Flying without a vertical stabilizer

[telnar1236]

I added in some drag rudders and ran the CFD model with them included. They should very easily be able to counteract the instability. On the 70 mm model, which is the more unstable of the two, the drag rudders are always at least 2.3 times as strong as the destabilizing moment from the airframe up to 12 degrees sideslip. By that sideslip angle, pitch stability would have also started to suffer very badly, so whatever flight rules I write won't ever let the plane exceed 8 degrees sideslip to ensure it remains stable, though I might also add in the ability to override this to give the ability to perform thrust vectoring maneuvers.

Interestingly, even with equal deflection on the top and bottom, the drag rudders also cause a small proverse rolling moment by killing the high-pressure region below the wing. Interestingly, the low-pressure region on top of the wing is mostly preserved even with full deflection.

The airflow isn't significantly disrupted anywhere except for over the drag rudder so fortunately it doesn't really kill much lift at all.

 

[telnar1236]

I don't expect it will fly, but I have nearly finished the 64mm version of the jet. The airframe was mostly printed already when I decided to run the CFD, so I figured I might as well finish it. It has essentially no stability, in that it is very nearly neither stable nor unstable, so a simple gyro might just about be enough to make it flyable. But still, I think there is at best a 10% chance of it flying. It is also very underpowered. The nozzle does successfully avoid any pitch up moment, but it kills almost 40% of the thrust, so instead of having a reasonable TWR of about 0.9, it has a TWR of about 0.5. Still enough to fly on paper, but maybe not practically.

 

[L Edge]

I don't expect it will fly, but I have nearly finished the 64mm version of the jet. The airframe was mostly printed already when I decided to run the CFD, so I figured I might as well finish it. It has essentially no stability, in that it is very nearly neither stable nor unstable, so a simple gyro might just about be enough to make it flyable. But still, I think there is at best a 10% chance of it flying. It is also very underpowered. The nozzle does successfully avoid any pitch up moment, but it kills almost 40% of the thrust, so instead of having a reasonable TWR of about 0.9, it has a TWR of about 0.5. Still enough to fly on paper, but maybe not practically.
View attachment 244614

Did you find that the CG is near the inlet area?
Please take a video and show it. You can glean a lot of info from it.
What is the angle of the exhaust nozzle?

 

[telnar1236]

Did you find that the CG is near the inlet area?
Please take a video and show it. You can glean a lot of info from it.
What is the angle of the exhaust nozzle?

The CG is maybe an inch behind, the inlet. It's what I've found to be the best CG by throwing a bunch of profile gliders, so hopefully it scales up to the 64mm version.

I'll do my best to get video. I want video too, since I should be able to verify how unstable it actually is by how fast it tries to enter a spin if, as is likely, it doesn't fly, but no promises. I suck at videoing, so what I capture might not be useful.

The exhaust nozzle is not angled relative to the airframe, with the thrust going almost through the CG, so it shouldn't produce any real moment. The losses come from me trying to cram a 64mm EDF into an airframe that is really too small for it.

I've also been playing around with pre-foamed LW PLA which means the airframe is about 30% heavier than it would be using a single wall of normal PLA and maybe 50% heavier than it could be using proper LW PLA. I will not be buying more rolls.

 

[telnar1236]

I'm still hoping the version I currently have built will be stable enough to fly, even I don't expect it to be, but I've also continued to play around with aerodynamics to try and make something stable, and I did it! Based on the CFD, the designs with inlets on top and swept wings were the most stable, and it showed that at least on top of the airframe, the sharp corners were hurting. Taking all of that into account results in this design.

It is actually more aerodynamically efficient than my previous designs, but more importantly... IT'S STABLE FOR MOST FLIGHT ATTITUDES! About 1/6 as stable as a more conventional layout to be more specific, which should be good enough to fly!

