# Will these heatsinks be sufficient for my mosfets?

I need to control 10 12VDC motors at ~25Amps, so I decided to make my own mosfet H bridges to control it ( with 2 n chan et 2 p chan ). I will need 10 H-Bridge, so 40 Mosfets. ( 20 n chan and 20 p chan ) The mosfets which I'd like to buy:

These mosfets have 62.5 degrees per watt thermal resistance to ambient, and the N-channel one has 0.045 ohm maximum resistance drain to source. The p-channel has 0.07 ohm

So I will have 0.045*25*25 = ~28 watts for N-channel and 0.07*25*25 = ~43 watts for p-channel

43 * 62.5 is a very big value so I was asking myself if a heatsink would be enough to absorb this heat ?

I'm gonna have 40 of these, knowing that I will use 8 mosfets maximum at the same time ( 2 motors max at the same time ) if I put all of the 40 heatsinks in contact the 8 mosfets would be able to dissipate on the 40 heatsinks. ( which makes one very big heatsink )

I could also put 1 or 2 cooler fans.

would this be possible ?

• I have a couple of comments. First, the thermal resistance you mention (62.5) is from junction to ambient. It applies when there is no heatsink. You want junction to case and case to sink (then add theta for sink, also). Second, you are assuming that all mosfets are on 100% of the time. If it is possible in your application for the mosfets to stay on for several seconds at a time, then I guess this is a good assumption. Finally, mosfets dissipate more power when they are being switched. Depending on switching rate and speed, this can be more than the I^2R dissipation. – mkeith Nov 2 '14 at 16:52
• I don't need to do PWM with it, I can keep at 100% all the time. But if it's possible I'd like to try PWM, is the rise of dissipation will be really important due to switching rate ? I'am going to change my mosfets to more powerful ones (under 10mOhm) like Mike suggested. – thp9 Nov 2 '14 at 18:01
• The idea is as follows: They don't switch instantly. During the time they are transitioning, they will have high current and high voltage drop which means high dissipation. I am sure it can be modeled or estimated but I don't know exactly how to do that off the top of my head. So, the switching frequency matters. Higher means more transitions which means more dissipation. Also, the drive strength of the gate voltage matters. More drive strength means faster transitions which means less dissipation. – mkeith Nov 3 '14 at 23:30

I would suggest using a beefier mosfet, like one of these for example: IRLB8748 or IRFB7546, these mosfets have very low Rds(on) of under 10mOhm.

I would also suggest using only N channel mosfets, this makes driving the high side harder but N channel mosfets have lower Rds(on).

With a 10mOhm mosfet you dissipate 6.25W at 25A, which is easy with a proper heatsink and there's no need for active cooling (a fan).

The heatsink you've linked on ebay doesn't specify it's thermal resistance so it may or may not be enough.

If my numbers are right, you need a heatsink with a thermal resistance of 1.49C/W.

These numbers were pulled from the FQP30N06L datasheet in the link you provided.

$$P_d = 28W$$ $$\theta_{jc} = 1.90C/W$$

$$T_{amb} = 30$$

Assuming a junction temperature of 125C and using those numbers the required thermal resistance is 1.49C/W or better (lower is better).

To get an idea of what kind of heatsinks you should looking at, have a look at these I found from digikey. Compare their size, and the number of fins with your ebay page. It should give you a rough idea how well they will be.

If you have forced air flow, then the required heatsink size decreases so you can get away with a smaller heatsink IF you have good air flow.

• Note that picking a better mosfet would be more cost effective and would take less space. – Mike Nov 2 '14 at 11:48
• I agree with you – efox29 Nov 2 '14 at 21:43