Circular bellcrank

Over on the SH forum there was a discussion going on about the importance of lowweight models. On a sidetrack the subject of the advantage of a 4'' bellcrank over a 3'' bellcrank was noted giving more leverage in sharp manouvres/high wind conditions (no problems with the Netzeband wall). That made me think if it would be usefull to use a -partially- circular bellcrank so that higher leverage is maintained even at full rudder deflection. It would not be very difficult to make, be only slighly heavier. Offcourse it would only be usefull if and when we actually run out of leverage with the current 4'' bellcranks and that I don't know. So the question is, why don't we use them, has anyone tried?

Rgds, Bram

I used a 4 inch in a plane for 7 years. Yeah I crashed it.
Greg L. Bahrman

[photo not recovered: 32927.jpg]


[photo not recovered: 32928.jpg]

Attachment #1, (jpg file)
Attachment #2, (jpg file)

OK, but how did it compare to a standard bellcrank? Ofcourse the controls will be a little faster especially with greater deflections. Did you get used to it, was it an improvement in controlability, did it keep you away from the infames N-wall? Or are modern ships with large stab/elevators and longer moments inherently more capable to stay within the envelope. I know Henk de Jong's latest stunter has a max deflection of +/-25 degr, that reduces loads considerably.

Rgds, Bram

The circular belcrank was not done to address the N-wall as you put it. As I understand the N-wall, it is purely the amount of torque required to move or deflect the controls. A circular belcrank is not much different from a normal one in that respect.

The size, or diameter of the belcrank, and height of the control horns is what determines the mechanical leverage and force required to move the controls.

A contributing factor in the modern large control systems, which is fairly crucial, is the more aft CGs developed by Paul Wlaker. Given a further aft CG, center of lift is closer the the center of pressure on the wing, in theory requires less force along the lateral axis to pitch the nose. Less force, less control deflection, less drag all = less effort at the handle. A larger control system helps reduce the forces even further to the point that finger tip control can be easily had and you never approach the N-wall. Which, as a review, is not enough leverage at the handle to aerodynamically deflect the controls enough to maneuver the plane. You can ocassionaly see this in an hourglass or verticle 8 when you see the plane literally hanging by only one of the lines, while the other is slack.

So, in Ted's case, the fact that it was 4" helped, but wasn't the primary reason for the experiment.

David,

I was thinking that once you start turning a standard bellcrank the moment-arm decreases. I did some calculation but a 10 cm (4") bellcrank is composed of two arms of 5 cm (2") each which efectively work as 2" arms around the neutral position but once deflected these arms 'shrink'. At 10degr=4,9 cm, 20degr=4,7 cm, 30degr=4,3 cm, 40degr=3,8 cm. So after rotating 40degr your 4" bellcrank has effectively become a 3" bellcrank with the associated decrease in force that can be applied at a given line tension.
Again, as you describe, current setups with large stab/elevators, long moment arms and rearward CG's only need limited rudder deflection to turn and therefore there may not be enough gain to go through the trouble of making a more complicated circular bellcrank. I agree with your other post that much more care must be taken to make an adequately large bearing surface to prevent the circular bellcrank from tilting and developing too much slop but that could be solved by making an aluminium bellcrank bushed by a brass/brons centerpiece with flange.

Rgds, Bram

>
>A contributing factor in the modern large control systems,
>which is fairly crucial, is the more aft CGs developed by Paul
>Wlaker (who ever that is). Given a further aft CG, center of lift is closer the
>the center of pressure on the wing, in theory requires less
>force along the lateral axis to pitch the nose. Less force,
>less control deflection, less drag all = less effort at the
>handle. A larger control system helps reduce the forces even
>further to the point that finger tip control can be easily had
>and you never approach the N-wall. Which, as a review, is not
>enough leverage at the handle to aerodynamically deflect the
>controls enough to maneuver the plane. You can ocassionaly see
>this in an hourglass or verticle 8 when you see the plane
>literally hanging by only one of the lines, while the other is
>slack.

