need new evap, some ideas to discuss

[|RickY|]

I’m trying to desigh a new evap for my cascade to help handling quad core massive load.

1 thing i like on a stepper is a central big piece of cooper with a lot of mass being cooled arround, the cold are centered on the same piece of cooper.

Trying to improve surface on a steeper, with big mass of copper centered being cooled arround on 2 steps:

Sorry the desighn quick desighned on a coffe:
http://img201.imageshack.us/img201/9935/69956271ps6.jpg

on left the plant, right 3d.
Trying to make it the most solid possible to lathe, on right, you see, 1st level is solid, central piece are solid too, 2nd level are soldered to 1st level and to big central piece.

Please tell-me if i’m making understand. I’m desighing it on autocad so i’ll post better pics.

Another way my brain goes is +- this: thinking on liquid like cold water on an watercooling block, it enter cold and exit hot, when we talk about air cooling, the air coolers dont work this way, it dont have only 1 enter and 1 exit, it enter new cold air everywere, more surface with new cold air.

Going to air cooled ideia, we have somthing like this:
http://img251.imageshack.us/img251/9710/80658187ok6.jpg

aplying to phase-change it may meens capilar beind divided on 4 before evap, and 4 independed channels inside the evap, will be and evap being floded by 4 quick passes of cold liquid and not 1 long channel. i think this last paint desighn show the ideia

My ideia is to think a litle on evaps on group, more brains, discuss some ideias.
I’m trying to desighn a new evap for my cascade but i can decide, the evap will be cnc maded, i can't make some, thats the problem, i need to decide for 1 hrhrhr

sorr the english :S
:toast:





Edit: finished evap


http://img399.imageshack.us/img399/7290/87750166zi9.jpg

http://img177.imageshack.us/img177/3544/60986394ap3.jpg

http://img177.imageshack.us/img177/7677/96857974by9.jpg

http://img177.imageshack.us/img177/3683/74051380gp5.jpg

http://img399.imageshack.us/img399/2853/56655134ie6.jpg

http://img177.imageshack.us/img177/1810/55186961ya4.jpg

http://img177.imageshack.us/img177/4405/60616681dm7.jpg

http://img399.imageshack.us/img399/285/38377834rc9.jpg

special thanks to P.P in da house, without you this couln't be possible ;)

[LardArse]

I like the 4 cap tube split idea, I don't know how equal flow rate will be even if same length though, you'd probably be crimping a pipe to round up the 4 tubes into 1.

Your evap, I think as long as you have ample of mass, especially at base area, it will do good for Quad Core benching as you know only 10% of time you will be running full quad load so why tune cap tube for that 10% of time when you can use the base to buffer you over the CPU Tests?

[Gray Mole]

I realise I've said all of this before, but it bears repeating so I'll say it again.

Mass, in and of itself, is useless in vapor phase. For LN2 it's essential, and for Dry Ice it's useful but a double edged sword.

There are 2 types of heat exchangers that we normally use, tube and shell (or tube in tube) and brazed plate heat exchangers.

It's a well known fact that the plated ones are anywhere from marginally better (in cases where the tubed version is either professionally made, or VERY well designed by the fabricator) and VASTLY better in most if not all cases.

An evaporator is a heat exchanger, pure and simple. Keeping that in mind makes the idea of design easier to translate.

Enter @itor's/Ssilencer's evaps. I'm not saying you should copy it (you really shouldn't, and likely won't have the greatest success doing so without a specifically suitable lathe, and custom cutters) or that you should run out and buy one (though for the best, it's the best choice) but if you want to make a VERY good evap there are some principles that you can glean from it's design.

@itor's work is very much like the BPHE. Thin plates, adequate but not overly massive, and very tightly spaced. Like a plate heat exchanger.

In vapor phase mass always loses because of the inherent limitations of the liquid's ability to be translated through said mass. Having an evap with a large amount of metal between the refrigerant and the item to be cooled may look 'shiny' but it's weakness shows quite badly in use, generally in terms of the cpu's actual temperature under load.

LN2 has a virtually 'limitless' tranfer capability and this is due mainly to it's being a complete liquid, and that it's far colder than the required temperature. You therefore can use the mass of the vessel to 'carry' the temperature you need, and too little mass can make it impossible to do so.

Refrigerant of any kind is never fully liquid and never fully flooded as you might think in terms of a full tube of LN2. It's in partial vapor phase, and using it in that way give you it's best and most complete heat transfer.

Other than the shell of the plated heat exchanger, the plates themselves are incredibly thin, and have little to no mass whatsoever. So why does it work? Large surface area, very tightly spaced plates. Sound familiar? It is, exactly like the @itor design. It wasn't 'luck' it was very good design work.

So, I can't really tell you what to make for a 'better evap' than you've been using, but I CAN tell you that your projected design is poor on a few levels.

Stop thinking mass, increase surface area, and don't have to much 'wasted room'.

1. Mass. A 3-5mm base has been proven time and again to yeild the most effective heat transfer. Increasing mass with a central column can be effective, but only if the supporting surface area is conducive to cooling it adequately.

2. Surface area. Thinner the better. Lots of it. Any time you go over about 2mm of thickness in the supporting mass of the evap you may as well not have it there, as the heat moves too quickly into it for the refrigerant to effectively remove it from the metal, and you lose cooling to a 'heat buffer zone' in the mass especially under load.

3. Wasted room. Having an area for the refrigerant to stabilise prior to the primary heat transfer zone has proven effective in most if not all cases. 'Preparing' it for the area that will utilise it is one of the key design flaws in many evaps. However, the larger spaces between plates/steps/maze-walls is a complete waste of internal surface area. If the refrigerant isn't in contact with metal that's full of heat when it goes through your evap, it's going to come out exactly as it went in. Tightening the channels to a point where the refrigerant is not restricted, but is restrained into the position you want for ideal forced contact is probably the most important and often forgotten of all design implementations.

