Results 1 to 6 of 6
  1. #1
    My Rusty Knee
    Reputation: Stuart B's Avatar
    Join Date
    Mar 2005
    Posts
    1,167

    Thermal management questions...inspired by mhaskell's thread

    Hi

    This thread is aimed both at mhaskell AND anyone familar with these issues. Its been a long while since I did thermal management. I know you said you are an enginner in the thermal sides of life and I didn't want to go off topic. I think this might be useful in it own right.

    mhaskell again thanks for posting up....its made me think about thermal issues again. Something I want to consider when I do my own housing.

    I think I have reminded my self of the basics. Its been a few years since I did it at uni hehe. My calcs are coming out close enough to yours to give me a warm feeling. I used this to remind me.

    Just to make sure I am reading things right. Are the figues in your table in your thread are for convection only? worst case?

    Also, I am learning to read the heatsink curves described here

    such as on this page. I have some 2288BG I want ot play with. As I understand it, the airflow curves are for a ideal air slow....which I guess means throgh the fins. for convection I assume the best orientation is to have the fins going verticaly, and not horizoantally which would be ideal for airflow on a bike. Is there a rule of thumb to estimate the non ideal air flow thermal resitance....or is it just too complex to put a fudge factor on?

    Is it best to have the heatsink oriented for best convection to prevent over heating damage or for best thermal resitance in airflow and risk over heating? The controller has lower power settings...which is fine for rme....but if I build any for my mates they not be so careful.

    also....is there a way to estimate the thermal resistance of the housing its self? from its surcae area and material and so on.
    Cheers

    Stu
    What exactly is a rigid hard tail?

  2. #2
    mtbr member
    Reputation:
    Join Date
    Aug 2004
    Posts
    248
    Hi Stuart, In my last post I'm afraid I oversimplifed things a bit. I will post tonight in much more detail, when I have more time, including how I measure Rjb values.


  3. #3
    My Rusty Knee
    Reputation: Stuart B's Avatar
    Join Date
    Mar 2005
    Posts
    1,167
    excellent. cheers fella.

    I might post my calc method later...you can be teacher if you like lol. I think I am on track as I got the same numbers as in chrisms reply.

    Stu
    What exactly is a rigid hard tail?

  4. #4
    My Rusty Knee
    Reputation: Stuart B's Avatar
    Join Date
    Mar 2005
    Posts
    1,167
    The calcs I used were

    Tj = Ta + NPd(Rjs/N + Rsh/N + Rha)

    N is the number of LEDs
    Pd Power dissipated by each LED
    Rjs Junstion to sulug Thermal res
    Rsh slug to heastsink thermal res
    Rha heatsink to air thermal res

    I transposed it to find the required hestsink for a given the other parameters.

    Rha = (Tj - Ta - Pdtot(Rjs/N) - Pdtot(Rsh/N))/Pdtot

    I have found the multiple path issue quite interesting. I think the light I am going to build is going to have maybe 6 leds driven less hard. My previous light calcs had been for 3 or 4 seouls to compete with a HID. I knew that if the LEDs were driven less hard they are more efficient. Now if more leds the heatsink can be smaller due to the parallel thermal paths as well as the leds dispating less power for a given brightness.

    I have used the Tj calc to correct my lumen calcs too. I used excel to provide the polynomial fit for Tj verses brightness.

    I reckon I can get 500 lumens from about 8 watts (include led driver efficiency). My HID needs about 12-13 to get the same level of brightness .

    The light could have a turbo mode to provide about 800 or 1000 lumens. Not sure if the heatsinking requirents are possible. I think for 1000 the housing/heatsink would need to be sub 1W/k to keep the Tj down to 80C or so. The Seouls light out put drops to about 90% at 80C. They also blue shift when they get hot. The 6th LED will be red for my light...these are even more sensitive to temperature, its output halving for a temps around 70C.

    Is it possible to estimate the Thermal resistance of a housing before bolting heatsinks on?

    Stu
    Last edited by Stuart B; 05-21-2007 at 02:09 PM.
    What exactly is a rigid hard tail?

  5. #5
    mtbr member
    Reputation:
    Join Date
    Nov 2005
    Posts
    500
    I think you should consider the movement of air across the body of your light into the equation. With the first light that I built, I incorporated fairly large heatsink fins into the design, only to find that as long as I kept moving 8-10mph (my minimum normal riding speed) the heatsink fins were unnecessary (as tested on later models). This might allow you to reduce the cooling surface area of your formula. Of course, if it is a requirement that the light is to be used on full power in a stationary fashion for any longer than a minute or two, you must consider that as well.

    Good luck.
    - Allen

  6. #6
    My Rusty Knee
    Reputation: Stuart B's Avatar
    Join Date
    Mar 2005
    Posts
    1,167
    Quote Originally Posted by achesalot
    I think you should consider the movement of air across the body of your light into the equation. With the first light that I built, I incorporated fairly large heatsink fins into the design, only to find that as long as I kept moving 8-10mph (my minimum normal riding speed) the heatsink fins were unnecessary (as tested on later models). This might allow you to reduce the cooling surface area of your formula. Of course, if it is a requirement that the light is to be used on full power in a stationary fashion for any longer than a minute or two, you must consider that as well.

    Good luck.
    - Allen
    My calcs do include it on my spreadsheet. I have been looking at both cases. I have neen debating it in my my head. For me its fine to not be able to use full power when still for long. I would just switch it down to min to save battery life and prevent heat up. If it turns out that the heatsink (include housing when i say heatsink) isn't excessive it might be worth having full power available at 0mph. If I build any for my mates I don't want them cooking it hehe.

    The 6 led will need a bigger housing than a 3 or 4 led housing. The 6 led needs a smaller heatsink than a 3 or 4 light. the bigger housing should mean that if any extra heatsinking is required...it will be smaller than required by a 3 or 4.

    if heatsinking isn't required for a 3 or 4 it should be ok for the 6 too. I am hoping extra heatsinking isn't required if sensible junction temps can be acheived. the housing is already bigger than a 3 or 4 led light.

    I am going to use the nflex controllers. I think if were to build my own (buck puck controlled by a pic probably) it would use a thermister or something in the housing to dim the light when it gets hot like the top lupine led lights do.

    I will probably only really know when I actually make a housing instead of spreadsheet jockeying hehe. although i do enjoy jockeying this spreadsheet hehe.

    Stu
    What exactly is a rigid hard tail?

Posting Permissions

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