The Zero's Maneuverability

[Holtzauge]

Some more info on the Zero's cable stretch solution for the elevator:

While the solution as such may sound simple, in the book Horikoshi mentions that it was not the first thing he thought of and that he actually had to fight for it because just as can be expected, the Japanese design practice at the time was to design the control circuit as stiff as possible using rod controls if possible. So he actually had an uphill struggle before he managed to convince the customer to even try it out. And connected to this is a funny story: Apparently Horikoshi was away on other business when they planned the first flight trials with the smaller cable diameters. But when the test pilots heard about the modification they were afraid it was too flimsy and would not fly it! So he had to go to the airfield and convince them personally before they would take off! However, the flight trials were a success and showed it worked just as he had hoped for, and the solution was accepted for inclusion in the Zero's series deliveries.

And I can really recommend the book: It's not only a good read because of Horikoshi's memoirs about his life and work at Mitsubishi, but there is also a lot of technical details, figures and photos, flight trial and accident investigation data etc.

 

[Engineman]

About the Zero's elevator responsiveness and the cable stretch solutions: I noticed that this was mentioned earlier on in this thread and that someone was looking for a reference. There may of course be Allied reports saying the same thing, but the information can also be had from the horse's mouth so to speak :

In his memoir book Eagles of Mitsubishi, the Zero's Chief Designer Jiro Horikoshi actually dedicates a number of pages (pages 72 to 83) on this subject, and begins by stating that it was though that the Zero was actually TOO responsive in pitch at high speeds. That the roll response at high speed was poor is acknowledged in other parts of the book, but apparently the elevator was considered to be too responsive even at higher speeds. So while he confirms that stick forces in roll were too high, apparently they were not it pitch.

And those pages in the book also cover the solution, which was reducing the diameter of the elevator control cables from 4 to 3.5 and 3 mm in diameter. This introduced more flex in the elevator control circuit at high dynamic pressures, and it seems to have worked, because when they did flight trials, they concluded that not only did this fix the excessive responsiveness at higher speeds, but at the same time it did not mess with the low speed handling which still remained crisp.

Picture below from Jiro Horikoshi's book Eagles of Mitsubishi


View attachment 795618

This is an unusual description of flying control cable stretch function. Twisted steel Wire control cables are usually pre-stretched and installed with a set preload. The cables do have a linear elastic stretch rate in service and that might be approx 0.1% of nominal length at a working load around half of the max working load. So, for a say 15 foot cable, that would be less than a fifth of an inch elastic stretch of an installed cable, not much.
The elastic stretch is at a fairly linear rate between zero and 60% of proof load, so the elastic elongation would not occur at a set load, it gradually stretches and the slight elongation would not be noticable. However, above about 60% proof load, the cable will stretch with permanent deformation and damage.
So, an interesting subject. The described "stretch" above a certain load would seem to more describe a spring-box with a pre-loaded spring motor.

Eng

 

[Holtzauge]

This is an unusual description of flying control cable stretch function. Twisted steel Wire control cables are usually pre-stretched and installed with a set preload. The cables do have a linear elastic stretch rate in service and that might be approx 0.1% of nominal length at a working load around half of the max working load. So, for a say 15 foot cable, that would be less than a fifth of an inch elastic stretch of an installed cable, not much.
The elastic stretch is at a fairly linear rate between zero and 60% of proof load, so the elastic elongation would not occur at a set load, it gradually stretches and the slight elongation would not be noticable. However, above about 60% proof load, the cable will stretch with permanent deformation and damage.
So, an interesting subject. The described "stretch" above a certain load would seem to more describe a spring-box with a pre-loaded spring motor.

Eng

Well maybe it can be described as an "unusual description", but the only thing we know from his book is that he reduced the diameter. However, a book is a book and not a paper, and that was probably why he left out details like if it had been pre-treated in any way and the exact type of braiding and material used in the wire. But that this was a well thought through solution and not just something slapped together in a garage is underscored by the fact that he actually published a paper about the idea after the war: A Proposal for Improvement of Flying Qualities of Piloted Airplanes, by Jiro Horikoshi

TBH, one has to sympathize and identify with the IJN technical staff who were confronted with this crazy idea by a Mitsubishi engineer: What! You want to make the control cables stretch on purpose!? The design rules calls for solutions to minimize play in the control system and now you want to INCREASE it?

