Heatsink usage

For use with a TO220 package device, the LA6/150 12V heatsink would be an overkill and expensive too. From what I have seen, this type heatsinks are used on SMD ICs like processors or FPGAs. Basically on devices that can get quite hot. If you have multiple TO220 packages for example incase of inverter circuits or any other power electronics circuits, then I think it's feasible to use and yes drilling and hole and screwing the TO220 in the LA6/150 is fine. I have used simple aluminum from the waste material of frames for french windows in my house. They work fine as a heat sink.

For use with a TO220 package device, the LA6/150 12V heatsink would be an overkill and expensive too. From what I have seen, this type heatsinks are used on SMD ICs like processors or FPGAs. Basically on devices that can get quite hot. If you have multiple TO220 packages for example incase of inverter circuits or any other power electronics circuits, then I think it's feasible to use and yes drilling and hole and screwing the TO220 in the LA6/150 is fine. I have used simple aluminum from the waste material of frames for french windows in my house. They work fine as a heat sink.

I'm no thermodynamic expert, but agree with your advice. Unless a device is being driven to its max limits, the important thing is to get a thermally conductive piece of metal on a TO220 into a puff of fresh airflow, in an enclosure that an air intake path and an exhaust path (to minimize the enclosure ambient temperature) .   

I Was more interested in sussing out how that series worked and whether the design intrinsically required the IC to be attached to the bottom 1.2mm plate or not.  Have you used one or, like me, looked at it and thought that devices could be attached to the side: it certainly looks feasible.

Hi Andrew,

I remember a little from my thermodynamics classes that there are special materials/coatings that will have high thermal conductivity/capacity. I'm not sure if these heat sinks have that or work uniformly from all side (device attached to any of the four side) but the inner fin structure seems to be an important parameter to decide on which side the device should be mounted.

I‘d go from what the datasheet told me in terms of thermal efficiency.  I actually don’t need one that big but it’s smaller companions would be useful for my purpose.  Anyway, it’s one for my options list.

From the data sheet it seems clear that Fischer expect the devices being cooled to be attached to the bottom (thick) face.

It's the only face with flatness and roughness specified and the text says:

"semiconductor mounting surface for milled flat"

I would expect the thermal impedance for devices attached to the side to be a lot worse, and to the sliding plate pretty awful and unpredictable.

It isn't a very good shape for TO220 devices.

How much power are you trying to shift from how many devices ?

MK

Hi Michael,

The (little) info here also agrees with that: https://www.fischerelektronik.com/en/latest-news/press-releases/2020-fischer-pressemitteilungen/thermally-optimized-holl…     I think from memory my DC load internally has something similar, although they have attached the semiconductors to the sides, which is likely not good but maybe it was practical for the layout. Quite good that there is the panel to make the heatsink independent of the enclosure, so that the circuit can be tested even with the project enclosure lid off, since that would normally cause issues with the fan ducting. 

Edit: on the other hand, I wonder if they also sell the same product without the fan attachment, in which case the panel might not be supplied, so that it just radiates vertically out through holes at the top of the chassis.

That's what those square root like symbols refer to is it?  I've not come across those before.  So it's for a DC load which I've rated to 90W and there would be - in the simplest design - a TO-220 MOSFET, a TO-220 sense resistor (say, up to 12.5W) and a TO-220 temperature sensor.  The thermal calculation I've made is, taking to different heatsink ratings:

Clearly a 0.5C/W heatsink gives me very little wiggle room.  Can I just add here that I'm still working on this and there are other approaches I might take, e.g. parallelising the MOSFET to share the load.  At this point I was interested in the feasibility of keeping things simple so I was looking at what was available, hence the query about this particular heatsink. 

That press release is useful and states the info I was looking for!  Thanks.

shabaz  wrote:

..?

Edit: on the other hand, I wonder if they also sell the same product without the fan attachment, in which case the panel might not be supplied, so that it just radiates vertically out through holes at the top of the chassis.

The aluminium heatsink and the fans are available separately. Also available in custom sizes, cut to customer specs.

Unless the MOSFETs are very very expensive you will do much better to improve the weak link in the chain, ie the Rjmb and the coupling to the heatsink.

Suppose you put 3 MOSFETS in // - now you get 15 + 12 + 18 = 45 above ambient for a total of 80C.

Now you can use a much cheaper 0.5W/C heatsink for 15 + 12 + 45 + 35 = 107.

MK