feel it is best just to think about it and through trial and error evolve your designs to perform more effectively.

* Large surface area which is in contact with the refrigerant vapor. This will allow more power to be transferred for a given temperature difference between the evaporator and the refrigerant.

* Turbulence in your design. Whilst refrigerant will normally boil very violently, I believe that the vapor in the centre of the flow path may remain reasonably unaffected, especially if your flow channel is large. Creating harsh angles or rough surfaces in you design will cause disruption to the normal flow pattern and mix the inner and outer refrigerant allowing more of the inner layers contact to the evaporator surface.

* A small difference between the pressure of the refrigerant entering the evaporator from the metering device (typically capillary tube) and the suction line. Higher pressures in the evaporator with respect to the suction line will lead to higher evaporator temperatures and the possibility of more liquid refrigerant passing to the suction line.

Nearly every change you will make to an evaporator design will trade off one (or more) of these factors against another; for example, higher surface area will almost always mean a more resistive conduction path and greater turbulence will cause increased pressure difference. You should aim to produce a good balance and target your design to your application.

Occasionally someone will mention evaporator "mass". In my opinion this is completely irrelevant to performance and merely a by-product of the other factors and usually comes from low resistance conduction path(s).

Only dimensions that I have of the chilly1 spiral evaporator.
http://img34.imageshack.us/img34/1998/g21e.jpg

3.) Most people use corrugated stainless steel tubing, however even the CSST tends to expand and or elongate with pressure. So most people use a stainless steel braid placed over top of the CSST to control the expansion. I buy pre-built suction lines from under the ice myself. http://www.under-the-ice.com/index.php?cPath=38

[SoddemFX]

Christ that was written a long time ago...


A small difference between the pressure of the refrigerant entering the evaporator from the metering device (typically capillary tube) and the suction line. Higher pressures in the evaporator with respect to the suction line will lead to higher evaporator temperatures and the possibility of more liquid refrigerant passing to the suction line.

A larger pressure drop will just increas the final temperatures as it increases the saturation pressure.


Partial evaporation before the refrigerant reaches the point where the conduction path and surface area combine to allow maximum energy transfer. Typically this is at a point closest to the heat source. The maximum capacity of the refrigerant is reached when ~30% of the liquid has turned to vapor; whilst this may have been utilized in commercial refrigeration for many years, to the best of my knowledge the first person to apply this to CPU evaporator design was Chilly1 of www.xtremesystems.org.


Ignore. The whole paragraph above is complete and utter nonsense, it just doesn't work this way.

Tom

[Sgrios]

Thanks for clearing that up Tom. Was just a copy/paste to get him started, I'll remove the parts that do not pertain.

Side note: You should write another guide Tom, the original was great when I was starting out.

[gosmeyer]

Truth be told, Tom forgot more about evaporators than most will ever know.

[mytekcontrols]

This seems like as good a place as any to post this question.

Has anybody ever done a good comparison of some of the different evaporator designs on an actual refrigeration system? Or in other words; used a known reference (same unit, same exact charge, calibrated heat load) to compare evaporator designs ranging from simplest to most complex.

Theory is all well and good, but real life test results are usually the best way to weed out what works from what doesn't. For instance, it would be a shame to go through all the trouble to design and machine some extremely complex channeled design, just to find out that you only get a few or no percent advantage over something of a much simpler to construct design.

I ask this question, because I am looking down the road at having to create an evaporator of my own for a special project (non CPU related, but still a cold head of sorts). I would just like to get off to a good start.

[gosmeyer]

( same exact charge)

Not to my knowledge :shrug:

[loonym]

It would be very interesting to see a head to head comparison of evaps under controlled conditions. Seems like it would also be a very expensive and time consuming venture.

[DonNiger]

Just a theoretical ... not even idea

Has anybody used or thought about distributors? Using design like Prometeia and placeing distributor on top.
Or was I supposed to be silent?

http://www.xtremesystems.org/forums/showpost.php?p=697745&postcount=4

[quintus]

Thinking of a big ass evap like 3inch diametre stepper...

