In researching evaporator designs I ran across a study a few months back detailing the effective of micro pin evaporators in phase change applications. I ordered a Swiftech Apogee GTZ last week as it features a rather nicely milled micro pin base and a mount that can be adapted for our application.
I plan on milling the lip down to the height of the pin matrix so that when a copper plate is brazed on top the pin matrix is flush with the top of the plate. I plan on injecting refrigerant into the pin matrix and having two suction ports on opposite corners of the waterblock base.
I'd appreciate any feedback or suggestions just in case I forgot to consider something. I need to find a machine shop that will mill the base flat, worse comes to worse I'll effectively lap it smooth till the edges are level with the internals.
edit: this was also brought about after talking to Bazx about his F1EE conversion, sounds like it was less than ideal and I rather like my F1EE unmolested
__________________
Team PURE is about kickass pirates and unicorns defeating ninjas.
Last edited by [XC] gomeler; 09-18-2009 at 08:48 AM.
I really like the idea gomeler. The two suction ports should be interesting. What unit do you plan to test this evaporator on? Any estimate as to when you'll be able to start testing this design?
I really like the idea gomeler. The two suction ports should be interesting. What unit do you plan to test this evaporator on? Any estimate as to when you'll be able to start testing this design?
The pile of parts sitting at my feet has two hand valves that I want to use to isolate the evaporator and make it removable. Once the GTZ comes in I'll see how difficult it would be to just sand it smooth, if that fails I'm going to see if Teyber can do the work As with all my plans this will probably take 6-12 months I have just enough money right now to assemble the pile of parts and purchase the refrigerants.
__________________
Team PURE is about kickass pirates and unicorns defeating ninjas.
What are you thinking of using it on.? If a cascade then aren't you going to need a bit more weight/mass in your evap if it is going to have to hold any load.
__________________
Quote:
Originally Posted by To(V)bo Co(V)bo
I think ATI's big surprise is probally a FUNCTIONAL DRIVER!
Swing issues are caused by the inability to remove heat fast enough, a function of surface area. I'm hoping that the pin matrix will reduce swings to just the swing in temperature due to the pressure at evaporation of the refrigerant and a few degrees due to inefficiencies in the thermal transfer. Lots of mass works well for LN2 pots as we humans aren't that adept at pouring a thin stream of LN2 in order to remove all the heat. Aaron's LN2/LHe4 pots have very thin bases and intricate base designs as they rely on the chips not having cold bug but more importantly having excess amounts of capacity so that there is minimal amounts of swing. Hoping this GTZ will give me the same freedoms but with a metering device(cpev) limiting the refrigerant flow.
This won't be the only evap I test, if the system works well I'm going to play with a few other evaporators. This is just the easiest way to get a pin matrix evaporator.
__________________
Team PURE is about kickass pirates and unicorns defeating ninjas.
Why not pass over the core of it as per it's current layout, across the block.
__________________ .:Melted's Hardware:.
.:My Workstation::Sorta Done:.
.:Intel E4300@3ghz(375*8):2gb G.Skill HZ 6400@500mhz:Sapphire X1950pro:Gigabyte GA-G33M-DS2R:550w FSP Blue Storm:80gb and 160gb WD SATA2:Cheap MATX Case:.
.:Panasonic Toughbook CF-T2:.
.:Pentium M 753 (900mhz ULV):256MB:40GB:12.1" Touchscreen:. .:Not too shabby for a 2.4lbs of pwn:.
I've been using Intel ever since AMD put an end to my socket A fun. I was almost going to go back when I saw the cheap AM2 chips, but they suck Viva La Chipzilla!
Why not pass over the core of it as per it's current layout, across the block.
I'm worried about the refrigerant taking the easy route around the pin matrix rather than flowing through it. If you look at the third to last image they use some sort of o-ring to force the water from the inlet into the pin matrix and towards the outlet. If I could fill in the channels around the pin matrix then I would force the refrigerant from one end to the other but I don't know of a suitable way of doing this that will withstand the temperature of brazing besides having thin metal stubs made to lay in the channels and obstruct them. Maybe I can have this one on a second GTZ and compare if the first is a failure.
Any ideas on how to fill those channels?
__________________
Team PURE is about kickass pirates and unicorns defeating ninjas.
You are going too braze this up then? A bit of copper that's a few thou of a mm over and the item compressed together and brazed shut might work.
__________________ .:Melted's Hardware:.
.:My Workstation::Sorta Done:.
.:Intel E4300@3ghz(375*8):2gb G.Skill HZ 6400@500mhz:Sapphire X1950pro:Gigabyte GA-G33M-DS2R:550w FSP Blue Storm:80gb and 160gb WD SATA2:Cheap MATX Case:.
.:Panasonic Toughbook CF-T2:.
.:Pentium M 753 (900mhz ULV):256MB:40GB:12.1" Touchscreen:. .:Not too shabby for a 2.4lbs of pwn:.
I've been using Intel ever since AMD put an end to my socket A fun. I was almost going to go back when I saw the cheap AM2 chips, but they suck Viva La Chipzilla!
