problem with MOSFET high-side switch

Thread Starter

watchdogtimer

Joined Feb 23, 2018
1
I'm using a PIC microcontroller and four P-channel power MOSFETs to switch four 12-volt automotive power sockets on and off. The PIC also measures the current through each of the sockets by measuring the voltage drop across shunt resistors. I've attached an image showing a portion of the circuit for one of those outlets.

The circuit works fine in most cases. The main exception is when it tries to switch on a 75W AC inverter. When it does, it usually causes the microcontroller to reset. This occurs even with no load attached to the inverter.

The odd thing is, if I first switch on a small load (like a 12 V DC fan), unplug the fan with the socket switched on, then plug the inverter in, the PIC won't reset and everything will work normally.

I'm presuming the inverter has a high capacitance that causes a high inrush current when the MOSFET is switched on, but I'm not certain of this.

How could I avoid this problem? What if I turned on the MOSFET using a PWM signal, to reduce the inrush current?

If I need to modify the circuit, I would prefer to stick with through-hole parts if possible.
 

Attachments

dl324

Joined Mar 30, 2015
12,826
Welcome to AAC!

Add a beefy capacitor on the 3.3V supply.

You could also try an inrush limiter on the AC inverter.
 

Sensacell

Joined Jun 19, 2012
2,792
I would be suspicious of the PCB layout.

When switching huge currents, voltage drops on the traces can show up in places you didn't consider.

I had an LED dimmer design once that worked fine, except it would crash and reset if you connected a large load.

Soldering a big fat copper wire over some of the ground traces fixed the problem, the PCB layout was amended with heavy copper and wider traces.
 
I'm using a PIC microcontroller and four P-channel power MOSFETs to switch four 12-volt automotive power sockets on and off. The PIC also measures the current through each of the sockets by measuring the voltage drop across shunt resistors. I've attached an image showing a portion of the circuit for one of those outlets.

The circuit works fine in most cases. The main exception is when it tries to switch on a 75W AC inverter. When it does, it usually causes the microcontroller to reset. This occurs even with no load attached to the inverter.

The odd thing is, if I first switch on a small load (like a 12 V DC fan), unplug the fan with the socket switched on, then plug the inverter in, the PIC won't reset and everything will work normally.

I'm presuming the inverter has a high capacitance that causes a high inrush current when the MOSFET is switched on, but I'm not certain of this.

How could I avoid this problem? What if I turned on the MOSFET using a PWM signal, to reduce the inrush current?

If I need to modify the circuit, I would prefer to stick with through-hole parts if possible.
 
Hello,
I recently came across your post.
I was wondering if you remedied your problem, and if so how?

I was curious as to where the 3.3 power was coming from to feed your PIC?
If you are regulating down from 12 volts I would check to see if the inrush current to the inverter was not glitching the input to the voltage regulator that is producing the 3.3 volts causing your PIC to reset. I would consider a large cap on the 3.3 volt regulators input voltage side. Also consider feeding the 3.3 regulators input voltage( and large input cap, >1000 uf) with a diode between the supply and these components to help insure the 12 volt feed voltage will not dip to low when the inverter powers on. The the regulators input cap should be supplying voltage to the regulator to get past the surge, and the diode will be preventing this cap from being drained to feed the surge. Also does the pic have a reset line that you should connect a pull up resistor too, on top of any internal high ohm pull up the reset pin?
 
Top