Differential ailerons question

[harveyweewax]

Is is possible to set up a differential control for ailerons using a pull-pull (closed loop) system? I know some aircraft, such as the Tiger Moth do it by switching to pushrods at the very end but is there a way to do it only using cables? Any input would be greatly appreciated.

Thanks,

Sven

[Sparky Paul]

The Tiger Moth cam is probably the only practical way to do that.
Pull-pulls rely on equal tension in both cables, so the control motions to each horn, upper and lower would have to be designed and attached to the horns at the appropriate locations.
Just offhand I can't "see" a way to do this.

[harveyweewax]

That's sort of what I thought. I've been racking my brains trying to figure out the geometry but I couldn't make it work. I was hoping that brighter minds than mine would have the answer.

sven

[Sparky Paul]

For the curious, the Tiger Moth has an extreme example of differential...
Equal deflections of the stick result in 1/4" -down- aileron with 3-3/4" -up- on the other side..
This is accomplished with the pull-pull cables rotating a cam set in the lower wing lower surface, with the pushrod to the aileron offset on the cam so rotation in the down direction results in almost no motion to the pushrod; the up direction pushing -a lot- on the rod.
The method to do this with an r/c servo would be similar..

[Flight Engineer]

Just rig both ailerons up a couple of degrees. now th down going aileron goes through zero. I used to fly the nimrod and its ailerons were rigged this way.

[Sparky Paul]

Differential is required on very slow planes such as the Tiger Moth because the down-going aileron adds a lot of drag to that side, pulling the nose -out- of the turn.. in some cases yawing the plane to an uncontrollable attitude.
On faster planes with their more efficient control systems, differential as extreme as the TM is seldom seen.
The Piper Cub for one could benefit by having differential. It's a notorious yawer, due to the impossibility of having a pull-pull system all by itself.
The Cub -needs- rudder assist to keep the nose following the turn.. Properly built models exhibit this effect.. :)

[harveyweewax]

How about this thought. Rig the upper control horn of the aileron to a spring. The lower control horn is connected by cable to the servo control arm. The spring pulls the aileron up while the servo pulls the aileron down. The servo travel can then be used to adjust the downward travel and, therefore the differential. Any flaws in this idea?

Sven

[Sparky Paul]

It has to work on both sides, remember..
You need a small amount of motion in the down direction, and a larger amount for up on the other at the same time.
Your scheme might require seperate servos.
I'm not a fan of springs on control surfaces.. stretching under load and snapping under load comes to mind. :(
Offsetting the pushrod on the servo wheel is the simplest.
Choosing DIFF in the computer transmitter also works.
I fly my SAE lifter planes with 100% differential using the computer radio. No down at all!

[harveyweewax]

I was thinking of two springs, one for each aileron. I agree with you about the springs. Springs strong enough not to stretch would require a pretty strong servo to pull against them. I believe you are right. Keep it simple. computer control or use the tiger moth setup.

sven

[BMatthews]

... or just use pushrods in the first place. Pull-pull has become some sort of stylish Holy Grail in the modeling world. For rudders and perhaps elevators it has it place. But for cases where differential is required just use pushrods.

[biber]

For differential travel just tilt the lever axes as shown in the attached iamnotyetabletooperatethegimpdrawingsoftwareproper ly sketch. ;)

It´s basically the same principle as in the pushrod solution. The only thing that must be taken care of is, that the pivoting point (axis) of the flap/aileron has to be exactly halfway and inline between the two points where the wires are attached to the levers (assuming the same for the servo levers, of course).

biber

[harveyweewax]

Biber,

Thanks! I'll give this a shot.

Sven

[Robert May]

Do you want to use 1 servo per aileron or 1 per pair?

Self tensioning is just a matter of crossing the wires, no springs needed.

[Sparky Paul]

A graphic solution...

[harveyweewax]

Robert:

Was hoping to use 1 servo per pair.

Sparky:

if I'm understanding your drawing correctly, All I would need to do is to make the bottom aileron control horn twice as long as the top. Is this correct?

