I wanted to open up a discussion about how we are powering our avionics, meaning servos and APM, on helicopters as they present a unique challenge in the APM world, and must be treated carefully. After crashing my Octocopter last weekend because I let the avionics battery drain, this has come front-and-center for me. I want to make sure it doesn't happen to my helicopter.
Helicopters present a particular problem because we don't have to worry only about the battery dying, but we have to worry about the servos doing something which drags down the APM supply voltage. While looking into the situation, I found this really interesting thread:
http://www.helifreak.com/showthread.php?t=398142
Long story short here, these guys are using a new Savox servo which is very fast and powerful, and it turns out it can draw a LOT of amps. A problem occurs when using a BEC powered by the main flight battery as the avionics power supply. If the servos move too fast, they draw a LOT of amps, it's a spike load, and the switching regulator cannot respond fast enough, and the voltage drops momentarily. It's not a big problem for the servos, but it is a problem for their FBL controllers. In some cases it causes a reboot.
Now, my servos are not that fast or powerful, but it's still a concern. Something to be aware of and watch out for.
This is one of the reasons I have decided to use HV servos, and power them directly from a battery. It does help the situation, but I think we should all consider using some capacitance in the system to prevent the problem. That was the solution in the case of these Savox servos.
This weekend I am going to be working on trying to improve the code to be able to monitor a second battery voltage, and also give some warnings about any problems
Replies
Hi,
there is some interesting information
here Power supply
and here Hints regarding Servos
This is excellent info! Thanks for posting!
Looks like even a 2S Lipo direct can benefit from caps, though they don't really have a risk of brown-out, but just to stabilize the supply. And they do help to soak up the voltage spikes.
I have some pretty good proof of the effectiveness of the ideal diode controller on the Pixhawk.
I'm setting up a new heli. I'm using a linear 5V regulator, powered by a 2S battery, wired into the PM input. The same 2S battery powers my servos, they are powered completely separately from the Pixhawk using a breakout board. While setting it up, I was getting the board rebooting every time I moved the servos. While troubleshooting, I plugged in the motor battery to the ESC. The ESC has a BEC connection, which puts power on the input rail. While the motor battery is plugged in, the brown-outs stopped.
Huh.
Check the voltage of the 2S battery... ooops, it's dropped below 3.0V/cell. Voltage probably went very low every time I moved the servos. But when the system has a secondary supply on the input rail, it will not be affected. The IDC selects the best power source live, on the fly, and can switch fast enough to avoid a brownout.
It looks you have plugged two 5V BEC in the rail?
As far as I can remember when I got into quadcopter, I read some posts about people cooked their ESCs when plugging all their ESCs into the rail without removing the 5v pin and the unstable 5v output has influence between each other.
I have ever measured the BEC output of some of my ESCs and they all have different voltage output, from 5.2v to 5.5v. I'm a bit concerned whether this voltage difference will damage them if running parallel.
No, I don't have two BECs on one rail. I have one linear voltage regulator going into the Power Module input. And I have an ESC BEC on the servo rail. The Pixhawk has an Ideal Diode Controller inside it which automatically switches between sourcing either from the USB, PM Port, or Servo Rail, in that order. It is an intelligent device, and it only switches on a power supply if it is suitable, and it never switches on two at the same time.
So the supplies are not in parallel really. And it switches fast. Fast enough to keep the Pixhawk alive if you have a regulator fail in flight. And I just saw a comment that somebody has already had a power supply failure in flight, and the Pixhawk stayed alive on the BEC input. Sweet.
That's cool. I'm sure we all need this and we'll have a very reliable UAV in the air.
Digital Servos can produce a nasty spike well over the voltage on the rail.
This has been replicated on a few brands of servos so far.
By putting a Zener across the servo rail, this effect can be clipped
Select a Zener that is of a higher voltage than your peek measured servo rail voltage when the servos are still.
On Pixhawk, running off a 5v servo rail, with the Brick removed, these servos will cause an inflight reboot.
Please be careful
Philip, when you say "brick removed" you mean no Power Module, thus the ideal diode controller is drawing from the servo rail, correct?
So I wonder... if we are using the Power Module, or some other supply on the PM port, the Ideal Diode controller prefers that port right? So then we can connect servos on the rail, and even if these spikes hit, it won't affect the Pixhawk negatively?
-
1
-
2
-
3
of 3 Next