Fet Drive Circuit With Pic Micro

[gramo]

I've read in the past of issues with Fets getting excessively hot due to not being able to switch off fast enough, as the pic can only sink around 25mA of current, and doesn’t provide a "fast enough" drive circuit for the FET to turn on and off.

For example - the FET takes slightly more time to turn off (or on), and as a result, it will get to its bias point, where it has resistance, and has a voltage drop over it (for this example say its around 5V at this point in time). Now remembering that there’s alot of current (say around 20A) being passed through the FET, its now dissipating alot of power in the form of heat, in this case 5 volts * 20 amps = 100 Watts.

I cant for the life of me find the thread that elaborated on how to build the drive circuit, but from memory, I believe it was something like this;

**broken link removed**

That allows the FET to turn on and off much faster, thus prolonging the life of the FET..

My problem is, how do you know what resistance the drive resistor should be

 

[Nigel Goodwin]

Do you need a drive resistor at all?.

 

[gramo]

Not sure Nigel, you think it would run safely without it?

 

[Nigel Goodwin]

Not sure Nigel, you think it would run safely without it?

I've never been much of an FET user, but you see a lot of FET circuits without it.

 

[cadstarsucks]

I've read in the past of issues with Fets getting excessively hot due to not being able to switch off fast enough, as the pic can only sink around 25mA of current, and doesn’t provide a "fast enough" drive circuit for the FET to turn on and off.

For example - the FET takes slightly more time to turn off (or on), and as a result, it will get to its bias point, where it has resistance, and has a voltage drop over it (for this example say its around 5V at this point in time). Now remembering that there’s alot of current (say around 20A) being passed through the FET, its now dissipating alot of power in the form of heat, in this case 5 volts * 20 amps = 100 Watts.

I cant for the life of me find the thread that elaborated on how to build the drive circuit, but from memory, I believe it was something like this;

**broken link removed**

That allows the FET to turn on and off much faster, thus prolonging the life of the FET..

My problem is, how do you know what resistance the drive resistor should be

The PIC output would cause heating in the FET for the reasons others have mentioned.

The gate resistor limits the turn on speed which will slow turn on. The limits EMI and voltage spikes from the motor inductance. You need a schottky diode across the motor as well.

One of the reasons for high current drive on FETs used this way is the miller or reverse transfer capacitance from drain to gate. As the FET is turning on this capacitance is charging from 0 to 12V in this case. If you look at the spec sheet you often see a turn on curve you can see where the output current stops rising as that charging is taking place. See fig 6. Since this is only running at 12V this is not such an issue.

D.

 

[gramo]

The PIC output would cause heating in the FET for the reasons others have mentioned.

Could you explain that for me please

The gate resistor limits the turn on speed which will slow turn on.

Doesnt that negate the use of the push pull amp? I thought we were trying to drive the FET faster to reduce power dissapation by it

 

[cadstarsucks]

Could you explain that for me please



Doesnt that negate the use of the push pull amp? I thought we were trying to drive the FET faster to reduce power dissapation by it

Typical gate resistor values are on the order of 10 ohms for most FETs. Logic level ones are lower, the spec on the one we are looking at for the purposes of discussion lists it's speed with 4.5V gate drive through a 1.8hm: resistor.

The result is control of the turn on while providing over 2A to the gate or two orders of magnitude more than the PIC pin spec.

D.

 

[gramo]

Typical gate resistor values are on the order of 10 ohms for most FETs. Logic level ones are lower, the spec on the one we are looking at for the purposes of discussion lists it's speed with 4.5V gate drive through a 1.8hm: resistor.

The result is control of the turn on while providing over 2A to the gate or two orders of magnitude more than the PIC pin spec.

D.

I'm sorry for delving further into this, jsut want to get it right the first time

I cant find where it lists its speed with 4.5V gate drive through a 1.8ohm resistor, where did these values come from on the datasheet?

 

[Ayne]

https://www.electro-tech-online.com/custompdfs/2007/02/MAX5054-MAX5057.pdf

MOSFET drivers

 

[dknguyen]

You only really need a FET driver if you are switching a high-side FET or if you are switching something high power. If it's something like a speaker you may not need a driver depending on how much power it is.

The gate resistor on a FET is to actually SLOW DOWN the speed that a FET switches on at to reduce noise and transients and others things like that. You probably don't need one.

And you might as well buy a driver chip from International Rectifier or Allegro Microsystems or whever rather than build your own. It's smaller and works better anyways.

 

[Ayne]

It is good practice to add optoisolator between Microcontroller & Driver chip.
PIC > optoisolator > MOSFET drivers
OR
There is no need of it???

 

[cadstarsucks]

It is good practice to add optoisolator between Microcontroller & Driver chip.
PIC > optoisolator > MOSFET drivers
OR
There is no need of it???

It is never required. Optos can be helpful for level shifting but are only required for safety isolation.

D.

 

[cadstarsucks]

You only really need a FET driver if you are switching a high-side FET or if you are switching something high power. If it's something like a speaker you may not need a driver depending on how much power it is.

The gate resistor on a FET is to actually SLOW DOWN the speed that a FET switches on at to reduce noise and transients and others things like that. You probably don't need one.

And you might as well buy a driver chip from International Rectifier or Allegro Microsystems or whever rather than build your own. It's smaller and works better anyways.

Drivers are ALWAYS needed to PWM with the FET to supply the instantaneous AMPS required to switch the FET in a reasonable amount of time.

D.

 

[cadstarsucks]

I'm sorry for delving further into this, jsut want to get it right the first time

I cant find where it lists its speed with 4.5V gate drive through a 1.8ohm resistor, where did these values come from on the datasheet?

Not a problem. Page two of the data sheet about halfway down there are the rise time, fall time and delay time specs. On the far right column it lists the test conditions. This is often a good idea if you feel the need to slow it down just a touch.

D.

 

[dknguyen]

Drivers are ALWAYS needed to PWM with the FET to supply the instantaneous AMPS required to switch the FET in a reasonable amount of time.

D.

THen take a look at the PCB in an RC servo.

 

[Nigel Goodwin]

THen take a look at the PCB in an RC servo.

Certainly historically they didn't use FET's, they used bipolar transistors.

 

[cadstarsucks]

THen take a look at the PCB in an RC servo.

Is that not a timed as opposed to a PWMed function?

D.

 

[dknguyen]

Not sure what you mean by timed but the internal circuitry controls the motor via PWM. Whether or not BJTs or FETs are used in the commercial servos (I'm positive that a bunch of newer ones do), there is a group out there that makes a custom PCB to replace the one inside an RC servo and it drives the FETs with an AVR. Works just fine (obviously you can't switch giant FETs at insane speeds without a driver).

 

[cadstarsucks]

Not sure what you mean by timed but the internal circuitry controls the motor via PWM. Whether or not BJTs or FETs are used in the commercial servos (I'm positive that a bunch of newer ones do), there is a group out there that makes a custom PCB to replace the one inside an RC servo and it drives the FETs with an AVR. Works just fine (obviously you can't switch giant FETs at insane speeds without a driver).

Timed, as in you turn it on for 5 seconds and then turn it off, as opposed to PWMed which is a speed control technique that turns it on and off 20000 time a second or more.

D.

 

[Nigel Goodwin]

Modern servos use PWM to slow the motor down as it nears the target position, this helps reduce overshoot and makes then faster and more accurate. It also gives you a little speed control over a modified one used as a motor.