I've got a very simple circuit. A MOSFET (TK15D60U) is powering a 3 to 4 Ohm heat bed by 19.6 volts. From my calculations this results in 6.5Amps. The power adapter seams to have its specs at max 6.32A but seems to function fine. The MOSFET should be able to handle 15Amps continous drain. The drain should be able to handle 600v, so that 19.6Vdc should be fine.

The gate is driven by a 25Vdc via a 10k resistor to the gate and 100k resistor from the gate to ground (to allow to leak away to discharge the gate, I thought it was a good thing). So effectively there is (25vdc/110k)*100k=22.7Vdc voltage at the gate. The Gate-source voltage is 30v so that is below max.

What I can read from the datasheet is that the Drain-source ON resistance @Vgs 10Vdc @ 10Amps is about 0.15Ohm. Lets take 0.2. I never really thought about the ON resistance. I thought, the gate would be fully open, so resistance should be minimum and will handle it without a problem. But when the MOSFET is active, it becomes too hot to touch within 10 seconds or so.

Now this part is something I am not so familiar with. Hence the question. Is it true that the formula P=(I2)*R (Watt=Ampere squared * resistance) yields (6.5A*6.5A)*0.2Ohm=6.5Watt that is dissipated via the MOSFET?

6.5Watt is a lot if you ask me on a TO-220. Is it not? How large should the heatsink be to handle that? Should I actively cool it using a fan? All I have now is this very tiny heatsink. This topic is probably a laugh for the experts around here.

TO-220 heatsink

  • \$\begingroup\$ 3/4 you mean 0.75 or 3R to 4R? \$\endgroup\$ – Trevor_G Oct 20 '17 at 18:50
  • 1
    \$\begingroup\$ Way cool that stackexchange adds a link to the datasheet automatically! The datasheet from Farnell (that SE added) states with a BIG watermark: Not recommended for new design. The one I have doesn't state that. I see now, a RsdOn of 0.02Ohm is more common (obtained from a quick search)? \$\endgroup\$ – Mike de Klerk Oct 20 '17 at 18:58
  • 4
    \$\begingroup\$ It was not SE that added the link.. it was me ;) \$\endgroup\$ – Trevor_G Oct 20 '17 at 18:59
  • 3
    \$\begingroup\$ @Trevor Correction: Way cool that SE members are so helpful in adding a link to the datasheet! Thanks Trevor. \$\endgroup\$ – Mike de Klerk Oct 20 '17 at 18:59
  • 1
    \$\begingroup\$ Btw I calculate 7.5W worst case with Rload =3 and Ron = .2 I = 6.125. That squared *.2 =7.5W. And yes that is way too high for a bare To220. \$\endgroup\$ – Trevor_G Oct 20 '17 at 19:04

I calculate 7.5W worst case with Rload =3 and Ron = .2

I = 6.125A.

That squared *.2R = 7.5W.

Yes that is way too high for a bare To220. At 83.3C/W channel to ambient of the device that's a whopping 625C temperature rise.. So it is going to melt.

You could add a whopping heat sink, but really, you need to find a different MOSFET that has an on resistance in the milliohm range.

| improve this answer | |
  • 1
    \$\begingroup\$ I never thought about the C/W value. Thanks for adding that calculation. I understand more now how to pick a heatsink. I guess it comes down to taking my time to answer every question I have instead of 'rushing' to the information that I think I need. But thats with every thing when learning I guess. \$\endgroup\$ – Mike de Klerk Oct 22 '17 at 11:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.