The inlet placement was also changed and moved a bit farther forward since the inlet placement on some of my earlier designs seemed to be disrupting the vortices generated by the fuselage/leading edge extension in a way that hurt flight performance and stability a lot. On the other hand, the wing is designed with 3 degrees positive incidence relative to the fuselage, so there is essentially no vortex generated when the plane is just cruising around which reduces drag. The first picture is at an angle of attack of 5 degrees which is the trim condition for steady flight at 45 mph and the second picture is at the same speed but at 25 degrees angle of attack (the aircraft pulling a hard 6g turn).

Since it's designed to be powered by a single 64 mm fan, I designed it to use a more conventional inlet and outlet for the EDF which should reduce losses and increase thrust by a ton.

In the very likely even my current build doesn't fly, or crashes for one reason or another, this is the next version I plan to print.

 

[L Edge]

I'm still hoping the version I currently have built will be stable enough to fly, even I don't expect it to be, but I've also continued to play around with aerodynamics to try and make something stable, and I did it! Based on the CFD, the designs with inlets on top and swept wings were the most stable, and it showed that at least on top of the airframe, the sharp corners were hurting. Taking all of that into account results in this design.
View attachment 244705 View attachment 244702
View attachment 244703
View attachment 244704
It is actually more aerodynamically efficient than my previous designs, but more importantly... IT'S STABLE FOR MOST FLIGHT ATTITUDES! About 1/6 as stable as a more conventional layout to be more specific, which should be good enough to fly!
View attachment 244709
The inlet placement was also changed and moved a bit farther forward since the inlet placement on some of my earlier designs seemed to be disrupting the vortices generated by the fuselage/leading edge extension in a way that hurt flight performance and stability a lot. On the other hand, the wing is designed with 3 degrees positive incidence relative to the fuselage, so there is essentially no vortex generated when the plane is just cruising around which reduces drag. The first picture is at an angle of attack of 5 degrees which is the trim condition for steady flight at 45 mph and the second picture is at the same speed but at 25 degrees angle of attack (the aircraft pulling a hard 6g turn).
View attachment 244708
View attachment 244707
Since it's designed to be powered by a single 64 mm fan, I designed it to use a more conventional inlet and outlet for the EDF which should reduce losses and increase thrust by a ton.

In the very likely even my current build doesn't fly, or crashes for one reason or another, this is the next version I plan to print.

Are you going to put landing gear on it? If you do, suggest you have it set it about 1 -3 degrees nose up.

 

[telnar1236]

Are you going to put landing gear on it? If you do, suggest you have it set it about 1 -3 degrees nose up.

I'll probably have to. I'd really like some kind of launch dolly or maybe a bungee launcher or similar so the gear don't mess with the aerodynamics but it might get too complex and anything larger will need to be capable of ROG takeoffs, so why not start small. That's a good suggestion, I'll almost certainly be taking it

 

[L Edge]

Found my spare landing gear that would match your system and real cheap.

It weighs 53.1 grams and cost $12.69 that includes a servo for nose wheel steering(MotionRc store) and is designed strong. I used one for my SR-71 and can help you out of trouble on tracking and landing. It is a Freewing Panther 64mm F9F Panther and actually used on 3 or 4 other 64 mm planes.





To get a nose up attitude on the plane, either shorten and rebend the mains or my case, I straighten the nose wheel. I used basswood inside to stiffen the planes frame and super glued it. Use the unglue bottle and you can get the gear back.

Go to #36:

FTFC19 Lockheed X-44 Manta

Any updates?

forum.flitetest.com

go to thread #36, finally found my sled. That was early attempt, never tried mix of aileron/rudder like I did for my X-47B due to TVNozzle. I did a lot of crashing trying no vert stuff. I think it would have worked.

At least with gear, and use the Dark Star approach of never going high that a foot, you would get to do reruns, after look at videos.
Just a thought.

 

[badpilot27]

I think you could try adding some dihedral and twist the wings so that they have a little bit of induced thrust. if the plane yaws one wing will have a lower angle of attack then the other so it crates more induced thrust which stabilizes the plane.

