Paralleling Multiple RC Servos, with line buffers

Thread Starter

djsfantasi

Joined Apr 11, 2010
9,156
I want to drive several RC mini-servos in parallel with one control signal. The servos are similar to Pololu Sub-Micro Servo 3.7g (Generic)

I understand that I could drive two with a Y cable. But what if I want to drive 6? 8?

I am planning on connecting multiple buffers (SN74LVC1G34) in parallel with one servo driver output. The servo driver is the Adafruit 16-Channel 12-bit PWM/Servo Shield - I2C interface
Product ID: 1411).

The power supply is strong enough to drive multiple servos. It’s a 5V 20A supply.

But what about the signal? How much current is required? I’ve been searching for this information to design drive buffers, but have not had any luck.

The SN74LVC1G34 looks like each driver needs 1uA...
A) Will this work and B) what am I missing?

Thanks,
djsfantasi
 

Papabravo

Joined Feb 24, 2006
21,159
From the brief description it sounds dubious. I know you do animatronics so we are not talking about huge distances but some kind of diagram might be helpful. I think the LVC1G family is primarily designed for use on a single board which limits the length and capacitative drive capability to about 18" and 500pf. This is where the rubber meets the road.

Edit: with a +5VDC supply the outputs can source or sink 32 mA. What I dont' know are the distances from control signal to the part and from the output to the servo.
 
Last edited:

Deleted member 115935

Joined Dec 31, 1969
0
one caution, RC servos tend to be open loop, i.e. there is no feedback between you saying move this far and has it moved that far.
result, each servo will move a different amount for the same control signal,
I'd suggest you have some sort of break point in case one or more is a lot different to the others, so the link breaks rather than what you are moving.
 

Thread Starter

djsfantasi

Joined Apr 11, 2010
9,156
From the brief description it sounds dubious. I know you do animatronics so we are not talking about huge distances but some kind of diagram might be helpful. I think the LVC1G family is primarily designed for use on a single board which limits the length and capacitative drive capability to about 18" and 500pf. This is where the rubber meets the road.

Edit: with a +5VDC supply the outputs can source or sink 32 mA. What I dont' know are the distances from control signal to the part and from the output to the servo.
The distances are 1ft at a maximum. 3” from the control signal source to the driver. The driver is 1” away from the servo. The servo control signals will be daisy chained 1 ft between them. If necessary, I can use a buffer there as well.
 

Thread Starter

djsfantasi

Joined Apr 11, 2010
9,156
one caution, RC servos tend to be open loop, i.e. there is no feedback between you saying move this far and has it moved that far.
result, each servo will move a different amount for the same control signal,
I'd suggest you have some sort of break point in case one or more is a lot different to the others, so the link breaks rather than what you are moving.
The application does not require consistency. It only requires a direction and a distance within 10-20% of the commanded position.

Specifically, each servo will move a mouth relative to its last position. The distance again is coarse and dependent on an MP3 soundtrack. The multiple servos are for bass, tenor and soprano members of a choir.
 

jpanhalt

Joined Jan 18, 2008
11,087
Signal input current specification, if available, is hard to find. Newer chips will probably have less than the older M51660L (Mitsubishi).

I would take a representative servo and put an in-series big pot (or fixed resistors) and see when it drops out or becomes erratic. Be sure your resistors are clean. PIC comparators require 5 uA (20 max). I did find an input offset spec for the M51660L, which was a few nA.
 

MrSoftware

Joined Oct 29, 2013
2,188
Servos used for RC toys typically take a PWM input signal for control and have separate lines for power. I'm not sure what the input impedance is but I would guess it's pretty high so the current requirement is likely very small, meaning you can probably drive quite a few without anything special. If you've got a really good multi-meter, maybe try measuring the impedance between the signal line and ground. Also given that it's PWM, the solution may be as simple as just using a mosfet to drive the servos, unless your buffer solution is easier.

If you've got a handful of those servos laying around, why not put 6 or 8 in parallel and see if your driver will drive them all in parallel. Let them run overnight and see what happens.
 

