I’m building an very large robot and I need to design a seriously bad ass motor controller to handle the current requirement–two motors, 15 to 30 amps each at 12 to 24 volts. (The robot’s purpose is outdoor navigation, not combat, but its size and construction is comparable to a Battle Bot.)
I want to use your dual serial motor controller because I’ve been very impressed with the product from the experience I had using the micro version on a smaller robot and the price is very cheap. However, I do not wish to use your 30 amp version because it is $99 and controls only one motor. If I use those for this project, I’ll need two of them. Then I have $200 invested and if I connect something backwards then -poof- $200 goes up in smoke.
Instead, I want to use International Rectifiers Hex-Fet power MOSFETs for the H bridge drivers. They have one that can handle 180 amps. Also I want to use optoisolators to drive the gates of the MOSFETs to keep voltage spikes from the motors from damaging the microcontroller. By using discrete transistors for my H bridge, if I blow one up, all I have to do is replace one $5 discrete component, instead of an entire motor controller.
The main issue I have is with the switching overlap. Obviously I don’t want both transistors in the same totem pole turned on at the same time. I know that the SN74410 motor controller chip has built in circuitry to handle this for its internal H bridge but can my discrete H bridges gain the benefit of it indirectly?
Would I be better off to use the kit with the SN74410 included, or just forget about the SN74410 and simply drive the optoisolators directly off the PIC’s logic outputs?
Dennisf486,
I use external FETs to increase the current handling capability of the Dual Serial Motor control. See my web site. “Excuse II” has some schematics that may be of use.
The SN74410 is used to drive the FET gates. I did use optical isolators originaly, but found that they were not needed.
The 180 amp FET measurement is for impulse current handling capability. You would have to do the heat dissipation calculations to get the actual steady state current capability. I’d expect more like 10 amp limit on the FETs without a heatsink.
Thanks. I had a look at your schematic and I ordered some MOSFETs from International Rectifier, but I decided to go with the less-powerful IRFZ24N and IRF9Z24N because they are substantially cheaper. That way I could afford to buy 16 of each so that I can bank more together, won’t have to cry if I blow up 1 or 2, and maybe I’ll even have some left over for the running lights on my robot.
I’ll probably also leave out the Zeners you have protecting the gates of the MOSFETs from over-voltage, since I won’t be running a high enough motor voltage for that to be an issue.
I had to study the characteristics of the resistor-divider network you have connecting the MOSFETs to the output of the motor controller IC to fully appreciate the ingenuity of the design. It appears that if for some reason the inputs to the circuit are left floating, the top transistors in the H bridge will turn on and the bottom ones will turn off - making a motor braking effect. I’m also guessing the schottky diodes allow the charge to be sucked off the MOSFET gates more quickly than if they were emptied via the resistors alone. Have I got that right?
However, what if I want the motor to float intead of brake? It seems that sending the float command to the motor controller will only result in a brake situation in this circuit. Could this increase power consumption during forward or reverse operation as well? If the normal pulse width modulated operation of the Pololu motor controller pic we’re using here is On - Brake - On - Brake - On, etc., then the MOSFET add-on circuit will be behaving identically to the way the chip is. But I thought it was more power efficient to operate on - float - on - float, etc. (because you’re not trying to brake the motor speed in between each on cycle.) Then the add-on circuit will be converting those float periods to brake periods. I just don’t know what the Pololu chip is programmed to do.
Dennis,
The IRFZ24N and IRF9Z24N should be good for about 4 amps continuous. You have to calculate the heat dissipation to get maximum steady state current capacity. Sure, they will pull 18 amps continuous with a lot of heat sink attached.
If your supply voltage is under 18 volts, the circuit can be simplified. Omit D5 thru D12. Keep R1 and R2, and replace R3 thru R6 with solid wire. The problem is the FETs Gate to Source voltage limit of 20 volts. There is no room for “just a little more”. The Gates are destroyed at 20 volts.
There has been a lot of discussions about PWM and what to do in the “off” state. Some provide reverse voltage, some provide brake, and some provide open circuit. The standard Pololu motor controllers provide brake. There is no open circuit option. The outputs are either +voltage or ground…no open circuit output.
I tried the open circuit option some time ago and I could not control the robot. My sumo robots run a lot of patterns. I need reliable, repeatable constant radius turns. With open circuit condition during the “off” state, the turn radius kept on getting larger and larger. I needed the brake phase to keep the inside wheels at a constant speed and the turn at a constant radius.
Good luck with your project.