These are all reasons why the @itor system is better than most if not all evaps out there, and why I recommend it so highly. The actual temperature of the evap is normally a little better under load than other evaps, but it also means that the optimal transfer gives greatly reduced cpu temps, and brings 'high load floodback issues' under control because of the designs ability to not only do high load (ie Quad core) but also to maintain low load cooling and use that heat to counter the floodback issues seen in less capable evaps.

When you look at designing a new evap, think about the technology that works and try to emulate that into your own design. You'll end up with a much better evap. The rest of the system can be amazing but if you limit it with a substandard evap you're losing so much of the cooling that you could achieve.

This is why I recommend the @itor evap to pretty much everyone if they want the best. I don't get them for free and they're not cheap, but what they ARE is the best evap for ANY application that I've ever used.

That's me off my soapbox about it (again) but I hope that some of this sinks in for everyone.


Cheers

Gray

[boshuter]

Good explanation GM... I just finished reading a PM from Ssilencer that was saying almost the exact same thing; I was telling him how I had just tested a 1/5th hp single stage with a cap-in-cap evap using a Maze4 GPU block base plate. That evap was performing better than the commercial spiral evaps I normally use. I have a probe inside my load tester as well as a probe on the evap itself and the c in c evap had better temps transferred to the actual load than the more massive spiral evap.

[|RickY|]

I like the 4 cap tube split idea, I don't know how equal flow rate will be even if same length though, you'd probably be crimping a pipe to round up the 4 tubes into 1.

Your evap, I think as long as you have ample of mass, especially at base area, it will do good for Quad Core benching as you know only 10% of time you will be running full quad load so why tune cap tube for that 10% of time when you can use the base to buffer you over the CPU Tests?

yes i'm pointed for what you say.

about cap tubes... i dont know what must be better 2 cap's like it was dual evap sistem, or the ideia inside the evap, 1 evap on top, top layer divide the liquid for 4 holes that feed the lower stepp.




Gray, i understand what you are telling, but i dont think aitor evap are the perfect one so the better we can do is buy it or copy it, if we all think like that we still have Aristóteles computer.

i think we must use that we know form hx and ln2 pots, comparing aitor evap to an hx dont make to mutch sence for me, i see aitor evap like all other steppers, but smaller it maybe be better, but the ideia for me is the same: cool down a central mass off cooper. on hx you have to exange heat between 2 gases, on evap you are changing heat between 1 solid and 1 liquid or smi-liquid.

need to go, my girlfriend kill-me but i'll be back :P

thanks gray, lardharse :)

ha, there is the last evap i'll use on a single stage :) i'll test it soon, nothing new, but pure copeer and cnc maded more 1 channel gained. I'v used this desighn lot, i like it, dont know why it performs a litle better than my other 1 way mazes, i'v thinked on it and maybe becouse the 2 ways of cold liquid. It is from there i start thinking on multi feeded evap.

http://img68.imageshack.us/img68/9512/evapbt9.jpg

http://img134.imageshack.us/img134/5273/evapgt2.jpg

[n00b 0f l337]

I think more then anything we need longer refrigerant paths as well as more direct cooling. The path based mazes and ring like evaps like that seem to perform great becuase of that. THe chilly1 spiral is decent but really requires more surface area to peak out on its potential.

[Freddie123]

What do you guys think of:

[|RickY|]

I think more then anything we need longer refrigerant paths as well as more direct cooling. The path based mazes and ring like evaps like that seem to perform great becuase of that. THe chilly1 spiral is decent but really requires more surface area to peak out on its potential.

yes you are right, desighn i have on top of thred are trying to increese surface area on a steper, not round center but another tipe of massive coper center for more surface

freddie123, maybe to mutch coper for the chanells it have.

[Freddie123]

Redid it: More channels, less wasted copper.

[Ssilencer]

I think more then anything we need longer refrigerant paths as well as more direct cooling. The path based mazes and ring like evaps like that seem to perform great becuase of that. THe chilly1 spiral is decent but really requires more surface area to peak out on its potential.

I think the last time I posted an evap insides was the first revision, and we are now on the 5th iirc, and I'm not allowed to post the insides anymore, but it had changed a lot from the first one.
Gray got one of the latest revs about a week ago, and he don't even know that :), no more 3mm "thick" bases there gray...;)

[Freddie123]

Its a shame that people steal ideas because it means we never get to see porn pics of LN2 pots and cool evaps ;(

[SoddemFX]

I realise I've said all of this before, but it bears repeating so I'll say it again.

Mass, in and of itself, is useless in vapor phase. For LN2 it's essential, and for Dry Ice it's useful but a double edged sword.

There are 2 types of heat exchangers that we normally use, tube and shell (or tube in tube) and brazed plate heat exchangers.

It's a well known fact that the plated ones are anywhere from marginally better (in cases where the tubed version is either professionally made, or VERY well designed by the fabricator) and VASTLY better in most if not all cases.

An evaporator is a heat exchanger, pure and simple. Keeping that in mind makes the idea of design easier to translate.

Enter @itor's/Ssilencer's evaps. I'm not saying you should copy it (you really shouldn't, and likely won't have the greatest success doing so without a specifically suitable lathe, and custom cutters) or that you should run out and buy one (though for the best, it's the best choice) but if you want to make a VERY good evap there are some principles that you can glean from it's design.

@itor's work is very much like the BPHE. Thin plates, adequate but not overly massive, and very tightly spaced. Like a plate heat exchanger.

In vapor phase mass always loses because of the inherent limitations of the liquid's ability to be translated through said mass. Having an evap with a large amount of metal between the refrigerant and the item to be cooled may look 'shiny' but it's weakness shows quite badly in use, generally in terms of the cpu's actual temperature under load.

LN2 has a virtually 'limitless' tranfer capability and this is due mainly to it's being a complete liquid, and that it's far colder than the required temperature. You therefore can use the mass of the vessel to 'carry' the temperature you need, and too little mass can make it impossible to do so.

Refrigerant of any kind is never fully liquid and never fully flooded as you might think in terms of a full tube of LN2. It's in partial vapor phase, and using it in that way give you it's best and most complete heat transfer.

