Woops. The 5ms is suppose to come from frequency to time which should actually be 2 ms for a 500Hz PWM signal. I see what yours saying though. I am still not super clear on the gates of mosfets as you can see. Can you explain the switching loss a little more in depth (i don't see where this comes from)? The only power calculations I did were for the load power dissipation needed to drive the motor. 27A rated at 22mOhm when the FET is on give me 16.038 Watts for power. I plan on running two FETs in parallel like mentioned before to give me a max of 8 Watts of power for normal use per FET.
The other thing is I am using 5V TTL MOSFET so I wouldn't switch at the 10V so the current would be higher. So this gives me three viable options that I believe I have components for and I am capable of.
1.) MOSFET driver. I don't have the driver chips so I would have to wait and buy one. I have never used one before but have read up on them when I was planning my original H-bridge design for forward and reverse. I found a lot of information on using these on Electric Vehicles – Equals Zero where he has built multiple projects. It is pretty stable and can easily be driven by the arduino. So that option is there is needed but not preferred since I can't implement it without buying the hardware. I would probably go back to regular logic MOSFETs as well rather than the TTL for those. I have some of them but this would be a redesign of my current circuit.
2.) PNP BJT driver. I could use some PNP to connect the regulated 5V to the gates of the FETs. I can drive the BJT with the arduino and that seems like it would be pretty simple. I am only going to be able to get about 100mA out of the BJT most likely though. I also have to make sure the voltage drop of the PNP doesn't interfere with the TTL logic or I will have to add more circuitry and implement the 12V signal to drive the TTL. This I have parts for and it is cheap and simple. The question comes down to is how much current for nice smooth switching.
3.) SN754410 H-bridge. I have this chip as well. I can use simple TTL logic from the arduino to toggle one of the quad outputs (or two) and connect the gates to the H-bridge. This is able to supply 1A continuous if needed for a lot more power. I can connect either the 5V or 12V gate to this and be able to control either. This maybe a good option but I am not sure if there is a restriction on using an H-bridge to drive a FET. Any ideas on this?
I will look into the zener if need be for more protection. I guess I can explain a little bit better how ideally this project works. I attached a drawing of my current setup. The drawing is crowded but here are the basics.
There is a key switch that connects the 12V to the regulator that is smoothed with caps. 12V is supplied to the 24V relay coil that is hooked to another switch to make or break that for an estop. There is a third switch that turns on the controller. The controller handles the following:
Reads pot for amount of throttle (alters PWM to the FETs).
Read that the relay has voltage
Turn the relay on by activating the BJT
Turn on the FETs but PWM controlling the gates
So the way the software works is that when there is a throttle given by the pot, the controller reads to make sure the relay has voltage. If it does then the relay is activated while the FETs are still off (switches 0 volts). The controller then PWMs the FETs to conduct the main path to the motor. When the drive cycle is over, there is timeout period that will wait to make sure the user is done driving the motor. Once that is complete the FETs are verified to be 0 and the relay opens. This enables to arcing in the relay for safe operation (except for Estop). I have fuses on both the 24V line (30A) and the 12V line (3A) (I didn't include). I am hoping that with this architecture and control I can keep the current pulses down and a safe way to operate the motor. I have a bunch more indicators that I left out as well to determine what stay the software is in.
Currently everything works except the FETs. The FETs currently won't turn on the LED when I PWM the gates. That means that the conduction path is not complete. So hints why I have so many questions about FET. I have already broke a few too (measured diode voltage drops and resistance of the pins to verify) so I know I am not perfect and implementing them yet. So any suggestion are always appreciated.
Sorry the long and lengthy post but I just want to make sure I understand what is happening so I can control and modify it better.