Sven

[Robert May]

Using one servo per pair,

I would try equal length, horns top and bottom of each aileron.

Cross the wires so that the pulling wire becomes tight

set the diffential at the servo as you would for a rigid system. i.e set the servo arm at 80/260 degrees to the servo centre line.

I have not done this for real but it should give you what you want.

[Sparky Paul]

Robert:

Was hoping to use 1 servo per pair.

Sparky:

if I'm understanding your drawing correctly, All I would need to do is to make the bottom aileron control horn twice as long as the top. Is this correct?

Sven
.
Yes. to get 2:1 differential.
But with 2 servos and a mixing transmitter, selecting Flaperons is much easier. :)

[harveyweewax]

Thanks all!

I don't have a mixing transmitter. Can't afford one. The wife keeps me on an allowance :)

sven

[Robert May]

Sven,

All the married guys here are on an allowance. Have you not explained that a decent radio is actually mentioned in the small print of the marriage contract. (Forsaking all others, sub clause 2)

In summary; in return for your undying love and attention forever you the husband will be allowed to spend 10% of your annual salary on your hobby without any questions being asked.

If you were not made aware of this rule, you are allowed to calculate your hobby underspend for the period of your marriage and blow it on either, a weekend Stag party in Amsterdam with all your old buddies or on a decent computer radio. Your wife gets to choose

[fhhuber506771]

do some experiments with a servo, string and a mock-up of the control surface... I'm pretty certain that just altering the length of the upper control horn will be a mistake. maybe longer arm on one side of the servo AND longer horn AND crossed wires would do it... but I think that has problems too.

the offset connections (in rotation of the servo arm AND offestting the angles of the control horns at the aileron I think has the best chance. (of not needing the TM cam-pushrod)

What I keep coming up with is either equal movement up and down... or uneven cable movement.

********

Computer radios that will do the differential for you are comming down in price...

In 1978 a basic 4-Ch AM system was $500. (and a cheap servo was $50) Now a Futaba 7C system is under $350... and the value of the dollar is WAY down. This stuff is CHEAP now.

[biber]

Like huber, I´m pretty sure that "just altering the length of the upper control horn will be a mistake". But I also don´t see any special advantage in crossing wires.

So a simple mock-up will help you to find out wich approach may work and wich doesn´t.

biber

[Robert May]

Its standard practice on closed loop rudders to cross the wires, the "Pulling" wire tensions and the "pushing" wire slackens, removing any need for a tensioning sytem.

[Sparky Paul]

With the requirement for differential, and a seperate servo for each side, crossing the wires will not accomplish anything.
Lengthening the arm for the surface with the lower deflection of course will work!
The servo motion, i.e. amount of distance moved with deflection, providing the connections are the same distance from the center of the rotation of the servo will be the same amount of distance at the end of each cable.
The difference in arm length will then result in the desired proportional difference in the deflection of the surface.

[fhhuber506771]

But... you end up with a slack cable when you pull on the long arm, and you attempt to stretch BOTH cables when you pull on the short arm if both arms on the servo are the same length... (that won't work)

with the differential arm lengths on both the servo and the control surface... you end up back at even deflections.

With offsetting the angle of the control horns at the control surface but not at the servo you end up with differential but you also get the slack-stretch routine again.

Offsetting the angles at both ends goes back to even deflection.

Looks like the Tiger Moth "Cam" is the only way to do it with pull-pull and no computer.

[harveyweewax]

Sven,

All the married guys here are on an allowance. Have you not explained that a decent radio is actually mentioned in the small print of the marriage contract. (Forsaking all others, sub clause 2)

In summary; in return for your undying love and attention forever you the husband will be allowed to spend 10% of your annual salary on your hobby without any questions being asked.