There you go David, spilling the beans again. Now Sparky is going to read this and understand what the deal is, and why he crashed. Oh well, we try to keep secrets, but they always seem to slip out!!

>There you go David, spilling the beans again. Now Sparky is
>going to read this and understand what the deal is, and why he
>crashed. Oh well, we try to keep secrets, but they always seem
>to slip out!!

P Walker


I already knew what caused the crash. But ill keeep it in mind.

Ted flew a lot of airplanes with circular bellcranks (quadrants), and David still flies airplanes with circular bellcranks. I know the guys who make the airplanes.

Howard Rush
Bellevue, WA

>The circular belcrank was not done to address the N-wall as
>you put it. As I understand the N-wall, it is purely the
>amount of torque required to move or deflect the controls. A
>circular belcrank is not much different from a normal one in
>that respect.
>

Oh, I think it does help. As the conventional crank deflects, the torque available for given amount of tension goes down as a cosine of the angle. A circular bellcrank maintains a constant ratio of torque to line tension. So it does help with the Netzeband wall, a lot in some cases, depending on the other ratios.

What is tricky about it is that we are using a "straight" handle with a circular crank. With a conventional 4" bellcrank and a 4" handle spacing, the bellcrank angle changes 1:1 with the handle angle. With a circular bellcrank, the motion of the bellcrank falls off with respect to the handle angle as the deflection gets larger. The net effect is a "reverse" exponential effect (really, just another cosine effect). With a 4" circular crank and a 4" handle spacing, around neutral it's the same as a 4" straight crank.

Brett

Another mechanism to do this is Igor's flap control unit. It has the advantage of being able to change the ratio between flaps and elevator as a function of control position.

Howard Rush
Bellevue, WA

>Another mechanism to do this is Igor's flap control unit. It
>has the advantage of being able to change the ratio between
>flaps and elevator as a function of control position.
>

I thought that was pretty slick. The mechanism I came up with for a similar purpose (i.e. what amounted to a deadband) was incredibly crude.

Brett

>I used a 4 inch in a plane for 7 years. Yeah I crashed it.
>Greg L. Bahrman
>
>[photo not recovered: 32927.jpg]
>
>
>[photo not recovered: 32928.jpg]
>
I am getting ready to start construction of Don Hutchison's Stearman, the full fuse one from FM plan# CD122. It uses a circular bellcrank very similar to the ones in your picture, but it is a homemade version. Are the ones you are showing here commercially available?


Allan Perret

Ted Fancher has extensive expeience with circular belcranks. I think he liked them. Unfortunately, his particular belcrank wore out the pivot bushing relatively soon.

When that happens, the belcrank starts to wobble. This introduces a lot of slop in the control movement and shows up as hunting and difficult to exit corners accurately.

If the belcrank bushing was made of durable enough material, it should work very well. It is a different feel. The main idea was to try to keep a more linear control movement for a given handle deflection.

Call or E-mail Ted. I'm sure he would be glad to talk about it. Ted?

How do you keep the leadout wire in the groove? I would think it would tend to come out.

You could easily make up a Teamrace style housing in which the belcrank turns with minimal spacing between the bellcrank and the housing so the cable can not get out of the groove.

Rgds, Bram

I used a 3" circular crank in my scratchbuilt Travelaire Mystery Ship. It was ideal, as the leadouts came out of small holes in the fuse, as with the circular inline travel of leadouts I didn't have to 'oblong' the holes.
(I found out later that John Brodak had the oblong eyelets, so it would have been more 'unique' with them, and a conventional bell crank!)
But as for flying, I tried circulars on some stunt profiles,and they felt like bellcranks. (dem R ducks?)