Other than the shell of the plated heat exchanger, the plates themselves are incredibly thin, and have little to no mass whatsoever. So why does it work? Large surface area, very tightly spaced plates. Sound familiar? It is, exactly like the @itor design. It wasn't 'luck' it was very good design work.

So, I can't really tell you what to make for a 'better evap' than you've been using, but I CAN tell you that your projected design is poor on a few levels.

Stop thinking mass, increase surface area, and don't have to much 'wasted room'.

1. Mass. A 3-5mm base has been proven time and again to yeild the most effective heat transfer. Increasing mass with a central column can be effective, but only if the supporting surface area is conducive to cooling it adequately.

2. Surface area. Thinner the better. Lots of it. Any time you go over about 2mm of thickness in the supporting mass of the evap you may as well not have it there, as the heat moves too quickly into it for the refrigerant to effectively remove it from the metal, and you lose cooling to a 'heat buffer zone' in the mass especially under load.

3. Wasted room. Having an area for the refrigerant to stabilise prior to the primary heat transfer zone has proven effective in most if not all cases. 'Preparing' it for the area that will utilise it is one of the key design flaws in many evaps. However, the larger spaces between plates/steps/maze-walls is a complete waste of internal surface area. If the refrigerant isn't in contact with metal that's full of heat when it goes through your evap, it's going to come out exactly as it went in. Tightening the channels to a point where the refrigerant is not restricted, but is restrained into the position you want for ideal forced contact is probably the most important and often forgotten of all design implementations.

These are all reasons why the @itor system is better than most if not all evaps out there, and why I recommend it so highly. The actual temperature of the evap is normally a little better under load than other evaps, but it also means that the optimal transfer gives greatly reduced cpu temps, and brings 'high load floodback issues' under control because of the designs ability to not only do high load (ie Quad core) but also to maintain low load cooling and use that heat to counter the floodback issues seen in less capable evaps.

When you look at designing a new evap, think about the technology that works and try to emulate that into your own design. You'll end up with a much better evap. The rest of the system can be amazing but if you limit it with a substandard evap you're losing so much of the cooling that you could achieve.

This is why I recommend the @itor evap to pretty much everyone if they want the best. I don't get them for free and they're not cheap, but what they ARE is the best evap for ANY application that I've ever used.

That's me off my soapbox about it (again) but I hope that some of this sinks in for everyone.


Cheers

Gray

1. Comparing a plate heat exchanger to a CPU evaporator is a very bad analogy. A plate heat exchanger is removing heat from a very large surface area and transfering it to an equally large surface area with a very small ~1mm usually nickel plated copper barrier in between. A CPU evaporator is removing heat from a very small surface area and transfering heat to a limited surface area via multiple paths of varying area and length.

The main thing is that the difference in heat flux at the source (between a plate heat exchanger and a CPU evaporator) could well be three orders of magnitude. Compared to a CPU evaporator a plate heat exchanger is very simple.

2. Liquid nitrogen poured from a dewar will have a quality (mass gas / mass liquid) of 0 in that the refrigerant is in total liquid phase on entering the tube. Depending on the cooling recieved from the suction line the quality of the refrigerant entering the evaporator in a single stage might be more like 0.5 or higher as in half of the refrigerant has already evaporated before it enters the evaporator. The transfer capability of a liquid nitrogen tube isn't unlimited, it depends on the design of the tube and is finite like any evaporator although i think with the lower quality of refrigerant in the liquid nitrogen tube the capacity of the evaporator (ignoring enthalpy) will be higher due to the higher conductivity of liquid compared to gas.

The available enthalpy is double what it is with refrigerant quality of 0.5 which will give double the capacity to a liquid nitrogen tube v's an evaporator with the same mass flow. With a bottomless dewar and fast arms the available mass flow to the liquid nitrogen tube is unlimited but this isn't the same as the tubes capacity itself.

I have no idea at all why refrigerant with a quality of 0.5 would give better heat transfer than refrigerant with a quality of 0, it doesn't make any sense at all :confused:

3. I pretty much agree although i think the base should be the last thing to concentrate on because on evaporators like the @itor etc because it won't have that much influence within reason assuming the point of load is directly beneath the centre of the column.

Surface area is a big failure and feature sizes should be smaller, way smaller to allow this increase in surface area although machining limitations become a problem. I think evaporator pressure drop isn't anywhere near worrying about at the moment. I think a good addition to the evaporator |Ricky| posted would be small pins or threaded pins drilled or tapped into the base, i think roughing the evaporator side walls would also work well.

4. Gray - what you've said drifts away from the topic and kind of reads as a sales promotion for the @itor evaporator. I'd rather see a good picture of a bad evaporator than read a closed appraisal of a good evaporator and that's the truth. Also Ssilencer or @itor didn't come out of the womb with charging manifolds or cutting tools in hand... :D

Tom

[Freddie123]

Also Ssilencer or @itor didn't come out of the womb with charging manifolds or cutting tools in hand...

Stolen for sig :D

[Coyote]

... Also Ssilencer or @itor didn't come out of the womb with charging manifolds or cutting tools in hand... :D

Tom

no they didnt, but again ... MA1 was @itors first evap so...

i think what GM meant to say, is basically that if you desing an evap, based on some specific concepts, you wont end up too far away from an @itor's like evap.

take a peak at @itor's web site, specifically his waterblocks, then you'll see where his knoledge comes from.

still, ppl should be able to make different evaps, based on similar principles. in the end, @itor its only using a 50 year old lathe. that limits the evap's design. surely someone who knows what he is doing, and with better tools should be able to come up with a more efficient desing.