[teyber]

i made a big ass 2.5" stepper :) it was bead blasted too. i never got around to completely finishing it because i realized it might be too big to work with a normal size mounting, and they don't make 1/16" snap rings that bigt in diameter

[mytekcontrols]

Well I am a bit disappointed at the response to my question :(

loonym stated:
It would be very interesting to see a head to head comparison of evaps under controlled conditions. Seems like it would also be a very expensive and time consuming venture.

Hmmm... Time consuming yes (for conducting the tests), but probably more expensive in the long run building unproven designs. Especially if something simple worked just as well as a complex machined maze in the proposed application.

I was really hoping that someone had seen proven design advantages when all other aspects other than the evaporator had remained the same. Of course this leads me to my next question. What led the charge in designing more complex evaporators in the first place? I would guess that it was something to do with thinking that there would be a heat transfer efficiency improvement, and hence colder CPU temperatures. So I still find it hard to believe that no one has any real life comparison data between evaporator design changes. Or has this just become a contest at who can design the coolest looking, most complex evaporator?

[Buckeye]

In just thinking about this a bit, would it be very hard to be exact in a comparision of Evaps ?

Given the same SS or Cascade and the only thing that would change is the Evap itself. In swaping out the Evap with a different one you have to remove the gas, unbraze the old Evap and braze in the new one. Then using the same gas as before you would then have to completly retune the system.

How can you make sure that the unit gets tuned to the exact way it was with the old Evap ?

Or is there another way to do this ?

[[XC] gomeler]

Use flare connections to attach the suction and discharge lines.

Use a refrigerant scale and charge to the same charge.

Record values from the same point on the heatload.

Repeat excessively.

[Foamy]

i made a big ass 2.5" stepper :) it was bead blasted too. i never got around to completely finishing it because i realized it might be too big to work with a normal size mounting, and they don't make 1/16" snap rings that bigt in diameter

you got pics?:D

[Buckeye]

gomeler;4093564']Use flare connections to attach the suction and discharge lines.

Use a refrigerant scale and charge to the same charge.

Record values from the same point on the heatload.

Repeat excessively.

You make it sounds so easy LOL :D

[wdrzal]

Well DetroitAC built a calorimeter to test evap performance ,but I don't think anybody wanted to pay for his investment and time. You need to carefully reach steady state and balance all the parameters to get a accurate result. And accurate calibrated equipment
does not come cheap. He told me he would measure the calories and provide a written report to the customer. Seems no one wanted to pay the price for a engineers work so I think he tested a few of his own designs then dismantled the calorimeter.

I'm just repeating what I believe to be true. DetroitAC could set the record straight. To my knowledge of being around XS. He was the only one who did evaluate the performance of different evaps scientifically. But only a few at most.

DetroitAC feel free to correct me if my statement is incorrect.

Walt

[quintus]

I understand DetroitAC perfectly well. Staying around this or another forum and picking up SS builds or cascades doesn't even provide a normal income. Why should he invest in this? when with his know how he can get a verry well paid job.

Let's face it phase change cost money! And if you want to do it right it cost a lot of money.

[gosmeyer]

Of course this leads me to my next question. What led the charge in designing more complex evaporators in the first place? I would guess that it was something to do with thinking that there would be a heat transfer efficiency improvement, and hence colder CPU temperatures. ?

You answered that !!



So I still find it hard to believe that no one has any real life comparison data between evaporator design changes. Or has this just become a contest at who can design the coolest looking, most complex evaporator?

For the most part comparison data is not available and has not been available,
this has been going on for years :yepp:

Designing and manufacturing are two different things, I can only speak for myself but I can assure you the bling factor was never a goal, only performance. We just need an investment casted evap that's a hybrid stepper/maze made from sterling silver at a low cost :shrug:

[mytekcontrols]

Use flare connections to attach the suction and discharge lines.

