Elevators: Balance, Springs and Bungees

[Blue on Top]

For reference, this thread originated within (dated 1-5-20 posts) the Aileron-Rudder Interconnect thread.  We got a little off topic .  I'm an engineer; I talk too much . 

[Blue on Top]

We'll start with balance weights since it should be close to complete with the discussion in Rudder-Aileron Interconnect.  

Depending on highest speed, shape of the airfoil, shape of the elevator, etc. designers will balance these surfaces from not balanced at all (most slow, early airplanes) to over balanced in transonic and supersonic airplanes, and everywhere in between.  Clear as mud.

The compromises are between aeroelastic effects (flutter), overall weight and aerodynamic balance, including control forces.  There are pros and cons to each of those.  Well, okay, there are no pros to flutter, which can be both bending and torsion.

[Blue on Top]

Springs and bungees (often used interchangeably) are not changed by airspeed; they are (sometimes) changed by geometry/positioning.  A good example of this is the elevator down spring on the K and later airplanes.  This spring puts a constant force on the elevator control to push the nose down (by pulling the trailing edge of the elevator down).  Ingeniously, the force it applies to the elevator changes with empennage (trim, horizontal stabilizer) position to be more force when the trim is positioned more nose up (slower flight).  Now, for that reason and the fact that it's force is fixed with position, aerodynamics are much stronger at higher airspeeds so it does not do as much when the airplane is flying fast.

[Blue on Top]

Centering springs on the other hand, try to keep the surface in the trimmed position (or, in the case of Mooneys, faired with the stabilizer).  The advantage is that the control forces can be tailored to be higher when flying off trimmed condition.  In addition (and if strong enough) can enhance the stick-free (hands off the controls) stability of the airplane.  The tradeoff/compromise here is part count, weight and complexity (and failure modes).

[Blue on Top]

The fixed tab is only aerodynamic and not airspeed dependent.  IOW, if nothing else was attached to the elevator (control tubes, springs, bungees, etc.), moving a fixed tab will move the elevator to a set position … independent of airspeed.

So there you have the rest of the story.  All are there for different reasons (and all are required, legally).  They all are effective during different conditions.  Some are fixed forces (which may change with geometry/position) and some vary their force with airspeed.  Clear as mud.

Let the learning begin .. for ALL of us.  Thanks, Ron

PS. It's going to be great to meet all y'all in person at Sun-N-Fun, Oshkosh, Mooney Summit, etc.  I am honored to be a part of this family. 

[PT20J]

1 hour ago, Blue on Top said:

Let the learning begin .. for ALL of us.  Thanks, Ron

Ron, thank you. You have added much value to the discussions on MS, and by patiently answering so many questions, helped a bunch of curious pilots better understand our airplanes. With a legacy design that has undergone so many changes over the years by so many different engineers (many no longer living), it's a challenge to answer the question every 2 year-old asks: why? There are bits and pieces of the puzzle floating around and a lot of lore and misinformation. It's great fun to separate reality from conjecture and understand more about this airplane.

Skip

[PT20J]

2 hours ago, Blue on Top said:

Centering springs on the other hand, try to keep the surface in the trimmed position (or, in the case of Mooneys, faired with the stabilizer). 

Let me try to answer a question that comes up every so often, "When I'm trimmed in cruise flight and look back at the tail I can see the balance weight and I notice that the elevator is not in trail with the stabilizer. Doesn't this add a lot of drag? Is there something wrong with my airplane? If I somehow 'fixed' this, wouldn't it fly faster?"

I started another thread a while back to gather data on this and from the responses, the short bodies (Al's original design) seem to trim with the elevator in trail, the mid-bodies (except for the K) trim with the elevator slightly trailing edge down. From the K on the long-used trim assist bungees were changed out for a bobweight and downspring. The M20K and all the long bodies use this system and trim with elevator slightly trailing edge up. So, there is nothing wrong with your airplane -- that's just the way it is.

I was told by an engineer at Mooney back around 1990 that Lopresti noticed this when creating the J but determined that it didn't create enough drag to make it worth fixing. (Ron can weigh in here, but I believe to fix it would require changing the fixed incidence of the stabilizer -- big deal). I can see how this makes sense because the offset is small and so the drag would also likely be small compared to induced drag components of trim drag (the tail generates a tail down force which creates drag and the wing has to generate extra lift to carry the tail down force which creates drag).