Thread Starter

djsfantasi

Joined Apr 11, 2010
9,156
Servos used for RC toys typically take a PWM input signal for control and have separate lines for power. I'm not sure what the input impedance is but I would guess it's pretty high so the current requirement is likely very small, meaning you can probably drive quite a few without anything special. If you've got a really good multi-meter, maybe try measuring the impedance between the signal line and ground. Also given that it's PWM, the solution may be as simple as just using a mosfet to drive the servos, unless your buffer solution is easier.

If you've got a handful of those servos laying around, why not put 6 or 8 in parallel and see if your driver will drive them all in parallel. Let them run overnight and see what happens.
I don’t. Yet.
 

MrSoftware

Joined Oct 29, 2013
2,188
Another idea; if your multi-meter is RMS type with mA or uA capability, put it in-series with the control line and see how much current you measure. It won't be perfectly accurate but should give you a ballpark to work with.
 

Thread Starter

djsfantasi

Joined Apr 11, 2010
9,156
The distinction was less than clear from the original post. I trust that you know what you are doing.
Of course. This is but one more animatronic application. I’ve built several.

My first post explicitly stated the power available. I did omit “to drive the motors”. +/- connects to the servo on two wires.

The next paragraph asks about the current requirements for the control signal. The signal is sent to a third wire on the servo. In an RC servo, the signal is a 50Hz PWM signal that connects to a control board inside the RC servo. The width of the pulses controls the position, by comparing the width to an internally generated pulse, controlled by a pot in the drive train.

My original question was how to determine the current requirements of this control signal.
 

jpanhalt

Joined Jan 18, 2008
11,087
Rather than trying to measure the current directly (it will take a sensitive meter), measuring voltage drop across a large resistor (at least 100 k) with a scope might be more feasible.
 

Papabravo

Joined Feb 24, 2006
21,159
Of course. This is but one more animatronic application. I’ve built several.

My first post explicitly stated the power available. I did omit “to drive the motors”. +/- connects to the servo on two wires.

The next paragraph asks about the current requirements for the control signal. The signal is sent to a third wire on the servo. In an RC servo, the signal is a 50Hz PWM signal that connects to a control board inside the RC servo. The width of the pulses controls the position, by comparing the width to an internally generated pulse, controlled by a pot in the drive train.

My original question was how to determine the current requirements of this control signal.
Current requirements for a "receiver" are modest because the input impedance is usually fairly large. Normally "line drivers" are used when the required load looks like a capacitor. Given the distances I don't think this will be a problem. The test will be to breadboard the arrangement and confirm that you get the "expected" voltage levels without distortion.
 

Thread Starter

djsfantasi

Joined Apr 11, 2010
9,156
Current requirements for a "receiver" are modest because the input impedance is usually fairly large. Normally "line drivers" are used when the required load looks like a capacitor. Given the distances I don't think this will be a problem. The test will be to breadboard the arrangement and confirm that you get the "expected" voltage levels without distortion.
This is useful information. I’m concerned that without using a driver when daisy-chaining servos over 10-15 feet, the control signal will be too weak or too distorted.

This distance was not explicitly mentioned before because I thought it would not be a concern when using a daisy chain with line drivers between servos at several feet.
 

jpanhalt

Joined Jan 18, 2008
11,087
Six foot runs (12' wingspan) sailplanes are not unusual. All we did was twist the servo wire. I think 10 to 15 feet -- maybe with a ferrite bead -- may work. I'd give it a try.
 

Thread Starter

djsfantasi

Joined Apr 11, 2010
9,156
Six foot runs (12' wingspan) sailplanes are not unusual. All we did was twist the servo wire. I think 10 to 15 feet -- maybe with a ferrite bead -- may work. I'd give it a try.
Thanks. Can you be more specific. I’m assuming the “servo wire” is the control signal. And you would twist it with the servo power ground? Running the motor power alongside the twisted pair. Or are all three wires twisted.

Where would you place the ferrite beads?

My original question was how to control many servos with one control signal? With all the replies, I am not sure this question has been answered.

I’ve read that two servos are not a problem, but once you get to three and more, all of the servos cannot be controlled by a single control signal.

Thus, for simplicity’s sake, I was going to use a line driver for each servo (instead of grouping the servos in pairs).
 
Top