Use a refrigerant scale and charge to the same charge.

Record values from the same point on the heatload.

Repeat excessively.
Thanks Chris. I would add packless isolation valves prior to the flares, in order to minimize refrigerant loss during connection and disconnection. Being sure to do a quick purge of air by slightly loosening one flared connection with only one isolation valve open at the time (or add an additional port for connection to a vacuum pump). This might eliminate the need for recharging and the scale.

gosmeyer -- Sorry about seeming to suggest that this was being done for the bling factor, but without good comparison testing and data, it is difficult to say what level of complexity is required, and what is not. I do understand the goal of trying to maximize heat transfer between the external surface and the internal refrigerant stream, and the idea of using multiple channels to try to enhance this effect. But without good real life data for comparison, some of the extra work required to do this may be a wasted effort and have unwarranted extra cost.

As you guys probably noticed with my own AutoC projects, I tend to run multiple tests with varying refrigerant charge compositions in order to zero in on what works best for a given heat load. And to help in my comparisons, I collect temperature data from multiple points across the system. The AC-2 unit went through at least 2 dozen charges before I settled on the one that took it down to -140C, as well as what would yield the best operating parameters for the compressor's longevity. The AC-3 unit went through 3 HXC designs to discover what was optimal.

And yes doing all of this does take time and money. But it often times saves time and money down the road, knowing what is really required to build an optimal design.

So back to my original question...
Has anybody ever done a good comparison of some of the different evaporator designs on an actual refrigeration system?
I know many of you are members of other forums besides this one. If any one finds the answer to this question, could you please share it with us here. I think all of us would benefit from knowing.

[sdumper]

Thanks Chris. I would add packless isolation valves prior to the flares, in order to minimize refrigerant loss during connection and disconnection. Being sure to do a quick purge of air by slightly loosening one flared connection with only one isolation valve open at the time (or add an additional port for connection to a vacuum pump). This might eliminate the need for recharging and the scale.



Great ideas....

Now we just need some evaps and a volunteer :)

[gosmeyer]

gosmeyer- without good comparison testing and data, it is difficult to say what level of complexity is required, and what is not. I do understand the goal of trying to maximize heat transfer between the external surface and the internal refrigerant stream, and the idea of using multiple channels to try to enhance this effect. But without good real life data for comparison, some of the extra work required to do this may be a wasted effort and have unwarranted extra cost


:D 100% agreed

[Sweeper]

I've have a SS built for testing evaporators. I designed it so the evaporator can be removed without lossing the charge on the system. I wanted to compare my Crossflow design evaporator to other designs, but I don't have any. I made one Crossflow that is huge and really needs to be tested with a cascade, because of the mass. Later, I'm thinking of building a cascade for testing only.

98406

98407

98408

[gosmeyer]

That works :up: NICE

[sdumper]

Thats a very nice looking evap do you have any more you would like tested or to sell? Send me a PM because I need a solid cascade evap...

[teyber]

you got pics?:D
http://i198.photobucket.com/albums/aa15/teyber/DSCN2975.jpg

Nobody has tested a bunch of evaps together as there isn't really a overwhelming supply of evaps at all, it seems in the past people buy the cheapest they can as long as it performs decently well

[mytekcontrols]

I just came across something interesting in the way of finned tubing at this site (http://www.tex-fin.com/solid.htm). It is a piece of tubing with a helical bonded fin on the outside (see image below). It would seem to me that adding a cap to the outside, and a flange with an offset inlet and centered outlet hole to the bottom would create a fairly decent evaporator.

According to the website it can be made in a variety of fin and tube sizes:
Tube Sizes: 3/4" O.D. to 8" NPS (8.625" O.D.)
Fin Heights: 3/8" to 1 ½"
Fin Thicknesses: 22 ga. (.035") to 12 ga. (.102")
Fin Pitch: 1 to 7 fins per inch
Material: Any material or combination that can be resistance or arc welded.

