Problem with heatsink C/W calculation?

Discussion in 'General Electronics Chat' started by Porkchop, Feb 13, 2007.

  1. Porkchop

    Thread Starter Member

    Feb 5, 2007
    17
    0
    I seem to get a odd result? :confused:

    Here is what im doing

    θsa = (Tj - Ta) / Q - (θjc + θcs)

    So for my rectifier it follows

    (150 - 50) / 8.4W - (1.5 + .09) = 10.31 C/W min ok fine

    now another

    (125 - 50) / 50W - (4 + .09 + .8) = -2.89 C/W :confused:

    I keep getting negative numbers?

    Does this mean i have no chance of keeping the junction temp below <125oC of this T0220 package?

    Cheers peeps
     
  2. Papabravo

    Expert

    Feb 24, 2006
    10,179
    1,800
    Where did the number 4 and the 0.8 come from. It would help if you would explain what you are trying to do and then applied the equation consistently changing only one thing at a time.
     
  3. Porkchop

    Thread Starter Member

    Feb 5, 2007
    17
    0
    The first equation is a working example

    The second is for a different component.



    (4 + .09 + .8)

    That part is (θjc + θcs)

    Since this is a diffrent component the datasheet specifies a junction to case C/W of 4 the .09 and .8 is my mica washer and grease mounting resistances.
     
  4. n9352527

    AAC Fanatic!

    Oct 14, 2005
    1,198
    4
    That is exactly what it means. You have no chance of getting the component at 125degC with 50W of power dissipation. If you notice, the junction to case is already 4degC/W, with 50W and case temperature of 50degC (assuming zero case to air thermal resistance), this would give you minimum 200degC junction temperature. Add to that the washer, grease and heatsink to air thermal resistance, then you are looking at way more than 200degC.

    BTW, what kind of device and circuit that you have in mind? 50W dissipated power is high for TO220.
     
  5. Porkchop

    Thread Starter Member

    Feb 5, 2007
    17
    0
    Its a LM338 Reg i was originally using a TO3 package with a 1.5 C/W but in order to make room i moved to the TO220 package so the IC could sit on the edge of the pcb at right angles.

    With the TO3 i had it flat on the pcb which also nab 60% of the area for the components.


    Its just what i feared :( darn

    Cheers bud
     
  6. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
    There's no rule that says a TO3 has to lie flat. I've seen several applications with TO3 regulators mounted vertically. There are different varieties of heat sink available for TO3 vertical mounting as well.
     
  7. Porkchop

    Thread Starter Member

    Feb 5, 2007
    17
    0
    Is there anything that can be cliped on the pins ? or do i have to solder and heat shrink wire?

    Just seems a bit ghetto looking.

    Can anyone show me a piccy of a pro install using a TO3 package id be interested in how the wiring looks :)
     
  8. thingmaker3

    Retired Moderator

    May 16, 2005
    5,072
    6
  9. Porkchop

    Thread Starter Member

    Feb 5, 2007
    17
    0
    Cheers for the help !

    I finally got one of those connectors for the TO3 its great!

    I have another question about the heat sink resistance calculation.

    How do you account for air flow say from a fan in the formula?

    Forced convection cooling

    What do i need to know please ?

    Could some one point me in the right direction please :)
     
  10. n9352527

    AAC Fanatic!

    Oct 14, 2005
    1,198
    4
    Unless the heatsink manufacturer give a figure for forced convection, then you are out of luck. You could measure it, but you would need to be equipped for that.

    Or you could guess and lower the thermal resistance a bit and have a go with a constant maximum load and see the temperature.
     
Loading...