Reversing Servos - and how do servos work (you ask).

[Plummet]

If you look on the web about how to reverse a servo it shows that you should desolder and resolder wires from the circuit board. This is not all that easy. With small servos this is very fiddley, and to make it worse on the servo I have just been playing with, the wires at the circuit board are coated in gunge - presumably to stress-relieve the joint. Trying to solder with the gunge in place does not work well. Trying to clean off the gunge is hard.

The obvious solution - but I have not seen it suggested before - is to simply cut the wires to the motor (red and black in my case) and to resolder them, red to black, black to red. I then covered the join with a tiny bit of heat shrink. Then do the same thing with the end wires from the potentiometer. (In my case the three wires from the potentiometer were green, yellow and red.) That is, green to red and red to green.

The slider wire from the potentiometer (yellow) should not be changed. Also there was also another wire to the motor, connecting to its case. This should also be left untouched.

Cutting, stripping,soldering and insulating the wires is fiddley. It requires good specs, a steady hand, and a fine pointed soldering iron. It was a lot easier that working on the circuit board.

What follows is about why I needed to reverse the servo, and about how servos work. You can ignore it of course.

Plummet

In my humble opinion, it is a design fault on the Ripmax FunCub that has set me off on this one...

There are four servos in the wings of the FunCub. Two for the ailerons, and two for the flaps. There are cut-outs into which the servos are glued. The cutouts for both the aileron and the flap servos in one wing are mirror images of those in the other wing. This is fine for the ailerons, where mirroring the servo means that the direction is reversed. This means that a simple y-lead is all that is needed to get the correct aileron motion.

The design fault is that the flap servos are mirrored as well. We do not want the left flap going up when the right one comes down which is what happens if you just use one rx channel and a Y-lead. My wife's transmitter is short of channels, and anyway, I want to keep transmitter programming problems at a minimum. (Been there before. Keep it simple. Not casting aspersions at her, but as I tend to have to do the programming - it is me that makes the mistakes.)

So the solution is to either:-

Hack about the servo cutout in the wing so that the servo is reversed, or:-

Modify one servo so that the two flap servos move in opposite directions.

I have looked on the web, and reversing the servo should be fairly easy. For those who do not know about such things, the way in which analogue servos work is :-

A simple dc motor drives a chain of gears that drives the splined shaft on which you put the servo arm. Also connected to the splined shaft is a potentiometer. This is a resistor, with a moving contact so that the voltage on the contact varies as the contact moves from one end of the resistor to the other. Thus as the splined shaft moves, the contact moves, and the voltage on the contact varies.

The signal from the receiver is a series of pulses, but these are converted into a voltage that represents the desired position of the shaft.

If the servo is in the correct position the voltage from the receiver signal and from the potentiometer is the same.

However if the voltage from the potentiometer is different from the voltage from the receiver signal, this indicates that the receiver wants the servo to move. The motor is given current so that it turns until the two voltages are the same.

Would it make it clearer if I used some numbers. Suppose we are powering our servo with 5 volts. One side of the potentiometer might be connected to +5v, the other to 0v. The midpoint of it will be at 2.5v.

The pulses from the receiver vary in length from 100 ms (milliseconds) to 200 ms. For the servo to be centred the pulse length is 150 ms.

Inside the servo the 150 ms pulse is converted into a voltage of 2.5v. Suppose the servo is not centred. Say the signal from the potentiometer is only 1v. The servo compares the 2.5 and the 1v and turns the motor in the correct direction to reduce the voltage difference. When the servo reaches the centre the voltage difference dissappears, and the motor stops.

If the servo had been off centre in the opposite direction then the potentiometer signal might have been 4v. In this case the servo would have driven the motor in the opposite direction in order to centre it.

Of course in normal operation it is changes in the signal from the receiver that prompt the servo to drive the motor to correct the "error" between the signal and the servo position.

We want to reverse the servo. We want its motor to turn in the opposite direction. With its simple dc motor we do this by exchanging the two wires to it.

But this means that when the servo turns the motor it will drive it in the opposite direction than that needed to move the potentiometer to the correct position. What we need to do is to reverse the "direction" of the potentiometer as well. We need to connect the end that was at 5v to 0v, and connect the 5v to the other end instead.

[Alan Cantwell]

either alternate between futaba, or hi tec, cos they go differant ways, or buy a servo reverser lead, no-one manual reverses a servo these days

[Plummet]

In that case - I am a no-body

P

[Alan Cantwell]

when its so easy to avoid the tedium of doing a servo reverse yourself, and i have done them myself, it seems strange to not take the easy and inexpensive way, its up to you, i am only trying to help, the servo reverse leads work well, and takes little room up

[GONZO]

Plummet,

The description of how a servo works is for the old genuine 'analogue' servos. What we now call an analogue servo only started to be called such when 'digital' servos (those with a micro processor inside) arived on the modelling scene.