Now, a puzzle for Ron. At least one person has reported trying to simulate what would happen if the stabilizer incidence were changed such that the elevator was in trail by trimming until the elevator is in trail while holding pressure on the yoke to maintain altitude. (I keep forgetting to try this -- I need to make a note of things to try out next time I'm on a cross country) Anyway, the reported result was that the airplane flew a couple of knots slower. I can't see how this would be as the tail down force (as discussed above - the largest contributor to drag) would have to be the same and the profile drag should be less. But......????

Skip

[Blue on Top]

3 hours ago, PT20J said:

I was told by an engineer at Mooney back around 1990 that Lopresti noticed this when creating the J but determined that it didn't create enough drag to make it worth fixing. (Ron can weigh in here, but I believe to fix it would require changing the fixed incidence of the stabilizer -- big deal). I can see how this makes sense because the offset is small and so the drag would also likely be small compared to induced drag components of trim drag (the tail generates a tail down force which creates drag and the wing has to generate extra lift to carry the tail down force which creates drag).

Skip(minor) 

@PT20J  Skip:  I'll try to do this by paragraph, but I am a little confused by the second part of the first paragraph.  First, I agree with the 1990 Mooney engineer and LoPresti … fixing it would not be worth a drag reduction (if there would even be one).  Where I am confused is "changing the fixed incidence of the stabilizer".  The pilot does that every time he/she moves the pitch trim.  The float (stick free) position of the elevator is determined by all the forces/hinge moments acting around the elevator hinge line, summed to zero: aerodynamics, bob weights, elevator control system, springs, bent trailing edge (or dinged TE due to hangar rash), etc.  The position won't change with (minor) stabilizer (trim) position changes.  It will change with major trim changes because the stabilizer position changes the amount of force the down spring applies.   My understanding is that the horizontal stabilizer does not have any camber (it replicates the flat wood tail).  With that said, a little camber is more efficient than a flat plate at a higher AOA.  The Ralph Harmon metal-metal gap seals are awesome!  So being not in trail is not like airplanes without a trimmable stabilizer.  The fixed stabilizer on those airplanes are set at a minimum drag for a typical cruise flight.

The amount of down force required of the tail is fixed by the pitching moment of the wing.  How efficiently the tail produces that moment determines drag.

Now for the more fun one  

[Blue on Top]

4 hours ago, PT20J said:

Now, a puzzle for Ron. At least one person has reported trying to simulate what would happen if the stabilizer incidence were changed such that the elevator was in trail by trimming until the elevator is in trail while holding pressure on the yoke to maintain altitude. (I keep forgetting to try this -- I need to make a note of things to try out next time I'm on a cross country) Anyway, the reported result was that the airplane flew a couple of knots slower. I can't see how this would be as the tail down force (as discussed above - the largest contributor to drag) would have to be the same and the profile drag should be less. But......????

Skip

So, this sounds like a fun test, but very difficult to get good results (too many variables … including the pilot).

So get a good, trimmed (hands off) cruise speed.   Since your J is flying with the elevator a little TE down, you will need to pull back on the yoke slightly to align the elevator to the stabilizer.   The airplane will start to climb.  You'll need to trim nose down slightly.  Then readjust your force to align the elevator again.  Repeat this iteration until you're within the elevator flight control hysteresis band.  After the airplane is again stabilized, see if the airspeed has changed.  Good luck!

Remember that when you are done with the test, the airplane will be slightly out of trim.

[PT20J]

18 minutes ago, Blue on Top said:

So, this sounds like a fun test, but very difficult to get good results (too many variables … including the pilot).

So get a good, trimmed (hands off) cruise speed.   Since your J is flying with the elevator a little TE down, you will need to pull back on the yoke slightly to align the elevator to the stabilizer.   The airplane will start to climb.  You'll need to trim nose down slightly.  Then readjust your force to align the elevator again.  Repeat this iteration until you're within the elevator flight control hysteresis band.  After the airplane is again stabilized, see if the airspeed has changed.  Good luck!