If you were not made aware of this rule, you are allowed to calculate your hobby underspend for the period of your marriage and blow it on either, a weekend Stag party in Amsterdam with all your old buddies or on a decent computer radio. Your wife gets to choose

LOL I'll have to print this out and show the wife. I might be getting a new radio soon or maybe a real fun time in Las Vegas (Amsterdam is a little too far away for me). :D )

Sven

[harveyweewax]

I also tried a mock up using some bits of wood and some string. It appears that I get the same result as Huber. I don't observe any differential using unequal control arms. I do observe some differential using the methodology that Biber recommended. I haven't tried crossing the control wires.

Huber, did you incorperate a little bit of Ackerman in your setup (i.e. offsetting the control horns a small amount from the hinge line)? I didn't and I'm wondering if that would help.

sven

[Sparky Paul]

After playing with any number of pure cable setups, the much less strainful system is what I proposed originally, the Tiger Moth cable to the cam, pushrod to the aileron.
To simulate a cable system, the unused horn can have a fake pushrod similar to the real one going into the covering, with a tensioner inside to maintain a semblance of stiffness.
Redrew the original with nominal servo motion.
Note there's potential "overcenter" lockup situation on the pull side of the cam.
Much more than 35 degrees offset from servo neutral may move the pushrod past the servo center, and lock the thing up.
Use your ATV (or EPA) to control the endpoints.

[fhhuber506771]

offsetting the horn from the hinge pivot line is essentially the same as the one biber posted a pic of... possibly also with a bit of the different arm length effect.

I can't see that working any better than either of those... but it might. I didn't try combining those.

[biber]

But... you end up with a slack cable when you pull on the long arm, and you attempt to stretch BOTH cables when you pull on the short arm if both arms on the servo are the same length... (that won't work)Yes. (Still assuming a closed loop system as it was desired by the thread starter, since he wants to use only one servo vor both ailerons)

with the differential arm lengths on both the servo and the control surface... you end up back at even deflections.Exactly.

With offsetting the angle of the control horns at the control surface but not at the servo you end up with differential but you also get the slack-stretch routine again.Not exactly what´s happening. There won´t be slacking wires, but yes, the diff. will be there.

Offsetting the angles at both ends goes back to even deflection.Yes, if both are offset in the same direction. If the servo arm is offset in the opposit direction of the control horn, there will be even more diff.

Looks like the Tiger Moth "Cam" is the only way to do it with pull-pull and no computer.Definately no! Playing with a simple mockup is a rather quick way compared to any computerfiddling and you will get quite soon a good imagination of what is really happening there.

I think I´m going to do and post another little gimpdrawing soon, showing how it is possible to do the offset on both, servo and control.

biber

[biber]

That would be my approach to do a diff. with closed loop wires and only one central servo. The servo arms are effectively tilted/offset to the opposite direction of the control horn(s) to further increase the diff.

BTW, I still don´t get the crossing wires thing, could someone please post a pic of what exactely is mean by that and how it´s supposed to work.

biber

[biber]

Of course, the one wire method with a spring countering the single wire is also possible. For the Servo there virtually won´t exist much forces induced by the springs, as one spring more or less compensates the other one (at least for neutral aileron they exactly do, due to the symetry). Should be a workable solution, one of the lightest possible and without any play in it except for the servo´s gears part. I personally never had a good feeling with that spring thing. But that´s merely a mental problem, not reasonable. So, why not?...

biber
(trying not to hijack this thread :rolleyes: ...)

[Sparky Paul]

Biber, something like this...

[Sparky Paul]

Moving things around, the end result at the aileron doesn't change.

[Robert May]

Apologees, I can't get a drawing posted but try this;

What happens if you have a 40mm servo arm with 2 wires attached 19mm from a centre pivot point.

The servo horn is set at 100/280 degrees to the centre line.

Equal length aileron horns aligned with the hinge line crossed wires join the aileron horns at 15 mm centres.

Draw a line from the top of the servo horn to the lower aileron horn and vice versa. (crossed wires)

When the upper servo arm pulls through 10 degrees to 110 degrees the aileron is pulled down.

When the lower servo arm is allowed to pull back to 270 degrees the aileron raises more than the aileron was lowered.

If you complete the drawing for the oppostite wing you get differential movement for a pair of wings with one servo.

