Motor/Coil Current Limiting with OpAmp & MOSFET

[Solveer]

I'm working with a DC motor controller that requires an additional current limiting circuit. The value of the current doesn't have to be very exact as long as I can control it so it doesn't drastically overrun the specs (5A, 1ohm, 3mH).
What I have thought of is a opamp that has a potentiometer on the non-inverting input and feedback through a 0.1ohm resistor on the inverting input. The Output goes directly to the gate of a MOSFET.
Having put together a quick breadboard prototype using a LM724 and a IRFZ44 with a non-inductive 0.1ohm, 3W resistor, I have found that the MOSFET heats up too much for my liking (and easy implementation of a heatsink). I took a look at the voltage on the gate and found that it doesn't rise/fall nearly as fast as expected (the Gate charge is 43n and the opamp outputs 40m typical, so the rise time should be <2ns... I measure in the range of 10us)... what am I missing?
Thanks

http://www.solveering.com:

Multidisciplinary Engineering Services

 

[itsmoked]

Boy.. I would do this differently. You are placing that device into linear control. That's asking for lots of heat and associated headaches.

You should just ON-OFF the circuit to protect it and not try to regulate it.

Far better yet would be to just PWM the motor and monitor the average current. There are even PWM drivers that will feed back the current for you and help limit the Iavg.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[Solveer]

The idea behind the opamp is really to use it a bit like a PWM circuit. The Op Amp is being used as a comparator here, not as an amplifier. Take a look at the attached images of the circuit (parts aren't exactly what I'm using but fairly close) and plot that I got from LTSpice (green is the voltage at the gate, blue the current in the sense resistor). I think the rise time is ok, but the fall time gets messed up by the plateau around 2.4V... why does it go through that and not just right back to Gnd? The gate capacitance/charge is really small for this and the Opamp can put out a fairly high current.

Circuit:

http://files.engineering.com/getfile.aspx?folder=ccab4b10-3cf4-48b4-b7da-c75ecbb86365&file=trace.jpg

Traces:

http://files.engineering.com/getfil...-b4e5-4ddf-ba6d-ee999342a0b4&file=circuit.jpg

http://www.solveering.com:

Multidisciplinary Engineering Services

 

[benta]

Do you mean LM324?
In that case, why would you expect that this part can switch in 2 ns?
Second, I do not think it likes a capacitive load very much.

Benta.

 

[Solveer]

Thanks... I guess you nailed the reason why this won't work like this. The LM324 (yes, that's what I meant) has a fairly low frequency response.
Now... what if I added a Schmitt trigger gate (like the 74AC14, and reversing the + and - inputs on the LM324)? That should get the switching times down to the levels that the gate can achieve, correct?
Any other ideas on how I could do this? I'd like to stay with a opamp/comparator and MOSFET configuration...

Thanks.

http://www.solveering.com:

Multidisciplinary Engineering Services

 

[Solveer]

Ok... some time passed and I actually got the chance to work on this a bit more. So I've added a 74HCT175 flip flop and 10kHz square wave to the F/F Clock and Master Reset inputs (I might try adding a higher freqency to the clock and a lower one to the MR). I then ran the output from the comparator to the Data input and the output from the F/F to the MOSFET. So essentially, the Comparator will reset the F/F once the current (voltage across the 0.1ohm resistor) exceeds a preset value. Then once the clock pulse (connected to the Master reset) goes high, the output goes high again...
Am I missing something? The Mosfet still gets hot when I run this at currents higher than about 1A... The rise and fall times of the MOSFET seem fairly good (<600ns).

On a related question, how do I calculate the heat generated by a MOSFET switching in this situation? I've got the Equation Vs*Id/2*fsw*(tr+tf)... if I combine that with the Id^2*Ron*D, I get a value that is too low... any comments on that?

Thanks (EE unfortunately was/is lacking attention in a ME degree...)

http://www.solveering.com:

Multidisciplinary Engineering Services

 

[benta]

It seems you are a bit in over your head here, but anyway:
* The IRFZ44 will not turn on hard with only 5 V gate drive, it needs 10 V.
* A 74HCT has far too little current drive for a MOSFET this size.
* Use a commercial gate driver, there is a reason semiconductor companies make them.

Benta.

 

[Solveer]

... over my head? Understatement
As I was burning my fingers, I figured that turning on the IRF might be the issue and tried the same setup using a TC1427 inbetween the 74HCT175 and the IRFZ44 with the driver's Vdd at 10V. But still the same problem. I checked the gate voltage on my DSO and found it to rise quite nicely...
Now what? Would swapping out the IRFZ44 for say a IRLZ24 (I have a couple of those on hand) do me any favors?

Thanks for the support.

http://www.solveering.com:

Multidisciplinary Engineering Services

 

[RajeevCell]

Are you driving the MOSFET gate with +5 volt? That's not enough. The MOSFET will be in the linear region with this gate voltage. Use +10 or 12 volts for gate drive. That will solve your heating problem.