Just a thought, and it certainly would eliminate a bunch of machining :shrug:

[teyber]

nice find there mytek! would be worth getting a quote on for 1.5"/1.75" copper!

[mytekcontrols]

nice find there mytek! would be worth getting a quote on for 1.5"/1.75" copper!
Or get a quote on a 12" piece (probably more economical when factoring in shipping) which can be chopped up to create possibly 8 evaporators.

[DonNiger]

I'm a bit lack of imagination - how would it look like mounted on CPU :confused:


I just came across something interesting in the way of finned tubing at this site (http://www.tex-fin.com/solid.htm). It is a piece of tubing with a helical bonded fin on the outside (see image below). It would seem to me that adding a cap to the outside, and a flange with an offset inlet and centered outlet hole to the bottom would create a fairly decent evaporator.

According to the website it can be made in a variety of fin and tube sizes:
Tube Sizes: 3/4" O.D. to 8" NPS (8.625" O.D.)
Fin Heights: 3/8" to 1 ½"
Fin Thicknesses: 22 ga. (.035") to 12 ga. (.102")
Fin Pitch: 1 to 7 fins per inch
Material: Any material or combination that can be resistance or arc welded.

Just a thought, and it certainly would eliminate a bunch of machining :shrug:

[Jack]

Jinu did some evap testing, but he was a commercial builder so he never gave away the results.

Also I vaguely remember one of the German builders doing a comparison test. Kayl send the guy one of his evaps IIRC. Maybe Kayl remembers...

[gosmeyer]

I just came across something interesting in the way of finned tubing

Now there is a lot of options :up:

[mytekcontrols]

DonNiger asked...
I'm a bit lack of imagination - how would it look like mounted on CPU :confused:
How about model it after a CHILLY1 Evap and use a snap ring based mount? I drew this up in CAD to get a feel for what this might look like using the Tex-Fin finned tubing (see below). Of course this is only one possibility, but it does give you a good idea of what could be done based on the Tex-Fin minimum tube and fin sizes available.

The refrigerant would travel around the central tube following the helical finned path, and then return directly through the central tube (via the notch) back to the compressor.

Edit: Or the flow could be reversed, which might yield colder temperatures at the snap ring end.

[Sweeper]

I don't think it would work very good, because the fin core is not attached to the base. Heat would not transfer through the core, so refrigerant circulating around the fins would be pointless.

[[XC] gomeler]

The fin core would have to be welded/brazed/superglued to the base to form a thermal bond. Problem is this will add thermal resistance versus a one piece evap and may negate any benefits gained through the higher surface area provided by the Tex-Fin. An interesting concept that needs to be tested by somebody.. dibs not :D

[mytekcontrols]

Heat would not transfer through the core, so refrigerant circulating around the fins would be pointless.
What if we make the outer area (the shell) a little thicker, then it would seem that this would take the place of a solid core. Think of it as inverted from the normal method. The hollow core in this case being formed by the central tube, is merely a passage way for the refrigerant. And the fins are just there to form a channel in direct contact with the outer shell. It really isn't much different than making a small coil from tubing and bonding a smooth conductive shell to the outside (something that is done quite often in the industry I came from).

Anyway this takes us back to my original query (once again). Any evaporator design can be ruled in favor of or ruled out by conjecture based on a theoretical presumption. Real life testing and comparison is the only 100% accurate way to do this, that is until enough data has been gathered to create and test a simulated design model that always concurs with the real life test.

[Jack]

Also I vaguely remember one of the German builders doing a comparison test. Kayl send the guy one of his evaps IIRC. Maybe Kayl remembers...