The "pushing" wire will slacken but that doesn't matter, it is not doing anything.

[Sparky Paul]

Looking thru the MAN book of antique airplane scale drawings, the aileron connections vary a lot.
Some use a simple quadrant with a single cable from the top or bottom out to the lower wing lower horn, then a follower to the upper wing lower horn, and then a cross-span connection between the upper wing upper horns.
On some pthers, the cable takeoff from the control rod is offset.. something like 45 degrees, (Fokker D-VII) with the cable running up to the ailerons in the upper wing.
There's no indication if differential aileron was the intent, or just an accidental result for riggings that would provide it.

[biber]

Hm, sparky, did you check (measure) if the length of each wire at the shown control/servo deflections is really constant? My ruler tells something different :rolleyes: . So, how did you actually figure out the control angle values you give in the drawing?

The "pushing" wire will slacken but that doesn't matter, it is not doing anything.I still don´t get it right pictured :o . But that last sentence made me wonder for what purpose there is a slackening '"pushing wire"[...], that is not doing anything'. That would mean the control having play in that situation wich is certainly not a desireable thing :confused: .

biber (more than ever confused)

[Sparky Paul]

Biber, I measure the length of each wire, using a constant angluar deflection for the servo horn.
Within the limits of the drawing, the wire lengths, servo to horn remain the same.
What I've found is the surface horn offset is really the only control for getting differential with cables.
I measure the chord of the arc of the servo arm deflection.
For these examples it's .34"
Then strike a arc at the surface horn to the last hole, and measure a chord of the same distance along that arc.
I rotate the surface and horn as one to that angle.
And draw in the cables.
Then add up the cable dimensions from arm to horn to see if anything changes a whole lot.
Nothing appears to change, any differences are usually too small for a modeller to be concerned with.
I did a drawing with Robert's crossed cables, and got an overlay of the uncrossed layout.

[harveyweewax]

Hm, sparky, did you check (measure) if the length of each wire at the shown control/servo deflections is really constant? My ruler tells something different :rolleyes: . So, how did you actually figure out the control angle values you give in the drawing?

I still don´t get it right pictured :o . But that last sentence made me wonder for what purpose there is a slackening '"pushing wire"[...], that is not doing anything'. That would mean the control having play in that situation wich is certainly not a desireable thing :confused: .

biber (more than ever confused)

As I understand it is somewhat desirable to have a slight slackening (with a strong emphasis on "slight") in the "pushing" wire. If there is a constant tension in the wires all the way through the control arm deflection the controls would bind up. Hence you would need some sort of Ackerman incorporated in the geometry.

For those following this thread and are not familiar with the term "Ackerman" here's a link that describes it:

http://members.cox.net/bdfelice/Ackerman/ackerman.htm

Sven (also confused and more than slightly dazed)

[Sparky Paul]

To make it as unconfusing as possible, I ungrouped the cables,.. straight shot from the servo arm to the surface horn..
Nothing changed.

[Robert May]

Excellent diagrams Paul, just one last request from me please.

If you ungroup the crossed cables and have the Aileron horns perpendicular to the centreline, A big "T", what happens to the aileron deflection when you increase the servo throw to 90 degrees as in a live situation.

The slackening of the pushing wire does not cause any control slop, as you have formed a triangle with 3 fixed sides and the pressue caused by the forward movement of the plane is what keeps the pulling wire tight.

[biber]

Now that I´ve read the Ackerman article, I have to admit that a little Ackerman will be ok in the very most cases. But that sentence quoted below does not right hit the point for all cases of ailerons."It also does not matter which side has the slack- as the airflow will always force the deflected surface toward the neutral position, the nature of the forces involves actually chooses which cable is taut and which is slack."Depending on the airfoil used on a wing (and the actual lift coefficient) , the aileron is not forced towards neutral position by the airflow in most cases.

Sparky, I´m glad you finally ungrouped the wires, because before that, there was some sort of Ackerman in it effectively, because the wires were not even close to beeing parallel when approaching the control arms. The point where the wires meat each other should at least be far enough away from the horns, so that at least the absolute angle of the wire is nearby constant for any moves of the horns. And if the wires are grouped, crossing them won´t have any effect at all.

