Page 1 of 2 12 LastLast
Results 1 to 13 of 18
  1. #1
    Join Date
    Jul 2004
    Posts
    9

    Question

    Hi everyone! I'll greatly appreciate the time and patience put into my newbie questions

    My project consists of a refrigeration unit chilling the water of a computer watercooling system (I know...insulation EVERYWHERE to prevent condensation). I will be modifying a small refrigeration unit, originally designed to cool down a laser diode, by installing an evaporator consisting of a 10-15' 1/4" ID long copper tube installed in an insulated 4-5 gallon reservoir with a lid. The unit has a 1/6hp compressor using R134

    I would like to have the water down to anywhere beteen 20-30f. Yes, I'll be using an antifreeze mixture. Mind you, the water will have a heatload on it of about 80-90f since it will be cooling down the CPU and the graphics processing unit. Ambient in the room is ususally about 75F.

    And now for the questions
    1) Is it better to use a TXV or a capillary tube?

    2) If I should use a TXV, how do I know which one to get?
    A) Adjustability of the evaportor temp would be nice to
    fine tune the unit. Is this possible with a TXV?

    3) If a capillary tube is better, how do I know how long it should be and what ID?

    4) When I construct my evaporator, which is more efficient? Just one long copper tube spiralling around or a grid? By a
    grid I'm thinking of 1 long copper tube on each side with several Ts on it connecting to Ts on the other tube. I guess you could say it will look something like this:

    In
    |
    |
    |-------|
    |-------|
    |-------|
    |-------|
    |-------|
    |-------|-----------out


    5) When torching the pipes together, should I use flux AND solder or just solder?

    and last but not least:
    6) When I repressurize the system, how do I know what pressures are best on the low and high side, and what do the differance in pressures mean? For example: which gives you a colder evaporator temp, a higher or lower pressure on the low side? Yes, I'll be getting some cheap little R134 gauges just for this project.

    Thanks guys!

    [Edited by flybye on 07-17-2004 at 11:29 AM]

  2. #2
    Hi There,

    What you really need is a consultant. You need someone to give you the plans for this project. The you need someone who work with this stuff on a regular basis.


    BOSS wouldnt let me post all the questions on this site cause it would occupy too much "system resources".

    You dont need a shade tree mechanic helping you experiment here.


    If this project is worth doing, then it is worth doing properly.



  3. #3
    Join Date
    Jul 2004
    Posts
    9
    Originally posted by R12rules
    Hi There,

    What you really need is a consultant. You need someone to give you the plans for this project. The you need someone who work with this stuff on a regular basis.


    BOSS wouldnt let me post all the questions on this site cause it would occupy too much "system resources".

    You dont need a shade tree mechanic helping you experiment here.


    If this project is worth doing, then it is worth doing properly.
    Thing is, I don't KNOW anyone that deals with this on a regular basis! If I did, I wouldn't be posting on a forum

    In the PC world, all anyone ever does with their PC water cooling system is experiment. I've searched and searched and searched the internet and can't find any info on this. I'm hoping someone who does this for a living will pass by and give me some answers

    I even went to a MAJOR refrigeration distributor in my area, and they couldn't even help me either I even tried emailing a few engineers and never got a reply back!

    I mean, are these questions THAT difficult to answer?

  4. #4
    Join Date
    May 2004
    Posts
    11,858

    Yes flybye, they are...

    Here is the problem.

    You want a 1/6HP condenser to cool 4 to 5 gallons of water to 20*-30* using an antifreeze mixture.

    You have no design and no experience in refrigeration and no EPA license to experiment with the handling of refrigerant. You don’t even know what solder to use or what kind of copper to be used.

    What R-12 is saying, and very politely I might add, is that you will need someone in this field with experience in research and development that may also have a license to deal with a Federally controlled substance.

    Please don't let one MAJOR refrigeration distributor in your area stop you from finding the right guy to help you design your system. As a matter of fact I love to deal with computer systems and an IT for Harley Davison is my next door neighbor.

    If I can offer this... look up a couple of numbers for supply houses (not distributors, they are all salesmen) or companies that offer commercial refrigeration service or even a tech school near by. Tell them you would like to have a guy that is into computers help you design and build a kick A** system that is water cooled. Then when you find this guy and if he needs help, give us a ring on this site. I, for one, would love to help someone that can deal with your issues but experiance and license is a must.