I found the thread. Johann did make a start, but no results unfortunately...

http://www.xtremesystems.org/forums/showthread.php?t=114646

[mytekcontrols]

Jack -- I checked out the link, but as you said no results, and no updates in 2 years. Also is it just me, or did all the posted pictures disappear? In fact there are absolutely no pictures in the whole thread, but the text content does infer that some should be present. Very odd :confused:

[[XC] gomeler]

Jack -- I checked out the link, but as you said no results, and no updates in 2 years. Also is it just me, or did all the posted pictures disappear? In fact there are absolutely no pictures in the whole thread, but the text content does infer that some should be present. Very odd :confused:

The lack of images is most likely due to the poster using imageshack and such and removing the images later. Just another reason why the images should be hosted on the forum.

[bazx]

here are some of my experiences with evap design


a design that works well on a single stage @ -50 may not work well @ -130
so you need to be clear which temp range you want to develop an evaporator for

refrigerants ability to remove heat needs to be included in the evaporators design some are better then others and may well be why some evaporators do better @ -50 and not so good @-130

load testing with resistors is a flawed:

they apply a load across the entire evaporator surface this can lead to a false impression the part is performing well when in fact it has no load holding ability in the center where “real load” will be from the cpu cores

load from a CPU is in the center of the evaporator not at the outer edges this is where you need to test the evaporators ability to hold load

loads are not constant when benching, cascades/single stage need time to recover from a sudden loads like when you begin to run wprime this time lag where the unit tries to stabilize is a big problem and ultimately decides the mhz you will be able to bench at

in my opinion to have a better evaporator we will also have to look at the way in which the refrigerant is delivered to the evaporator, the quality of that refrigerant and the units ability to recover and keep up with the evaporators needs

[[XC] gomeler]

Here is an observation I have made and been using for a while on cascades and wPrime. Temperatures are not precise, just there to illustrate the point.

3D06/Vantage load the CPU to around 25% during the 3D tests and then hit it with 100% for the CPU test. This pre-loading helps prep the cascade by having it stabilize closer to a loaded state from what I've observed. I do the same thing for wPrime by loading the CPU to 50% capacity and let the cascade temps dip and then recover. For example this may require running wPrime1024M for ~30 seconds during which the evap temps will drop from -95C idle to -85C and then creep back up to -93C or so. Then I kill wPrime1024M, fire it back up again with all threads enabled and hit it with max clocks via setFSB/eleet. The evap will droop a degree or so but will not nosedive into the deep-end.

I'm not sure if this is due to evaporator design or refrigerant puddling in less than ideal areas while idle but I know it is a repeatable behavior observed across the last 2 cascades I've built. Both were R507/R1150 cascades, one being dual 1/2hp(Loonym's) and the other dual 7/8hp(the pile of pieces scattered across my workbench). Both used Chilly1 spiral evaps so maybe that is the central flaw?

[sdumper]

gomeler;4104110']Here is an observation I have made and been using for a while on cascades and wPrime. Temperatures are not precise, just there to illustrate the point.

3D06/Vantage load the CPU to around 25% during the 3D tests and then hit it with 100% for the CPU test. This pre-loading helps prep the cascade by having it stabilize closer to a loaded state from what I've observed. I do the same thing for wPrime by loading the CPU to 50% capacity and let the cascade temps dip and then recover. For example this may require running wPrime1024M for ~30 seconds during which the evap temps will drop from -95C idle to -85C and then creep back up to -93C or so. Then I kill wPrime1024M, fire it back up again with all threads enabled and hit it with max clocks via setFSB/eleet. The evap will droop a degree or so but will not nosedive into the deep-end.

I'm not sure if this is due to evaporator design or refrigerant puddling in less than ideal areas while idle but I know it is a repeatable behavior observed across the last 2 cascades I've built. Both were R507/R1150 cascades, one being dual 1/2hp(Loonym's) and the other dual 7/8hp(the pile of pieces scattered across my workbench). Both used Chilly1 spiral evaps so maybe that is the central flaw?