Remember that when you are done with the test, the airplane will be slightly out of trim.

Yeah, what I'm worried about is that if the effect is only a couple of knots, would I be able to hold altitude precisely enough to observe the effect? The bungees increase stick forces when out of trim so there's a bit of a pull to maintain.

Edited January 6, 2020 by PT20J

[PT20J]

 

1 hour ago, Blue on Top said:

@PT20J  Skip:  I'll try to do this by paragraph, but I am a little confused by the second part of the first paragraph.  First, I agree with the 1990 Mooney engineer and LoPresti … fixing it would not be worth a drag reduction (if there would even be one).  Where I am confused is "changing the fixed incidence of the stabilizer".  The pilot does that every time he/she moves the pitch trim.  The float (stick free) position of the elevator is determined by all the forces/hinge moments acting around the elevator hinge line, summed to zero: aerodynamics, bob weights, elevator control system, springs, bent trailing edge (or dinged TE due to hangar rash), etc.  The position won't change with (minor) stabilizer (trim) position changes.  It will change with major trim changes because the stabilizer position changes the amount of force the down spring applies.   My understanding is that the horizontal stabilizer does not have any camber (it replicates the flat wood tail).  With that said, a little camber is more efficient than a flat plate at a higher AOA.  The Ralph Harmon metal-metal gap seals are awesome!  So being not in trail is not like airplanes without a trimmable stabilizer.  The fixed stabilizer on those airplanes are set at a minimum drag for a typical cruise flight.

The amount of down force required of the tail is fixed by the pitching moment of the wing.  How efficiently the tail produces that moment determines drag.

Now for the more fun one  

The horizontal stabilizer is symmetrical -- I think it might be a NACA 0012. My thought about changing the tail incidence being necessary to make the elevator trail follows this logic: The amount of tail down force is fixed by the airplane pitching moment. If you were to mechanically lock the elevator in trail with the stabilizer, there would be an incidence that would provide the angle of attack for the stabilizer/elevator to generate the appropriate tail down force. The center point of the bungees would also have to be adjusted to make the hinge moments zero when the mechanical elevator lock was removed. The reason for changing the fixed incidence of the stabilizer rather than just adjusting trim is to preserve the trim range in both directions.

If the original design trailed, then something changed when the fuselage was lengthened that causes it no longer be in trail.

I didn't consider the difference in drag from camber change vs AoA change. But, on the J, the trailing edge is down making for positive camber which means that the camber is decreasing the tail down force generated by negative AoA.

BTW, why are symmetrical airfoils chosen for horizontal stabilizers? Is it aerodynamics or structural (like a helicopter rotor blade)?

Skip

Edited January 6, 2020 by PT20J

[takair]

For what it’s worth, my short, 64 E has always flown with elevator trailing edge up, even with aft CG.  While I haven’t done a rig check recently, it was within spec last time I checked.  I would also consider it a fast, relatively stock, E.  From my informal survey here on MS, I think you might find that even the short bodies were a mixed bunch of trailing edge high, in trail and maybe even one or two trailing edge down.  
 

I’ve considered trying to rig the bungee such that it provides more in trail, while being within limits, but I guess I’ve been too lazy....as then other things, like the trim indicator need to be re-rigged as well.  
 

i like the idea suggested to just adjust the yoke to in-trail and re-trim to see the results, but , like Skip, I’m not convinced I could hold things steady long enough to get good numbers.  Might take two pilots working as a crew.  In my experience, it takes multiple minutes for the Mooney to settle on a new speed when trying to measure a couple of knots.  Had recently done a number of test points with the retractable step and one needs to be in smooth air and very patient to see the new number after a change.

[Andy95W]

My short body C flies like Rob's E- elevator trailing edge slightly up.  It gets a little better at aft CG, but not much.

I've tried flying with the trim set so the elevator is in trail and didn't notice a significant difference in speed.  And the 1-2 knots I did notice were probably due to not holding altitude very well from the out-of-trim control force needed- just like Skip said.