I haven't tried this on any of my planes since I don't know how to do math but I think it could work. The horten flying wings had wing twist and they did make a glider which had a 60 degree wing sweep that was stable.

 

[telnar1236]

Found my spare landing gear that would match your system and real cheap.

View attachment 244761

It weighs 53.1 grams and cost $12.69 that includes a servo for nose wheel steering(MotionRc store) and is designed strong. I used one for my SR-71 and can help you out of trouble on tracking and landing. It is a Freewing Panther 64mm F9F Panther and actually used on 3 or 4 other 64 mm planes.





To get a nose up attitude on the plane, either shorten and rebend the mains or my case, I straighten the nose wheel. I used basswood inside to stiffen the planes frame and super glued it. Use the unglue bottle and you can get the gear back.

Go to #36:

FTFC19 Lockheed X-44 Manta

Any updates?

forum.flitetest.com

go to thread #36, finally found my sled. That was early attempt, never tried mix of aileron/rudder like I did for my X-47B due to TVNozzle. I did a lot of crashing trying no vert stuff. I think it would have worked.

At least with gear, and use the Dark Star approach of never going high that a foot, you would get to do reruns, after look at videos.
Just a thought.

I'll probably just end up 3D printing it, but will definitely keep this in mind as a backup

 

[telnar1236]

I think you could try adding some dihedral and twist the wings so that they have a little bit of induced thrust. if the plane yaws one wing will have a lower angle of attack then the other so it crates more induced thrust which stabilizes the plane.

I haven't tried this on any of my planes since I don't know how to do math but I think it could work. The horten flying wings had wing twist and they did make a glider which had a 60 degree wing sweep that was stable.

I've looked at doing both and the wing twist definitely helps. You can kind of see it in the pictures of the most recent version, but the root of the wing has an incidence angle of +3 degrees and the tip and -8 degrees. I don't think it will be enough to give me induced thrust with full span elevons on a wingspan this short, though, so I will still need drag rudders unfortunately. This seems to have been proven by a recent test I did with a foam board design.

It had plenty of power (a TWR of more than 1) and so long as it was flying straight it was fine, but the roll control was very unpredictable and would change between being reversed or not which made it unflyable. Like the design I plan to print, this wing had 11 degrees of twist from root to tip which seems to have helped in that the was not always reversed but not enough to make it flyable. I think mixing drag rudders into the ailerons should fix this in the final version.

As for the dihedral, my first version had a few degrees positive dihedral, but it turns out that with a tailless design like this, dihedral is actually destabilizing and what you want is anhedral. I don't really understand why this is, but both my initial XFLR5 predictions and later and more accurate Simflow runs show that even a bit of dihedral kills the stability of these designs.

On this plot, the steeper the slope, the more stable the aircraft is in yaw. The pink line is with dihedral while the grey line is without and the very dark blue line is with anhedral.

 

[badpilot27]

Just wondering, what cfd and simulator do you use?

 

[telnar1236]

Just wondering, what cfd and simulator do you use?

Just replied to your thread asking about that, XFLR5 for most stuff, SimFlow for anything more complex

 

[badpilot27]

oh ok, thanks

 

[L Edge]

Just replied to your thread asking about that, XFLR5 for most stuff, SimFlow for anything more complex

Have you ever video your models that you either hand chuck or launch in windy conditions?
I usually try 2 conditions:

1) Try to get a 10mph+ straight wind to simulate a moving run down the runway. I throw underhand so it's no more than a ft off the ground. Worked for my A-10 Warthog. Figured out from videos.
2) Then try a gusty 5-10mph to see what happens. Especially since there is no rudder. Did that for the X-47B and once I homed in, moved to tall grass area for tuning where crashes into the weeds did only cosmetic damage.

Good way to show stability of your model.

