Using mosfets as switches

[nobody special]

This is a little off topic, but I know that someone here can help me out...


I am trying to use an irf-540 as a switch for a heater circuit (high current- 40 amps!) but I'm having trouble trying to turn the thing on all the way. Even when I take the gate to the supply, it shows a significant voltage drop across the FET. I have tried putting the load on either end of the fet- Drain or Source, and the voltage drop remains the same. When I parallel fets, it dosen't make a difference. Now, this is where it gets really strange- if I take the gate to an external supply, it goes to where I would think it should be, and will remain there, even with the gate disconnected!?? What the... heck... is going on? Can anyone help me out? Thanks.
-NS

[stadams]

At what voltage is the element energized?

You may want to try something more industrial, possibly in a four layer device. A triac may be more applicable to your situation. After looking at the datasheet the 5th gen. 540 is only capable of 33 A or 100 V. If you insist on using a MOSFET, find one with exceptionally low drain resistance, even more so than the IRF 540.

Later,

[peranders]

Power loss approx. 200 watts! Hot! Hot chip = even hotter chip!

40 ampere is a little too much I think.

Choose a Fet with < 10 mohms Rdson

[nobody special]

I should have been more specific... for now we are just testing at a lower current- trying to find a lower cost solution to using a solid state relay. Unfortunately, it needs to be isolated. I understand that the fet needs to be as low as possible of an Rds when turned on. We were still getting much too much of a voltage drop across the fet, even considering the RdsOn. The guy trying this is using a bridge rectifier without a filter cap to drive the heater wire, which is how they are currently doing it with a solid state relay. It has a 120hz ripple that swings from 0-20Volts. I don't think they are going to acheive what they want with a fet- there's no way to isolate it and turn it on, and the bridge is causing it to turn off for a certain amount of time during each pulse of the recified ac (when it gets below the turn on voltage). I guess that was where we were getting the resistance at, most likely.
It was just really odd, the way that it was reacting. When he put a cap on it to smooth the ripple out, it still would not turn on all the way when the gate was pulled high. When you added in a seperate isolated supply to the gate, bam- it turned on hard, and would actually stay on when the gate was removed from the drain! I thought that was odd. Although, maybe that's perfectly normal... I've never tried to use one as a switch before.
Anywho... a bad idea in this case! THanks for the replies
-NS

[dlharmon]

A Fairchild RFG70N06 is what you need. 60V 70A 0.01ohm on resistance at VGS = 10V TO-247. Digikey has it for cheap -- $2.17 I would have expected it to be much more expensive. I just had some delivered today and have not had a chance to test them yet. I am using them as switches also.

Darrell Harmon

[crown300]

if you are really truely trying to isolate it - use a relay. When it finally gets down to it a relay still has its place. My 2 cents worth.

[dkemppai]

Originally posted by nobody special


When you added in a seperate isolated supply to the gate, bam- it turned on hard, and would actually stay on when the gate was removed from the drain! I thought that was odd. Although, maybe that's perfectly normal... I've never tried to use one as a switch before.

-NS

Fets make good switches, given you turn them on properly. I turn on and off 275Amps at 50Khz with them... ...so I suspect that what you're trying to do can be done.

The IRF540 is a N channel device, so the Source needs to go to ground and Drain to the load. Next you need to ensure that the gate voltage goes to 10 or 12 Volts above the source voltage.
(Since it is the gate to source voltage that turns the device on.
(0 volts for off, and 10 volts for on) Don't let the gate to source voltage go above 15... ...you'll burn the fet up...

You probably want a bigger fet. Your conduction losses (i.e. current x Rdson) are probably low, but you need some margin of error for the current rating. The other possibility is to parallel two or more fets. Simply the drains and sources together, and use a couple of hundred ohms in series with the gates before tying them together. (You're runing them in switching mode, so no need for balancing resistors such as in linear amplifiers)

As for the Fet Staying on, they are a voltage controlled device, and the presence voltage turns them on, not current into the gate. The gate is basically a low value capacitor, once you charge the capacitor, it can hold voltage for a long time, and the fet will stay on for a correspondingly long time.

You can also check a fet to see if it is good or not by using this property. Use the diode checker to charge the gate, then watch the Drain to Source resistance (or voltage drop). Reverse the polarity, and you can check that it is off. You only need one meter to do this...

Hope this helps...
-Dan

[jackinnj]

and you also want to make sure the gates go low -- and quickly -- I remember this from my first switching power supply -- tie a resistor from gate to ground for a discharge path.

another good idea is to use a diode in the path to the gate, while you get a diode drop you provide the controller with an extra element of protection.

lastly, if you parallel MOSFET's you lower the effective Rds - but make sure to read International Rectifier's application note on paralleling MOSFET's.

[phase_accurate]

nobody special wrote:

It has a 120hz ripple that swings from 0-20Volts. I don't think they are going to acheive what they want with a fet- there's no way to isolate it and turn it on, and the bridge is causing it to turn off for a certain amount of time during each pulse of the recified ac (when it gets below the turn on voltage).

This isolated drive can be achievedby the use of a cheap, isolated DC/DC converter, an optocoupler and a gate driver IC. If the switch on/off time is not very fast (or critical) and the switching frequency isn't high either, then the latter can be just a CMOS gate.

When you added in a seperate isolated supply to the gate, bam- it turned on hard, and would actually stay on when the gate was removed from the drain!

Remember that the input impedance of a MOSFET is very high. A voltage that was applied to the gate will most probably be stored and the Mosfet will stay on. While there are in fact circuits that make use of this effect, in most of the cases the gate of a mosfet should not be left floating.

Regards

Charles