5
\$\begingroup\$

Normally, we connect an IC to the other side of a heatsink than shown here:

enter image description here

Is it bad to place the IC on the 'unusual' side of the heatsink (and if so, why?) or are both sides each equally functional? In the end I will place the heatsink upside down in the case, so with the ICs on top of the heatsink.

\$\endgroup\$
11
  • 5
    \$\begingroup\$ Better crop and scale down your image a bit before Olin sees it ;o) \$\endgroup\$
    – jippie
    Commented Feb 20, 2013 at 12:41
  • 1
    \$\begingroup\$ Also you probably need to show how it fits in the final enclosure, otherwise that arrangement could go from far better to far worse so there's no generic answer. \$\endgroup\$
    – PeterJ
    Commented Feb 20, 2013 at 12:46
  • 1
    \$\begingroup\$ "Is the style in the image bad?" - YES, but mainly because of the horrific soldering and mass of uninsulated contacts. Whatever you are doing, you're not doing it very well (and I mean "well" in terms of care, neatness, quality, reliability, safety...) \$\endgroup\$
    – John U
    Commented Feb 20, 2013 at 16:02
  • 1
    \$\begingroup\$ On the other hand, the "unusual" way is more open in the sense that the heat is being pulled farther away from the IC. The IC isn't cradled in the heat sink. The heat sink is in actual fact pulling heat farther away from the IC. The aggregate is more of an unfolded structure in this configuration. \$\endgroup\$
    – Kaz
    Commented Feb 26, 2013 at 1:53
  • 1
    \$\begingroup\$ @Kaz can you make that an answer? :-) \$\endgroup\$
    – user17592
    Commented Feb 26, 2013 at 7:04

2 Answers 2

3
\$\begingroup\$

The job of a heatsink is to move heat away from some component that would otherwise get too hot and dissipate that heat into the air. This means the surfaces of the heatsink needs to be open to free air. Putting anything other than the part that produces the heat onto a heatsink will decrease its effectiveness. Putting something that produces its own heat there makes it even worse. Then there is also the additional issue of whether the extra part can handle the temperature. So, the short answer it that this is not a good idea.

However, this does not meant it is guaranteed to fail. If everything isn't running at its thermal limit, then the heatsink has extra capacity, and some degradation of that capacity can be tolerated. Still, this is not something I'd do except as a last resort, and then with some careful analisys to make sure everything is OK.

\$\endgroup\$
4
  • 3
    \$\begingroup\$ It's not clear that the question proposes to put parts on both faces of the heatsink, I'm reading it more about asking if one face is better than the other. My gut instinct would be that the side away from the fins could actually be a tiny bit better, since the package itself wouldn't be in the pocket of hot air between them - on the other hand, if there cooling air aimed at the fin side, then that side might be better. \$\endgroup\$ Commented Feb 20, 2013 at 19:19
  • 1
    \$\begingroup\$ That's more what I meant, yes. Sorry for any misunderstanding and Olin: thanks for your explanation which was sure very useful! @ChrisStratton, can you make this an answer? \$\endgroup\$
    – user17592
    Commented Feb 23, 2013 at 9:04
  • \$\begingroup\$ In particular, in this picture the electrolytic caps are going to be at almost the same temperature as the ICs. It's not kind to electrolytics to cook them like this. \$\endgroup\$
    – markrages
    Commented Feb 25, 2013 at 22:01
  • \$\begingroup\$ @markrages good point. I'll change it asap. \$\endgroup\$
    – user17592
    Commented Feb 26, 2013 at 7:05
0
\$\begingroup\$

Since there is natural cooling convection , the position of heat source (the device), as well as the orientation of heatsink-device system plays important role, some times in a magnitude of 10 oC. So if your thermal design it is critical and close to Tj max, then you have to put the devices in the recomended place in order to achive maximum cooling performance. For temperatures up to 150 oC the life expectancy of semiconductor halves for every 10~15 oC and up to 175 oC halves for every 5~10 oC

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.