 

[Piotrsko]

Very impressive, but I'm beginning to suspect your software is fudging some data back to you since you are "down in the weeds" as we used to say where the numbers are so tiny as to be unrealistic.

 

[telnar1236]

Have you ever video your models that you either hand chuck or launch in windy conditions?
I usually try 2 conditions:

1) Try to get a 10mph+ straight wind to simulate a moving run down the runway. I throw underhand so it's no more than a ft off the ground. Worked for my A-10 Warthog. Figured out from videos.
2) Then try a gusty 5-10mph to see what happens. Especially since there is no rudder. Did that for the X-47B and once I homed in, moved to tall grass area for tuning where crashes into the weeds did only cosmetic damage.

Good way to show stability of your model.

Definitely did some of that earlier on, playing around with various models till I got something that glided well fairly reliably. I think there are some pictures of them in the first few posts in this thread. My problem has mostly been going from foam board chuck gliders to something with power that I can control. My guess is the gliders naturally fly at a higher angle of attack which makes them more stable. Not shown in the video was the about 20 glide tests of this EDF I did too prior to trying to launch with power, but it just doesn't fly for long enough or with enough speed for the issues to show up.

 

[telnar1236]

Very impressive, but I'm beginning to suspect your software is fudging some data back to you since you are "down in the weeds" as we used to say where the numbers are so tiny as to be unrealistic.

There is definitely some of that going on. After several failed attempts, I got to the point I decided I didn't trust XFLR5 anymore which was why I switched to SimFlow, and I have continued to build foamboard chuck gliders to verify the predictions as much as possible. With XFLR5, it was definitely the issue you were talking about, where the small stability of the wing got lost under things I just wasn't modeling. However, since I switched to SimFlow, all it's predictions have pretty much been right, and where I run into a problem, like the adverse yaw issue on my most recent version, I can go back in and see that had I looked for it, SimFlow would have predicted it. There are certainly some issues with my mesh since the free version has a limit on the number of nodes which produces some noise, but by running a sideslip angle sweep and using common sense, I can mostly get around it. It's also worth noting that the numbers aren't actually that small, a Cn beta value for a normal plane could very realistically be 0.05, and my value is only 0.009, so the aircraft is pretty comparable to a normal plane. The numbers just look small because of the way they're calculated.
So far as my comment about dihedral, I should clarify that only applies to yaw stability, adding dihedral still increases roll stability a fair bit. It also has been proven right repeatedly by chuck gliders I built, since when I originally got the result, I didn't trust it at all.
Getting into the weeds is the purpose of this project, to a certain degree. The eventual goal is to figure out a technique that lets me design tailless low aspect ratio planes reliably without needing to go through a crazy design process each time, and CFD is just one of a few tools I'm using to do it. While I don't trust it completely, without it, I could never have gotten from a plane that spins out in seconds to a plane like the one in my earlier video which is stable except for problems with adverse yaw. Some of its more counter-intuitive/less obvious predictions (anhedral, getting rid of the sharp edges on the top of the fuselage, and putting the inlets on top and further forwards) were the key things that made the difference.

 

[Piotrsko]

Sorry: into the weeds is more meant to be the weedy section of this data not the design weeds. Still not sure if this is a model or full scale which used to make a difference. No longer fluent in wind tunnel data or sim software.

Actually, I expect after you get one that works, you will be able to design many without the iterations.

 

[telnar1236]

Sorry: into the weeds is more meant to be the weedy section of this data not the design weeds. Still not sure if this is a model or full scale which used to make a difference. No longer fluent in wind tunnel data or sim software.

Actually, I expect after you get one that works, you will be able to design many without the iterations.

Makes sense. The chuck gliders are by necessity mostly not full scale, though I have built several ones at the size of the rc plane I plan to build. The CFD is also run at the scale and speeds, pressures, and air densities I plan to fly, so it should have aerodynamic similarity so far as Reynolds number goes. The scale definitely still does make a difference, but if you input the right data in the software, you should get results that apply to your scale.