Mechanical Exponential

[FlyingW]

Is it possible through the use of linkage geometry, bellcranks or other means to achieve exponential controls - that is, without using the electronic feature of modern transmitters?

Thanks,

Paul

[JoeSantora]

Hi Paul;
The only way I know of to EASLY get expo without programing is with a one way deflection surface ie. flaps. You set up the flap horn clevis hole, the servo arm clivis hole and the servo shaft in a straight line. You get negligable movement to start, but it ramps up quickly to where the servo arm is at 90deg to the linkage. This works great for getting max travel from the flaps and gives the most torque and least play at neutral flap to minamize flutter. The down side is that the flap has the most leverage at 90deg deployment which makes it easyer to strip the servos when landing.
You can also get negative expo from the setup by aligning the linkages when the flaps are deployed 90deg.

Joe

[eflightray]

Exponential can be set at either the servo, (preferably with something like a disc output), or at the control horn by moving it back or forward from the hinge centerline.

Picture helps. It's all about angles.

[eflightray]

There was me thinking single servo for aileron, (my age is showing).

For two servo ailerons or for just one servo requiring expo, as Joe said, just angle the output arm on the servo.

[FlyingW]

Ok, Joe and eflightray,

Thanks for the input so far. I need to clarify my request to specify that I want negative exponential.

I can see about angling the horn on either the servo or the control surface itself - we used to do that to get more up than down throw on ailerons.

I also understand your interesting diagram using the geometry of the servo wheel; however, as Joe suggested earlier, I do not believe this approach can be used to deflect a control surface with negative expo in both directions.

The reason I am asking about this is because I was thinking about getting the radio gear from a micro RTF like a Vapor, and using it in another airframe (the "Nutball" - see large thread) that I know requires negative expo to tame it.

[Rodney]

Do you mean exponential or differential. If you mean more throw in one direction than the other, you want differential, usually more up than down. If you truly mean exponential, it is defined differently by different manufacturers; i.e. Futaba uses negative for the same effect that JR uses positive. The geometry shown in the above posts are for differential throw, not exponetial.

[Brandano]

I can think of a few ways to achieve this, but they are all mechanically complex and bound to fail when less expected

[FlyingW]

I want exponential - Futaba's negative style.

[BMatthews]

What eflightray showed is how to get differential. Exponential where you have a "soft" response around neutral and a hard response as the wheel or arm or horn moves further isn't as simple.

There is one way to get a form of exponential. And that is to go with a very short control horn on the surface and a long travel large diameter wheel or arm at the servo. The long arm at the servo will more closely simulate a linear operation and the travel at the contro horn as it passes the 25'ish degree mark will provide a much faster travel at the tail. Keep in mind that this relies on you wanting a large throw such as at least + or - 45 degrees or more. At lesser angles the amount of expo would be negligable.

I tried to imagine how to EASILY do this using a mechanical converter. There isn't a way that doesn't involve way too many arms, levers and pivots with connecting rods. All of which produce slop in the system.

I'm afraid this is one case where it's best to just lay down until the feeling goes away and then go pay for a mixing transmitter. Back in the day some of us, me included, used mechanical mixers for flap to elevator or V tails or elevons. It was a happy day when I was able to buy a mixing transmitter and say goodbye to the Rube Goldberg mixers. I can't imagine why you'd want to go back to that Dark Age.

[FlyingW]

All,

Given no simple mechanical solution, I concede to modern electronics.

Thanks,

Paul

[Allan S]

Hi Paul,

A rounded trailing edge on your control surface should add exponential. I think the Great Planes Dazzler had rounded ailerons for this reason.

Regards,
Allan

[BMatthews]

I've not heard of that one. However there's lots of writeups to indicate that big blunt rounded trailing edges are good flutter inducers.