 

[tyrodtom]

Having control cables that would stretch sounds like a recipe for metal fatigue.
Wonder if there was note in maintenance procedures for extra checks on control wire tautness?
The Zero recovered in the Aleutians was disassembled, and the Zero had fuselage joint just behind the cockpit where it was meant to be separated. So if they did separate the fuselage at that section, the control cables were disconnected.
The USA no assembly manuals for the Zero, nothing, while I'm sure they had their best airframe men on it, but I have doubts about their ability, at that time, to restore the Zero to Mitsubishi factory acceptable condition.
Sometimes the SWAG method isn't good enough.

 

[special ed]

Those of us who fly controlline models mostly use .015 inch or .018 inch diameter seven strand stainless flying lines between the model and handle. . If there should happen that a kink in a line occurs, that line must be discarded, as under the tension of flight the strands of the line less kinked, now hold the full load. The lines do stretch slightly in flight, but not enough to measure without special scientific equipment. At a 1980 Nationals contest my .15 cu in Bf 109 was pull tested by officials, who flew R/C and were math challenged, at 135 pounds. The lines sounded like guitar strings. When finished, I asked why 135 pounds and their answer was "Sorry. We miscalculated." I build 'em like Grumman.

 

[Holtzauge]

Having control cables that would stretch sounds like a recipe for metal fatigue.
Wonder if there was note in maintenance procedures for extra checks on control wire tautness?
The Zero recovered in the Aleutians was disassembled, and the Zero had fuselage joint just behind the cockpit where it was meant to be separated. So if they did separate the fuselage at that section, the control cables were disconnected.
The USA no assembly manuals for the Zero, nothing, while I'm sure they had their best airframe men on it, but I have doubts about their ability, at that time, to restore the Zero to Mitsubishi factory acceptable condition.
Sometimes the SWAG method isn't good enough.

As long as the cable is being stretched elastically, and the stretching is more due to the wire strands moving in relation to each other, I wonder if fatigue would be much of a problem? But sure, maybe they did have a maintenance schedule like replacing the elevator cables every X hours. But even so, that seems like a manageable issue if it solved the high speed problems they were battling.

One thing that's strange though, is that some posts in this thread mention that Allied pilots who test flew the Zero in the war, and that some pilots who get to fly them today describe the elevator as "set in concrete" at higher speeds, while Horikoshi's book and Japanese pilots do not seem to agree as I understand it? OTOH, if we compare this to the Bf-109, it's exactly the same story: Allied pilots thought the Messerschmitt's elevator was far to heavy while AFAIK there were very few German pilots who complained. Maybe familiarity breeds comfort as the saying goes?

Regarding SWAG, if that does not work out, KISS-solutions sometimes works wonders!

 

[Holtzauge]

Those of us who fly controlline models mostly use .015 inch or .018 inch diameter seven strand stainless flying lines between the model and handle. . If there should happen that a kink in a line occurs, that line must be discarded, as under the tension of flight the strands of the line less kinked, now hold the full load. The lines do stretch slightly in flight, but not enough to measure without special scientific equipment. At a 1980 Nationals contest my .15 cu in Bf 109 was pull tested by officials, who flew R/C and were math challenged, at 135 pounds. The lines sounded like guitar strings. When finished, I asked why 135 pounds and their answer was "Sorry. We miscalculated." I build 'em like Grumman.

Used to fly line-controlled combat with streamers (about half a century ago!) so I recognize the problem with kinks! . But in the wires I used, these were just a few strands and were soldered together so more rod-like with basically no play at all.

But TBH, I can't figure out why so many in this thread are insisting that wire cables don't stretch? They do. And just to take an example, the Spitfire's ailerons were wire controlled, and you can find stories of them rising up visibly on both sides in high speed dives from this very effect.

 

[GregP]

So I guess it's: Stretchy cables-Myth Busted!

Not too sure it is myth busted, but reducing the diameter would seem to also reduce the strength, and the last thing any pilot wants is to lose elevator control due to snapped control cables while he is pointed at the ground and is also at high speed. I surely wouldn't want to be there!

A servo tab is one that assists the movement of the surface, and is also called a balance tab and is used whenever the surface is too hard to move at speed.

The standard "fix" for a control surface that is too sensitive at speed is an anti-servo tab. That is, a small surface that tends to work against the movement of the surface. There are no servo or anti-servo tabs on the Zero as far as I know. Unless I am mistaken, they are trim tabs; one per elevator half. That is, they do not move when the elevator moves ... they are put into position by a trim crank and they stay there unless moved deliberately with the trim crank.

But, hey, I can ask about this stuff. Someone knows, and they might tell me. Stranger things have happened.