Below is a copy of a description of how, what we call an analogue servo, works that I posted on the RCReviews forum. IMHO it is close enough, although others may wish to expand/correct as they see fit.

".............you have a slight misconception of how, what we RC'ers call, 'analogue' servos work. Please excuse me if I am teaching my grandmother to suck eggs here;
The Rx decoder outputs, on the various channel o/p's of the Rx, pulses that can vary in width between 1 and 2 mili seconds with the centre being 1.5. The width of the pulse, which varies in an analogue way, is related to the Tx stick position for that channel. Known as Pulse Width Modulation. A 'frame' of information consists of the number of channel pulses plus a longer pulse to re-set the decoder. The frame repetition rate (update of information relating to stick position) is at 50c/s (frame rate timing 20 mili secs). Full title for this system Pulse Width Time Division Modulation. (please excuse the terminology as the passing years mean all this was many decades back for me)
Taking one channel as an example: a pulse of between 1 to 2 mili seconds (which varies in width in an analogue way) and whos width is updated at the frame rate of 50c/s (every 20 m/s) is feed to the input of a servo. This pulse triggers the servos internal pulse generator to generate a pulse of opposite polarity to the incoming pulse. The width of the internally generated pulse is determined by the position of the feed back pot which is linked directly to the servo mechanical o/p. These two pulses are then feed to a suming junction - both the same width but opposite polarity = no o/p. If one should be wider than the other then there will be an o/p the polarity of which depends on which pulse is wider. If the difference is small (there is a small variable degree of pulse stretching applied) then the signal to the o/p bridge driving the servo motor will be pulsed causing the correct polarity of pulsed DC being feed to the servo motor to drive the o/p to a postion that causes the feed back pot/internal pulse generator to match the width of the incoming pulse. If the difference is over a predetermined amount then the o/p to the bridge will cause the bridge o/p to be fully switched on with no pulsing. As the servo nears the position of matching both pulses the bridge o/p will go back to a pulsed o/p allowing the servo to slow down and reach the desired position without overshoot. Along with the deadband, if correctly designed for the mechanical characteristics of the servo, this gives a servo without overshoot(bounce) and hunting that moves rapidly to the desired position. There are no micro processors in 'analogue' servos.
Therefore IMHO what we describe as an 'analogue' servo would more correctly be described as an 'analogue digital hybrid' "

HTH

[GONZO]

**LINK**

Thought I'd add a link to the thread on the other forum (mods permiiting) for completness

[Plummet]

Posted by Plummet on 22/07/2012 18:36:45:

In that case - I am a no-body

P

But seriously...

I have always used two 'matched' servos when they were driving two 'matched' control surfaces.

Are servo manufacturers well enough behaved that I can assume that the same signal will generate the same servo movement angle?

P

[Peter Beeney]

Plummet, - From experience some servos are very easy to to reverse, others are virtually impossible to even take apart in a sensible manner, let alone do any work on them. For a start, sometimes the motor is very much part of the circuit board and the pot driver connections are tiny bus bars that are part of the assembled components. I’ve not had very much experience with the small ones, but I can well imagine that they will be be a bit fiddly. As you say, if you do much of this I guess you need the right kit, such as miniature cutters and needle nosed pliers, plus the exact soldering iron and some very fine gauge solder etc. I have to say, cutting and changing the wires has been done before, I also look for the easiest way of doing anything! But I still like to have a little dabble occasionally, I’d reverse a servo if necessary and if it’s possible.

With the greatest respect, your description of the servo action does tend to be a little bit vague in parts. Also the pulse length will be 1 - 2 milliseconds long, centred on 1.5 milliseconds, repeated very approximately on a varying 20 milliseconds, it’s not hundreds of milliseconds. But in general you are overall describing what happens. How it actually does it is the slightly different bit. I think it would need a slightly more in depth description to accurately be able to get a clear picture of the whole operation.

And again with respect, I think there has been some possible issues with inline servo reversers and 2.4, has there not?

GONZO - I’d agree with you, this is all digital, it’s just described as analogue to make the difference. However, I’d considerer that the transmitter actually transmits a series of very short pulses, and this is known as Pulse Position Modulation, or PPM.

Just to muddy the waters a little bit more, so to speak……

PB

[GONZO]

Was called pulse width way back(70's) but got changed to pulse position with time. The fact that the Tx transmits a series of very short pulses is implied in my description by the desciption of the Rx o/p's. The description Plummet gave was for the v/old analogue servos that went out of use decades ago. I think Stavely (for those of you who can remember back that far) may have been the last to persever with these type of servos, IIRC.

