Motor speed control using H bridge

[p.gholap93]

I have been using L298 IC as motor driver and control motor's speed using the same. But the problem is it can supply peak current of 2A. Whereas I often need 6 to 7A of current. So I decided to go for H bridge driver using IRFZ 44n power mosfet.But there are 2 problems-
1) How should I control speed of motor? Should I directly give PWM output from MCU to gate? If so then there is problem of reduced drain current.Because as I reduce gate voltage drain current also reduces. And gate threshold voltage is 4V and square wave is of 5V peak.
2)Is there any problem of frequency of square wave also?

 

[Sceadwian]

You'll have to take the square wave of 5V and amplify it, this is easily done using a push pull switching transistor arrangement. Ideally the gate should be driven as stifly by the highest voltage that the gate will routinely tolerate. The threshold voltage has nothing to do with the optimal gate voltage which is always MUCH higher than the threshold voltage.

 

[misterT]

2)Is there any problem of frequency of square wave also?

The higher the better, but at some point the switching losses and the performance of the FET and the fet driver become a limit. Higher frequencies reduce the ripple current in the motor and it will run cooler and more efficiently.

 

[p.gholap93]

Ok. First let me confirm that to control speed I have to give PWM to the gate of FETs, right? Having said so the problem which is in my mind is if I give 50% duty cycle square with 5V then average gate voltage will be 2.5V which is well below threshold voltage so FET will no longer conduct. Even if I use 70 or 80 % duty cycle then average gate voltage will be around 4 or 4.5V thus reducing the drain current.

 

[Sceadwian]

The average gate voltage is completely irrelevant, unless the resistance to the gate is extremely high it will not integrate it will only see 5 volt pulses not an average of 2.5 volts. Also as stated above you can not directly drive the gate of the FET with a 5 volt source you will get only a fraction of the total current able and it will never properly reach a fully switched state, you need to use a gate driver, the easiest setup is a basic NPN and PNP transistor in a totem pole configuration, this will allow LARGE gate currents (fast switching) and can be run from a much higher voltage source (12 volts is generally a good voltage, typical gate breakdown voltage is 20 volts) the 5 volt logic would be used to drive the base of the NPN/PNP transistors which would provide the actual drive to the mosfet gate. Even so called logic level mosfets will benefit from a higher gate voltage.

 

[p.gholap93]

First thing I didn't get what does it mean-"unless the resistance to the gate is extremely high it will not integrate it will only see 5 volt pulses not an average of 2.5 volts."????????
If the gate still sees 5V pulses then I will directly drive it with PWM of any duty cycle from controller. That much voltage is sufficient to keep FET in ON state. Why should I amplify then? Also gate current is very small in case of FETs. So there is no problem with current also.

 

[ronv]

I'm not sure what your schematic looks like but the high side FETs may want to be PFETs or you might need a high side driver for the NFET.

 

[Sceadwian]

You might want to double check your on state resistance at a given current. FET's are NOT that simple.

 

[p.gholap93]

OK finally got you. Basically MOSFET does not turn ON completely for 5V supply so need to amplify it.Now my second question is why not use opto coupler (instead of transistor driver) whose output photo transistor is driven with 12V collector supply? On the input side LED I will feed PWM output from controller.

 

[Sceadwian]

Only use an opto coupler if you actually need the isolation, the switching speed of an opto is POOR. A basic NPN/PNP totem pole driver using switching transistors is the easiest solution, just need one gate resistor and two transistors, though some protection circuitry might not be a bad idea because if the gate ever fails short circuit the switching transistors will smoke, protection circuitry will however tend to reduce response time, direct connection to fast switching transistors will give you a very crisp on and off time resulting in the least amount of loss and can be run from any voltage rail compatible with the transistors and mosfet gate.