However you reminded me of something I saw years ago. If the trailing edge of the stabilizer is left quite thick and blunt and the elevator is thin enough that there's quite a big step then the elevator will be riding in a thin blanket of turbulence on each side. When this happens the control response for small movements around neutral is reduced. Only when the deflection angle is larger does it first deflect the turbulence and then poke up or down out of this and then you have full control. As for how big the step needs to be and how wide the elevator don't ask me. It was supposed to be an idea that was used in the old Pattern class to allow small unnoticable corrections to the model but still let it have full response for larger travel angles.

[biber]

Rounded TEs and steps fron thick to thinner parts are not very predictable and generally bad ideas.
Rather have a small stepup in thickness from stabiliser to elevator and only gradual thinning slopes to avoid separations.
The TE can be thick, but must be sqare, not rounded off.
You can even thicken the most aft part of the TE a bit to get more predictable response characteristics.
In any case where you don't need 3D capability, dial in as small throws as possible and enjoy a precise response.

biber

[mnowell129]

This produces mechanical expo.
Move the pivot to the center for linear output.
Pivot as is will make less motion around the center, both directions.
Cable needs to be taught.

[Brandano]

Hmm, interesting setup, very smart and simple too. No need to use a tube, a cradle that allows the disk through will work at higher deflections too. The cable should be fixed to the disk on the side opposite to the pivot to avoid it slipping, and the slack in the wire could be compensated by a fairly strong spring on one end of the cable, without really increasing the load on the servo.

[mnowell129]

Hmm, interesting setup, very smart and simple too. No need to use a tube, a cradle that allows the disk through will work at higher deflections too. The cable should be fixed to the disk on the side opposite to the pivot to avoid it slipping, and the slack in the wire could be compensated by a fairly strong spring on one end of the cable, without really increasing the load on the servo.
All true, but the cradle is hard to make (split a tube?). The cable never really goes slack. If you make one and play with it for a few seconds this becomes more apparent.

[Brandano]

Just doubling the pushrod for a short length and glueing two attachment plates for the cables to both ends of the pushrod should work fine. On a very small plane the wire could be attached to the pushrod in two places using small horns. I realize the cable never "goes" slack, it's always the same length. However unless there's something keeping it taut it will probably stretch with time or just change length with different temperature or humidity.

[mnowell129]

Just doubling the pushrod for a short length and glueing two attachment plates for the cables to both ends of the pushrod should work fine. On a very small plane the wire could be attached to the pushrod in two places using small horns. I realize the cable never "goes" slack, it's always the same length. However unless there's something keeping it taut it will probably stretch with time or just change length with different temperature or humidity.
I see your point. You could probably use kevlar or spider wire as well. Don't know about the temperature/time stability of those.
I think once someone sees the basic idea there will be many clever solutions that I couldn't think of....

[Brandano]

Hmm, i have to admit I made a mistake in visualizing this thing. You don't really need a cradle, the disk will never go through the pushrod since by design the pushrod will always be tangent to the disk. The simplest miniaturized version of this would be a disk with a groove cut into it large enough to accommodate the pushrod (either made on a lathe or built from a 3 disk sandwich, with the wire attached on the middle of pushrod, looped through the groove and then attached to the end of the same rod. On a small system a small flex in the rod will keep the wire taut, and the groove will ensure it won't derail. the clevis on the control surface horn should keep the rod from twisting. This sort of device could also be a really simple and effective way to make slop-free mechanical mixing devices. Properly made this would have less slop than a plain clevis arrangement. Come to think of it, the same setup could be used in place of the control horn, allowing for pushrod clearance.

[KraftyOne]

A few mechanical "soft neutral" systems were tried back in the early days of r/c, when rotary o/p servos started to replace the early linear o/p stuff.

One type that comes to mind used a slotted servo arm with a sliding pin pushrod clevis controlled by a secondary arm pivoting opposite the servo arm, so that as the servo arm rotated away from neutral, the clevis pin moved up the slotted arm, away from the servo.

