Motor control question...

[June]

Dear all,

I found a motor control question in the forum. Somebody suggest using FET transistor (which is voltage control) for PWM control.

I'm currently using BJT (which is current control) for my PWM. How do we adjust the output voltage of the PIC to control FET? What is the advantages of using FET to control motor speed?

Thanks for reading...


Regards: June,
___________________________
This message was ported from CCS's old forum
Original Post ID: 10181

[R.J.Hamlett]

:=Dear all,
:=
:= I found a motor control question in the forum. Somebody suggest using FET transistor (which is voltage control) for PWM control.
:=
:= I'm currently using BJT (which is current control) for my PWM. How do we adjust the output voltage of the PIC to control FET? What is the advantages of using FET to control motor speed?
:=
:= Thanks for reading...
You don't adjust the voltage. On both BJT's, and FET's, you are normally switching only on/off. The 'key' is that you switch at a frequency, which is fast enough, that the target device (motor/lamp etc.) behaves as though it is receiving a varying DC voltage.
Now the advantages of FETs are various. The first is that they can be operated in parallel to give extremely low voltage drops in the 'on' state (in fact the internal structure of many power FETs actually has several thousand smaller FETs all built onto the single die, operating in parallel). This works, because the devices don't experience the tendency to thermal runaway of a BJT (with a BJT, if is gets hotter, it tends to draw more load, and get yet hotter - hence if you parallel two devices, and one gets warmer than the other, it will tend to get progressively hotter, and fail).
The equivalent resistance when switched 'on', can be incredibly low. For instance the SUP75N0304, which is a fairly typical 'power' NMOS FET, has an Rds, of only 0.004ohms. Operating at say 20A, and the drop across the FET is only 0.08v, with a power dissipation of just 0.8W. For DC motor control at high power, larger modules like the SKM101RZR, have even lower on state resistances (.0035ohm), and maximum current handling capability of several hundred amps.
FETs also offer very fast switching times (which in RF terms can be a disadvantage), but the switching speed, can be controlled by shaping the waveform applied to the gate, hence this becomes an item the designer can control.
The 'downsides', are that at high frequencies, you need a significant drive ability to give fast transitions, given the gate capacitance, and this driver has to swing much larger voltages than are normally involved to drive bipolar transistors. It is also very difficult to design power MOSFETs, that can switch both high currents, and high voltages, and the drive requirements become harder.
Hence if you look at commercial PWM drivers, you will find power MOSFETs cover perhaps 90\%+ of the market, for voltages below 100V. Above this, bipolar switching becomes far more common.
Now I suspect I was the person who you refer to, and the original poster in this case, was dealing with low voltages. Hence it becomes possible to drive a FET H-bridge, directly from the PIC, without having to worry about special 'high side' drivers. In this case it also becomes possible to do some nice 'tricks'. For instance, if you have an H-bridge, with the two 'high side' transistors, called Q1, and Q2, and the low side transistors Q3, and Q4, then you can drive Q1, and Q2, directly from the PIC. For PWM, you can connect Q3, and Q4, to two simple CMOS 'AND' gates, feeding one input of each from the PWM output of a PIC, and the other from two more control pins. Again, given that the input impedance of such a CMOS gate is very high, you can have a 100k resistor to ground from each of these input pins.
Now the nice things about such a layout are these:
On switch on, when the PIC pins are programmed as inputs, the 100K resistors ensure the AND gate is 'off', and hence both the low side FETs are switched off. You can program your 'direction' bit patterns, and PWM rate, before enabling the TRIS, to give a smooth start.
You have independant control of direction, and speed.
If you use the PWM, driven from Timer2, and generate an interupt on (say) every sixteenth pulse from Timer2, you can have an ISR, that sets the new PWM pulse width, at this point. This allows smooth power changes.
If you select a PWM rate at perhaps 16KHz, this gives a speed update at 1KHz, and a PWM rate high enough that most motor designs will behave as if this is a smoothly varying DC voltage.
If you turn 'on' both high side drivers (disabling the low side ones..), this gives magnetic breaking on the motor.
There is a balancing act here, in that the efficiency of the switcher, declines slightly with faster switching, but at low frequencies the motor will produce audible noise.
The overall efficiency of this type of system can be excellent. For instance I use a variation of this, to control a 24v 1/4HP servo motor, off a PIC, and the power drivers do not even need heatsinks!.
Controllers used for applications like robotics, golf trolleys etc., will all use MOSFET switches.

Best Wishes
___________________________
This message was ported from CCS's old forum
Original Post ID: 10185

[Jason]

Just some suggestions:

I like to use IRL1004's (digitally controlled mosfets, only require digital high to turn on and gnd to turn off) for motor drive & H-bridge operations. When using an H-Bridge configuration you will have to use opto-isolators on the top 2 mosfets. I use PVI-5050N's for optoisolators.

Regards,
Jason
___________________________
This message was ported from CCS's old forum
Original Post ID: 10195