Here is a version used on an F2C racing plane. The aluminum bellcrank is housed in a very close fitting circular pocket built into the wing. This is made from ~1/16" plywood top and bottom and basswood(?) formed around the perimeter. In this photo, the bottom plywood plate has been removed for repairs. The shaft is machined into the bellcrank and rides on steel inserts glued to the plywood on both sides. The edges where the sector is removed from the bellcrank make contact with a wooden stop to limit travel. (Before you close up your plane, think what would happen if you had a line failure in pull test. Will the controls survive?) The wire is keyed to the bellcrank with a small diameter crimped tube. For scaling purposes, the fuselage is about 2" wide. The cable is .015" and the leadouts are .027 solids with wrapped and soldered connections. The pushrod is balsa, and is wrapped and coated. The system is subjected to a little over 19 lbf load during level flight.

Dave "McSlow" Hull
LA_Flyer

[photo not recovered: 32948.jpg]

Attachment #1, (jpg file)

Adding to Greg's reply...

Many years ago, metal circular bellcranks were available. Dynamic Model Products? ...consisted of three 'disks' riveted or spot-welded together. Middle one enough smaller that the leadout cable rode a groove far enough inside the outer rim to allow several small wire "twist-ties" to keep it in the groove, as on the Greg's white wheels.

Not really sure I remember how the leadout cable was kept from slipping, but it may have been like the friction keeper on EZ-Just handles, or a dip below a pin or bolt 180° away from the leadouts, at neutral.

By the way, the leverage/mechanical advantage picture is as follows:

Conventional TEE-layout bellcrank keeps the same leverage at all angles. The "torque radius" (at right angles to the forces involved - the pull down the leadouts, and pushrod resistance due to surface airloads) - shortens as the physical radius times Cosine of the rotation angle, for the leadout arms AND the pushrod output arm. (It's one piece; all arms turn the same angle.)

Offset TEE-layout bellcranks are very similar, but there is some bias due to the "droop angle" of, say, the SIG self-centering piece.

Greg's illustrated circular bellcranks give a NON-uniform response, moreso if you turn them far from neutral. The pushrod output does NOT maintain a constant radius in regard to the pushrod load it meets. It foreshortens, just like the standard TEE item, BUT the leadout radii remain constant.

---While the leadout radius remains the same, the pushrod output radius (perpendicular to the force required) grows smaller. THAT mechanical advantage ratio increases. BUT, if the wheel turns to a large angle, say past 45°, the pushrod travel per bellcrank degree decreases more rapidly.

---Yes, that increases the mechanical advantage ratio, but since the control surfaces also don't move as far per bellcrank degree, how useful is that?

---If the control surfaces were moved by pull-pull cables to drums instead of lever-shaped horns, BOTH the leverage and the rate of motion would remain the same. We'd feel the airload force feedback directly. A mite tricky to install? Baffling to make adjustable after installed? Anyone game to try it?

If we all do this right, we don't turn our bellcranks more than about 45° each way. That keeps things more nearly "linear," although pushrod travel per degree IS actually decreasing as angle off neutral increases. The rate of this decay is less between +45° and -45° than beyond those angles.

Think of the Sine and Cosine graph curves... quite linear over this range.

If all the pushrod arms are the same radius (output from bellcrank, flap horn, elevator horn) their mechanical advantage remains the same at all angles. Force required due to airloads still rises, and the larger angle the control surfaces are turned to, the greater the airload resistance... And, with the reduced (foreshortened) arm radii, the airload force changes probably feel more noticeable...

We seldom use the same arm radii at bellcrank pushrod output, flap horn and elevator horn (and I don't recommend it, necessarily)... However, around neutral, non-linearity is not very great, whatever the radius lengths. If it is dependable, and consistent both ways from neutral, it should be just as useful today as when Jim Walker filed for his original patent.

