X-3 STILETTO Is it possible to overcome pitch/roll coupling problem???

[Tumbler]

Hi,

I have been working on a twin ducted fan X-3 Stiletto and tried out my test plane made out blue foam and a pusher prop today. The real one had a bad pitch/roll coupling problem and so does this one. As the real one did, when I make a turn, the nose pitches up badly. The CG is right based on glide tests and the fact it is fine in straight and level flight.

Is there anyway to overcome this problem????

Thanks,

Brian

[BMatthews]

I suspect that this is one of those times where the overall oddness is fighting you.

The only thing I can think of is to make a second foamy test model. But this time forget the long fuselage. Use the same wings and tail surfaces and tail moment arm but on a short more or less "normal" fuselage with a tractor motor and see if the same issues crop up. At least then you'll find out if it's related to the long fuselage or pusher prop.

If you're not using a symetrical airfoil now you may want to change to one. The neutral pitching action of the symetrical airfoil may be all you need vs the strong tendency for a cambered lifting airfoil to want to nose up. If a switch to symetrical helps then consider using a self stable flying wing airfoil and forward CG. Then the stabilizer is relieved of it's stabilizing duties and becomes just a pitch control.

It may be that this is also one of those rare times when the wing and fuselage blanks out the tail for real. If so the self stable flying wing airfoil may provide the answer.

OK, so I surprised myself by coming up with more than one possibility... :D Best of luck with a very odd and interesting project.

[Tumbler]

Hey, thanks for the detailed reply. Its very wierd you are right!

I am just using a flat plat for testing but the wings i made for the real one are just depron wings sanded to an airfil shape. I would think if it will fly as a "foamy" they it will fly with a flat bottom aifoil...but who knows.

Your suggestions are very helpful thanks.

what do you think about positive incidence on the stab, or moving the wing foward.

Rioght now the CG is only about 1.5 inches forward of the leading edge on a 10 inch chord. I can't beliwve the pitchiness is due to being tail heavy do you?

Thanks again?

brian

[Tom Harper]

Forward of the LE?

[biber]

Well, that thing seems to be just right fitting for the PNF Jartists in the Slope Forum. :D

But seriously, I would guess the long nose is heavily counteracting the vertical stab. Maybe the C.G is just right for the top view planform, but to far aft for the lateral view planform. Or lets say the lateral NP is maybe not as far behind the C.G. as would be desired, so that it won't yaw into the turn.

The easiest solution, but not scale, would be to increase lateral surface behind the C.G. and decrease it in front of C.G.

biber

[biber]

Or, try an even more forward C.G. position, that would at least be scale. Be sure the fuse produces quite a portion of the overall lift of that thing, so the C.G. doesn't only depend on the Wing and Tail alone.

biber

[Sparky Paul]

Although some will deny it, the X-3 provided the aerodynamic data used to design the F-104. It moved the horizontal to the top of a much larger vertical to get rid of the pitch/roll coupling, and lessened the amount of fuselage volume ahead of the wing.

[BMatthews]

The huge fuselage in proportion to the wing is why I was suggesting a more or less normal version but that uses the exact same wing and tail configuration including spacing. If that one flies then it's probably due to the fuselage interaction.

I'm assuming that you actually mean 1.5 inches BEHIND the leading edge and not forward... :D

Given the taper that puts your CG at less than 15% and that means it should be stable. But that strongly forward a CG requires a lot of pitch control from the stabilizer to compensate and it may just end up being that the model is so pitch stable that you are seeing the nose lift in the turn just from the extra speed and corresponding downlift from the stabilizer. However if when the nose starts to lift you actually need to use down elevator then it definitley is a severe pitch up problem not related to the CG.

Is the fuselage on your test model fully shaped or just a profile?

[Tumbler]

Forward of the LE?


Sorry, I mean behind the LE......however from what I am seeing, it might actually need to be balanced ahead of the wing.

[Tumbler]

The huge fuselage in proportion to the wing is why I was suggesting a more or less normal version but that uses the exact same wing and tail configuration including spacing. If that one flies then it's probably due to the fuselage interaction.

I'm assuming that you actually mean 1.5 inches BEHIND the leading edge and not forward... :D

Given the taper that puts your CG at less than 15% and that means it should be stable. But that strongly forward a CG requires a lot of pitch control from the stabilizer to compensate and it may just end up being that the model is so pitch stable that you are seeing the nose lift in the turn just from the extra speed and corresponding downlift from the stabilizer. However if when the nose starts to lift you actually need to use down elevator then it definitley is a severe pitch up problem not related to the CG.

Is the fuselage on your test model fully shaped or just a profile?

Yes, I am sorry I meant nehind the CG. The test model is just a profile. Here is the acually fuse.

I am wondering now from reading posts from you guys. .....what about moving the ing forward a little????