[Unknown_road]

my latest (silver) evap:

http://www.overcooling.nl/images/evapv2/IMG_1101.jpg

it's got a cone shaped bottom, in the middel the bottom is 3mm and on the side it's 1mm thick. heat spreading capability where it is usefull.

the walls are 12mm high so very much contact area pretty close to the core and much room for liquid refrigerant.

haven't tested it yet but will do soon, but imo an evap should have the properties I just mentioned and more of it is better. From there it's just testing if it works as well as you hope. If you really want to know without testing it, you have to ask NASA if you can borrow there supercomputer. :lol:

designing evap's it a too complex science to do on the back of an envelope

[[XC] 2long4u]

Freddie I like the second one:up:
Unknown:slobber: :slobber: :rocker:

[SoddemFX]

Jurgen,

The casting on that evaporator is great! You must have had a lot of practice now :up:

Tom

[[XC] 2long4u]

Waiting for the diamond evap.
You know lab created diamonds are pretty cheap!

[Freddie123]

How in gods name can you afford the silver for that thing?

[[XC] 2long4u]

Silver is not as expensive as people think.

[Unknown_road]

Jurgen,

The casting on that evaporator is great! You must have had a lot of practice now :up:

Tom

thanks Tom, you have no idea how much practice :D I probably trew more then 20 evaps back into the furnace before it looked anywhere near to this.
This is done with vacuum assisted lost wax investment casting. A long way from sand casting ;) .

the amount of silver used is probably worth +- 35 euro. small amount compared to the 1000 euro I invested in equipment. It got a little out of hand I geuss :lol: but experimenting is still 10x more fun then building.

[Ssilencer]

3. I pretty much agree although i think the base should be the last thing to concentrate on because on evaporators like the @itor etc because it won't have that much influence within reason assuming the point of load is directly beneath the centre of the column.

4. Gray - what you've said drifts away from the topic and kind of reads as a sales promotion for the @itor evaporator. I'd rather see a good picture of a bad evaporator than read a closed appraisal of a good evaporator and that's the truth. Also Ssilencer or @itor didn't come out of the womb with charging manifolds or cutting tools in hand... :D

Tom

Coyote pretty much responded what I thought.

3_ Your assumptions are wrong, looks like your never saw a naked quad core processor. And I'd just made a reference to one of the changes.

4_Gray doesn't have any monetary interest in @itor's evap, as it would be a lot more profitable to use any other evap in his units.
I don't know nothing compared to you about refrigeration, so, if you can explain me why you did that evap in that way, and how it performed would be great to keep learning and sharing knowledge
http://gorajec.net/lukexe/v7%20evap/PICT9467.jpg

[SoddemFX]

I had a short length of copper and had just bought a vertical milling adaptor for the lathe which i wanted to try out. I didn't plan that evaporator and didn't want to keep it myself but i didn't want to waste it either, so i gave it to Luke in exchange for a r22 bottle adaptor, he got a good deal in my opinion...

I made this V6 for myself a bit before:

http://www.pclincs.co.uk/Forums/attachment.php?attachmentid=2277&d=1157753524

http://www.pclincs.co.uk/Forums/attachment.php?attachmentid=2278&d=1157753554

http://www.pclincs.co.uk/Forums/showthread.php?t=9972

I made this V8 for myself a bit after, i used a 1 5/8" copper end cap like you do on the @itor which, makes the hollow column evaporators much cheaper and easier to make as well as shorter:

http://img403.imageshack.us/img403/7640/v8dg4be8.jpg

There's some better pictures of that evaporator you posted (V7) here:

http://www.pclincs.co.uk/Forums/showthread.php?t=9971

Luke seemed to like it but i don't think the V7 is a good evaporator either and i never pretended it was. The V2, V6 and V8 are all very good evaporators. At least when there are documented pictures people can draw their own conclusions.

3. ...assuming...

4. If you like the idea of sharing knowledge so much then why don't you share some of yours? :)

That is if you want an enthusiast community and not an enthusiast marketplace...

Tom

[Freddie123]

Thats a really nice Evap Tom! So the refrigerant starts to evaporate in the hollow centre column then goes out into the plates? Is it stepper or spiral?

[SoddemFX]

Yes, that's right. The capillary tube passes through a small brass stopper in the centre top, the refrigerant enters the 7mm bore column which goes to the base of the evaporator and exits through three 2.7mm holes onto the base. After that it's a normal thin plate stepper with a tapered column.

The 1 5/8" end cap has about 6mm clearence over the top of what you see in the picture. Brasing the brass stopper is the hardest part, you need a high flow of nitrogen to avoid oxidisation when the 3/8" pipe stub is out of the pipe end cap for a few seconds. After that part the brasing is easy.

The picture of the V8 is taken before the slots are cut into the plates using a shaper machine (kind of like the predecessor to a mill) from the 1930's, the shaper machine scares the cr*p out of me :D

Tom

[[XC] 2long4u]

Tom when you take the pics use the setting that looks like a flower.

[SoddemFX]

Yeah it's a realy old camera, it focuses way too long even in macro. The best way to use it is to take pictures over 1m away in macro at max res with the camera sat on something and then crop them down on the PC.

Thanks for the advice tho :)

Tom

[Freddie123]

Ok thanks

[Freddie123]

No wasted copper, lots of surface area and a maze for the refrigerant says you?

How's this? Says I:

[[XC] 2long4u]

Um. That would be very hard to machine.
Would kick a$$ tho.

[Freddie123]

Meh i'm sure you could CNC it. We have a CNC mill and router at our school.

[berkut]

http://i129.photobucket.com/albums/p201/berkutelbrus/skanuj001001.jpg

A sketch of a very old Evaporator design of mine. When build properly it will outperform a maze type evaporator. The design can be improved by threading the holes, changing the base thickness and size, the amount of nozzles used.

What is crucial is the expansion chambers size and shape for even flow.

[n00b 0f l337]

So basically a swiftech storm but as an evaporator to create an impingement effect? I don't think it will work as well as you say unless you have some tests to back it up. Impingement creates a buffer often and we dont want stationary gas at the bottom of those holes.

[n00b 0f l337]

Freddie also remember, easier to work with a suction in the center.

[Freddie123]

Wouldnt cap in middle give better temps though?