>Adding to Greg's reply...
>
>Many years ago, metal circular bellcranks were available.
>Dynamic Model Products? ...consisted of three 'disks' riveted
>or spot-welded together. Middle one enough smaller that the
>leadout cable rode a groove far enough inside the outer rim to
>allow several small wire "twist-ties" to keep it in
>the groove, as on the Greg's white wheels.
>
>Not really sure I remember how the leadout cable was kept from
>slipping, but it may have been like the friction keeper on
>EZ-Just handles, or a dip below a pin or bolt 180° away from
>the leadouts, at neutral.
>
>By the way, the leverage/mechanical advantage picture is as
>follows:
>
>Conventional TEE-layout bellcrank keeps the same
>leverage at all angles. The "torque radius" (at
>right angles to the forces involved - the pull down the
>leadouts, and pushrod resistance due to surface airloads) -
>shortens as the physical radius times Cosine of the
>rotation angle, for the leadout arms AND the pushrod output
>arm. (It's one piece; all arms turn the same angle.)
>
>Offset TEE-layout bellcranks are very similar, but there is
>some bias due to the "droop angle" of, say, the SIG
>self-centering piece.
>
>Greg's illustrated circular bellcranks give a NON-uniform
>response, moreso if you turn them far from neutral. The
>pushrod output does NOT maintain a constant radius in regard
>to the pushrod load it meets. It foreshortens, just like the
>standard TEE item, BUT the leadout radii remain constant.
>
>---While the leadout radius remains the same, the pushrod
>output radius (perpendicular to the force required) grows
>smaller. THAT mechanical advantage ratio increases. BUT, if
>the wheel turns to a large angle, say past 45°, the pushrod
>travel per bellcrank degree decreases more rapidly.
>
>---Yes, that increases the mechanical advantage ratio, but
>since the control surfaces also don't move as far per
>bellcrank degree, how useful is that?
>
>---If the control surfaces were moved by pull-pull cables to
>drums instead of lever-shaped horns, BOTH the leverage and the
>rate of motion would remain the same. We'd feel the airload
>force feedback directly. A mite tricky to install? Baffling to
>make adjustable after installed? Anyone game to try it?
>
>If we all do this right, we don't turn our bellcranks more
>than about 45° each way. That keeps things more nearly
>"linear," although pushrod travel per degree IS
>actually decreasing as angle off neutral increases. The rate
>of this decay is less between +45° and -45° than beyond those
>angles.
>
>Think of the Sine and Cosine graph curves... quite linear over
>this range.
>
>If all the pushrod arms are the same radius (output from
>bellcrank, flap horn, elevator horn) their mechanical
>advantage remains the same at all angles. Force required due
>to airloads still rises, and the larger angle the control
>surfaces are turned to, the greater the airload resistance...
>And, with the reduced (foreshortened) arm radii, the airload
>force changes probably feel more noticeable...
>
>We seldom use the same arm radii at bellcrank pushrod output,
>flap horn and elevator horn (and I don't recommend it,
>necessarily)... However, around neutral, non-linearity is not
>very great, whatever the radius lengths. If it is dependable,
>and consistent both ways from neutral, it should be just as
>useful today as when Jim Walker filed for his original
>patent.

I'm game. I been thinking on this for several years now. Have a design in my head, just haven't gotten around to executing it yet. Are you saying that Jim Walker already has a patent on a circular bellcrank system?


Allan Perret

Allan,

No, the Walker patent drawings showed something very similar to the 'standard' bellcrank used for so long. Leadout connections in line with the pivot bearing, pushrod connection outboard to keep the pushrod out of the way as the bellcrank turned. The drawings appear in Charles Mackey's great book on the origins of CL...

Al Rabe wrote a few important articles (in the 1970's, Al?) about the effects of gyroscopic precession from the prop's rotation when the model turns sharply in pitch. A solution that's been popular and successful since has been to attach the pushrod inboard of the pivot.

The 'standard' precession picture (i.e., CCW rotation engine, CCW upright flight) brings a tendency for the nose to yaw out on inside-turning pitch change (UP input), and in on outside-turning pitching (DOWN input.)