[Tom Harper]

The long fuse is unstable simply because it wants to fly across the relative wind. The other problem is the amount of lateral area forward of the CG. That's all in thin profile which makes it a forward fin. I had problems with the LD4 until I boxed the forward fuselage.

I think Bruce has the right approach. Remove 75% of the forward fuselage and see what happens. Keep the CG at or a little forward of the 25% chord point. If you can adjust the model in that configuration - work your way forward. It may work better if you add a paper cone instead of a profile.

BTW: Great project!

[Tumbler]

The long fuse is unstable simply because it wants to fly across the relative wind. The other problem is the amount of lateral area forward of the CG. That's all in thin profile which makes it a forward fin. I had problems with the LD4 until I boxed the forward fuselage.

I think Bruce has the right approach. Remove 75% of the forward fuselage and see what happens. Keep the CG at or a little forward of the 25% chord point. If you can adjust the model in that configuration - work your way forward. It may work better if you add a paper cone instead of a profile.

BTW: Great project!


Well, roughly 75% of the forward fuse was removed when it hit a school yesterday haha. Ironically the only damage.

what do you mean keep the cg 25% forward of the chord point?

Thanks,

Brian

[BMatthews]

He's saying keep the CG at the 25% chord point or even a trifle further forward.

Thinking about this while reading the replies posted since my last one I got to remembering some profile scale jet model gliders I made in my teens. I do recall that a lot of them tended to do funny things when they turned even though they were stable gliding straight ahead. It wasn't that they pitched up but that once banked a little the long nose with all the side area tended to make them yaw strongly to the "nose up" side of the turn and once that happens they would more or less roll and fall into a tumble. I'm wondering if this is more what you're seeing than a true pitch up issue.

If so then it means that the newly shortened nose version may well fly more or less normally at this point. Give it a try and perhaps even shorten the nose some more. If that causes you to need a lot more nose weight then try replacing the entire nose from the leading edge forward with an X-3 length round boom with the nose weight on that and see how it flies. The idea being to restore the overall length and mass distribution but without the side area. That would truly confirm whether or not this is a side area distribution problem or a mass damping issue or a combination of both.

The mass damping thing just occured to me. It may be that with the mass spread out over such a long length that what you are seeing is the model nosing up and then not having enough aerodynamic leverage to dampen the movement of the mass. It's a very long fuselage and has the motor presently located in the tail and presumably the rest of the gear ahead of the wing to balance the whole issue. This creates a rather spread out mass distribution. As you know if you hold a brick with your arm extended it's harder to get it moving and harder to stop it than if you hold it close to your body. The X-3 may be partly suffering from this as well.

Going from a profile to the rounded fuselage will help a lot as the effective side area of the rounded shape will not be as high as the profile one but likely it will still be an issue. To get around it a larger than scale fin may be needed and perhaps even some extra tail fin area in the form of clear plastic ventral fins. Try the nose boom first. If this is successful then add back on the profile nose along with a larger fin and a bit of a ventral fin and see if you can maintain the stability. Once you can get the profile version flying stably then you're good to incorporate any changes to the scale model. And if the mass distribution is partly to blame then switching to center located DF units and keeping the rest of the gear in close proximity to the center of the model will reduce this effect greatly if not eliminate it.

All of which goes to say that I suspect strongly that if you can get the profile model to fly decently that the final model will be better in every way stability wise.

I really love these odd designs and problems. It's things like this that really keeps me hanging around here. Far more fun than silly crossword or MENSA puzzles... :D Hopefully "we" can come up with a solution to your X-3 issue.

[kristianb]

Or, try an even more forward C.G. position, that would at least be scale. Be sure the fuse produces quite a portion of the overall lift of that thing, so the C.G. doesn't only depend on the Wing and Tail alone.

biber

Hello,
Nice project Brian!
I believe Biber is quite right here, the fuse is responsible for a lot of lift generated, therefore your stiletto should have a C.G very much forward of the traditional location.
I would estimate that You need to have the C.G 2-3 inches in front of the wing leading edge.
Lateral point is just behind Your current C.G for the top view and quite in the same point for the side area so Your observation is correct. It should be impossible to maintain controlled flight with this set up.
vbr KristianB

[BMatthews]

Bear in mind that the test model under consideration at this point is using a profile fuselage. Not much lift off a 1/4 inch wide edge.

Later on with the full fuselage version there will be some fuselage lift and it would be wise to begin with a more forward CG but even then the X3 does not have a fuselage lift component like the current very wide twin engined fighters do. There will likely be a need to move it forward thanks to the top view area of the fuselage but I strongly suspect it will still be within the bounds of the wing area outline.

A new test glider model that uses a roughly carved but full section fuselage coupled with any changes to the tail areas determined by the testing of the present test model will soon tell you where the CG needs to be. This CG test model would be a very good idea to make as it would be a true shame to see the pointy nose of that final version come to harm over something that is so easily avoided by adding an evening's work to the program. It would also confirm the effect of any changes to the tail area or added ventral fins that came about from the current model's tests.

