External servo power source - revised

by Lubis | April 8, 2013 | (5) Posted in Tips

On my hexacopter with ArduPilot I have a camera gimbal with two MG995 servos. These servos are quite large and therefor draws a lot of amperage, especially when they move fast. This caused my flight controller board to reset and me loosing my OSD amongst others. Hence I needed an external power source for the servos. Since I have six Turnigy Plush 25A I have five BECs to spare, so why not use one of them?

 

Since I first published this article I have made a second version of the power distribution. But since de older version also works I figured I'd leave that tutorial on the bottom of this article.

############################## VERSION 2 ##################################

I didn't want to cut the servo leads so I had to come up with something to connect the power lead and the ground from the servo to my BEC whilst still having my signal connected to my flight controller (FC).

Here is what I was trying to accomplish.

On the left hand side id where I will connect my two servos. One for roll and one for tilt. On the top right is the two signal wires to my FC and on the bottom right is my BEC.

The wiring is all done. As you might see in the picture the bottom two pins in the bottom left doesn't seem to be connected, but to spare some room I wired those on the bottom of the board.

The two bottom pins are wired on the backside of the board.

Here is what it looks like when a servo is connected to the top left, a BEC to the bottom right and two signal wires to the top right.

To clarify a bit:

Top left yellow - servo signal
Top left red - servo power
Top left brown - servo ground

Top right yellow - receiver/FC signal to servo 1
Top right orange - receiver/FC signal to servo 2

Bottom right red - power from BEC
Bottom right brown - ground from BEC

In the comments on the bottom of this article you can read why I made these changes.

 

To make things a bit less comprehencable, here is what it looks like when the board is connected to my hexacopter. =)

 

 

Just to be clear. Both version 1 and version 2 works just fine. In version 2, the signal grounds through the BEC and in version 1 the servo power grounds through the receiver/FC/APM.

##############################  VERSION 1 #################################

I didn't want to cut the servo leads so I had to come up with something to connect the power lead from the servo to my BEC whilst still having my ground and signal connected to my APM.

Here is the oh so professional wiring diagram of what I wanted.

On the left side I connect my servos, two of them, on the right side I connect the APM and on the bottom right my power source.

The signal and ground wire is quite simple, it's just a direct connection.

The middle wire (the positive, red one) is connected to the power pin. Although I ultimately ended up removing this pin and replaced it with a wire.

All done. Now all it needs is a bit of shrink tube.

 

In hindsight, I should have used a 3-pin on the right side too. This way I could have used an ordinary male-to-male lead. Well, you live and learn.

COMMENTS

iLoveRC on April 10, 2013
Thanks!! I needed this for my gimbal too. I have a question. BEC has two lines come out and one is Red (power) which I think that is what you are referring to but what about black? I don't need to ground it?
Log In to reply
Lubis on April 10, 2013
The wires from the servo are usually black-red-white or brown-red-yellow. The white/yellow needs to be connected to your flight controller or receiver.
The red needs to connect to your BEC.
As for the black/brown one, you could connect it to either your receiver/flight controllers ground pin or your BECs ground wire. Both the external BEC and the receiver/FC uses a common ground so it really doesn't matter.

In my article the servo ground wire is connected to my flight controller. I did it this way because the signal to the servo might be a bit "cleaner", not using the BEC ground. I might be wrong in this, but that is my guess.

Are you using the build-in BEC in your ESC or an external one?
Log In to reply
Lubis on April 10, 2013
If you connect the servo ground to your receiver/FC that means your BEC ground wire wont be used.
Log In to reply
iLoveRC on April 10, 2013
Got it and I am planning to use External BEC. I will build this weekend if possible and update here and yes, I will use 3pins for right side as well as based on your lesson learn :)

Thank you!
Log In to reply
Lubis on April 10, 2013
Message me if you run in to any problems.
Also, make sure not to connect your external BEC to the receiver to avoid causing any interference. The middle pin on the right side should not be connected to anything.
Log In to reply
iLoveRC on April 10, 2013
Thank you!
Log In to reply
Lubis on April 11, 2013
I have now updated my board and my article to the better way of grounding the servo.
Log In to reply
ddrueding on April 10, 2013
I was under the impression that the ground wire should remain associated with the power, and that the signal (white/yellow) could be routed independently?
Log In to reply
Lubis on April 10, 2013
You might be right, in general I would say you are. I figured I wanted to keep the power source separated as much as possible, but now that you say it, associating the ground with the power might do that more since none of the power, not even the ground, needs to be routed through the FC/receiver.
Log In to reply
Lubis on April 10, 2013
Come to think of it I think you are very right.
Although you usually say current moves from positive to negative, the actual electrons move from negative to positive, meaning that the electrons creating the difference in potential in the servo motor actually runs through the receiver.