From memory it worked OK. Others tried various aerodynamic solutions, as has been mentioned, some seemed to work, some didn't, & some caused other problems. :D

EGB.

[mnowell129]

... since by design the pushrod will always be tangent to the disk.
yes, this is the key and how the non-linear behavior is obtained.

One other point. Using a tube that is not part of the pushrod allows you to drill a small hole in the tube for clearance on the grub screw of a wheel collar and then using the wheel collar to fix the pushrod in the tube, thus providing trim.

[FlyingW]

mnowell129,

I'm trying to understand your design. Is your pivot driven by the servo output?

I'm highly encouraged by this.

Thanks,

Paul

[Brandano]

Here is my simplified interpretation of mnowell's setup (again, I am not worthy... my original idea used at least 2 levers and a sliding cam!)
sorry for the crappy quality, I just jotted down a few solids in blender and added in a few details freehand. The red and black lines are actually the same wire, but I used different colors to higlight how they are laid out around the pulley. The cut in the pulley side and the set screw next to it allow the wire to be adjusted for trimming, while the slot for the servo screw allows the amount of exponential to be set. The overall travel has to be adjusted on the horn of the control surface, though. also, if the pushrod is allowed to move too much the wire could come out of the pulley, and, well, you can imagine the rest. The nice thing of this layout is that it can be scaled really well in any direction.

[mnowell129]

Paul,
The pivot is the servo screw or the center of the servo wheel. See the sketch in the previous post.
Brandano,
sketch is right on. Travel is adjustable, you just have to make a bigger or smaller wheel!
Seriously as an aside, if you mount in the linear position you still get expo due to the horn motion on the control surface.
And I can't take credit for this. Something like this was a commerical product back in the 80's, it was called a "line drive" or some such. It was marketed as a device for making a pushrod move linearly in a straight line and not have the typical side to side motion of a servo wheel.
I just figured out that by moving the center you could make mechanical expo out of it.

mickey

[BMatthews]

HA! I never thought of this method but it would work. And the output wheel doesn't need to be a wheel either. Even an D shaped arc segment with slightly rounded corners would work and the variety of size that would offer would allow for more or less expo.

Keeping the pushrod or tube from altering its angle due to flexing or other forces would be important. But still, this would work.

I lower my Tx antenna to you in respect and admiration..... :D

[RCAV8R13]

Gentlemen,
He said he was going to use the gear from a Vapor. Vapor servos are linear. He will already get the the negative expo he is looking for. When you convert linear to rotary motion the outcome is always exponential.
You are good to go.

Kip

[gerryndennis]

Hi All,

It's reasonably easy to get mechanical exponential from a rotary servo and a pull/pull system.

Arrange four pins/screws, or whatever, in a diamond pattern on a servo disc. Mount the disc with the long axis of the diamond pointing toward the control surface. Run the pull/pull cables from the control surface outside the two pins on the short axis and attach them to the long axis pin furtherest from the control surface (wish I could upload a diagram). During the first twenty degrees or so of servo travel (depends on the relative lengths of the diamond axes) the control surface is controlled by the short axis pins. With further servo movement the long axis pin closest to the control surface picks up one of the cables, while the other cable moves away from its short axis pin. The control surface is now being controlled by the long axis pins, and therefore has a greater deflection for the same servo movement.

To be fair this will only give dual rate not true exponential, however, if you replace the pins with an elliptical shape (plastic or ply perhaps) with a plate on top to keep the cables in place then you will get exponential motion at the control surface.

All seems a bit redundant with most radios having electronic expo, but there you go it can be done.

Wish I could say I thought of the idea but no, read it in a magazine from before the age of computer radios.

MNOWELL. I love the variation on the 'Line Drive', never would have thought of that.

[HX3D014]

can you not add an L shaped lever between the Servo and Control horn.

Connect the Servo to the bottom of the L and the Control to the Top of the L and have the L pivot around its corner.

Bryce.

Scratch that.

Brain fart :(