[kristianb]

ok if the profile fuelage only have the side projection, it will then suffer from the lateral instability and have limited possibilitys for an optimized C.G. Adding the upper projection to the test model (shock flyer type) will improve the C.G try out.
KristianB

[Tumbler]

ok if the profile fuelage only have the side projection, it will then suffer from the lateral instability and have limited possibilitys for an optimized C.G. Adding the upper projection to the test model (shock flyer type) will improve the C.G try out.
KristianB


Not sure if I understand what you are saying,but here is my test model. 62" length.

Brian

[BMatthews]

Ah HA! You added side strakes. That's going to complicate the CG location by providing lift ahead of the wing itself. So it's likely you need to move the CG ahead a bit. Here again the full body is not going to add so much lift as the side strakes do. The final CG location on the full bodied model will be a compromise between the one you find is needed for this model and the one needed for a similar model that is just wings with no side strakes.

[Tumbler]

Ah HA! You added side strakes. That's going to complicate the CG location by providing lift ahead of the wing itself. So it's likely you need to move the CG ahead a bit. Here again the full body is not going to add so much lift as the side strakes do. The final CG location on the full bodied model will be a compromise between the one you find is needed for this model and the one needed for a similar model that is just wings with no side strakes.


Weeeellll, actually I had them on there when I flew it the first time.....I probably should have posted a pic of it when I had originally started the thread. The CG was about 2 inches in back of where the wing meets the side of the fuse or "strake" as you called it. It flew great in a straight away but picthed up in a turn. So, I now have a CG about a half inch in front of the leading edge. we'll see how she goes after the rain stops one day haha.

I have made up weighs of 5 oz increments which i plan to add to it until I get a weight that is comparible to the final full-fused version.

You are right the final version of the fuse won't provide as much lift, but problem is it will need a lot of tension and speed off the bungy and can't have a lot of room for erro. A little tail heavy and it will snap and thats it. fun fun
Thanks for the help.

Brian

[Sparky Paul]

I had the same result with a profile scale F-22, with the sidestrakes and fuselage planform in profile.
Flew straight from launch, pitched up and departed when trying a turn.
Added a ridiculous amount of nose weight trying to get past that condition, but never could.

[Tumbler]

I had the same result with a profile scale F-22, with the sidestrakes and fuselage planform in profile.
Flew straight from launch, pitched up and departed when trying a turn.
Added a ridiculous amount of nose weight trying to get past that condition, but never could.


Well there plenty of f-22's out there flying so I am sure you can get it worked out. I am confident the X-3 will fly, it will just have to be turned gently or lots of big round split S's.

[Sparky Paul]

I suspect anything like that will need power to keep the speed up.
My sloper F-22 probably could have used power, but at gliding speeds it couldn't maintain pitch angle without pitching up and departing.
An F-117 model did the same thing. OK in a straight line, departed when trying to turn.
Neither of them made a successful flight.

[vintage1]

I suspect anything like that will need power to keep the speed up.
My sloper F-22 probably could have used power, but at gliding speeds it couldn't maintain pitch angle without pitching up and departing.
An F-117 model did the same thing. OK in a straight line, departed when trying to turn.
Neither of them made a successful flight.

...and yet a paperdart flies reasonably well. Hmm.

Try anhedral. I can't think why I said that...but something stirs in the distant recesses of what used to pass for a brain. Also use rudder at slow speeds

I built a jetex starfighter profile thing years ago. It used to be really vile..it almost flew ONCE - it hit the ground JUST before it had enough speed to climb. ISTR that also would get into fairly weird oscillations at slower speeds.

I just looked at your model and squeezed my eyes..and thouh of it as if the fsusealage were a wing...and the real wing was just a bit extra..I'd say a CG about 4 inches ahead off the LE might be a place to start :D

[BMatthews]

Just for the heck of it you may want to add some lightweight white styrofoam side blocks to the fuselage and more or less chew it away until it's sort of shaped like your final fuselage. I'm getting the impression that the aerodynamics of the fuselage are largely clouding the issues here. Reducing it's effect by adding in filler to replicate the final shape at least more or less will put the lift emphasis back on the wing and you'll see if the final product holds promise or not.

A few years back I built a little rubber band catapult F18 and it's a fine flyer, LEX's and all. Nice predictable clean stalls with recovery and all. So it can be done.

[kristianb]

Not sure if I understand what you are saying,but here is my test model. 62" length.