+1 I have seen the exact same thing across my last two cascades using r507/r1150 and chilly1 spiral. That said I am not seeing the same drastic fluctuation on the current cascade I am tuning but that may change as I get further along with my load testing.

[lalPOOO]

Despite having very little practical knowledge, I have followed these builds and systems for several years, and can contribute the following. At one point someone did build a block with a plexi or lucite top, so that it was possible to see inside the evaporator and see what was happening with the liquid refrigerant, but the block was just a simple maze, so not really too useful other then in concept. I searched but couldn't find the video he had recorded of it running, but your luck might be better then mine.
This is an idea which could prove useful in a way for prototyping designs even though the plexi/lucite/acrylic has a detrimental effect on the block and gives it a short lifespan. Also from what I have seen it doesn't appear that attempts have been made to utilize simulation software which exists and has been used in designing watercooling blocks, such as Cathar's and others that I have seen during the very brief time I've spent browsing procooling.com forums. I am of course unsure as to if software exists which is able to simulate the more complex evaporation process, but I would wager that it is out there somewhere. Use of software could very easily be more affordable and faster then physically manufacturing 100s of blocks and spending hours testing them.
With the mention of cathar, I feel it is also necessary to point out that it appears no jet impingement/nozzle design blocks similar to cathar's cascade, but designed with phasechange in mind, have been built or tested. (publicly at least) Finally, it appears that the only person to test a material other then copper in the design of a block was fugger/chilly1 with the alu block, so this leaves silver or any design using mixtures of silver and copper fully open for exploration.

Yes I realize that suggesting ideas for block designs and ideas for testing block designs are two different subjects, I suppose I am mainly pointing out that research in this direction has really only just begun and while none of the things I have suggested solve the issue at hand, it is out of my league by a substantial amount to begin with, so Sweeper or CaptianCascades testing setups are the only existing empirical data setups. But even with these setups, you could spend a lifetime and go through an unlimited number of blocks without ever finding/discovering the 'ultimate' design, due to the simple fact that you're not employing any real calculations, and lack the ability to employ a computer to tests 1000s of possibilities based of mathematical formulas. As said before, you could definitely build a lot of interesting blocks that you think would work well, but you're missing out on ever knowing that you undoubtedly have the ultimate block design.

[n00b 0f l337]

That is a good point, the chilly spiral is certainly not cutting it anymore, and never was that great at low temperatures. After testing a chilly evaporator with the most stupid/silly of techniques, I found that quite a large amount of liquid refrigerant escapes the spiral as seen in a sight glass.
At the time, and I could try and find the topic here (or maybe it was Teampuss...) it was pointed out that this on it's own was useless, but should the technique be applied to more then one evaporator we could see which evaporators allow the expansion of the most amount of refrigerant.

What I took from it was rather simple; we can shorten our capillaries and widen our metering devices, we can deliver more and more refrigerant, no problem. But if our evaporator cannot allow this liquid to evaporate, or is too short or restrictive, it doesn't help much, and we just end up with floodback issues.

I'll be posting a sample of a new piece of evaporator tech soon, I'm having one milled out in the next week.

Higher surface area, higher internal volume, Going to test extreme opposites if possible.

[Buckeye]

gomeler;4104110']Here is an observation I have made and been using for a while on cascades and wPrime. Temperatures are not precise, just there to illustrate the point.

3D06/Vantage load the CPU to around 25% during the 3D tests and then hit it with 100% for the CPU test. This pre-loading helps prep the cascade by having it stabilize closer to a loaded state from what I've observed. I do the same thing for wPrime by loading the CPU to 50% capacity and let the cascade temps dip and then recover. For example this may require running wPrime1024M for ~30 seconds during which the evap temps will drop from -95C idle to -85C and then creep back up to -93C or so. Then I kill wPrime1024M, fire it back up again with all threads enabled and hit it with max clocks via setFSB/eleet. The evap will droop a degree or so but will not nosedive into the deep-end.