[n00b 0f l337]

Not necesarily but youd have to test it. Its a bit of a pain in the butt having suction off to the side, cant really rotate the mount around then can you.

[Freddie123]

True enough, for you then NoL:

Cap tube comes in at the white circle, center is 1/4" (More space but can easily slot into 3/8" tubing). Outer wall 3mm thick, inner are 1mm

[n00b 0f l337]

Could work very well, for the cap over it I'd have a 3/8" hole, the 3/8" tube I'd use then would rest against the first inner wall and would be just fine. Again very price to mill tho.

[Freddie123]

Unless milled at school for free :D Are you saying you'd make the center hole 3/8"?

Edit: See what you mean now nvm

[.p4nzer]

True enough, for you then NoL:

Cap tube comes in at the white circle, center is 1/4" (More space but can easily slot into 3/8" tubing). Outer wall 3mm thick, inner are 1mm

I don't like the side entry of the captube...
U can use a central captube with the sunction line on its side.

I like the profile of the gosmeyer's evaporator, but that must be a little higher IMHO...

But this thread is so insane...
I'm not a physic's spacialist but I think that and evaporator can perform in different ways depending of the gas, the pressure, the load an so much more...
I can say that an evaporator is really better than another one if I have done a lot of test with both...
We can challange different evaporators with some different gases (for example r404, r1150), with same gas and different temperature, with same condition and try the efficiency at different load, see the response to load changes (idle//load) and much more...
Without a massive number of test we talk about nothing 'cause a core2 can use some hour for a simulation and our brain is so bad in this kind of computation...

[n00b 0f l337]

You can all say that we need more tests, but we can all say empirical data collected in our heads over time is much better in alot of situations ;)

[.p4nzer]

You can all say that we need more tests, but we can all say empirical data collected in our heads over time is much better in alot of situations ;)

Yes, u are right...
But an evaporator like chilly1 or @itor or other evaporators that are in this forum were created for maximize performance, cost and minimize the production problems...
Between a chilly1 or an @itor we have some little difference in performance...
Only with some real test we can say who are the best. We can say that the spiral of the chilly1 provide a massive load and that a maze can be the best way for large surface area and low temps (for example, don't matter my words) but the single stages that we can see in this forum are built with only one of these and the difference between the evaporator are so little (and the variables so high; like gases, compressor, piping, dsh and so much more) that we can't say who is the best.
This post can be the impulse to compare the evaporators and increase our knowledge.

[berkut]

The evaporator worked. Its crucial to have the distribution / expansion (however you like to call it) chamber and the nozzles (i made them out of capilary tubes) at a proper size related to the refrigerant flow or the flow wont be even.

The holes need to be sized properly too, too little and the gas will blow out all the liquid, too big and youre loosing efficiency... Adding a thread makes wonders here. Heat distribution over the block was excellent.

It took me a while to make it work :/

This was one of my "secret" projects, never shown anywhere, only mentioned once or twice. I have no hard numbers or pictures, sorry, i hope this hint helps.

[n00b 0f l337]

I understand the concept, again like a swiftech Storm, I'm surprised only that impingement (as you did seem to create impingement) helped at all and didnt hinder. Its possible that the expansion of the liquid made the buffer zone disapear, and then yes would probably work very well, create heat transfer in turbulence. And yep, threading helps too. I do that inside every straight hole I drill in a Chilly1 evap.

[gosmeyer]

I would like to see everyone make there own evap if for no other reason than the experience. However what looks good on paper or even a model may not be attainable in practice. Take any maze design, 1mm walls and 1mm channels would be fine if you were going to use a cnc mill and machine it to a depth of say 3mm. Deeper than that will require milling 2 or 3 graphite electrodes and using a edm sinker. I am a firm believer in anything can be done as long as you keep things in perspective and have the proper tooling and knowledge. Or the easy way, know someone who does.

[Freddie123]

Could a CNC router mill a maze type design>

[PPInDaHouse]

@ freddie123

your latest design is very hard to machine because the chanels are very thin and the tool that make them will break very easly..

[|RickY|]

well i'm hard on time to read and think now, i drink to mutch beer's with gf.

but i'v read it superficial and that's this way the thred must go, change ideias, discuss evaps, if you have secret's go out :P

i like solden fx improvements on a stepper, i need to study better baker evap to make and ideia.

that's bad we are just curious we make what we can, maybe a big company would make 100 diferent evaps and test them, here we need to simulate it working on our brain and test it after and compare with the others we have used hehe

need to desighn on cad the 1st i'v showed and the "air cooled style" ideia i'v on brain.

:toast:

[[XC] gomeler]

Perhaps with these very densely finned ideas could they be casted like the silver evaps above? Can such detail be accurately casted?

[gosmeyer]

Sure, but you still need the wax mold or resin from rapid prototype, Easy enough right, but then you need an investment powder like satin cast, the flask, vacuum chamber to remove any air after investment pour, oven to cure it and melt the wax or resin, furnace to hit 1640F for .925 silver to flow correct all under vacuum so the fine detail can be maintained. Or you could just pour it into Styrofoam and have a pile of goo.

[AllAgainstPaul]

I know this might sound stupid, but instead of milling out those long paths in maze style evaps, wouldn't it be easier to melt down the metal in a forge and pour it into a mold? Would this introduce impurities or flaws into the wall and base?

[SoddemFX]

Berkut,

Do you have any pictures of the evaporator you made? It's realy interesting :)

Between a chilly1 or an @itor we have some little difference in performance...

I think the Chilly1 isn't a good evaporator, i think the @itor probably is a very good evaporator, this is just my opinion :)

Tom

[Freddie123]

Read up Paul :D

[Unknown_road]

Sure, but you still need the wax mold or resin from rapid prototype, Easy enough right, but then you need an investment powder like satin cast, the flask, vacuum chamber to remove any air after investment pour, oven to cure it and melt the wax or resin, furnace to hit 1640F for .925 silver to flow correct all under vacuum so the fine detail can be maintained. Or you could just pour it into Styrofoam and have a pile of goo.

casting fins is very very hard. For silver casting the investment must be around 450 *C and the silver may not be above 1100*C else you get casting defects like brittle metal and holes in the metal. The silver solidifies at 960*C, so when you try to cast fins the silver rapidly solidifies because of the contact with the investment.