Brian

Hello Brian,
You made Your test model as I expected, I was a bit mis-lead by that Bruce believed it was only the side profile of the forward fuselage.
I dont believe there will be any significant difference by the amount of lift created by a rounded fuselage or Your test version with strakes. Lift will only depend on amount of air affected by the body and how much downward acceleration this air is given. A rounded fuselage might even give some more lift.
The Stiletto doesnt looks very smart at all, airplanes normally are quite stupid and dont understand us humans when we name different parts different. If we call a part wing for example this have no meaning to the airplane. It feels only lift distribution and aerodynamical forces from lateral points etc.
Difference to a "normal" design is that the forward part of the fuselage can be neglected in comparison with the wing.
So I´m convinced your Stiletto needs a C.G well ahead of the wing leading edge, probably Vintage is correct as he very often is. He probably studied the scale drawings more thorough than me so a C.G 4 inch ahead of the wing leading edge probably is correct!

Exciting project Brian! From the oddness of the design it would be a great achievement if You manage to make a complete flight with it!
You have choosen the correct way of sorting out it´s behaviour with a simple test model first!
Good Luck!
KristianB

[Tom Harper]

Brian,

My vote is for Bruce's scheme of filling in the nose section with foam or paper covering. Rounding the profiles will reduce their aerodynamic impact.

Hey, just what you needed - more advice!

Keep us informed.

[vintage1]

I din't study any scale darwings. I just looked at that enormous side area and thought 'if I designed that to fly on its side, where would I put the CG?'

With a hot ship, a forwad CG merely means it lands very hot and needs a LOT of up elevator which in turn means it glides like a brick due to the drag.

I'd play around with way forward CG's just to see what happens..moight be worth strapping a servo on the elevator too. And bungee or catapulting it up high.

[burkefj]

I think you hit the nail on the head, I just completed a profile X-3, which does not have a top view, just side profile, therefore the flight CG was even further aft, first test flight, yawed under a side gust(very windy conditions)
and pitched/snapped. I increased the vertical tail because I was sure this was the issue, and moved the cg just slightly forward and it flies perfect.

Normally we don't need to worry about lateral stability, but in this case we do.

Frank


Well, that thing seems to be just right fitting for the PNF Jartists in the Slope Forum. :D

But seriously, I would guess the long nose is heavily counteracting the vertical stab. Maybe the C.G is just right for the top view planform, but to far aft for the lateral view planform. Or lets say the lateral NP is maybe not as far behind the C.G. as would be desired, so that it won't yaw into the turn.

The easiest solution, but not scale, would be to increase lateral surface behind the C.G. and decrease it in front of C.G.

biber

[Brandano]

Are you sure the original had a pitch-roll coupling issue? I believe the issue was with jaw-roll coupling, and was caused by inertial coupling caused by the long fuselage. The 104 suffered from the same problem, though in a lesser measure, and probaly the tip mounted missiles or tanks mitigated this quite a lot. You might want to add some mass at the wingtips and lighten the tail and nose substantially to have the overall mass of the plane better distributed.
[edit:] I stand corrected. Any roll input on a plane with inertial coupling issues will cause out of control pitching and yawing oscillations, http://en.wikipedia.org/wiki/Inertia_coupling has a good explanation of the issue. Adding more mass to the wings and less mass to the fuse will still fix the problem, other solutions require visible changes to the plane, which is not desirable in a scale ship. adding a yaw and a pitch gyro as dampeners might help too, some full size planes went this route after all

[burkefj]

My comment was to the lateral stability, ie more forward area, than rear area relative to the CG, and that definitely was the problem and was fixed with increased tail area. There is very little mass forward of the CG, the battery is only about 4" ahead of the CG, and the motor is very light, so I don't think there is a yaw/roll issue with my profile model, just a yaw stability issue.

I believe some of the learnings were applied directly to the F-100 which was having problems at the same time and was fixed by the increased tail. My tail increase doesn't affect the scale look enough to be noticed by most.

Frank


Are you sure the original had a pitch-roll coupling issue? I believe the issue was with jaw-roll coupling, and was caused by inertial coupling caused by the long fuselage. The 104 suffered from the same problem, though in a lesser measure, and probaly the tip mounted missiles or tanks mitigated this quite a lot. You might want to add some mass at the wingtips and lighten the tail and nose substantially to have the overall mass of the plane better distributed.

[Brandano]

Well, a visualization of the problem is this: you have the plane, flying straight at a slightly positive AOA, then you suddenly apply a roll input. the tail is "lower" than the nose, and the plane is rolling on an axis that is not exactly aligned with the mass. centrifugal forces act on the tail and nose, and the plane pitches up wildly. If it was also side-slipping at the time it will also yaw. It's like an out of balance engine shaft, any roll input will cause the plane to move on the other axes. Lifting the horizontal stab to the top of the tail fin on the 104 meant that more mass was away from the centerline, and also moved some mass on the tail above the centerline so that the centrifugal forces at a positive AOA would make the nose pitch down instead.

[burkefj]

Got it, that's a good explanation.