Thread Tools
May 01, 2014, 10:45 PM
Registered User
Thread OP

DIY Brushless ESC woes...

Alright, so for the last few months, I've been trying to build my own brushless ESC, and I've had absolutely no luck whatsoever in the hardware. I've scoured app notes and schematics, but I'm having a lot of trouble still.

My main problem is that I can't build up the correct transistor array to actually control the energizing of the coils.

I started by getting a bunch of N and P mosfets (the PSMN050 and the P12PF06) but those didn't even turn on when I tried controlling them with an Arduino microcontroller board. Turns out they weren't logic level, so I tried using a mosfet driver (the UCC27321), which somehow ended in a puff of smoke… I then tried a triple half bridge (the L6234) which resulted in some weird jumpy movement when I passed current through a single coil to test it, and it fried my Arduino.

My real problem is that I don't fully understand mosfets and the schematic required to control these motors. I hate to ask this, but would someone be able to draw up a really basic schematic of the control circuit (i.e. the transistor end, ignoring all the back-emf sensing and whatnot), or recommend which mosfets I should use to control a smallish out runner from a microcontroller? I know I could just go out and buy an ESC, but I need to build one that I can reprogram for a high school project.

Thanks greatly in advance for any help, I'm at my wit's end
Sign up now
to remove ads between posts
May 02, 2014, 07:37 AM
Registered User
jpparisy's Avatar
To drive a MOSFET having a non-logic input, we need a little N-channel MOSFET having logic-input, acting as a "driver". Here is the schema of this driver :

The basic schemas for N-channel and P-channel MOSFETs using this driver are as follows :

This is the basics, of course, in real life , we don't use little MOSFETs, but we use instead integrated drivers.

There are, at least, one thousand references, I prefer those produced by Microchip because they have a more recent design, and particularly these ones : TC4426, TC4427 and TC4428, their schema is the following :

We find the N-channel MOSFET input, the resistor is replaced by a constant current generator, a whole bunch of other components (let designers justify their salaries) and a CMOS output. There are two drivers per package.

Using an half TC4426, the previous basic schema for the P-channel MOSFET becomes :

Note that, if the other half of the package is not used, the input 2 must be tied to ground (0 V).
May 02, 2014, 08:58 AM
Registered User
Thread OP
Thanks a lot for the reply - I went out and bought some logic level mosfets with a low resistance (the P channels cost $4 a pop...), and I have success

I managed to get my 750kV outrunner spinning at decent RPM, fairly smoothly! All that's left now is to implement a positional feedback loop, which is going to be tough on the code
May 02, 2014, 09:42 PM
Registered User
It sounds like you got the answer you were looking for.

Basically a FET is a variable resistor, low resistance when the gate to source voltage is high and infinite when the gate to source voltage is zero.

It is very important to turn the FET off and on very fast which is hard to do as the capacitance of the gate is large. Just using a pull up resistor will not work. Most people just use a FET driver module like the IR2110 which has all the fast, high power drivers needed.

There are special devices like the on semi MDC1000A turn off device that can be used.

Here are two circuits for driving a brushed motor. There has to be a flywheel diode across the motor and a capacitor to damp the arcing from the brushes.

Both circuits use logic level input but the first circuit can use a higher level input. I have used the first circuit but not the second. I plan to use it on a DIY quad using a cheap control board.

The second circuit uses an SCR to turn off the FET and a fast 7407 logic gate to turn on the FET.

On the second circuit the power supply voltage would be limited to the gate to source voltage of the FET. Sometimes it is possible to push the gate to source voltage way beyond the spec so 30 volts is possible if you want to live dangerously.

Quick Reply

Thread Tools