Cheap heat-sinking of 250W

On June 17, 2012, in Misc, by Cmc

PCL as Programmable Current Load is my small project for some extra lab equipment that will be little more than just dummy load.

Project itself is rather easy, at least from the viewpoint of the electronics. I need voltage measuring circuitry (operational amplifiers and avr), current measuring circuitry (ACS712) and some way for changing resistance of the current path (MOSFET in linear region).

Where the thing gets interesting is this 250W of energy that needs to be removed from the MOSFET-s. That is what the post is actually about.

So, my first idea was to find a way to do it with heat-sink and natural air convection. Sadly I would need a heat-sink that has a size of a door. Also such things are insanely expensive.

The next idea was to try my luck with active cooling. The prices in Farnell are really insane. 35€ for a sink only? I do not think so.

What I did find out was a normal CPU heatsink and fan. For  2.85€ I got myself  an  XILENCE CPU COOLER S775.

S775 xilence

S775 xilence CPU fan and heat-sink.

The nice thing about CPU cooling systems is that CPU core suggested temperature is around 60C and the design is planned for long working hours and dust collecting. That means that my MOSFET which should be ok for up to 140C die temperature can disspate loads and loads of heat.

heatsink and temperature @ 10W

heatsink and temperature @ 10W MOSFET in linear limiting region.

The ambient temperature here is 26.6C and with 10W limiting (12V and 0.8333A) temperature inside the heat-sink rised 5C. Since heat resistance was minimum (very good thermal paste and mosfet was clamped to the heatsink) the die temperature should have rised about 10C from ambient. Sadly right now I have no real knowledge of heat conduction calculations. So I will be back with real calculations soon not just some random numbers and measurements.

This little experiment did show me that it is possible to do what I am hoping to do.



2 Responses to Cheap heat-sinking of 250W

  1. Maciej Pijanka says:

    You can use spice to model thermal parameters too, ability to store heat ie mass * thermal capacity is just capacitor, and resistance of patch to infinite capacitance called ambient is just resistor, i think i even saw such simulations in some appnote.

    In some smps appnotes there are paragraphs regarding thermal calculation, ie
    when you dissipate p watts and have Rjc (junction case) of 0.3K/W, and Rca (case->ambient) like anoter 0.4K/W.
    you have something like that in steady state 0.3K/W 0.4K/W , so rise of junction will be like 7K, and case will be 4K hotter than ambient.

    best regards

    • Cmc says:


      I never thought about using spice for physical properties although I did notice the similarities about different parameters.

      Thank you a lot for this input :)

      best regards,

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>