[Former Member]

[This posting has been removed]

[Doug Ireland]

Me either! Buy a reversing lead.

[PatMc]

Don't use a Y lead. Use separete channels for each flap & programe the Tx to give opposite movement from the sevos. It's the arrangement used with most 4 servo glider wings.

Edited By PatMc on 22/07/2012 21:08:16

[John Olsen 1]

Some digital servos can be programmed. On the ones I have, that includes the ability to reverse the servo itself. Of course such servos are more expensive plus you need to get the programming interface as well, but it could be handy if you are running short of channels or don't have a suitable TX.

John

[John Sheehy]

I'm a pre-digital servo user but before computer radios we simply reversed the red and black leads by lifting the flap on each brass contact of the plug, pulling them out and sticking them in the other hole.

[Martin Harris - Moderator]

Pre-digital or pre-digital proportional?

[John Sheehy]

Pre those kind of questions.

[Ralph Day]

I tried the changing polarity on a servo for my Fun Cub. I'm waiting for a replacement unit. The magic smoke escaped. Melted the casing and a little of the foam. Live and learn.

Ralph

[Ian Jones]

Be very carefull with servo reversers because:

• When I tried to use one the range of movement still didn't match
• the servos jittered
• the reveresed servo centred wherever it felt like each time the Rx was switched on
• the failsafe positon was random and couldn't be set

Servos were matched sets of Futaba and Hitec - that is, no mixed makes at the time of testing.

My fix has been mentioned - channel mixing via the TX but make sure that the slave channel repsonds to any trimming applied to the master channel.

[Martin Harris - Moderator]

Posted by John Sheehy on 13/08/2012 01:07:50:

Pre those kind of questions.

For the reason I asked what was meant to be a serious question, see Ralph's post...

All servos that I know of back to the advent of digital proportional PPM systems in the late 60s/early 70s will protest with a puff of expensive smoke if you simply reverse the polarity of the red and black wires!

You need to reverse the motor connections and the 2 outside feedback pot wires internally to reverse a servo.

Some non-proportional "servos" used typically on single channel systems would be reversible from outside and perhaps there were some unusual proportional systems around in the early days which I'm unfamiliar with but I wouldn't advise anyone to do what you mentioned with one of today's servos!

I've assumed that by "digital servos" you're referring to modern non-analogue servos with digital amplifiers?

Edited By Martin Harris on 13/08/2012 11:31:45

[Mark Powell 2]

I use Multiplex radio. You can assign any number of servos you want, each on its own channel, and any channel you want. -

CH1 Elevator

CH2 Aileron

CH3 Throttle

CH4 Landing Gear

CH5 Elevator

CH6 Flap

CH7 Landing gear

Any channel can be reversed, and have everything else changed. So two flap channels working together or in opposition. No external reversers or Y leads needed.

All channels can have a 'curve', exactly like a helicopter pitch curve. You will see I have assigned two channels to 'Landing gear'. Use differentnt curves, and servo slow on these two channels and you have a door sequencer, and don't have to buy one. Add another 'Landing Gear channel and you could even make the nose door work differently to the main doors, etc.

All the mixers have up to 5 inputs, so each Multiplex mixer can replace five mixers on other radios, and they can be 'daisychained' too..

Here endeth the lesson on why I use Multiplex!

PS: 'Landing gear' always amuses me, Presumably the Germans (and the Americans) use something else to take off with.

 

Edited By Mark Powell 2 on 13/08/2012 12:07:03

Edited By Mark Powell 2 on 13/08/2012 12:14:36

[Frank Skilbeck]

Posted by Mark Powell 2 on 13/08/2012 12:03:39:

I use Multiplex radio. You can assign any number of servos you want, each on its own channel, and any channel you want. -

Not on the Cockpit Sx though, just the Royal Pro range and above.

But quite a few high end radios allow you to select the channels you want to use but it's always interesting to see how many fliers with high end radios still use Y leads because they find the programming too daunting.

[Andy Green]

Going back to the OP about reversing servos, back in the day, I had an old Acoms 35mz am set, and I fitted tiny plugs on the wires from the pots in the tx, and simply swapped them there to reverse a servo.

Andy

[Mark Powell 2]

Frank, you can do it on the SX, but only as far as coupling two servos, and you can do that twice, but of course you only have seven channels in total to use. But it is enough to overcome the problem for flaps, as in aileron, elevator, throttle, rudder, flap, flap, and a spare channel.

Don't buy the matt black one. It get vey hot lying around on the field on a hot sunny day. I chose a black one, in addition to the Royal,. and found that out.