Pwm and mosfet question

Couple of beginner questions:

When pwming the gate of a n channel mosfet at 50% duty 1khz for example will that mosfet will dissipate more or less power for the same load at 100% duty ?


When hooking up a logic level mosfet to a pic pwm port to be used for control of a 12v motor for example what kind of mosfet driver is needed ?

As the FET spends all its time in the low resistance state, the one at 100% will be cooler.

The driver is hard to say much about without more input. In general, it must make the gate voltage swing quickly. The less time the FET is in the ohmic region, the less heat you get. The driver must also be able to the the gate at least 10 volts above the source while on.

Is it possible to make a mosfet operate a load at 50% duty half power on the load without suffering the penalty of having higher power dissipation on the mosfet due to lower vgs ?

The main reason one uses pwm is that the FET is either ON or OFF. It only dissipates a little during ON due to slight ohmic resistance, and dissipates very little when OFF (just a tiny amount of leakage.)

With a positive source into a positive voltage motor, a p-channel device will turn ON when pulled to GND, which eases the drive situation considerably! The Source goes to the source of power, and the Drain goes to the motor.

Some kind of driver would be needed out of a digital device. Remember that a logical 0 turns into GND which turns the motor on, so you get a logical inversion: 10% on becomes 90% ON at the motor....

You might need 1K of Gate to Source res to turn the FET actually OFF when the logical output goes high = 1.

billbehen,
I don't know why you're talking about a P-ch MOSFET when our OP asked about an N-ch MOSFET.

P-ch MOSFETs have a built-in handicap; hole flow doesn't work as well as electron flow, so the gate area has to be about 2.5x as large as an N-ch MOSFET to get the same Rds(on) vs Vdss; so the Qg (gate charge) is ~2.5x as high. This means that much more current is required to charge/discharge the gate.

It's also more difficult to control a high-side MOSFET than a low-side MOSFET. One can get low-side MOSFET drivers capable of surge currents of several Amperes pretty inexpensively.

curry87,
You haven't mentioned what MOSFET you're using, nor the current of your motor.

You might look at Microchip's offerrings. a TC4420 or TC4429 can sink/source up to 6A.

However, we don't know your switching frequency, your motor supply voltage, current, or anything else.

There are many other choices that could be made.

*added diagram*

Im using this STP36NF06L MOSFET LOGIC N 30A the problem is when duty cycle drops the voltage on the gate drops and mosfet heats up alot more than at 100 % duty.What mosfet driver do I need to operate this mosfet with a load of 10 amp at different pwm levels from a pic without it burning the mosfet at low duty ?

You could use a variety of gate drivers.
Take a look at Microchip's offerings:
http://www.microchip.com/ParamChartSearch/chart.aspx?branchID=9010&mid=11&lang=en&pageId=79
TC1410N, TC1411N, TC4405 are some possibilities for you that are available in DIP packages.

Right now, your 1k resistor from the I/O pin to the gate is causing very slow turn-on and turn-off times. The gate driver IC should be located as close as possible to the MOSFET gate terminal, and it will need bypass capacitors across it's Vdd/GND pins. The PIC I/O pin should be fairly close to the gate driver input.

Is there a simple transistor combination equivalent driver that could be used instead of a driver ic ?

At the end of the day im only trying to control the brightness of the 12v lamp using the pic pwm port doesn't have to be perfect just work is enough without cooking the mosfet.

There is a transistorized MOSFET gate driver on this page:
http://www.picotech.com/applications/pwm_drivers/#chap9

There are voltage follower/totem pole designs out there, but the resulting gate voltage is reduced.

Here is an applications guide for MOSFET drivers: http://focus.ti.com/lit/ml/slup169/slup169.pdf

Note that figure 10 is what I mentioned above; if your logic input is 0v and 5v, your MOSFET gate voltage will be ~0.7v to ~4.3v; it won't turn on all the way.

Is the lower the resistor value between the pic pwm port and mosfet gate the faster i can switch the mosfet ?

For this particular mosfet how fast should i be able to switch it on and off at 10 amp without using a gate driver and without it heating because of being in the switching phase alot ?

Even if you switch the MOSFET on and off very quickly, you will wind up with 4 Watts of power dissipation in the MOSFET at 10A current flow, as the MOSFET you are using has an Rds(on) of 40m Ohms/ 0.04 Ohms.

10a^2*0.04 Ohms = 4 Watts.
Decreasing the 1k Ohm resistor to 250 Ohms will switch the MOSFET on and off about 4x as fast, helping to keep the MOSFET out of the linear region more. Decreasing your PWM frequency will help even more, but your losses in an inductive load will increase. You might try decreasing the 1k resistor below 250 Ohms, but you risk burning out the I/O port, as 5v/250 Ohms = 20mA.

You will also need a fast diode across an inductive load to take care of the reverse-EMF when the MOSFET turns off. Otherwise, the polarity across the inductor will reverse, and cause a breakdown of the MOSFET's internal diode. The diode across the inductive load should have a suitable current rating.

Will 10 x 1amp 1N4001 diodes in parallel for the flyback be okay for switching at 1khz 10 amp?

1N400x diodes are very old, and they don't turn off very fast. They were made for 50/60Hz rectification, and when you try to use them with a rapid PWM, they wind up shorting things out for a brief period of time, causing heat and "stealing" power.

In general, you don't want to try to use diodes nor LEDs in parallel, because their Vf is not matched. One diode has a lower Vf and gets more current than the others, which causes it to heat up, which causes the Vf to decrease even further, and soon it goes up in smoke.

So even for protecting against back emf from the inductive load a diode still has to fit the load specification ie current rating and be able to switch fast not the just provable "one way valve" ?

When the MOSFET turns off, there will still be 10A flowing through the motor. Where else do you propose that current go?

The power dissipation in the reverse-EMF diode will be quite a bit more than in the MOSFET. 10A across 0.04 Ohms = 400mV. A diode will have a Vf of around 0.7v to 1.3v, depending upon its' rating vs forward current. It's fairly common nowadays to use a MOSFET as an "ideal diode" instead of an actual diode, as the power dissipation in a properly-selected MOSFET would be much lower than a diode.