I also use vacuum but still don't have a 100% succes rate.

[gosmeyer]

I also use vacuum but still don't have a 100% succes rate.

Jurgen,
vacuum @ both investment and pour?
Hell, nobody will hit 100%, Do you use nitrogen at any time?

[[XC] 2long4u]

Could you preheat the cast so it doesn't harden when it hits it?

[gosmeyer]

Everything is kept at very specific temps, everything. You remove the investment from the oven and pour liquid immediately.

[Freddie123]

Basically, its really hard. Unknowns silver evaps are the most complex evaps made on here imo purely because he has to cast it. Props to him for that ;D

[Unknown_road]

Jurgen,
vacuum @ both investment and pour?
Hell, nobody will hit 100%, Do you use nitrogen at any time?

yes at both else it got a 100% failure rate :D . I know nobody will hit 100% succes but if I wanted to cast fins with the equipment I've got atm I probably get a 5% succes rate. :D . I don't use nitrogen.

the investment is cooled down to casting temperature by the way before casting. the final step of the wax burn out cycle heats the investment to 730*C. Before the silver can be poured it needs to be cooled down to 450*C.

[.p4nzer]

I think the Chilly1 isn't a good evaporator, i think the @itor probably is a very good evaporator, this is just my opinion :)

Tom

I can think the same but the point in my post is an other.
Answer me a question:
If we compare two different systems with different compressor (even with the same power), different condensors, different gases, different pipes and captubes (or different expansion valve settings) and different evaporators how can u say what evaporator perform better? U can say what system can perform better in minimum temperature and maximum load but nothing about the evaporator.

U have done a comparaison between your evaporators and find that v8/v6/v2 perform better than the other. Chilly1 evaporator can be too simply, without enough loop, or everything else but it can afford the load of a dual core processor well and can achieve good temperature (i don't know about quad core). The @itor evaporator can do better but how much? 2/3 degrees? Even 5 degrees can be a lot but, with the important differences that can be done by changing the other parts of a single stage, it doesn't shine so much... And almost all the custom evaporator can handle a dual core load...

So I personally need (to talk about the evaporator design) real information from practice.
Obviously I like so much an evaporator with maze and a lot of fins, and I have a mine design in my head, but without any test we have only words.

Sorry for my not very technical english.

[Gray Mole]

Sorry if I came of sounding like an advertisement (Tom :p: ) but it was only to illustrate the point.

I could have used that revision of your evap as well, as it shows the same kind of idea.

Using a plate heat exchanger as a comparison is far more ideal that you think. The basis for it is identical in principle and the reasoning for thinner plates is exactly what I've stated. Too much metal and it's just storing heat that never goes away. Under no load it will end up being a storage for 'cold' but on load it reverses that and as far as I'm concerned trying to create storage within an evap is false economy.

Keep in mind Tom that anything that exchanges heat is a heat exchanger. Keeping that basic idea in mind, and ensuring that your design is based on the same principles and ideals will make for a better evaporator. Just because you have to approach how the heat is transferred doesn't make it any less a heat exchanger. More so, I think, because you've got to fight for every bit of efficiency you can as you've only got the base surface primarily, and the supporting column and plates (or whatever you're using) secondarily.

The maze pattern in theory could end up better than a stepper.

No wasted copper, lots of surface area and a maze for the refrigerant says you?

How's this? Says I:

http://www.xtremesystems.org/forums/showpost.php?p=2381602&postcount=29 see that post for pics.

That would make for an incredibly effective evap, and as the plates themselves are both thin, and all directly attached to the base, would likely be more effective than any stepper could be. 2-3mm base, 1-2mm walls, and 1-2mm channels, about 4mm high (to allow for sufficient mass flow) 'should' make for an excellent evap. Getting it machined that way may prove difficult but in theory that's an excellent evap.

I only used @itor's design as an example because it's designed well, and works INCREDIBLY well. Chilly's spiral system was previously the best I'd used, due to the foundation behind the spiral maximising forced contact of refrigerant to inner walls. I like to think of it as the 'swish factor' but that's just me :rolleyes:

@itor was born with a colorimeter in one hand, and an artists paint brush in the other ;)

Ssilencer is right. I don't get freebies, and other evaps don't cost as much. However, with dealing on levels of higher load (250-275w) these evaps save me a lot of grief in tuning, because of the higher natural efficency. Makes my job easier. Besides, the clear enclosure is amazing to see.

Now I sound like I'm selling it again don't I? Hard not to I guess. They really ARE that good, and not just because of colder overall temps, but I find they have a natural capacity for 230 to 250w depending on the revision, from 180 to 200 on most other evaps I've worked with. That means that you don't have to fight as hard against floodback on tuning, don't have to do as much with heat exchange on the suction return, etc etc to get high capacity.

If someone else had an evap just as good, I'd be saying the same about it. Now that I'm back at work, I'll be seeing about getting an Eleven evap to try out and see how they hold up as well. Looks very promising but until I test one out, I just don't know. I've seen many evaps that looked like they should work, but weren't as good as I'd hoped.

To shorten the story, though, the principles behind a good heat exchanger and a good evap are the same. Now that we've got high load to deal with as 'the norm', we've got to look at increasing efficiency as much as possible within the evap, or deal with ensuring the heat exchange is enough on the suction return to avoid floodback when tuning for high load systems. Probably both.

Cheers

Gray

[Ssilencer]

I had a short length of copper and had just bought a vertical milling adaptor for the lathe which i wanted to try out. I didn't plan that evaporator and didn't want to keep it myself but i didn't want to waste it either, so i gave it to Luke in exchange for a r22 bottle adaptor, he got a good deal in my opinion...

