High power MOSFETs soldered circuit board.

[AndJuan]

Hello,

I'm doing another project to control a engine at H Bridge with 8 Mosfets, about +/- 1200W

Engine Features:
RPM= 3000
Voltage: 36V
current 35.5A


P =Rds(on) x 35.5A2=0,0014Ohms*35.5A(2) =1.76W (in the best case) Mosfet disipation.

I selected N Channel MOSFETs (IRFS7530-7PPBF) package =
'D2 Pak-7PIN ' for its low RDS (on).


My question is this type of package goes straight to the board and pass high current MOSFETs.
and warmed. I'm afraid that board was burned.

I advise to make board and dissipate well.
and characteristics must have the board for to withstand the temperature. etc..
multilayered and below all copper mosfets. other better ideas?

Thanks,

[AndJuan]

[robertferanec]

Hello AndJuan,

if you are not sure, I would probably build a simple 1 layer PCB with the transistor on it and would test the circuit before using it on a bigger board.

Be very careful, 35A flowing through PCB may require some special care - e.g. wide copper. There are number of calculators to check maximum current through a track, for example this: http://www.eeweb.com/toolbox/externa...ce-max-current

[AndJuan]

Hi Robert,

Maybe I have not explained well. I was not referring to the width of the track by the amount of current,

it is for the dissipation of this type encapsulated, it is not the typical TO220, since this is welded on the circuit board.
As I can make a better heat dissipation circuit board to resist.

THX.

[mairomaster]

Hi AndJuan,

As you have stated, 1.76W is the power dissipation in the best possible case. The actual dissipation could be significantly more, depending on different factors:

- Are you sure the transistor is properly switched on? Are you supplying enough voltage to the gate?
- Are you statically driving the motor in an always on position of the MOSFET? If you are using PWM, switching losses can significantly increase the power dissipation.

You can do a few things during the PCB layout to improve the heat dissipation capabilities of the board:

- Make the drain pad on the board as big as possible, surface area increases the radiation (hence dissipation).
- Create a mirrored pad on the opposite side of the board and connect it with many vias to the drain pad. You can increase the size of the mirrored pad even further, if possible.
- Remove the solder mask from both pads.
- Consider using thicker copper layers, especially if you have problems with the current capacity.

If that is not enough (it should be if you are doing everything right), you might consider using a different transistors and mounting a heat sink.

I didn't quite get it - did you already made one board? If that is the case, what problems exactly did you have with it?

[robertferanec]

AndJuan I was referring to complete testing on the 1 layer PCB - including power dissipation and possible way how to cool it down. As mairomaster explained, in real board the heat can be different from datasheet. So you could just create a small PCB, soldered the transistor down and played with it. You would see how hot really can get and how effective your footprint will be (of course the result may be better if in the real application you use GND planes). Possibly, you can create one test board for different packages.

We test a lot of small circuits, here is an example
(the boards may look ugly, but they save a lot of time and money)