    I bet you will get a fish on the hook that will become a great friend of yours. The next thing what will happen is you will be charging to build your computer buddies systems and be bigger than Dell.

    Take this with a grain of salt and show a little respect and gratitude when a seasoned professional, such as R-12, tries to politely give you advice and don't act like a 3 year old that can't get his candy.
    I mean, are these questions THAT difficult to answer?
    This will go a very long way in your future as a great computer wiz and becoming a respected young man.

    Good luck in all things.

  5. #5
    Join Date
    Jul 2004
    Posts
    9

    Re: Yes flybye, they are...

    Originally posted by lusker
    Here is the problem.

    You want a 1/6HP condenser to cool 4 to 5 gallons of water to 20*-30* using an antifreeze mixture.
    Please tell me why it is a problem.
    Is the HP rating TOO small?
    Actually, the 4-5 gallon reservoir might be a little overdoing it. I'm thinking more along the lines of 2-3.

    Originally posted by lusker
    You have no design and no experience in refrigeration and no EPA license to experiment with the handling of refrigerant.
    Did you actually think I was going to vent the stuff into atmosphere? You probably did.

    I was planning on taking the unit to a shop, having them dispose of the R134 properly, then I'll begin to take it apart and construct the evaporator I want. Finally, I'll pressure/vacuum test it to ensure I have no leaks, and then I'll have it filled up again. In my area, you can buy R134 off the shelf of any regular autoparts stores.

    Originally posted by lusker
    You don’t even know what solder to use or what kind of copper to be used.
    I guess that is why I am asking all these questions
    I have found sources for refrigerant grade copper.

    Originally posted by lusker
    .....Take this with a grain of salt and show a little respect and gratitude when a seasoned professional, such as R-12, tries to politely give you advice and don't act like a 3 year old that can't get his candy. .....
    In NO way did I intend to disrespect him, and if it seems as if I did, then I apologize. I was merely stating that I have emailed questions regarding this project to several engineers and I have never received a reply back. Which is when I asked if these questions are really that difficult to answer. It feels as if noone even wants to take the time to just help a regular guy with a few simple questions for a small home project.

    Please understand, there are only several (actually I think only one) manufacturers that produces a chiller for a computer at an affordable level for a power computer user. Almost everyone else is basically constructing something from scratch. I've seen people modify anything from a 20 year old refrigerator to a scrapped beer chiller running different mixtures of antifreeze and whatever else anyone can find that prevents water from freezing at 0C.

    There are absolutly no sources that tell us power users EXACTLY what to build and what components we will need. Which is why I came here. I don't want to build my system and be experimenting along the entire way. This is why I did a search for HVAC forums, found this place, and hoped to possibly get some of my questions answered in order to be able to build a system which will work as efficiently as possible. I am in no monetary situation to hire a specilist for such a tiny home project, and if I was, I wouldn't even be here. I would have gone to a shop and told them "I need this and this and that done. I don't care how much just do it." Again, this is why I came here hoping to get the right answers from the pros.

  6. #6
    Join Date
    Jul 2004
    Location
    ABQ. NM
    Posts
    35
    Your gonna need a water agitator, a 3000 btu unit for 5 gal. tank, and go w/a txv. Or go buy a used Haskris or Mltiplex unit.

  7. #7
    Join Date
    May 2004
    Location
    Caldwell, ID
    Posts
    395
    The only people who don't experiemnt are the people who build piles of these to make a living, therefore they are not giving out info on how to do it.

    The people who are giving out the information have only built a few, each one of them different. They don't spend a lot of time fine tuning to the gnat's ass because it is not worth it pricewise. With equipment this small, the parts are so much cheaper than the labor it is ridiculous. You spend as much time on a 5 ton as a 1/6 ton (not reall but you get my point). It is not worth the time to experiment to find a smaller cheaper, more efficient XYZ part, because it costs you 8 hours to do so and saves you $.25 on the part.

    On top of the previous correct statements about your project, here is some more input:

    1. What is the heatload? You never mentioned that. A 1/6 HP unit will not do 1/6 tons. Less than half of that or so @ 20 F with glycol.

    2. The size of the reservoir is mostly going to be determined by how big your coil is, based on the surface area you need. The size of the tank generally does not impact the heat load. If you take 1 ton out of 5 gallon tank, and a heat load puts in 1 ton, the temp stays the same. Same thing if you have a 2000 gallon tank. The difference is the increased surface area of the tank will absorb more heat.