I might have to rethink my wiring.
Log In to reply
FlyingSloth on April 10, 2013
Going on my experience with Computers and case fans, the signal wire doesn't need a ground. Many case fans you can get have a power lead that connects direct to the Power supply, and a second lead with one wire that goes to a fan header on the Motherboard. This wire allows the computer to read the fans RPM. Alternatively the fans can be connected direct to the Motherboard with all 3 wires, just like a servo and gets it's power through the motherboard.
Seems like the same theory to me. For safety I'd move the ground to the BEC ;)
Log In to reply
Lubis on April 11, 2013
I stand corrected.

I will change my wiring to this: http://www.norrbyteknik.se/stockholmrc/wp-content/uploads/2013/04/wpid-20130411_1249561.jpg

I will also revise the article
Log In to reply
Lubis on April 11, 2013
I have now updated my board and my article to the better way of grounding the servo.
Log In to reply
Jebnor on April 17, 2013
Now that you've updated it to V2, might I suggest a V3? I would hybrid both of your versions. That is, have one ground wire, along with the signal lines, from the receiver and the +/- from the separate BEC. The PWM signal for the servos are ground referenced on the receiver. Cross connecting the grounds into one common ground doesn't hurt anything.


In this image the GREEN section is the Signal grouping. The signal is transmitted from the the Rx to the servos (S1, S2,... Sn). The power for the servos is provided by the BEC. So the servos move with the POWER from there, indicated by the RED group. The critical point is that the signal doesn't mean anything if it doesn't have a ground reference, in this case a '-' from the Rx. So to have everything _always_ work properly, connect the all the '-' (ground) paths together and connect the '+' (Power) from the BEC to the Servos.
Log In to reply
Jebnor on April 17, 2013
Further, a normal signal is pulses of +5V(ish) on a 0V line, where the 0V is the '-' wire. These numbers are purely arbitrary. The idea is that there is a 5V difference. The '-' wire needs to be connected so when the pulse is transmitted it's at a relative 5V. If there is no '-' wire, the signal could be from -25V to -20V, which is just garbage to the servo.

To clarify, if there is no ground connection from the Receiver to the servos, the signal can be garbage.

Think of Power like air pressure differences. If the pressure between inside and outside is the same, the door opens no problem (pressure difference is 0.) But if one side has more pressure, then the door opens by itself and the wind blows in (+5 outside.) It's not a perfect analogy, but it works.
Log In to reply
Lubis on April 18, 2013
That's a good clarification, what you're talking about is electric potential. I usually make the analogy with water and yours it principally the same.
I made this diagram to further explain the common ground.



You are correct in that it might be a good idea to cross connect the ground, but for others I kinda wanted to explain why it works either way.
Although it might look like the incoming ground is separate from the outgoing, in reallity they are connected within the component. I tried visualize this in my diagram with the dotted lines "inside" the BEC or ESC.

In my diagram I have two BECs. Servo 1 is connected to illustrate the usual way of connecting a servo to a receiver. Servo 2 is connected the way I have explained in my article, with a stand-alone BEC.

We have already established that the signal needs a ground to work and this ground also need to be associated with both the Rx and the servo. On servo 1 this is easy to see, but on servo 2 it might not be as easy.
But taking a look on how the BEC actually looks inside it might be a bit clearer. I have tried drawing up a respresentation in my BECs, so let's trace the route of the signal.
1. The signal travels from the Rs signal lead to servo 2
2. From servo 2 the signal travels in the ground lead to BEC 2 and through BEC 2 to both the negative side on the BATT and BEC 1.
3. The signal then travels through the negative lead in BEC 1 to the negative lead in the Rx.

I hope this makes any sense.
Log In to reply
Jebnor on April 18, 2013
The ground lead needs to be connected AFTER the BECs. It is possible ( but unlikely) for the BEC to truly isolate the power planes. Without the ground connection from the Rx to Servo2 a reliable connection is not possible.
Log In to reply

You need to log-in to comment on articles.


External servo power source - revised