Servers by jDrones

Making ESC PWM pins redundant


(piningforfjords) #1

We had an aux pin on the pixhawk2 act flaky. We use it to drive our LEDs. Under the microscope it mechanically looks fine and after seating the connector a few times it seems to have sorted itself out. Still I couldn’t ship this vehicle/flight controller to a customer. We have made our own carrier board and this is roughly vehicle 10 in our development process. This is the first issue we have had with the connector. However this is a very difficult connector to weatherproof and vehicles do take a beating sometimes. Who knows what sort of issues will crop up with corrosion over time. If this was an ESC PWM line the vehicle would have crashed. I would like to see redundant pins for the ESC lines. One way it could be done via firmware for a quadrotor configuration (not completely sure this doesn’t have some serious downsides) is just duplicate the PWM signals out on the remaining 4 ESC pins. Right now this is the weakest link we know about on our machine and I would like to see how we could make it more robust.


(Peter Hall) #2

you can output the same motor output on multiple channels, ie set SERVO1_FUNCTION to motor 1 and set SERVO2_FUNCTION also to motor 1.

Not sure how you will get both signals into your esc tho. Probably not a good idea to just connect the two outputs in parallel. If you after redundancy your better to add more props, ie go to a hex or octo, you get full redundancy of motors, props, esc’s and pwm pins.


(piningforfjords) #3

Hi Thanks for the reply. I would tend to agree with you if you are using budget esc/motors/props. In ten years of designing and operating drones in salt water (Alaska) and dusty (Iraq) conditions if you pick/test your components well and obey the MTBF for those components you just don’t have failures of the ESC/motor/props. Increase the number of motors you lose efficiency and your machine is significantly bigger and harder to pack. The failures you do see are in electrical connections that can’t be soldered. The Pixhawk 80 pin connector is the only connector we can’t choose. Every other connector critical for flight has either redundancy or a double blade self cleaning connector. Every single aerial vehicle out there has to use the first 4 aux lines. And many will just drop out of the sky if those connections fail. The Pixhawk seems to fail in this one aspect as an industrial component. Makes sense to me to have an ‘official’ way to have redundancy. I would just like to see some discussion on the best way to achieve that.


(Mike Boland) #4

Redundancy is big issue, especially as Ardupilot/Pixhawk is pretty much mainstream now.

I am glad you have broached the subject.

FC to ESC/Servo connection is a major point on all aircraft.
So how do you feel about the move to UAVCAN?
Here we have a single connection from the FC to ALL devices.

Do you run multiple batteries with a sensor on each and ability to run off one?
GPS’s? One or two? Compasses?
Now CASA in this country are demanding redundancy in any RPAS, even though they don’t actually understand what that entails.

So I see this subject as a real can of worms and am glad you have raised the discussion.
I hope we get lots of worthwhile input.


(piningforfjords) #5

Hi Mike
Definitely a worthy question. Where do you draw the line? It is really where do you want to place your bets and what risks can be anticipated or reduced to acceptable levels through maintenance. We do happen to have two batteries and each is able to sustain the vehicle in hover with full payload weight as can their redundant / shared connection to the vehicle. There are other benefits to dual battery arrangements. I am fine with a single line for UAVCAN and a single GPS provided the connector is robust. Those components don’t physically degrade much over time. With a GPS the performance is measurable and loss might cause it to land in an unexpected place that might be catastrophic if you can’t manually take control but you can mitigate/understand that risk at the time. The best thing dual GPSs gets you if your vehicle is big enough is the ability to eliminate the chief sensor vulnerability. The compass is particularly onerous but all you can do in most cases is minimize the risks through electrically coupling power lines and component / field arrangement. I think Pixhawk has done a great job with the compass by having a redundant compasses and locating them on the flight controller in a shielded case. I don’t use an external compass because I2C is terribly vulnerable to rf interference. With an ESC aux pin failure there will be no warning you simply have a vulnerability that can take down your vehicle at any time. Not really acceptable. It likely will be a deal killer in terms of staying with the cube. Our philosophy is to apply redundancy for all physical connections that sustain flight that can degrade. This one can’t even be inspected. For me electrical connections are gigantic risks when they are exposed to wet/dusty environments with airflow and vibration.

Tom