The 4A magnetic switch is all well and good, but not of much use to me. I'm flying mostly electric gliders that use 3S or 4S LiPo batteries with a 30A ESC to control the motor. I decided to pursue something that might be more useful for my purposes. Here's what I came up with as a first pass:
The LDO:
U1 is a low-dropout regulator (LDO). The MCP1702 can't handle 16.8V from a fully charged 4S LiPo battery. I needed something better that could withstand at least 18V without going up in smoke and not drain the battery in an afternoon. After an exhaustive search through DIgikey, I settled on the LT1761 LDO regulator: VIN = 20V, quiescent current = 20µA typ., Iout = 100mA. There were other LDO regulators available, but they did not handle the required input voltage or the quiescent current, or cost was unacceptable. The LT1761 is about 2.5x more expensive than the MCP1702.
The FET Switches:
I poked around Digikey trying to find a reasonably priced Nchannel FET, with really low RDSon @ VGS = 4.5V. The best one that I found was the IRLHM620PbF: RDSon < 2.5mΩ @ VGS = 4.5V. The package will be a bit of a problem -- 3.3mm square DFL, with the leads under the package. This might be a problem when soldering, but if it works it will be great. I'm using two of them to get 30A load current, 40A in a pinch.
Other Stuff:
Almost all the rest of the circuit is unchanged. I paralleled R5 to reduce the stress on a single resistor, given the expected power dissipation.
R2 is there to protect the LDO regulator from damage if the input/output voltage spikes above 20V when the EXC is running the motor. I expect there will be significant noise when the motor is running. R2 is a little bit of insurance.
I've tried to minimize the trace resistance on the PCB. After I evaluated the trace resistance on the first 4A prototypes, I made an effort to minimize it on this version, especially since the load current is so much higher. My goal is for the trace resistance to be less than 1mΩ. Using 2oz. copper will help, along with keeping traces short and wide.
Cost:
This circuit will cost more than 2X the 4A version:
A Still pretty cheap. Almost ready to pull the trigger to get boards made.
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.