Well, for a more or less serious scaled drawing I will have to use a pencil (CAD is not really in my repertoire yet), reinstall the scanner driver (we had a hard disc crash some weeks ago) and then I´ll get it posted.

biber (still not wanting to believe :p :D ;) :rolleyes: )

[fhhuber506771]

any condition where you have a direction where the "pushing" wire slackens means the opposing direction will need one (or both) wires to stretch. This indicates a pull-pull system that is NON-FUNCTIONAL... you're gonna break something

[Sparky Paul]

Excellent diagrams Paul, just one last request from me please.

If you ungroup the crossed cables and have the Aileron horns perpendicular to the centreline, A big "T", what happens to the aileron deflection when you increase the servo throw to 90 degrees as in a live situation.

The slackening of the pushing wire does not cause any control slop, as you have formed a triangle with 3 fixed sides and the pressue caused by the forward movement of the plane is what keeps the pulling wire tight.
.
As long as the pulling wire is up to holding the load, the system is functional.
There are some wing profiles where the "neutral float point" of the aileron will be other than streamlined, and the load in the pull wire diminished, but the load in the opposing wire will work against any system instability.
These unusual configurations are not likely to use pull-pull in the first place.
It is rarely used in models, because of the geometry problems.
A couple of real old timey planes before WWI had no cables connecting the ailerons to each other at all. There was a single cable to each from the control stick, attached to the lower side of the surface. The ailerons drooped straight down from the hinge, and floated up until the single cable stopped the upward motion. The cable then pulled the aileron down, with the cable to the other side slackening and that aileron floating up. There was still tension in that cable though. (Slackening doesn't mean going limp.)

[biber]

Of course diff is also possible by only offsetting the servoarms...

biber (packing the pencil back into case)

[Robert May]

Any chance of putting the measurements on all the wires, in all 3 drawings before you put the pencil away? Please

Robert

[biber]

The measures that are put on the wires of the bottom part of the drawing apply also to the two upper parts, as I basically took that measures with a circle from the bottom part to draw the upper ones.

biber

[Sparky Paul]

From way back when.. and a real installation seen here on E-Zone,,,
With the surface horn vertical, not slanted forward or back, choosing a point on the servo wheel off the center of the travel will give differential.

[Sparky Paul]

By circumstance our 18th Annual Pattern Meet was held this weekend.
Here's some vertical views of the rudders on the FAI winner, the Masters winner, and a plane that's been around for a number of years.
The FAI winner went to the trouble to get the horn connection ON the hinge line.
The Masters winner missed by 1/16" of an inch.
The older plane isn't nearly that precise, but all work well.

[Sparky Paul]

Excellent diagrams Paul, just one last request from me please.

If you ungroup the crossed cables and have the Aileron horns perpendicular to the centreline, A big "T", what happens to the aileron deflection when you increase the servo throw to 90 degrees as in a live situation.

The slackening of the pushing wire does not cause any control slop, as you have formed a triangle with 3 fixed sides and the pressue caused by the forward movement of the plane is what keeps the pulling wire tight.
.
If both the servo arm and the surface horn have the same dimension.. servo axis-to-pushrod, surface hinge line-to-pushrod, the surface deflection will be the servo deflection.
The amount of differential can be computed using the posted diagram above in #47.

[Sparky Paul]

Biber, grouping with its effect on Ackermann is inevitable on full-scales, with the horns typically taller than the wing is thick.
I don't see any really practical way to avoid it.

[harveyweewax]

I was playing around with this using a cad program. In each picture the cable lengths are constant. It appears to me that you have to have ackerman built into the geometry in order for this to work. Also, I think the angle between the control horn and the cable is important. I played around with crossing the wires and it wouldn't work. The "pushing" cable would of had to stretch. Please don't misunderstand this. In the real world I don't think that there is a problem with crossing the cables but in my drawing I think that crossing the cables made the angle between the cable and the control horn too acute and that bound up the system.