I'm not sure if this is due to evaporator design or refrigerant puddling in less than ideal areas while idle but I know it is a repeatable behavior observed across the last 2 cascades I've built. Both were R507/R1150 cascades, one being dual 1/2hp(Loonym's) and the other dual 7/8hp(the pile of pieces scattered across my workbench). Both used Chilly1 spiral evaps so maybe that is the central flaw?

That is exactly what I do Gom. My Cascade will nose dive just like to said and that goes to about 1/4 of the way into the run, it will then creap back up just like you said. At this point the Cascade is catching up with the load and will even out at the temp you posted, or about there.

I am sure this is more than just a Evap problem, but rather how the whole system reacts to loads.

As I am not a builder but just throwing this out there to think about. When you guys build a system and tune it for a load, say 300watts and we see the effect you talked about. Would tunning it for a higher load, say 350watts lessen this effect ?

So in effect what I am trying to say is, this is not just an Evap problem but how the whole system reacts to loads.

[mytekcontrols]

This is definitely getting much more interesting with all the recent feedback.


...from what I have seen it doesn't appear that attempts have been made to utilize simulation software which exists and has been used in designing watercooling blocks, such as Cathar's and others that I have seen during the very brief time I've spent browsing procooling.com forums. I am of course unsure as to if software exists which is able to simulate the more complex evaporation process, but I would wager that it is out there somewhere. Use of software could very easily be more affordable and faster then physically manufacturing 100s of blocks and spending hours testing them.
Very good points by lalPOOO, and I would agree that using simulation software would be an ideal way to evaluate and develop evaporator designs. But only if that software was proven to be accurate by yielding data that when cross checked against real life tests, was found to be true.


load testing with resistors is a flawed:
they apply a load across the entire evaporator surface this can lead to a false impression the part is performing well when in fact it has no load holding ability in the center where “real load” will be from the cpu cores
This points to a design flaw in the mass or conductivity of the material being used on the face of the evaporator, and not so much the fault of using resistors for the test load. Or in other words an excessive gradient exists between the center to the edges of the evaporator face. Seeing the test results as described, would dictate that either a material with better conductive properties be used, and/or to increase the thickness of the material being used on the evaporator face.


a design that works well on a single stage @ -50 may not work well @ -130
so you need to be clear which temp range you want to develop an evaporator for
refrigerants ability to remove heat needs to be included in the evaporators design some are better then others and may well be why some evaporators do better @ -50 and not so good @-130
This probably has more to do with where all the evaporation is taking place, either mostly in the face, or spread across a much larger area. I would imagine that a system designed for -50C would have a higher percentage of condensate available and probably a higher mass flow of said refrigerant than one running at -130C (that is unless we are talking about a tremendous increase in system size and refrigerant storage). So for -50C operation an evaporator having its evaporative surface area spread out over distance, but still made with good conductivity in mind, could be quite successful at removing heat so long as there is a good supply of condensate feeding it (e.g.; high mass flow). However putting this on a system with poor mass flow would be a performance killer.

So it would appear to me that good evaporator design should always attempt to maximize the percentage of the face that is in contact with the refrigerant, and minimize the gradient between the two. (Both the Chilly1 and my proposed Tex-Fin design don't do a very good job at following this rule)

As for handling different loads encountered on the same system, this would be a limited when using a fixed flow device such as a cap tube, and respond much better if the flow regulating device was active and responding to the change in heat load (e.g.; thermally regulated expansion valve, or an additional cap tube in parallel coupled with a temperature activated solenoid valve -- opens when temperature rises).

[n00b 0f l337]

Or say a CPEV where you've tuned the CPEV at the load that is going to be applied through the benching run, at this point the evaporator will at idle be flooding with refrigerant and be ready for the heat load about to be applied.

