which P channel mosfet

[bitnick]

My dew heaters are fed from phono plugs and sockets. With a low side, I have to isolate the socket body from the panel its mounted on. With high side switching, the phono body would always be at ground potential. A much more elegant solution.

 

Huw

I see. However, it might be a good idea to use an isolated connector anyway, to avoid ground loops and noise being fed into other equipment. If you run several amperes through the cable, there will be a small voltage over the cable on the "inside" of the panel, which will feed current and thereby voltages through all your other equipment that is not isolated from the panel.

Also, the electrical side of things becomes more complicated since you will need a high-side driver for your P-MOSFET. And, P-MOSFETs have a higher RDS(on) than an equivalent-die-size N-MOSFET, so will be somewhat more expensive for the same performance, although this might be moot in a one-off case like this.

[bitnick]

... In conclusion, I'd say that the more elegant solution would be a low-side MOSFET and an isolated connector!

[Gina]

I agree.  Phono plugs and sockets are not the best form of connector - cheap, yes, but the cost of decent connectors is insignificant compared with the likes of cameras and scopes etc.

[hughgilhespie]

Can some kind soul explain how to do this?

'slew limiting the switching is the way to go.'

Will a say 10uF capacitor across the output be enough to stop the switching noise?

Thanks again, Hugh

[bitnick]

A capacitor across the output might dampen the noise somewhat. Make sure it's a low-ESR type - probably ceramic? - and connected very close to the MOSFET.

A better way is to make sure the noise does not appear in the first place - i.e. slew limiting. To do this, you need to control the time it takes for the MOSFET gate capacitance to charge and discharge. Look up the capacitance in the MOSFET datasheet, and use resistors to control the current into and out of the gate. Unless you use some dedicated gate drive circuit for the MOSFET, chances are that you already have "sane" rise/fall times, but it doesn't hurt do do the calculations to make sure.

There is a trade off between power loss and noise here - the longer the rise & fall times, the higher the power loss in the transistor. So don't go too slow. I'm guessing a few microseconds might be fine (but it really depends on a lot of factors, like the characteristics of your MOSFET). See e.g. this StackExchange question for more info.

[Ajohn]

Not really. It's not that easy a 

Can some kind soul explain how to do this?

'slew limiting the switching is the way to go.'

Will a say 10uF capacitor across the output be enough to stop the switching noise?

Thanks again, Hugh

Not really. It isn't an easy thing to do that way. A better option is to drive a low side switch via a resistor. The gates have significant capacitance so this will slow the switching down but also increase the dissipation in the switch. The usual aim is to switch them as fast as possible to minimise that. It can actually be tricky to make these things switch as fast as the data sheets suggest.

It could be done by driving the dew heater through a resistor with a capacitor on the other side of that. This could give what ever time constant some one desires but it wont be very efficient when amps of drive are needed. The resistance value would have to be chosen to keep the turn on current within suitable limits - just sticking a 10uF capacitor across the wires to the dew heater might well destroy the FET as in principle the current that can be taken out of the capacitor would be very high. Actually capacitor have ripple current limits so that could cause problems any way.

If some has had interference problems due to switching I would hazard a guess that it has nothing to do with the switching speed - more likely to be a problem with supply decoupling to the various bits and pieces being powered from a single supply. The box of tricks with the switches in it should have a substantial capacitor in it to help hold up the supply when sudden pulses of current are taken.

John

-