[Blue on Top]

9 hours ago, PT20J said:

1) The horizontal stabilizer is symmetrical -- I think it might be a NACA 0012. My thought about changing the tail incidence being necessary to make the elevator trail follows this logic: The amount of tail down force is fixed by the airplane pitching moment. If you were to mechanically lock the elevator in trail with the stabilizer, there would be an incidence that would provide the angle of attack for the stabilizer/elevator to generate the appropriate tail down force. The center point of the bungees would also have to be adjusted to make the hinge moments zero when the mechanical elevator lock was removed. The reason for changing the fixed incidence of the stabilizer rather than just adjusting trim is to preserve the trim range in both directions.

2) If the original design trailed, then something changed when the fuselage was lengthened that causes it no longer be in trail.

3) I didn't consider the difference in drag from camber change vs AoA change. But, on the J, the trailing edge is down making for positive camber which means that the camber is decreasing the tail down force generated by negative AoA.

4) BTW, why are symmetrical airfoils chosen for horizontal stabilizers? Is it aerodynamics or structural (like a helicopter rotor blade)?

Skip

1) Thanks for the symmetrical affirmation.  12% sounds really thick for a stabilizer; half of that would sound better.  I actually doubt it is an airfoil per say.  My guess is that it is more of a flat plate with a rounded (or elliptical) leading edge and a tapered elevator (to simulate the wood tail).  The beading/forming on the elevator also causes drag, but it is easier to assemble when the skins can be rivetted together without ribs.  

2) Yes.  All the springs, bungees, bob weights, bent trailing edges, control system weights, etc. that cause any elevator hinge moment.

3) It's all about total down force.  Yes, it is relieving some of the downforce, but it is also acting like a reflexed airfoil (remember it is up side down) which is lower drag in cruise.

4) Symmetrical tail surfaces are easier to build and can be flipped right to left (less part count).  But, on that note, many tail surfaces are not symmetrical, especially horizontals that are more typically cambered … up side down.  Verticals are more likely to be symmetrical, but typically they are either offset (to offset the engine torque and slipstream) or they are cambered, too.

[Blue on Top]

3 hours ago, takair said:

1) For what it’s worth, my short, 64 E has always flown with elevator trailing edge up, even with aft CG.  While I haven’t done a rig check recently, it was within spec last time I checked.  I would also consider it a fast, relatively stock, E.  From my informal survey here on MS, I think you might find that even the short bodies were a mixed bunch of trailing edge high, in trail and maybe even one or two trailing edge down.  

2) I’ve considered trying to rig the bungee such that it provides more in trail, while being within limits, but I guess I’ve been too lazy....as then other things, like the trim indicator need to be re-rigged as well.  

3) i like the idea suggested to just adjust the yoke to in-trail and re-trim to see the results, but , like Skip, I’m not convinced I could hold things steady long enough to get good numbers.  Might take two pilots working as a crew.  In my experience, it takes multiple minutes for the Mooney to settle on a new speed when trying to measure a couple of knots.  Had recently done a number of test points with the retractable step and one needs to be in smooth air and very patient to see the new number after a change.

1) @takair  CG should not affect elevator trail position (stabilizer position yes).  You have done awesome work!  A little up or down is probably correct ... and with little to no affect.

2) Right on!

3) 100% in agreement with all @takair has said!  The higher one flies the longer you need to stabilize, too.  In the jets at FL410, we would wait 10 minutes before starting any data … and then another 10-15 minutes to stabilize.  Boeing also does potential energy changes up high to determine if they should take data at all (of course they weigh a bit more .  You're right on. 

[takair]

On the older bungee equipped aircraft, I think CG has a small Indirect impact on the elevator in trail position. When trimming nose down to compensate aft CG, the bungee arrangement mechanically pushes the elevator down a little. One can see it in the yoke position and roughly by looking at the counterbalance position. 

[PT20J]

OK, so now I am not so sure about where the elevators rest when trimmed on all models But, it doesn't really matter -- the airplanes fly just fine and we decided that the drag is not significant 

Now, an interesting question is, "Why are the bungees there?" First, they are not an Al Mooney invention. The earlier Piper Supercub has a trimmable stabilizer with up and down bungees that are variable with trim position similar to the Mooney design (but much less elegantly implemented!). There are probably other examples. In another thread I recall a comment attributed to Bill Wheat that Al Mooney had said that the design allowed making the stabilizer (20%?) smaller which reduced drag. I've been thinking about that. The purpose of the tail is stability and control. Looking only at stability for the time being, the larger the tail the more stable the airplane is (and the wider its CG range). If the elevator floats (like say a C-172) then the tail loses some of its effectiveness and the stability is reduced. If the bungees acted as centering springs, they would reduce elevator float improving stability. This could allow for a smaller stabilizer.