[bazx]

This points to a design flaw in the mass or conductivity of the material being used on the face of the evaporator, and not so much the fault of using resistors for the test load. Or in other words an excessive gradient exists between the center to the edges of the evaporator face. Seeing the test results as described, would dictate that either a material with better conductive properties be used, and/or to increase the thickness of the material being used on the evaporator face.


my point is a resistor based load tester would not reveal the design flaw you just out lined and as such would give you false data

where as a cpu would reveal the flaw

[quintus]

Resistor based load tester and most of the load testers are made only to charge the sistem right, not realy for testing evaps.
A real cpu and only 1 and the same for all the evps testing will be best.

Just curious, is the nose dive happening also in normal ss? I have a feeling that is something particular to cascades and/or the use of low moleculare mass freons(r1150)
Never seen it in my ss, not even on my small cascade(the only built by me, charged with a mixture of r22/r744/r50, -75c at 1bar)

Beter evaps should be taller, larger diametre resulting in more mass with more surface area. Also sand blasted.

[SoddemFX]

a design that works well on a single stage @ -50 may not work well @ -130

The first evaporator to have this floor in an obvious way was the Cryostar, which was a pretty good single stage evaporator but an abysmal evaporator with a two stage cascade - where the mass flow was substaintially lower.

In my opinion, the failure of this evaporator was it's inability to transfer heat into superheated gas which seems to be critically important in a cascaded system. Spefically its failure were the paths with low resistance to the useful surface area, and this is of critical importance to any evaporator.

Tom

[EvoCarlos]

for my next evap sution line im not even going to use a spiral /stepper imo its not needed
just a copper pot with 2mm holes with 2mm spacings it would be simller to a duniek ln2 pot but with out the alloy bit
also that way i would be able to reproduce the evap from silver by means of casting :)

i intend to let the gas do its job of cooling the evap and not the evap cooling the cpu as most people see it :)

[[XC] gomeler]

I would like to see how pin-matrix evaporators would work. I have a Swiftech Apogee GTZ with a pin-matrix base that I will adapt for use as an evaporator but I suspect it will be too small and have too thin of a base.

Would be nice to have some simulation software in which we could dump in cad files and see how the evaporators would work at low temperatures. Someone with free time get in on that ;)

[Jack]

Pin matrix Apogee or any evap with one big evaporating chamber will probably have orientation issues. I can imagine it will work fine when the evap is horizontal, but will perform worse when positioned vertical (such as in a case).

[n00b 0f l337]

The Mach series of phase change systems have used a pin style evaporator, approximately off the top of my head.
The evap is like 2" by 1.5" containing large ~3/8" pins in a 4x8 matrix or something like that. It does actually work rather well.

Orientation is an issue though; and thus needs to be accounted for by use of a pathway with pins.

[teyber]

gomeler;4105385']I would like to see how pin-matrix evaporators would work. I have a Swiftech Apogee GTZ with a pin-matrix base that I will adapt for use as an evaporator but I suspect it will be too small and have too thin of a base.

Would be nice to have some simulation software in which we could dump in cad files and see how the evaporators would work at low temperatures. Someone with free time get in on that ;)

hey if you wanted to test one i could make one pretty easily

edit: really easily, i could just jog it around

[n00b 0f l337]

The problem is establishing a flow pattern as well, not just a straight pin setup, and it must be nearly coherent in all orientations ;)
Teyber do you have your CNC fully working? PM me if you'd like to discuss a design.

[teyber]

edit: oops

[DonNiger]

A bit odd but te same topic

How much R1150 is actually needed to keep evap under -100c @ 250W? Other words, how much liquid has to evaporate?

[SoddemFX]

A bit odd but te same topic

How much R1150 is actually needed to keep evap under -100c @ 250W? Other words, how much liquid has to evaporate?

With a -45C liquid line, no suction superheat and a -100C saturation temperature ~0.74 grams per second.

Obviously you need a temperature difference (delta) between the refrigerant and the heatsource for any transfer to happen.

Tom