Vishay SI7141DP P-Channel MOSFET

[raptor1956]

This will be the first time using this transistor and I'm a bit uncertain about the use of the bottom pad.  The device, a PChannel MOSFET, has 8-pins configured as 1-Gate, 3-Source and 4-Drain so all 8 pins are spoken for, and in addition it has a single large pad underneath the device that is also connected to Drain.  So, it appears that the drain connection can be made to any of the four drain pins or to the large pad underneath, but is there anything else I need to know about this large pad?  Can I ignore it and just connect to the drain pins?  Given my desire to operate at a max of 5A it makes sense to connect to ALL the drain pins and all the source pins to spread the current load -- the pins are not real big.  The N-Channel MOSFET (SI4186DY) I chose as a complement for my H-bridge is also 8 pin with 1-Gate, 3-Source and 4-Drrain but is has no pad underneath.  So, I think I should be able to avoid the large pad altogether but am open to advise here...

The reason I chose these MOSFET's is that they are large enough to see and work with but it also has appropriate voltage and current ratings as well as very low RdsON.  I find it amazing that these devices are both well below 3mOhm RdsON.


Brian

[DBecker]

It's a thermal pad.  While it probably has good electrical properties, don't count on it.  The bond wires attach close to gate and may have a lower resistance than the bulk substrate.  Route the circuit as if all of the current flows through the pins, and all of the heat exits through the pad.

[ejeffrey]

It isn't really an "or".  You should connect all the drain and source pins as well as the large pad unless the datasheet says otherwise.

The multiple drain and source pins are used to handle the current.  You should connect them all.

The large drain pad is for heat sinking.  Usually the data sheet will give some guidelines on this, but you want to connect it to something to dissipate heat.  Often it is ideal to connect to a power or ground plane, but if that is not possible, just a square of copper on an outer layer will help.

[raptor1956]

Since my max current is about 5A and the RdsON is less than 3mOhm there isn't much heat to worry about -- round about 65mW so heat should not be a problem.  Actually, since this is a PWM controlled device the peak current could be about 10A but the duty cycle less the 50% so the weighted average power over the cycle would max out at about 130mW which is well within the un-heatsinked rating.


Brian

[Benta]

Since my max current is about 5A and the RdsON is less than 3mOhm there isn't much heat to worry about -- round about 65mW so heat should not be a problem.  Actually, since this is a PWM controlled device the peak current could be about 10A but the duty cycle less the 50% so the weighted average power over the cycle would max out at about 130mW which is well within the un-heatsinked rating.


Brian

Careful! Your analysis is true for DC operation. PWM is a different story.
Depending on your switching frequency and the quality of your gate driver, a lot of power can be dissipated during the on-off and off-on transitions.

[raptor1956]

I'm going to use one of my Arduino compatible boards (Mega2560, Teensy3.6, ESP32) as the controlling element but the gate driver will be an NPN transistor.  Since the P-Channel MOSFET is switching the high side and that's 12.5VDC you don't want to drive the gate directly from the microcontroller.  So, the switching time should be quite short.  But, in addition, I intend to have a decent flow of cooling air washing over the PCB...


Brian

[Monkeh]

...gate driver will be an NPN transistor. ....  So, the switching time should be quite short.

Only in one direction.

[raptor1956]

...gate driver will be an NPN transistor. ....  So, the switching time should be quite short.

Only in one direction.

Yes that's true, but by using a more powerful driver I can also use a lower value resistor to pull it the other way. 


Brian