 

[Engineman]

Well maybe it can be described as an "unusual description", but the only thing we know from his book is that he reduced the diameter. However, a book is a book and not a paper, and that was probably why he left out details like if it had been pre-treated in any way and the exact type of braiding and material used in the wire. But that this was a well thought through solution and not just something slapped together in a garage is underscored by the fact that he actually published a paper about the idea after the war: A Proposal for Improvement of Flying Qualities of Piloted Airplanes, by Jiro Horikoshi

TBH, one has to sympathize and identify with the IJN technical staff who were confronted with this crazy idea by a Mitsubishi engineer: What! You want to make the control cables stretch on purpose!? The design rules calls for solutions to minimize play in the control system and now you want to INCREASE it?

I read and google-translated the link, but it is impossible to make a fully detailed understanding of the result, due to poor and ambiguous translation. For example, the illustration of the aircraft control column translates as "disaster lever". However, I can see the pitch control system diagram illustrates a control linkage and a spring in series to the control surface.
Overall, it seems that the proposal is to change the ratio of stick movement to surface deflection under increasing air loads, by the action of deflecting the spring. This might certainly allow further movement of the control stick but, the further movement would be mostly absorbed as loading energy into the spring. A far better result might be obtained by linkage geometry that gives a suitable change in relative movement, giving an increase in mechanical advantage when the control resistance is high . In fact, such use of variable linkage geometry is common practice in control systems.
Having a spring in the system to allow further control stick movement when the control surface has high resistance to deflection may give the deception of control authority, but really, it is just stretching the spring.

Eng

 

[Holtzauge]

Not too sure it is myth busted, but reducing the diameter would seem to also reduce the strength, and the last thing any pilot wants is to lose elevator control due to snapped control cables while he is pointed at the ground and is also at high speed. I surely wouldn't want to be there!

A servo tab is one that assists the movement of the surface, and is also called a balance tab and is used whenever the surface is too hard to move at speed.

The standard "fix" for a control surface that is too sensitive at speed is an anti-servo tab. That is, a small surface that tends to work against the movement of the surface. There are no servo or anti-servo tabs on the Zero as far as I know. Unless I am mistaken, they are trim tabs; one per elevator half. That is, they do not move when the elevator moves ... they are put into position by a trim crank and they stay there unless moved deliberately with the trim crank.

But, hey, I can ask about this stuff. Someone knows, and they might tell me. Stranger things have happened.

Horikoshi in his book mentions that he calculated the breaking strain on the reduced diameter elevator cables for the Zero and that they were well within strength. And the problem he was attempting to solve was not that the stick forces were too high and that elevator was not deflecting enough, but rather the contrary: The Zero was TOO sensitive in pitch at high speeds and responded to quickly and abruptly to elevator input according to the book. And if this was the case, one can IMHO even intuitively understand how the cable stretch solution works even without any equations: The pilot pulls the stick and the elevator does not move as far as it does at lower speeds. Just like in the figure I posted earlier. But for sure, an anti-servo tab would certainly work as well. And Horikoshi mentions that other solutions to the problem were considered, but that the cable stretch option was the simplest and easiest to implement, and I find it hard to fault him on that conclusion.

I read and google-translated the link, but it is impossible to make a fully detailed understanding of the result, due to poor and ambiguous translation. For example, the illustration of the aircraft control column translates as "disaster lever". However, I can see the pitch control system diagram illustrates a control linkage and a spring in series to the control surface.
Overall, it seems that the proposal is to change the ratio of stick movement to surface deflection under increasing air loads, by the action of deflecting the spring. This might certainly allow further movement of the control stick but, the further movement would be mostly absorbed as loading energy into the spring. A far better result might be obtained by linkage geometry that gives a suitable change in relative movement, giving an increase in mechanical advantage when the control resistance is high . In fact, such use of variable linkage geometry is common practice in control systems.
Having a spring in the system to allow further control stick movement when the control surface has high resistance to deflection may give the deception of control authority, but really, it is just stretching the spring.

Eng

Absolutely. But given that the cable stretch solution seems to have done the job and required basically no new design features and was easy to implement, I can see why they went for it. The anti-servo tab solution GregP mentions is probably a better way to go and was probably even considered given that Horikoshi tried a servo tab solution for the heavy ailerons that unfortunately did not work out. But that's another story.....






Edit: On second thoughts, maybe the anti-servo tab would not work so well after all since Horikoshi did mention that the Zero was TOO sensitive at high speed and even though an anti-servo tab would increase stick forces at higher speeds, even a small stick deflection seems to have generated a too rapid pitch response. So a better solution would probably be to have the mechanical advantage in pitch vary with speed, and IIRC then Horikoshi did actually mention that solution in his book.