I made this V6 for myself a bit before:

http://www.pclincs.co.uk/Forums/attachment.php?attachmentid=2277&d=1157753524

http://www.pclincs.co.uk/Forums/attachment.php?attachmentid=2278&d=1157753554

http://www.pclincs.co.uk/Forums/showthread.php?t=9972

I made this V8 for myself a bit after, i used a 1 5/8" copper end cap like you do on the @itor which, makes the hollow column evaporators much cheaper and easier to make as well as shorter:

http://img403.imageshack.us/img403/7640/v8dg4be8.jpg

There's some better pictures of that evaporator you posted (V7) here:

http://www.pclincs.co.uk/Forums/showthread.php?t=9971

Luke seemed to like it but i don't think the V7 is a good evaporator either and i never pretended it was. The V2, V6 and V8 are all very good evaporators. At least when there are documented pictures people can draw their own conclusions.

3.

4. If you like the idea of sharing knowledge so much then why don't you share some of yours? :)

That is if you want an enthusiast community and not an enthusiast marketplace...

Tom


So basically you don't have any idea why you did it that way, that's fine

4_I think I already did some sharing in the last 3 years, if you ignored my post s, it is your problem, not mine.

Anyway:

@itor dice:
decile que aparte de analizar las formas del evaporador analice secciones y formas de expansión y la velocidad de transferencia
@itor dice:
junto con las restricciones
@itor dice:
y que si quiere pensar en un evaporador siempre parta del punto de análisis de que lo importante a considerar es que como 1er objetivo debe sacar el calor del micro y transferirlo al elemento de transporte

Now I will translate him for you:
Tell them that apart from analyze the shape, analyze the sections, expansion ways, transfer speed, along with the restrictions, and if he wants to design an evaporator, always start from the point that the most important thing to consider as first objective is how to remove the heat from the processor and transfer to the transport element



About your V8 evap, that design was also made by @itor in 2002, proved not to be a good performer, as the base is too big and too little transfer speed.
The thread is in procooling, the image is lost.

http://forums.procooling.com/vbb/showthread.php?t=4627


Yes, u are right...
But an evaporator like chilly1 or @itor or other evaporators that are in this forum were created for maximize performance, cost and minimize the production problems...
Between a chilly1 or an @itor we have some little difference in performance...
Only with some real test we can say who are the best. We can say that the spiral of the chilly1 provide a massive load and that a maze can be the best way for large surface area and low temps (for example, don't matter my words) but the single stages that we can see in this forum are built with only one of these and the difference between the evaporator are so little (and the variables so high; like gases, compressor, piping, dsh and so much more) that we can't say who is the best.
This post can be the impulse to compare the evaporators and increase our knowledge.

Not at all
@itor evap takes about a day to be made (just 1 evap), it maximizes performance over cost (ask Gosmeyer how much it costs a CNC/machining service per hour...).
Sorry, I tried both, @itor and chilly evaps, and if it was for just a pair of degrees, I didn't pushed @itor for more than a year to design an Evap.
Check Gray moles thread, he tried almost all the "commercial" evaps.

[SoddemFX]

Ssilencer,

That looks like a waterblock thread...? Although from the posts in it, that thread could be about anything...

Could you ask @itor for some pictures of his waterblock since the thread has none? Could you also ask @itor to define "transfer speed"? For an unlagged bar surrounded by a transfer medium with a heat source at one end, what is the profile of heat tranfer? :)

Tom

[Freddie123]

Getting it machined that way may prove difficult but in theory that's an excellent evap.

Oh ye of little faith! Where there's a will theres a way, possibly... Anyone who finds a way to machine it is more then welcome to use the design. Its to scale there. Channels and inner walls are 1mm thick and the whole thing is 1 3/4" in diameter.

[[XC] 2long4u]

Oh ye of little faith! Where there's a will theres a way, possibly... Anyone who finds a way to machine it is more then welcome to use the design. Its to scale there. Channels and inner walls are 1mm thick and the whole thing is 1 3/4" in diameter.

Damn captube would be wider than a channel.:eek:

[Freddie123]

1mm captubes?? Even 0.032 inch cap tube is less than 1mm.

[[XC] 2long4u]

Huh thought mine was bigger. 1mm is .040 but when you get a .032 captube is id or od?

[Freddie123]

If you look on the evap Gray linked there is probably around 1.2mm where the cap goes anyway (it goes behind the central brown circle)

[|RickY|]

http://img393.imageshack.us/img393/8787/64711599xn8.jpg

i'm waiting for response if the channels can be like the pic on left, if yes, i can make 2 layers macisse just solder a division. on this desighn we have a central piece of cooper macisse cooled by all sides

[Freddie123]

I dont think that would be an effective design, way too little surface area and no direct cooling over the core

[|RickY|]

on quad i think direct cooling over the core isnt the way, surface on a macisse 3 layers desighn will be very good

the ideia of this evap will be 3 layers macisse, divided by soldered floors. the ideia os to cool the central piece of cooper by all sides.

[Freddie123]

Well, i'm not so sure. Give it a go though, you might aswell try :D

[Unknown_road]

layers are never good...

the heat conduction to the soldered layers is horrible in comparison to solid metal.