[harveyweewax]

here's the drawing with the cables stretched.

[Sparky Paul]

Harvey, crossing the cables is common on rudder pull-pulls, it lets the cable run from servo to horn be straighter.
My Eindecker has pull-pulls on the rudder and elevator. The rudder cables cross, the elevator cables don't.
In the PT-19, the cables run straight.
I would uss pull-pull on ailerons solely for scale. There's just too much fiddling around to get the geometry correct for a casual sports plane.

[fhhuber506771]

From way back when.. and a real installation seen here on E-Zone,,,
With the surface horn vertical, not slanted forward or back, choosing a point on the servo wheel off the center of the travel will give differential.

Yep.. that works for a PUSHROD.

When you go to pull-pull.... chosing the offset connections creates one direction that lets one line go slack... and the other direction trying to STRETCH both lines.

Build the little rig... connect it up with the offset pull-pull connections using a stretch resistant "cable". one direction you get one line slack (OK... we can deal with that) The other direction you may strip the servo gears when the control system binds up.

The Pattern rudder linkages which show matched ofset of the control horns froom lining up with the rudder hinge line... if you check they MIGHT have a line go slack as they move from centered... but its not to get differential. and you'd find that the "push" line goes slack in either direction, with the desired tension set at center.

If you rig the differential with the intentionally slacking line... then at neutral you have slack... which invites flutter. Neutral needs both lines to have tension.

[harveyweewax]

Harvey, crossing the cables is common on rudder pull-pulls, it lets the cable run from servo to horn be straighter..

I understand that. The point I was trying to make was that you need as straight cable run as possible regardless of whether you cross the cables or not.


I would uss pull-pull on ailerons solely for scale. There's just too much fiddling around to get the geometry correct for a casual sports plane.

Whole heartedly agree! :)

Sven

[biber]

Whole heartedly agree! :)

SvenSo do I!

biber

edit: Exception!: I´d definately also chose wires, when any excess mass is strictly forbidden.

[7car7]

Time to bring it back from the dead!

I've been doing some thought about this - Getting differential with a single servo, using Pull-Pull.

I made a model of strips of balsa, and string, and pins, and a building board.

Seems to work just fine.

Then I drew it up on a CAD program, and got the exact numbers.

There is a slight bit of stretch/sag going on, when deflected, but I think it's not enough to be a problem. I may use Kevlar/Spectra (no stretch) thread for the "Main Run" and use Upholstry Thread (slight stretch) for the "Idler" line.
Probably doesn't matter which one is used where, and you could probably use the stretchy kind in both.
This may not work for ALL planes, but for my basically slow flying biplanes, I think this will work.

Here's the drawing...

[fhhuber506771]

DeHaviland came up with the answer to differential and using pull-pull wires back in the 1930's...

The linkage for a full scale Tiger Moth's ailerons is interresting...

[Brandano]

...

[vintage1]

...deleted because..

When I was posting on the original pull pull cable thread, I did that about three times.. ;)

[MCarlton]

is there a way to do it only using cables?

There is an easy answer to that, which is that if you are using one servo per aileron, use a computer tx and set the differential that way :)

Otherwise, would it not work by just using a shorter horn on the upper surface of the aileron?

[fhhuber506771]

There is an easy answer to that, which is that if you are using one servo per aileron, use a computer tx and set the differential that way :)

Otherwise, would it not work by just using a shorter horn on the upper surface of the aileron?

No, the use of a short horn and a long horn won't work.
The result would be either tensioning or slacking cables with control movement... either of which is not desireable.

[Brandano]

I can only think of fairly complex ways to achieve this while keeping the cable lengths absolutely constant. Using pulleys on both the servo side and the control surface side, with the pivot offset from the center to compensate both sides. Difficult to tweak, and just too much work. Earlier on in the thread someone suggested using misaligned brackets on the servo and control surface sides. This will work, but it will still affect the cable stretch somewhat, though it will be much less than using different length control horns.