 

[SaparotRob]

I read and google-translated the link, but it is impossible to make a fully detailed understanding of the result, due to poor and ambiguous translation. For example, the illustration of the aircraft control column translates as "disaster lever". However, I can see the pitch control system diagram illustrates a control linkage and a spring in series to the control surface.
Overall, it seems that the proposal is to change the ratio of stick movement to surface deflection under increasing air loads, by the action of deflecting the spring. This might certainly allow further movement of the control stick but, the further movement would be mostly absorbed as loading energy into the spring. A far better result might be obtained by linkage geometry that gives a suitable change in relative movement, giving an increase in mechanical advantage when the control resistance is high . In fact, such use of variable linkage geometry is common practice in control systems.
Having a spring in the system to allow further control stick movement when the control surface has high resistance to deflection may give the deception of control authority, but really, it is just stretching the spring.

Eng

Can we still have a "disaster lever"?

 

[Engineman]

Yes, out of the mouths of google-translate.
Disaster-lever is a good term for the control column in the hands of some student pilots. I recall a fellow instructor being faced with a student facing IMC at low-level: Student (at a loss) "What shall I do Sir?", Instructor (bluntly) Oh, just go-ahead and crash!

Eng

 

[SaparotRob]

Yes, out of the mouths of google-translate.
Disaster-lever is a good term for the control column in the hands of some student pilots. I recall a fellow instructor being faced with a student facing IMC at low-level: Student (at a loss) "What shall I do Sir?", Instructor (bluntly) Oh, just go-ahead and crash!

Eng

Screwing up is what I do best.

 

[Admiral Beez]

I find all these comparisons pointless

Since when is striving to understand your enemy pointless?

 

[GrauGeist]

I recall a fellow instructor being faced with a student facing IMC at low-level: Student (at a loss) "What shall I do Sir?", Instructor (bluntly) Oh, just go-ahead and crash!

When I was a youngster, one of my instructors, who was a former WWII USMC Corsair pilot of VMF-212, took exception to my nervous, heavy hand on the stick.
He said "what the hell are you doing, boy? Are you trying to wreck my ship? You will not wreck my ship!"

Realizing I was effing up, I straightened up and I did not wreck his ship.

 

[MIflyer]

I have a vague memory of a patent being filed for the way that the A6M2 ailerons were controlled in a variable manner and speed was a key determinant.

In the Ercoupe the ailerons have a very differential action. At low roll rates they both move about the same, one up and one down. But as the amount of aileron applied increases the aileron on the high wing (which moves down) actually starts to go back the other way, so that at full travel it is almost flush with the trailing edge. Meanwhile the aileron on the low wing (which moves up) continues to move up until at full travel it is sticking almost straight up, and acts as a spoiler. This means that past a certain point the high wing, which is gaining lift, gains less lift, while the low wing, which is losing lift, keeps losing it. Seeing how this works is a marvelous thing to behold and I have no idea how they figured it out, but it is nothing exotic. The result is that you are less likely to stall the high wing as a result of "revising the wing camber" by the aileron action. I would not be at all surprised if the Zero did something similar, although how this would affect control forces I do not know.

 

[MIflyer]

The forward fuselage of the P-39 is a brick. The rear fuselage makes the Spitfire and A6M look like brick

Okay, that makes perfect sense. The aft fuselage is there to balance the weight of the engine by applying "negative lift" via the elevator. Aircraft loaded with aft CG's are easier to stall and spin but are more efficient because the elevator has to produce less "negative lift" to counterbalance the forward fuselage.

On the P-39 the engine is not way out there at the tip of the nose but is in the aft fuselage. Therefore the aft fuselage does not have to be as strong because less "negative lift" is required.

 

[MIflyer]

Since when is striving to understand your enemy pointless?

Since when is trying to understand people who were a hell of a lot smarter than us pointless?

 

[MIflyer]

When I was a youngster, one of my instructors, who was a former WWII USMC Corsair pilot of VMF-212, took exception to my nervous, heavy hand on the stick.

I have a WW2 USAAF Primary Flying manual that advises students to wiggle their toes in times of stress to help them attain a more calm attitude. I have no idea if this works.

 

[Engineman]

I have a WW2 USAAF Primary Flying manual that advises students to wiggle their toes in times of stress to help them attain a more calm attitude. I have no idea if this works.

Yes, and some people whistle a tune. Statistics showed 90% of accidents were preceded by the pilot whistling. So, pilot whistling is strictly prohibited!

Eng