[dinos22]

looks good :)
you going to have trouble machining that one or what? :D

some interesting discussions in this thread

shame that commercial nature has hit the enthusiast community so hard that people are hiding so much from each other nowdays :(

i think in the end we will suffer long terms for this



http://img393.imageshack.us/img393/8787/64711599xn8.jpg

i'm waiting for response if the channels can be like the pic on left, if yes, i can make 2 layers macisse just solder a division. on this desighn we have a central piece of cooper macisse cooled by all sides

[|RickY|]

http://img354.imageshack.us/img354/928/69390872lj2.jpg

almost ready to cnc :)

1: upside look, solid evap
2: this piece will fit inside evap @ 8mm to base and it will divide 1 to 2nd floor
3: this piece eill fit inside evap @ 16mm to base and will divide 2nd to 3rd floor
4: this will close the evap will stay @ 2mm from piece 3 and will create an open 3rd flood to permit the suction on center and an litle accumulator for evap.

i'm thinking on somthing from 3.5 to 4mm to base of evap 8 to 10mm (x 3mm large) for 1st flood channel, 6 to 8mm to 2nd flood (becouse it is larger, 4mm) and 2mm to 4mm open space for 3rd layer
it will be feeded by 2 capps, you an see it is like 2 evap together.


solid evap, 3 layers, 2 chambers

[runmc]

Okay my Portugese buddy :D - what is this word?? :confused: massice :shrug:

BTW - nice design on the evap :up:

[|RickY|]

Okay my Portugese buddy :D - what is this word?? :confused: massice :shrug:

BTW - nice design on the evap :up:


errrrr fixed :shrug:

:hitself:

[runmc]

macisse - no comprende :(

[|RickY|]

it is a mix from portuguese and english :P

i think i can have my evap ready next week :) maybe i have time to make some compare with my other maze and maybe a steeper

[Freddie123]

Massice - Mass ice - Mass flow of refrigerant?

[PPInDaHouse]

Massice - Mass ice - Mass flow of refrigerant?

LOL, no :D

Im portuguese too and Portuguese is very strange for non portugueses :D

Portuguese word "Maciço" means "bulk" in english

Hope you can understand now :up:

Ricky, good luck with the evap, I hope that it will be easy to make :D

[Freddie123]

Ahh so maybe the amount of copper that's there

[|RickY|]

well heheh here is the evap, i think it will be machinered very soon :)

http://img265.imageshack.us/img265/692/image1jd7.jpg

http://img265.imageshack.us/img265/5935/image3co4.jpg

http://img266.imageshack.us/img266/5294/image4nq2.jpg

:toast:

[PPInDaHouse]

awesome evap design :)

Hope that it can handle quads on LOAD very well mate ;)

[gosmeyer]

Looks nice:up: and am looking forward to this being machined. Looks like plate 2 is wire work. That or model shows no tool radius.

[|RickY|]

gosmeyer :)

well, final desighn to be machinered, i'll post the solid evap later, now the 3 pieces that enter on the solid evap. and make the 3 layers.

http://img267.imageshack.us/img267/2782/tampas3dpo5.jpg

lets put that cnc on his limts hehe

i'll try to test it on 300w load tester and compare temps on load tester with more evaps :) anyway i'll love some feedback, some ideias about how it will work hehe

[gosmeyer]

Looks good!!

[|RickY|]

well, i think i'll have it to test soon, sorry but i love it!! :)

http://img267.imageshack.us/img267/7185/evap1qt6.jpg

[tim-]

looks amazing!

[Fhqwhgads6680]

Very unique Ricky, Hope it works out well for ya!

[Freddie123]

What program do you use there Ricky?

[|RickY|]

What program do you use there Ricky?

well i dont know, i only desighn, cnc working and programing isn't with me :( wish i have one heheh

[Freddie123]

I mean for the designs/blueprints

[Pete]

Hey Ricky,

Look amazing on paper in the drawing and snipet of it being made. Must feel nice to see it taking shape!

Wish you all the best with it and good luck!

[|RickY|]

thank you Pete :)

I mean for the designs/blueprints

ok, my mistake :)

for 2d desighns i'v used autocad, blue prints are made by cnc friend, i bet on solidworks or mastercam

:toast:

[PPInDaHouse]

omg, look at this beauty, very nice evap ricky, that CNC is on fire :D

LOL

[|RickY|]

thanks p.p :)

well, only need to be maded the 1st to 2nd floor division and it is ready!!
awsome work on cooper :o

3 fotos: (imagehack is slow now, wait a bit :)
http://img105.imageshack.us/img105/8030/ready3rickyil2.jpg

http://img117.imageshack.us/img117/2127/ready2rickytp7.jpg

http://img381.imageshack.us/img381/6645/ready1rickynb6.jpg

today i'v had a good conversation with a man that works on phase-change for years, he tell-me the way to have less diference between load and idle is multy feeded evaps, he give-me an example of evaps used to cool breaks on a hi speed trains, dont know what kind of trayns but it was a multi feeded evaps. let's see if all this work will have an happy end hehe

[before]

Very interesting bro! :D I want one!

[|RickY|]

i'll try to test and compare with another on load tester :)

[tim-]

Hi ricky,
awesome evaps, hope it perform good. any dimensions.

drop me a pm with price :)

regards
Tim

[Unknown_road]

looks nice ricky!

[gosmeyer]

Hell of a job

[kayl]

very nice evap there ricky :yepp:
That should very nice

[|RickY|]

tomorrow i'll have it :D i'll take 100 photos before solder it! :D

[[XC] 2long4u]

You have to supply 2 captubes with that design? Atleast that is what it looks like. Very nice, I like the mass over the center.

[|RickY|]

yes 2 capps ;)

well finished pics :D

http://img399.imageshack.us/img399/7290/87750166zi9.jpg

http://img177.imageshack.us/img177/3544/60986394ap3.jpg

http://img177.imageshack.us/img177/7677/96857974by9.jpg

http://img177.imageshack.us/img177/3683/74051380gp5.jpg

http://img399.imageshack.us/img399/2853/56655134ie6.jpg

http://img177.imageshack.us/img177/1810/55186961ya4.jpg

http://img177.imageshack.us/img177/4405/60616681dm7.jpg

http://img399.imageshack.us/img399/285/38377834rc9.jpg

[Pete]

WOW, thats come about really fast dude.

Can't wait too see what it's like once built

[before]

Awesome design! :)

[|RickY|]

i simply love-it :D

[Pete]

http://img177.imageshack.us/img177/4405/60616681dm7.jpg

Dual captube?

[|RickY|]

yes, dual cap tube ;)

[Pete]

yes, dual cap tube ;)

Beast! Im really intrested to see what this will do dude.

[theclash]

Great design! What's the advantage to having two cap tubes? Can it handle more load?