Go Down

Topic: OK not insulating MOSFETs from their heatsinks? (Read 427 times) previous topic - next topic

DaleSchultz

I will be mounting four N-ch MOSFETs onto individual heatsinks on a solderable panel laid out like a breadboard. The MOSFETs will be handling 12 Volts.

I understand that the drain of each MOSFETs is connnecetd to the back plane of the TO-220 package.

The heatsinks came with the insulation pads, and insulating collars for the screws so that the drain of the MOSFET can be electrically insulated from the heat sink.

However, the heatsinks have a pin that fits into the breadboard directly opposite the drain pin of the mounted MOSFET. This means that no matter how well insulated the back of the MOSFET is from the heatsink, it will be electrically connected via the 'breadboard' anyway.

So, I am considering mounting the MOSFETs without the insulating pad, and using some thermal paste to improve the heat transfer. I will also mount the heatsinks so that there is a gap between the heatsink and the other rows of the 'breadboard'. I have some Dow Corning 340 silicone heat sink compound. I believe that the electrical insulation pad reduces the thermal transfer, so better thermal transfer will be achieved without it.

Since these are N-channel MOSFETs, I understand that doing this will mean that the heatsinks can be expected to have be the same as the ground when the the MOSFETs are passing current. Since the heatsinks will not be touching anything else, I can't see any electrical problems with this, but that is why I am asking here....

Does this make sense?

I am also aware that anodising on the heatsinks may provide some insulation, but I am ignoring that.

What dangers are there, if any, of bonding the drains to individual heatsinks?




larryd

#1
Jan 16, 2019, 06:01 am Last Edit: Jan 16, 2019, 06:02 am by larryd
You can do what you suggest.

Just remember the heat sink is now part of the circuit.

Do the MOS FETs really need a heatsink?


No technical PMs.
The last thing you did is where you should start looking.

MarkT

Modern MOSFETs seldom require much heatsinking, they are available down to a few milliohm on-resistances.

So what current levels are you handling?  More than a breadboard can handle?
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

DaleSchultz

Modern MOSFETs seldom require much heatsinking, they are available down to a few milliohm on-resistances.

So what current levels are you handling?  More than a breadboard can handle?
One of the four will handle 5 Amps.
The supply and drain connections will not rely on the traces on the breadboard. I plan on hefty wires there.

GeronimoDK

Constant switching (PWM) or occasional switching on demand?

DaleSchultz

One of the four will be simple on/off
The other three will use varying levels of dimming.

I am driving the MOSFETs using TC4427 MOSFET drivers.

I have been able to test one 5m length LED strip and the MOSFET did not get warm, but I am adding the heatsinks for the bigger loads I expect when it is in place. When I have it in place I will see how hot the heatsinks get when under the full 5A load and decide if I need more heatsink. I dont know of a way to create a 5A test load.



GeronimoDK

Which MOSFET?

At least you're using a MOSFET driver, this will insure fast switching which in turn means that you will loose as little power as possible when switching! This is important for high-current and high frequency PWM switching where you could otherwise loose a great deal of power in your MOSFETs!

Arduino PWM frequency is normally 490 Hz, which is quite low in terms of PWM. So with a driver maybe you could even do without heat sinks on any of your MOSFETS. But of course if you want to feel a little safer, a small heat sink should be sufficient.

If you were to drive the MOSFET directly from the PWM pin and at a much higher frequency you could run into problems with switching power loss.

DaleSchultz

I am using FQP30N06LE MOSFETs.

I initially tried to use them without MOSFET drivers but I could only get 7V out of the MOSFET instead of 12V, so I read what I could about MOSFET drivers and ordered the TC4427 drivers.

Here is how I am connecting things up. (showing one of the four, a second TC4427 also drives a pair of MOSFETs:


DaleSchultz

#8
Jan 17, 2019, 06:37 pm Last Edit: Jan 17, 2019, 07:18 pm by DaleSchultz
I previously ran these circuits through 7A fast blow fuses. (The circuits had manual PWM dimmers.)

Now that they are driven by MOSFETs does it make any sense to keep the fuses? Would the MOSFETs fail faster than the fuses anyway?

Does PWM current interact with fuses in any special way?


JohnRob

Dale,

The typical failure mode of a MosFet is fully ON.  However in your case the failure result would simply be your LED's on all the time.   In addition, most 12V sources have current limiting and will survive a shorted output.   I will admit some sources do better at handling a short than others.

Please do not PM me with thread based messages.  If your thoughts are worth responding,  the group should benefit from your insight.

DaleSchultz

thanks, my power source is one of those 12V 30A units fro China, so I have no idea what it would do if there was a short...  I have the LED strips running about in the ceiling of the basement, so I may leave the 7A fuses in line.

MorganS

Fuses protect wires. A MOSFET will blow faster than any fast-blow fuse. It is impossible to protect valuable components like that.

But if something happens, like the cat knocks it off your workbench and a piece of metal contacts the wrong thing then you really don't want to have 30A heating up the wires. That can start a fire and even a small fire on the workbench is much more costly than a MOSFET.
"The problem is in the code you didn't post."

GeronimoDK

Couldn't find the exact MOSFET with "LE" at the end, but it looks like the FQP30N06 has a RDS(on) of about 0.03 ohms typical at anything below 10A.

This means that you will dissipate R*I² = 0.03 * 5 * 5 = 0.75W in your MOSFET when fully on.
The thermal resistance to ambient, without a heat sink is 62.5 °C/W, which means you will heat up the MOSFET about 47°C (116°F) above ambient when fully on. If it sits in a well ventilated area that would actually be okay, but still it should feel quite warm or hot to the touch!

Even a small simple u-shaped heat sink would probably halve that temperature increase.

MarkT

Black anodized heatsinks are best, BTW - they radiate heat as well as passing it to the air by convection.
Shiny bare metal doesn't radiate much at all at lowish temperatures (which is why radiators are painted).
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

DaleSchultz

thanks for the calculations...  I got it all wired up last night and today I will test them under load.

I am using small black heatsinks... and they are in the open air below a train layout. The power supply that drives it all is mounted on the wall nearby and it has a variable speed fan that also moves air about so I think it  will be fine...


Go Up