    3. What kind of pump do you need? The power of the pump will be dissipated into the water. Normally this is not a big deal as the pump power is a small fraction of the refrigeration system power. But for 1/6 HP, it could be significant.


    4. Just build a coil. Forget the manifold thing. It is not worth the time and chance for leaks. Do increase the flow of the glycol across the evaporator, you might want to consider some time of counterflow heatexchanger versus just dropping a coil in the tank. Unfortunately, it might be hard if not impossible to find a catalog item for this small capacity. You might be able to dink around with a liquid line/suction line heat excanger.


    5. Cycle off of chilled water temp.

    6. Use a TXV. Cap tubes are cheaper, but it is easier to spend the extra ten bucks and find something that is easier to select out of a catalog.


    This may be none of my business, but why all the trouble to modify a used unit? Seems like you would be better off buying one. Unless this is a mental exercise, you are going to spend way too much time for something super cheap.

    I have people ask me all the time to build them a chiller for their fish tank. I tell them to go on EBAY and buy one for $250.00. There are enough six year old Chinese kids building them that I can't compete.

    I would recommend something from Peerlick, or Mutliplex (as gastech suggested). They make things this size for this temperature all the time.

    Good Luck,

    Clyde

  8. #8
    How are you with 3-D graphics?

    I have a project that needs a nerd.

    Maybe we can swap working on each other's projects.


    The price thing can be dealth with.
    First find some product, already developed, which does what you wish done.

    Take, for example, a soft drink system at the local conveinence store around the corner.
    This system has a compressor which removes heat from chill water. This chill water cools the drink lines feeding your syrup lines into the drink heads. It's called a post mix machine.
    There are literally tons of them in Restaurant Supply retailers. Look one up in the phone book.
    They plug into 115 volt.
    They are usually working systems. Just fill up the tank with water and your all set to go.



    There is one thing that concerns me. The actual cold connection with your chip.
    If you were to use metal to metal to transfer your heat over to the cold metal ... you would suffer severe moisture condensation.
    This would not be good around electronic gear.
    What might work better is to have an actual micro air handler to blow chilled air directly onto your micro chip.

  9. #9
    Join Date
    Jul 2004
    Posts
    9
    Originally posted by clydemule
    1. What is the heatload? You never mentioned that. A 1/6 HP unit will not do 1/6 tons. Less than half of that or so @ 20 F with glycol.
    Correct me if I'm wrong, but this is how heat load is calculated:
    Specific heat of H20 = 1 BTU/hr/F
    Weight of 1 gallon of H20 = about 8lbs/gallon
    Flow Rate (gph) x Coolant Weight (lbs/gal) x Specific Heat of Coolant x ^ToF = BTU/HR

    Even on a setup with a reservoir of just a couple of quarts, the temp differance between before the water enters the CPU cooling block and after it exits is only a fraction of a degree. I was wrong about the heatload before. I asked one of the guys at another site who is into these setups and he told me the temp differance is almost nill. These simple setups generally consist of a small radiator that fits in a computer case, radiator fan, water pump, and the cpu water cooling block.

    So lets say I were to use a 3 gallon reservoir; 3 x 8 = 24lbs
    Flow rate of 150gph
    Let's leave the temperature differance of about .5f
    So.....
    150 x 24 x 1btu x .5f differance = 1800 BTU/hr

    The water pump is rated at 10 watts.
    746 WATTS = 2545 BTU/HR so the pump generates another 34btu/hr. So for our simple little test calculations we have 1834 BTU/hr

    Heat Load = 1834 BTU/hr

    Originally posted by clydemule
    2. The size of the reservoir is mostly going to be determined by how big your coil is.....
    Ok. I figured 5 gallons will be too much and a 3 gallon reservoir will suit me just fine, so....
    1 gallon = 231 cubic inches

    So my 3 gallon reservoir = 693 cubic inches or .4 cubic feet

    Originally posted by clydemule
    3. What kind of pump do you need? The power of the pump will be dissipated into the water. Normally this is not a big deal as the pump power is a small fraction of the refrigeration system power. But for 1/6 HP, it could be significant.
    The water pump is rated at about 10 watts which I've allready included in the above heat load. Assuming I calculated the heat load correctly
    Originally posted by clydemule
    4. Just build a coil....
    That's what I'm planning

    Originally posted by clydemule
    5. Cycle off of chilled water temp.
    I'm also planning on connecting a digital thermostat with a relay to shut the entire thing off at the desired temperature. Assuming the desired temperature is even reached.....

