+ Post New Thread
Results 1 to 13 of 13
  1. #1
    Advanced Member level 4
    Points: 6,895, Level: 19

    Join Date
    Dec 2015
    Location
    Madrid, Spain
    Posts
    1,220
    Helped
    300 / 300
    Points
    6,895
    Level
    19

    Cooling SuperSO8 package

    Hello everyone,

    I am planning to use BSC117N08NS5 which is in a SuperSO8 package.

    Max power it will dissipate is 1 W and I want heatsinking via PCB itself (PCB horizontal position). I see that Rth(ja)=50 K/W at 6cm^2 area.
    I want to allow Tjunction=100ºC in a 25ºC ambient, so Rth(ja)max allowed is 75 K/W.

    My question is, how do I know how the area influences the thermal junction-to-ambient resistance ? Is it a linear relationship ? Anyone can provide some docummentation/explanation ?

    Thank you for your time !

    •   AltAdvertisment

        
       

  2. #2
    Full Member level 6
    Points: 3,150, Level: 13

    Join Date
    Sep 2013
    Posts
    368
    Helped
    145 / 145
    Points
    3,150
    Level
    13

    Re: Cooling SuperSO8 package

    Your selected device commanded cooling area is 6cm^2 with 75uM copper pad connected to drain. so the heat will transferred to
    PCB copper pad and dissipate in to air, if more area more deipation

    Click image for larger version. 

Name:	thermal-resistance-definition.png 
Views:	0 
Size:	2.2 KB 
ID:	147176


    Referance
    http://www.ti.com/lit/an/spra953c/spra953c.pdf


    1 members found this post helpful.

  3. #3
    Super Moderator
    Points: 248,981, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    43,336
    Helped
    13173 / 13173
    Points
    248,981
    Level
    100

    Re: Cooling SuperSO8 package

    The posted formula is applicable for the vertical heat flow through an interface with limited thermal conductivity, e.g. an insulation washer. It does not describe the horizontal heat flow through a copper plane, the topic is beyond the scope of the quoted application note.


    Obviously, you have an (inverse) linear relation between copper area and thermal resistance as long as the finite horizontal conductivity can be ignored. Above this range, the thermal resistance is converging against a constant number. You can estimate the overall thermal resistance by a simplified discrete model of concentric rings.


    1 members found this post helpful.

  4. #4
    Advanced Member level 4
    Points: 6,895, Level: 19

    Join Date
    Dec 2015
    Location
    Madrid, Spain
    Posts
    1,220
    Helped
    300 / 300
    Points
    6,895
    Level
    19

    Re: Cooling SuperSO8 package

    I recognize post #2 formula as the standard conduction thermal resistance for heat flow through that surface.

    I believe that is what I need because I did not understand what Infineon was saying with this (page 4, note 1):
    "PCB is vertical in still air."

    Does that reference the physical position of the PCB ? i.e. PCB is in vertical position ? Or that the flow of heat is vertical through the PCB and hence post #2 formula can be applied ?

    Formula #2 can be applied no matter whether the PCB is in vertical or horizontal position. That is why I am wondering why Infineon is saying "PCB is vertical in still air".
    Why vertical makes a difference with horizontal ?



  5. #5
    Super Moderator
    Points: 248,981, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    43,336
    Helped
    13173 / 13173
    Points
    248,981
    Level
    100

    Re: Cooling SuperSO8 package

    I believe that is what I need because I did not understand what Infineon was saying with this (page 4, note 1):
    "PCB is vertical in still air."
    That's related to natural convection heat transfer. You find empirical formulas for various configurations in technical thermodynamics text books.

    Vertical configuration is more effective due to hot air ascending and skimming along the surface.


    1 members found this post helpful.

  6. #6
    Advanced Member level 4
    Points: 8,221, Level: 21
    d123's Avatar
    Join Date
    Jun 2015
    Posts
    1,464
    Helped
    336 / 336
    Points
    8,221
    Level
    21

    Re: Cooling SuperSO8 package

    Hi,

    Related to the last line of the last post by FvM, have you noticed that routers are vertical these days, no longer horizontal?



    •   AltAdvertisment

        
       

  7. #7
    Advanced Member level 4
    Points: 6,895, Level: 19

    Join Date
    Dec 2015
    Location
    Madrid, Spain
    Posts
    1,220
    Helped
    300 / 300
    Points
    6,895
    Level
    19

    Re: Cooling SuperSO8 package

    have you noticed that routers are vertical these days, no longer horizontal?
    No , but good point.



  8. #8
    Super Moderator
    Points: 248,981, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    43,336
    Helped
    13173 / 13173
    Points
    248,981
    Level
    100

    Re: Cooling SuperSO8 package

    This Wikipedia article gives an idea about the math involved with calculating convection heat transfer https://en.wikipedia.org/wiki/Heat_t...vertical_plane


    1 members found this post helpful.

  9. #9
    Advanced Member level 4
    Points: 6,895, Level: 19

    Join Date
    Dec 2015
    Location
    Madrid, Spain
    Posts
    1,220
    Helped
    300 / 300
    Points
    6,895
    Level
    19

    Re: Cooling SuperSO8 package

    Thanks... I was looking for some app notes that directly apply it to this electronic problem, surely a lot of people designed with such a package.

    - - - Updated - - -

    In other words, I was hoping that some app notes already used those formulas (with those coefficients) and obtained accurate results.



  10. #10
    Super Moderator
    Points: 248,981, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    43,336
    Helped
    13173 / 13173
    Points
    248,981
    Level
    100

    Re: Cooling SuperSO8 package

    There are various examples of copper pour heatsinking for small SMD packages like DPAK in manufacturer datasheets.

    Accuracy is relative. As the quoted formulas show, heat transfer coefficient is a function of height and temperature difference. To get the thermal resistance, you also have to put in copper plane thermal conductivity and the geometry.

    May be someone knows a link to a calculation tool?



    •   AltAdvertisment

        
       

  11. #11
    Advanced Member level 4
    Points: 6,895, Level: 19

    Join Date
    Dec 2015
    Location
    Madrid, Spain
    Posts
    1,220
    Helped
    300 / 300
    Points
    6,895
    Level
    19

    Re: Cooling SuperSO8 package

    Speaking of calculation tool, I just found this:
    http://www.ti.com/adc/docs/midlevel.tsp?contentId=76735

    I hope it is more accurate than their average app notes formulas.


    1 members found this post helpful.

  12. #12
    Super Moderator
    Points: 248,981, Level: 100
    Awards:
    1st Helpful Member

    Join Date
    Jan 2008
    Location
    Bochum, Germany
    Posts
    43,336
    Helped
    13173 / 13173
    Points
    248,981
    Level
    100

    Re: Cooling SuperSO8 package

    Looks good.



  13. #13
    Advanced Member level 4
    Points: 6,895, Level: 19

    Join Date
    Dec 2015
    Location
    Madrid, Spain
    Posts
    1,220
    Helped
    300 / 300
    Points
    6,895
    Level
    19

    Re: Cooling SuperSO8 package

    The post #1 transistor (Optimos5 technology) has the same package as the older Optimos3 technology (e.g. BSC057N08NS3G).

    In the Optimos3 one, it is specified 62 K/W with minimal footprint area... so, the Optimos5 one must have the same.



--[[ ]]--