The 'standard' bellcrank layout, described above, is traditionally used with the elevator horn below the hinge. That puts the UP line aft of the pivot. In level flight, the pull force is divided evenly between the front and back leadouts. Pull is pull: the total force doesn't change much in the short length of time needed for a sharp corner. (Unless there's something seriously out of trim - e.g., hinging?)

When control force is applied, pull on the loaded line has to be greater, to move the surface out of "trail" and into the airstream. The mid-point of total pull force shifts according to the force needed. With the standard layout, UP input aims the mid-point a bit further aft of the CG than it was in balanced-pull level flight. That adds a tendency for the model to yaw nose-out.

Gyro precession also adds a tendency to yaw nose-out on sharp UP input moves. (The mirror image of this applies to DOWN inputs: precession yaw nose-in tendency; pull mid-point shift forward adds a nose-in tendency.)

So, Al wrote to the effect that flipping the bellcrank over made these two tendencies act opposite to each other. That had potential to reduce the total yaw tendency on sharp cornering. Whether we can completely cancel one with the other is theoretically possible - say by "designing" the fore-aft leadout spacing to aim the mid-point of maneuvering pull the right distance to cancel the worst case gyro precession yaw. Theoretically possible, but there's a bunch of factors involved about which we know too little to guarantee it would work.

Apparently, many of us believe that it helps, enough to use the flipped bellcrank idea.

Some side thoughts: If we could change the 'standard-traditional' layout a bit, we would not need to have the pushrod and leadouts on the same side of the pivot.

--- Reverse rotation, of either the engine or the upright flight direction. Either one reverses the 'sense' of the precession yaw tendency.**

--- Put the elevator horn above the hinge (flap horn, of course, would go below). That puts the UP line forward, giving the opposing load shift direction, with pushrod outboard. (Maybe the scoop under a P-51 fuselage could conceal an underside flap horn?)

**I have used a CW rotation engine setup: Fox 35, factory LH shaft, All American, Sr., flown CCW with traditional bellcrank system. Most noticeable effect was need for no outboard tip weight (and it's a wildly non-symmetrical wing.) Seems torque reaction at the motor mounts carried the weight of the heavier inboard wing, AND balanced the line weight well! Worst aspect was learning how to prop it for starting. With wheels on the ground, two out of three times I tended to punch the pavement. Ouch...

Lou mentioned the Sig self centering belcrank. I use that style of belcrank, but I install it the other way, so that leverage increases for the first 15 deg of movement (from .866 to 1, the cosine function), then decreases during the 15-30 movement (from 1 to .866). After 30 deg, the cosine drops off from .866. But for most of the movement, leverage is maintained above .866.
A T belcrank operates between 1 at 0 deg deflection to 0.707 at 45 deg deflection.
As for flying, these belcranks give a more positive and instantanious response.
Happy Landings

>Over on the SH forum there was a discussion going on about
>the importance of lowweight models. On a sidetrack the subject
>of the advantage of a 4'' bellcrank over a 3'' bellcrank was
>noted giving more leverage in sharp manouvres/high wind
>conditions (no problems with the Netzeband wall). That made me
>think if it would be usefull to use a -partially- circular
>bellcrank so that higher leverage is maintained even at full
>rudder deflection. It would not be very difficult to make, be
>only slighly heavier. Offcourse it would only be usefull if
>and when we actually run out of leverage with the current 4''
>bellcranks and that I don't know. So the question is, why
>don't we use them, has anyone tried?
>
>Rgds, Bram

Contact Randy Smith he has the New 4 1/2 inch round cast iron bell crank for sale.

Re: The round bellcrank in the Stearman model. I did that so the leadouts could exit the fuselage through a tiny hole rather than a big ugly slot. In flight I don't notice any difference from the "normal" bellcrank shape.
Don

>Re: The round bellcrank in the Stearman model. I did that so
>the leadouts could exit the fuselage through a tiny hole
>rather than a big ugly slot.

I seem to recall realizing, when I saw that, that you must be an engineer. That's an ideal application of the circular bellcrank.

Brett