OCN2
Results 1 to 7 of 7

Thread: How tall should fins be in a waterblock and how much does it cost to mill just 1?

  1. #1
    Xtreme Member
    Join Date
    Mar 2006
    Posts
    198

    How tall should fins be in a waterblock and how much does it cost to mill just 1?



    the above picture is a cross section of a waterblock i am designing, the vertical bars at the middle is the fins in a waterblock, 5.8mm tall

    the black square is the gasket that seals the waterblock, 1.5mm in height.

    Question 1. how deep should the groove be to hold the gasket in place?

    question 2. since the gasket will be compressed when the bolts/screws are tightened down, how tall should the fins be to compensate for the reduced height of gasket, assuming the top piece is completely flat..

    question 3. any idea how much it costs to mill just 1 or 4 waterblocks with top and assembly for prototypes?
    where can I find a AMD phenom keychain?

  2. #2
    If you're planning to use an o-ring their are groove design specs for each size rubber o-ring. If you're planning to use a square cut gasket similar to an o-ring but with square sides, that's a totally different consideration regarding how much compression you need to allow for and the groove dimensions.

    With a properly designed o-ring or square ring groove, the top plate will be in contact with the lower metal portion of the waterblock. The top should not "float" on the rubber gasket. Thus the fin height should be the same as the outer surfaces where the o-ring or square sided gasket is located. This makes machining easier as all of the the top surfaces including the fins will be the same height, so you don't need to try and compensate for gasket crush.

    The price to mill these waterblocks depends a lot on the material choosen (machinability and cost), the finish desired, the type of machinery used, how much programming is required if done on a CNC vs. manual machining, the shop rate, how busy the shop is, etc. The prices can vary drastically depending on the above. What you'd want to do is obtain a cost and delivery quote from any perspective machine shop once you have all of the machining details so they can provide you with an accurate quote that both of you should agree to in writing. Changes after the fact generally result in expensive surprises unless you find a retired machinists with little to do and an available mill. If you are in no hurry the price is usually lower as the machinists can do it at his leisure which generally means as "fill in" work.
    Last edited by AMDforME; 08-25-2013 at 11:30 PM.

  3. #3
    Xtreme Enthusiast
    Join Date
    Sep 2008
    Location
    washington, DC
    Posts
    507
    your gasket material should have a compressibility ratio - from that you can calculate your fin height. Personally, I would make the gasket groove deep enough so that when the gasket it compressed the top and sides of the block as close as possible without touching.

    We had some custom blocks made, and they ran us about $500 total. But, 'one-offs' are always expensive because you have all kinds of other fees - set-up etc.

    I have no idea of your age, but if you are in school and your school has a machine shop you could check with them to see what they have to say. We use the University's shop all the time. Also, the machinists there can review your drawings and let you know if they see any problems before anything gets started.

    good luck
    [SIGPIC][/SIGPIC]

    Ice Dragon on Facebook

  4. #4
    Xtreme Member
    Join Date
    Oct 2009
    Posts
    238
    1. For a 1.5mm gasket you'd be looking at a groove depth of 1mm and a width of 2.1mm.

    2. The top and bottom peice should lay flat on top of each other when fastened together, and the oring compressed. SO in the example drawing, you'd make the fins no higher than the base space you have, which is 4.3, 4.2 to be safe.

    3. I'm a fitter and turner, my company has made a few one off blocks, they're not cheap. Labour is around $88 an hour generally in trade + materials + set up time, for thin fins looking at 3+ hours of setup, have to do shallow passes with a thin cutter, which means higher RPM spindle which means more expensive machine again
    General calc for tool surface (cutter rotation) speed for carbide tips (most common) is 27000/diametre of cutter. For a 1mm cutter you'd need a spindle ideally, capable of 27,000rpm
    You'd be looking at 400-500 just for one block. It would have to be done on a CNC mill due to the oring groove's rounded edges, and the precision and speed needed for the fins.
    Plus if you're making it out of copper it has to be plated. Platers usually have a minimum order quantity.

    The first one is always the most expensive, the prototype. Once the lengthly setup is done, can run subsequent ones 10-20 minutes each depending on internal complexity. so about $15-30 for labour for subsequent blocks + material cost.
    Patience is the only talent.

  5. #5
    Xtreme Member
    Join Date
    Feb 2010
    Location
    Nashville, TN
    Posts
    155
    A few questions ...

    What type of pressure drop are you expecting in the WCing loop for this block. If it's really not much, you can tighten up the fins to 1mm each with a thickness of about 1.2mm to 1.4mm. The base should be no greater than 1.8mm thick. Reason, you need the heat transfer vs. having a hot-plate base.

    A few years back, my dad designed the Apple G5 LCS blocks, Dell H2C 700 series blocks and some of the Corsair blocks. Performance is in the knowledge of the pressure drop and the Delta-T. Now, picking a design, the right pump and the RAD is another question. haha
    Support Your Local Sheriff - At high noon

    My Heatware http://www.heatware.com/eval.php?id=63611

    The pump Maverick, BMaverick ... has a massive hoard of a Laing DDC pump stash


  6. #6
    Xtreme Mentor
    Join Date
    Feb 2007
    Location
    Victoria, British Colombia
    Posts
    2,677
    Its a function of many things thats non linear and very difficult to solve. There are simplified approximations involving hyperbolic trignometric functions.

    The short answer is longer the better. Most fins on waterblocks seem too short to me to keep costs down. Longer fins = lower base temperature for same heat transfer.

    Take a look at the parker o-ring handbook for help designing your oring gland.

    Cheers
    mentally confused and prone to wandering

  7. #7
    Xtreme Guru
    Join Date
    Dec 2009
    Location
    Latvia, Riga
    Posts
    3,959
    teyber: Longer fins - not just a bit worse heat transfer further from base, but also more cross section area => less flow speed/less turbulence, thus worsening heat transfer. It can even be taken to extreme, as in something like submersed normal air cooler with also lower performance then normal blocks, dispite enormous fin area/size.
    I'd say - not worth to reinvent bicycle, simpler to copy block params from mass produced vendor blocks for cooling what you need to. LC vendors had fine tuned optimum dimensions and proved that in real blocks and real performance, it just seems that they are too short & less efficient. Sign of dimensions being close to optimum - slowing down of temp drop progress for top block makers and them performing within one or two degrees.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •