Connect with us

#6 screw as heatsink

Discussion in 'Electronic Design' started by mkr5000, Mar 9, 2008.

Scroll to continue with content
  1. mkr5000

    mkr5000 Guest

    Just sharing....

    You know the little hole in a TO220 tab?

    Just a 1/2" to 1" machine screw with nut tightened to it does a very
    respectable job of sinking heat.

    Stays in place well and doesn't shift either like the press on types
    at a sub-penny price. Of course, you need to have the room for the
    screw to protrude.

    Anyone else do this?
     
  2. (snip)

    You should have a quantitative idea of how much additional
    power this bolt gets rid of, by doing some testing. I guess
    the 1/2 inch long one adds a quarter watt of capability and
    the 1 inch one adds a little more than a half watt, at
    maximum allowable temperature. Not much help for that much
    ugly. And not better than the watt or so capability of
    small clip on heat sinks.

    A TO-220 on no heat sink, at right angles to the board gets
    rid of about 2 watts if it is 100C above ambient, but
    something like 40 watts if it is mounted om a huge heat
    sink. So, I think your bolt idea is incrementally helpful.

    http://www.linear.com/designtools/packaging/therresist.pdf

    My favorite trick, like this, for axial lead diodes, is to
    solder a small loop in the lead, before soldering it into a
    board, giving it Mickey Mouse ears. Also incremental, but
    there is no real good alternative for heat sinking axial
    leaded parts, except for soldering them between copper planes.
     
  3. Eeyore

    Eeyore Guest

    It does nothing of the sort.

    Graham
     
  4. mkr5000

    mkr5000 Guest

    Well...not that much being spent in the LM317 anyway but rather than
    hot to the touch in 3 seconds it's at least continually "touchable".

    I know the 317 can run damn hot, but every "little bit" helps I
    suppose....I've seen these used with no heat sink in professional
    designs and they're hot as toasters.

    Can't bring myself to do that.
     
  5. John Larkin

    John Larkin Guest


    My "continuous touch" threshold for a metallic surface is about 52C.

    The bolt helps some. Bolting a fender washer to the can helps more;
    may as well use a long bolt for max effect. The added mass is a bad
    idea in a vibration environment.

    Incidentally, a 6-32 ss machine screw has a thermal resistance of
    roughly 100K/W per inch. So if you use a long one, use two nuts and
    jam the screw so that it pokes out equal distances from the TO-220 on
    both sides. Or use brass, about 1/5 the theta.

    John
     
  6. sycochkn

    sycochkn Guest

    I used to use the transistor as a mounting point for the board and use a
    spacer and place the heat sink side of the transistor touching the chassis
    with and insulator if needed.

    Bob
     
  7. mkr5000

    mkr5000 Guest

    I like the fender washer idea.

    There available in a zillion sizes and CHEAP.
     
  8. MooseFET

    MooseFET Guest


    I use threaded spacers as heatsinks fairly often. It is the same idea
    but lighter for how much heat it can carry way. If you put the screw
    in from below the PCB, it will works as the holding nut and the
    heatsink too.
     
  9. MooseFET

    MooseFET Guest

    You can solder brass washers onto their legs.

    You can stand them up and solder them to a brass object of some kind.

    You can put them on the back of the PCB and press them down onto
    Gappad
     
  10. MooseFET

    MooseFET Guest

    Aluminum is a lot lighter and thus less of a vibration problem.
     
  11. Eeyore

    Eeyore Guest

    I suggest you learn some stuff about thermal mangement.

    'Damn hot' is a pretty flaky way to spec anything. You'll find that
    Tj(max) is speciifed on the data sheet.

    Graham
     
  12. Somewhere I saw a board that had long slots routed for the
    diodes (generous air space on each side of the cylinder),
    with the rather long leads soldered onto large copper pours
    (maybe a square inch, each).

    I guess you can get something like 6 amps average from a
    large axial silicon diode that way, without mounting a heat
    sink. Cheap, but it eats board area.
     
  13. Tim Williams

    Tim Williams Guest

    Yeah, somewhere around 50-60C is where the "ten second rule" passes/fails.
    If you can't hold your finger on it for more than ten seconds, it's too
    hot.

    The disappointing thing about fingers is, they take a lot of energy to heat
    up, being mostly water, and "measuring" the temperature of a lone TO-220 --
    or even TO-92 -- is just about impossible to get an accurate value on.
    When you touch it, it's hot, but you can't feel it. By the time you can
    feel it, it's already near body temperature.

    How much can fingers dissipate, anyway?: I've held my 30W soldering iron in
    the palm of my hand for a minute (obviously from a cold start!), keeping it
    hot but not burning. Let go of it for ten seconds and it gets too hot to
    touch.

    Tim
     
  14. mpm

    mpm Guest

    I agree. A 6-32 screw and nut on the other hand will get you
    marginally more.
    However, for the price of an Avid Thermalloy 7136D, you can't go
    wrong...

    -mpm
     
  15. JosephKK

    JosephKK Guest

    Yes and no. Brass and bronze are a lot stronger against thread pull
    out. And not that big of cost in Theta.
     
  16. MooseFET

    MooseFET Guest


    I think it breaks in favor of aluminum when you do like this:

    Attempted ASCII art:


    -----
    ! !
    ! ! <- Threaded spacer
    ------ ! !
    ---! ! ! !
    ! =================== <- TO220 tab
    --- ! -----------------------
    --- ! ----------------------- <- PCB
    ! ! <- Split lock not shown
    (___) <- Bolt head

    I have used this basic design quite a few times. The lower weight of
    the aluminum, in this case, I think dominates the vibration question.

    Using the power of copy and paste, I will now show case with the big
    washer:

    ( ) <- Second bolt
    ! ! <- Split lock not shown
    ====================== <- big washer (www.bokers.com)
    -----

    ! !
    ! ! <- Threaded spacer
    ------ ! !
    ---! ! ! !
    ! =================== <- TO220 tab
    --- ! -----------------------
    --- ! ----------------------- <- PCB
    ! ! <- Split lock not shown
    (___) <- Bolt head

    In this case, it is likely that brass wins.
     
  17. JosephKK

    JosephKK Guest

    If something is working for you go for it. I just wanted to keep it
    clear that there are choices and alternatives and tradeoffs.
     
  18. GregS

    GregS Guest

    I would also attach a thin piece of aluminum I cut to whatever heat
    dissipation is needed. If its flat on a board, I would try a brass screw.

    greg
     
  19. MooseFET

    MooseFET Guest

    Agreed. There are people reading along who are just starting out. I
    hope this conversation has helped some of them to see the options and
    issues.
     
  20. John Larkin

    John Larkin Guest

    Don't skimp on the lockwasher. The FR-4 will cold-flow in these
    situations.

    John
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-