The standard method for comparing the stabilizer size  between airplanes is a dimensionless coefficient called a "volume coefficient." The beauty of this abstraction is that it normalizes physical dimensions so that airplanes of different sizes can be compared. It is simply calculated:

Horizontal tail volume coeff. = ((distance from CG to stabilizer 1/4 chord) x (stabilizer area))/((wing mean aerodynamic chord)x(wing area))

From the Mooney M20J service manual and POH:

Horizontal tail volume coeff. = (155 x 21.5)/(59.18 x 174.786) =  0.322

Different sources have different suggestions about what this coefficient should be for a Mooney-type airplane, but the range seems to be about 0.3 to 0.7 so this is definitely on the low end.

Edit: I looked at this again and I think there is an error here.  From some measurements I took, it appears that the stabilizer area listed in the manual does not include the elevator area. If the elevator area (13.0 ft^2) is added to the stabilizer area (21.5 ft^2) the above calculation comes out to a coefficient of 0.517 which seems more reasonable. I found a reference that calculates a C-182 horizontal tail volume coefficient at 0.7, so this is still low.

Ron @Blue on Top, any insights?

Skip

 

Edited January 7, 2020 by PT20J
Corrected that .7 is coeff for C-182 not C-172

[larrynimmo]

My plane is on annual so I took a picture

[PT20J]

9 hours ago, Blue on Top said:

Thanks for the symmetrical affirmation.  12% sounds really thick for a stabilizer; half of that would sound better.  I actually doubt it is an airfoil per say.  My guess is that it is more of a flat plate with a rounded (or elliptical) leading edge and a tapered elevator (to simulate the wood tail).  The beading/forming on the elevator also causes drag, but it is easier to assemble when the skins can be rivetted together without ribs.  

Ron, I was out by the hangar today, so I made some crude measurements. It's definitely a symmetrical airfoil, thinner at the tips and thicker at the root. I made the following measurements at the inboard rib rivet line:

Chord: 44"

Location of max thickness: 14" aft of LE

Max thickness: 5"

So, that would put the max thickness at 11.4% located at 32% of chord.. Given the crudeness of my measurements, that matches a NACA 0012 pretty well

http://airfoiltools.com/airfoil/details?airfoil=n0012-il

And, now when I look at it, it does look thick 

Skip

[Blue on Top]

4 hours ago, PT20J said:

1) OK, so now I am not so sure about where the elevators rest when trimmed on all models But, it doesn't really matter -- the airplanes fly just fine and we decided that the drag is not significant 

2) Now, an interesting question is, "Why are the bungees there?" First, they are not an Al Mooney invention. The earlier Piper Supercub has a trimmable stabilizer with up and down bungees that are variable with trim position similar to the Mooney design (but much less elegantly implemented!). There are probably other examples. In another thread I recall a comment attributed to Bill Wheat that Al Mooney had said that the design allowed making the stabilizer (20%?) smaller which reduced drag. I've been thinking about that. The purpose of the tail is stability and control. Looking only at stability for the time being, the larger the tail the more stable the airplane is (and the wider its CG range). If the elevator floats (like say a C-172) then the tail loses some of its effectiveness and the stability is reduced. If the bungees acted as centering springs, they would reduce elevator float improving stability. This could allow for a smaller stabilizer.

3) The standard method for comparing the stabilizer size  between airplanes is a dimensionless coefficient called a "volume coefficient." The beauty of this abstraction is that it normalizes physical dimensions so that airplanes of different sizes can be compared. It is simply calculated:

Horizontal tail volume coeff. = ((distance from CG to stabilizer 1/4 chord) x (stabilizer area))/((wing mean aerodynamic chord)x(wing area))

From the Mooney M20J service manual and POH:

Horizontal tail volume coeff. = (155 x 21.5)/(59.18 x 174.786) =  0.322

Different sources have different suggestions about what this coefficient should be for a Mooney-type airplane, but the range seems to be about 0.3 to 0.7 so this is definitely on the low end.