You are going too braze this up then? A bit of copper that's a few thou of a mm over and the item compressed together and brazed shut might work.
Guess what I can do is measure the height of the pin matrix and try to purchase some copper sheets that are roughly the same thickness and use some tin snips to cut it up. Will look into it when the GTZ gets here.
__________________
Team PURE is about kickass pirates and unicorns defeating ninjas.
gomeler,
What's the base thickness? Kinda agree with Dave-gota have mass. I like the idea of 2 suction lines, and then what "Y" fitting. Like you said, drop the lid on top of the pins and see what happens. Keep us posted
This was my last one and it had a 6mm base and was 170 grams.
My findings were each increase in mass = better performance staying around 6mm for the base.
This was a pair of suctions but it was the size of IHS and a 4mm base and just not enough mass
In your research/milling did you find an optimal base thickness? I imagine this base is under 5mm in thickness meaning I may run into issues if you are correct.
edit: looking at other images of the block it appears the base may be around 2-3mm in thickness. Wonder if I'll have issues with the evaporator bulging under static pressures.
__________________
Team PURE is about kickass pirates and unicorns defeating ninjas.
As far as MY results go as well as some help from Tom, 6mm is about the best and if I were going to error it would be on the thicker side--remember LardArse when he tested my GPU evaps, and how thick and massive his CPU evap was. It was bad large! But you never know, give it a shot and play with it. I want to know the results!
K, will keep that in mind. I like the idea of thin small yet effective evaporators so I'm going to definitely butcher this GTZ and if it goes well I'll look into having an evap made with the same pin matrix concept but a thicker base.
__________________
Team PURE is about kickass pirates and unicorns defeating ninjas.
raja asked me to convert one of these some months ago as he belived it was a very good water evap and he liked the design
we both just ran out of time and did not get around to it in the end
the shut off valves are a god send and imo worth the extra effort
i would never had been able to keep trying new things without them
and this project is a perfect example
the more data you can collect on what will and what wont work the closer we will get to a working efficient evap
its worth adding that some refrigerents are better then others at removing heat
so you have to think of your target gas/temp when making evaps
as what works well @ -50 may not do so well @ -130
this has been my experience
__________________
1. The first rule of Bench Club is, you do not talk about Bench Club
2. The second rule of Bench Club is, you do not talk about Bench Club
3. If someone says stop, goes limp, taps out, the the benching is over
4. No bummers in the batcave, apart from unkie dave
5. One Bench at a time
6. No air, no water
7. Benching will go on as long as they have to
8. If this is your first night at Bench Club, you have to Bench
As for the flowpaths in the GTZ, it's way more complicated than it looks (unfortunately). In stock trim, there are 3 main flowpaths with various subpaths through the main pathways (starting from a corner and going to an opposite corner).
Flowpath 1: to the left....options here include bypassing the fins and pins totally, or going through the fins.
Flowpath 2: to the right, same options as flowpath 1, just in the other direction.
Flowpath 3 (the interesting one): it starts as an "up and over" dropping the flow into the pin matrix. But what it does from there I really can't tell. All the exits are basically closed except for a few thousandths of an inch at the edges and corners of the matrix. Once in the pin matrix, there's no logical route for flow to go with the stock top. However, considering you'll be only using the base (if it's thick enough), you have other options to craft your own.
I'm not up on how refrigerant flows through an evap (is it like a waterblock?), but if you want to use the pin matrix to the fullest, I'd get rid of at least once bank of fins so that there's a logical exit for the flow...
I've tried sealing off the fins and recessed areas with silicone and testing with water...results were mixed. There's definitely room for improvement over stock config, I'll say that. And if you're willing to machine/sand down some of the fins, you should open up a bit more headroom as well
you could always take some copper, cut it nice and thin, almost like string, and make a little wall dividing the evap in 2 sections(diagonally), with the pins being the only way for the refrigerant to pass thru to the suction line,(cap tube on one side, and a single suction on the other)
like this, im not sure how well it would work... buuuut im going at it with like... a little common sense a little lack of knowledge lol.
red would be where i would mill in a small amount, and get some copper to braze in to create a wall
the blue is where i would dremel/sand paths thru the ridges, so refridgerant could flow thru the ridges and down the paths into the pin matrix(and out as well)
green is just showing direction of flow
the big black dot would be the suction line
the yellow would be the cap tube/entry for refrigerant
![[XC] gomeler's Avatar](image.php?u=17665&dateline=1215408777)
![Send a message via AIM to [XC] gomeler](http://www.xtremesystems.org/forums/images_black/misc/im_aim.gif)
Swiftech Apogee GTZ Evap Conversion?
![[XC] gomeler is online now](http://www.xtremesystems.org/forums/images_black/statusicon/user_online.gif)
As with all my plans this will probably take 6-12 months 
I have just enough money right now to assemble the pile of parts and purchase the refrigerants.
Viva La Chipzilla!
Linear Mode
