Trying To Slow Wiper Mower Via PWM

[toozie21]

I plan on using a wiper motor to run some things in my new mechanical display for Christmas this year, but I am having a few issues. I have a few P-Channel power MOSFETs from Fairchild (https://www.fairchild...FQ/FQP47P06.pdf)lying around that I thought would work fine. The wiper runs fine with a 5V source, so I thought that if I used a PIC and put out a 400Hz 50% duty cycle I could halve the speed.

I connected ground to the wiper's ground, then I connected 5V to the Source of the MOSFET, and I use the Gate as my input switch. That means the Drain becomes the power source for the wiper.

The input switch seems to work OK with a constant +5V/GND (with GND turning the motor on, and +5V turning it off), but when I hook up the Gate to PWM, things seems to go awry. The PIC goes haywire and the MOSFET seems to lock-up.

I am wondering if I need to be using a transistor to switch the MOSFET (not sure why), of if I need to use a snubber of some kind to knock down any EMI (I know nothing about snubbers).

Does anyone have any thoughts how I can clean this problem up some? I would prefer something simple (who wouldn't) that lets me run everything off of the 5V, but if I must, I can probably throw 12V into the circuit (though I would prefer not to).

 

[Ian Rogers]

A wiper motor running at 5v will draw a bit of current (even with PWM ) I suspect there isn't enough current to supply the motor and the pic.

 

[toozie21]

I think they quote it as drawing a little under an amp (I haven't checked it), and for now I was running it off of my benchtop supply (so should be enough), but that is something I better check out.

 

[Ian Rogers]

But that would be an amp at 12v... and that's continuous.. The start up current will be much higher!! (I take it this is off some car or van).

 

[toozie21]

That's a good point. I assume it was designed for some sort of foreign car, but it is actually sold as a prop (**broken link removed**).

 

[crutschow]

Do you have a diode across the motor (anode to ground). You need that to the keep the motor spikes from zapping the MOSFET.

 

[toozie21]

No I don't. When I was looking around last night for other people doing similar things, I saw some schematics showing that, but I hadn't tried it in my design yet. This image was the one I saved thinking it was worth a shot today: https://www.microchip.com/forums/download.axd?file=0;323806
I assume any old standard diode is probably good enough?

 

[Ian Rogers]

Most MOSFET's have them built in... Check the data sheet

 

[crutschow]

Most MOSFET's have them built in... Check the data sheet

That diode is in the wrong direction to suppress the spike. You need a diode across the motor.

Edit: Yes, just about any old diode should work but it should be rated for at least 10% of the motor current.

 

[aljamri]

Don't forget frequency handling of the proposed diode.

 

[crutschow]

Don't forget frequency handling of the proposed diode.

That's a common fallacy. You don't have to worry about the speed of the diode here. That's only of concern for high frequency RF or high power circuits where efficiency is of concern such as switching regulators.

All diodes turn on very fast, the fast ones also turn off fast, but a fast turn-off is not needed for inductive transient suppression. Thus any old diode will work for that.

 

[Mr RB]

That diode is in the wrong direction to suppress the spike. You need a diode across the motor.

Edit: Yes, just about any old diode should work but it should be rated for at least 10% of the motor current.

A large DC motor like a wiper motor at the OP's suggested 400Hz will possibly run in continuous switchmode, or very close to it so the diode should really be rated for the expected peak load motor current.

 

[toozie21]

I can lower the freq of the PWM, that was a somewhat arbitrary value pulled from the web in someone else's discussion. I see no reason why dropping it down to something like 40Hz (or less) wouldn't be smooth enough.