How to control the starter motor of the RC aircraft

Thread Starter

sumeryamaner

Joined May 29, 2017
111
I am trying to design a circuit to control the starter motor of a RC aircraft engine.
It is a 30 cc two stroke gasoline engine and has an electric motor attached as onboard starter. The original control module states that it can withstand 100 Amperes. I do not know the exact current required but especially during startup it will be very high. The system will be supplied by a 3S LiPo battery (that means 11.1 - 12.4 V).

My first choice would be a relay. I would use an Attiny85 to control a transistor which in turn would control the relay. The Attiny85 would receive PWM signal from the receiver for the ON/OFF commands. This part of the circuit is very straightforward and does not need any discussion. But the power section has problems.

It was very difficult (actually impossible) to find a suitable relay which should work under 12V and be not too big and heavy. So I decided to look for a MOSFET solution. The logic level N channel MOSFET IRL540N seemed to be a good choice (continuus drain current of 20+ A and peak drain current of 110 A. With four paralleled MOSFETs those values would be 80 A and 440 A respectively). I planned to use two or four MOSFETs in parallel to handle the high current and decided to use PWM to limit the startup current surge which would otherwise destroy the MOSFETs.

My code was sending a PWM signal of 50 % duty cycle for the first two seconds and then switched over to full power. But unfortunately the MOSFETs produced a nice smoke and then the motor was switched ON continuosly (that means the MOSFETs vere burnt out.

Afterwards I found a suitable relay with a max. current of 30 A and used two parallel relays and the control module works flawlessly now.

But...

I really want to know whether it would be possible to design a MOSFET driver for this task. Attached you can see the circuit diagram.

I need to know how I can manage the startup of the motor without damaging the driver MOSFETs.
 

Attachments

jpanhalt

Joined Jan 18, 2008
8,318
Why do you have a voltage divider on the gates? Most MCU's operate at 5V or less. That will not damage a logic level mosfet. Unfortunately, MCU's are current limited (about 25 to 40 mA), so I would advise a dedicated mosfet driver that can supply several amps for the short time needed to switch the mosfets.

I don't believe the mosfet you are using is logic level. Don't be confused by the gate "threshold" rating. That is the voltage at which the mosfet starts to turn on. You need to turn it fully on.

Edit: You also need a "main" capacitor for the configuration you show. Here is a good introductory link: http://www.4qdtec.com/pwm-01.html#simple The main cap may be enough, or you can use faster diodes as shown in Microchip AN898 (sorry, no direct link).
 
Last edited:

Thread Starter

sumeryamaner

Joined May 29, 2017
111
Thank you for the reply. My problem is not driving the MOSFETs. The MOSFET driver circuit is a very straightforward. IRL540N is a logic level MOSFET. The datasheet clearly states that it is a logic level MOSFET. ( https://www.vishay.com/docs/91300/91300.pdf )
The MOSFETs are not being damaged by the "high" Vgs. The damage is due to very high startup current.
 

Andrei Suditu

Joined Jul 27, 2016
51
A mosfet enters the hard on region only when it's 'gate capacitance' is fully charged otherwise it is a kind of resistor.
Fast switching is necessary to charge the gate near instantly so that the mosfet won't dissipate heat during switching periods.Also you need to calculate the power disipation and cool them if needed.
If you have a good switching time (t->0 during tranzitions) then it is something like r*i^2 where r is the mosfets resistance and i is the current.
Also note that at 100 degrees celsius it has other current capabilities.
Also mosfets are sensible to valtage spikes.Even 1Volt over the rating can be fatal.
 

AlbertHall

Joined Jun 4, 2014
8,762
You need the full 5V signal ftom the mcu on the gates of the MOSFETS ro ensure they are fully turned on. If they are not fully on there will be more voltage across them and more power disipation and so more heat.

The value of the top resistors in your diagram should be low (perhaps 47Ω). As the mcu drives its output to 0V this will connect the gate to 0V via that low resitance so the bottom resistors are not necessary. This will also ensure the maximum available voltage reaches the gate.
 

ebp

Joined Feb 8, 2018
2,332
There are a great many FETs with vastly better specs for this sort of application. 77 milliohms of ON resistance is dismal. I almost always ignore the current rating and consider only the ON resistance.
 

ebeowulf17

Joined Aug 12, 2014
2,954
I almost always ignore the current rating and consider only the ON resistance.
I'm still quite new at this, but I've started thinking in those terms too. More often than not, if I run the numbers, the ON resistance becomes a limiting factor long before the claimed current ratings, and they're often not proportional to each other in any obvious way.

Perhaps if I were using giant heat sinks things would be different, but I'm usually interested in seeing what I can do with little to no heat sink.
 

Thread Starter

sumeryamaner

Joined May 29, 2017
111
Thanks for the replies.
The series resistors at the gates are around 22 - 33 ohms and the resistors to ground are several kohms. I have tried it without the resistors to ground but I don't think that this is my main problem. Fortunately my solution with two relays is ok for now.

 
Top