    Originally posted by clydemule
    6. Use a TXV. Cap tubes are cheaper, but it is easier to spend the extra ten bucks and find something that is easier to select out of a catalog.
    Now how do I know which one to choose? Do they go by BTU rating as well?

    Originally posted by clydemule
    This may be none of my business, but why all the trouble to modify a used unit? Seems like you would be better off buying one. Unless this is a mental exercise, you are going to spend way too much time for something super cheap.
    I don't mind answering that
    I want to build a chiller that is completly self contained and fits perfectly under my computer. I've looked around at some of the chillers available out there. Not only are they a bit pricey, but I enjoy putting things together. My main hobbies are cars and computers, and I certainly don't mind tinkering around. This can not possibly be any harder than running my car run after run after run while adjusting the wastegate and turbo prespool solenoids via my programmble ECU and manually adjuting the flaps on the exhaust manifold in order to be able to get the perfect amount of desired boost when I'm switching over from turbo #1 to turbo #2 If you have a modified twin turbo sequential car, then you'd understand the traction issues that can go on exiting a corner on a track during transition from one turbo to another if boost is not finely tuned.

    Well Clyde, if my heat load is correct, can a 1/6hp R134 compressor do the job?

  10. #10
    Join Date
    Jul 2004
    Posts
    9
    Originally posted by R12rules
    How are you with 3-D graphics?
    Not Yet
    I've completed a course in AutoCAD 2D. I've been playing around with its 3D features, but am in no situation to say "Yes, I can draw that for you in 2 hours and it will look like a pro did it"

    Originally posted by R12rules
    I have a project that needs a nerd.
    Yep, people studying to be a ME usually end up being nerds

    Originally posted by R12rules
    Maybe we can swap working on each other's projects.
    If I were a master at AutoCAD 3D, I certainly wouldn't mind.

    Originally posted by R12rules
    ....It's called a post mix machine.....
    That is actually a VERY interesting idea. Kinda falls along the lines of the beer chiller I know someone is using.

    Originally posted by R12rules
    There is one thing that concerns me. The actual cold connection with your chip.
    If you were to use metal to metal to transfer your heat over to the cold metal ... you would suffer severe moisture condensation.
    This would not be good around electronic gear.
    What might work better is to have an actual micro air handler to blow chilled air directly onto your micro chip.
    Here's an exmaple of a CPU water block:


    As far as condensation prevention, there are a variety of ways we can prevent it from forming at and around the CPU. Neoprene rubber, conformal coating, die electric greases, etc. I've even been looking into a brush on type liquid insulation. Between the CPU and the block we generally use thermal paste.

    [Edited by flybye on 07-18-2004 at 07:11 PM]

  11. #11
    Join Date
    Dec 2003
    Location
    Nevada
    Posts
    1,815
    Why only.5 degree water diff?
    Quote
    “Engineers like to solve problems. If there are no problems handily available, they will create their own." Scott Adams

    "We can't solve problems by using the same kind of thinking we used when we created them."
    Albert Einstein

  12. #12
    Join Date
    Jul 2004
    Posts
    9
    Originally posted by hvacbear
    Why only.5 degree water diff?
    The actualy footprint of the waterblock that comes in contact with the CPU is about .0625 sqaure inches which is basically the square area you see in the middle of the water block.

    A CPU, at full load, can get as hot as 122F+ A typical setup will consist of a fan bolted to a radiator about 5.9" x 5" x 1.69" with a reservoir holding a few quarts of water. This entire setup is actually enough to keep the temperature of that CPU to ambient.

  13. #13
    Join Date
    Jul 2000
    Location
    Guayaquil EC
    Posts
    10,080
    You should check out this old thread. Go to where Chilly1 starts explaining all about what he's doing to cool his CPU........ http://hvac-talk.com/vbb/showthread.php?threadid=36554

    How about -150 Deg F at the CPU cooler?!!!

    [Edited by icemeister on 07-18-2004 at 08:01 PM]

Page 1 of 2 12 LastLast

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
Comfortech Show Promo Image

Related Forums

Plumbing Talks | Contractor Magazine
Forums | Electrical Construction & Maintenance (EC&M) Magazine
Comfortech365 Virtual Event