Edit: I looked at this again and I think there is an error here.  From some measurements I took, it appears that the stabilizer area listed in the manual does not include the elevator area. If the elevator area (13.0 ft^2) is added to the stabilizer area (21.5 ft^2) the above calculation comes out to a coefficient of 0.517 which seems more reasonable. I found a reference that calculates a C-172 horizontal tail volume coefficient at 0.7, so this is still low.

Ron @Blue on Top, any insights?

Skip

1) Agree (not something to worry about)

2) The bungees are there to "center" or "fix" the elevator.  They also add off trim force (good to a point as it will hold trim condition much better … more stable).  The C172 is not real applicable as its horizontal is fixed and the elevator is positioned with the trim tab.  The early C180/185 on the other hand is similar to the M20 with a trimmable stabilizer.

3) The tail volume equation is a rough guess and doesn't include things like dihedral, sweep, downwash and thrust.  The M10 POC looked good on paper but not so good in the real world.  We didn't do powered tunnel testing .  If you look at spin characteristics, I can tell you from personal experience that the NASA guidelines are worthless.  With that said, you would be surprised at what little items make or break spin testing.

Wind tunnels are great, but are limited to static conditions, which don't include dynamics and inertial effects.  CFD is getting closer, but again, it gets better when put into full simulation with inertial and dynamic effects.  Now, with that said, $$$ get really high for these simulators (close to or more than the cost of a POC airplane).  For $$MMM, Boeing "flies" their airplanes 2 years before the real prototype flies … It's a risk reduction process that pays off (if they trust the simulation … which is only as good as the CFD).

Bottom Line: We still need flight test.

[PT20J]

17 minutes ago, Blue on Top said:

Bottom Line: We still need flight test.

Oh, I don't think we'll ever run out of work for test pilots 

But my purpose in using tail volume is simply to compare the Mooney stabilizer sizing with other airplanes to test the hypothesis that the horizontal tail is smaller for less drag. So far, I have found it to be smaller than a C-182 (0.5 compared to 0.7). The Mooney stabilizer also appears smaller than average as shown in this excerpt from DeRaymer, Aircraft Design: A Conceptual Approach. So, I think there may be something to it.

Skip

[Blue on Top]

@PT20J  Skip:  There is definitely something to it.  The vertical stabilizer is much smaller, too, due to the forward sweep.  It's less drag and less effective at high speeds (which both are good) and it's more effective at low speeds (which is also good).

[PT20J]

According to some, the Piper Comanche was copied from the M20. They certainly have almost identical wings and planforms, so it is interesting to compare them. One notable difference is the tail. The Comanche has a stabilator. In John Thorpe’s patent, he claims advantages for the stabilator. So, I’m wondering what are the pros and cons of the stabilator versus the Mooney design?

Skip

US2563757(1).pdf

[EricJ]

4 hours ago, PT20J said:

So, I’m wondering what are the pros and cons of the stabilator versus the Mooney design?

Stabilators are preferred for transonic and supersonic flight to help prevent control reversal and other undesirable effects.   Mooneys aren't quite that fast.

[PT20J]

23 hours ago, Blue on Top said:

@PT20J  Skip:  There is definitely something to it.  The vertical stabilizer is much smaller, too, due to the forward sweep.  It's less drag and less effective at high speeds (which both are good) and it's more effective at low speeds (which is also good).

This is interesting. I ran down the specs on the PA24-250 at https://www.skytamer.com/Piper_PA-24-250.html.

Since the dimensions of the M20J and the PA24-250 are almost identical except for the tail feathers, it is interesting to compare.

The Mooney horizontal tail is 2 ft² larger than the Comanche (Mooney 34.5 ft², Comanche 32.5 ft²)

The Mooney vertical stabilizer + rudder is slightly larger than the Comanche (Mooney 14.15 ft², Comanche 13.4 ft²)

So, the Mooney's empennage design (swept forward surfaces and trimmable stabilizer with trim assist bungees) didn't end up smaller (less drag) than the Comanche's empennage (stabilator and swept-back fin). 

Skip