Flying without a vertical stabilizer

[telnar1236]

Any video?
Interesting
View from rear of plane?

No video, unfortunately, not that there was much to video. Mostly just a lot of spinning into the ground. I tend not to video the maiden flight so I can focus purely on flying, and then video the second or third flight.
Here is the rear view of the plane. The 3D printed version isn't in great shape so this is a picture of the CAD.

It doesn't look like much, but it has 3 degrees dihedral. In comparison, here is the rear view of the XFLR5 model with 5 degrees anhedral.

 

[L Edge]

Question on pic #11, could you explain the EDF, exhaust tube and exhaust location to the outside of the plane?
From pic, looks like EDF in front of coned exhaust and dump's into chamber, then out of 6 sided shape with this view?

 

[telnar1236]

Question on pic #11, could you explain the EDF, exhaust tube and exhaust location to the outside of the plane?
From pic, looks like EDF in front of coned exhaust and dump's into chamber, then out of 6 sided shape with this view?

It's nothing too fancy. Since I was trying to just make a quick prototype, I didn't really focus of getting the EDF thrust tube to go in the locations a lot of the concept art I've seen shows. It just exhausts directly out the back of the plane. I didn't bother to fair in the back too well around the thrust tube for the same reason, so there's a bit of a gap to either side of it, but nothing too wild.

 

[telnar1236]

Doing some more investigation with simulations, the dynamic stability is very weird. It doesn't show any oscillatory modes (no Dutch roll). It might also explain a bit more of the discrepancy between the chuck gliders and the RC version. Long story short, there is a rapidly divergent condition that gets worse the heavier the plane gets. Red is for a plane that weighs 60 grams and gray for a plane that weighs 600 grams.

I somewhat doubt the accuracy of this analysis since it seems strange that I wouldn't get a Dutch roll mode. However, I do think the divergent mode is real and kind of a reverse spiral divergence where the plane rolls a bit, starts to side slip, and then yaws more dramatically into that side slip due to a lack of vertical surface area to resist it. The roll stability then makes it roll more and it gets progressively worse.

 

[L Edge]

It's nothing too fancy. Since I was trying to just make a quick prototype, I didn't really focus of getting the EDF thrust tube to go in the locations a lot of the concept art I've seen shows. It just exhausts directly out the back of the plane. I didn't bother to fair in the back too well around the thrust tube for the same reason, so there's a bit of a gap to either side of it, but nothing too wild.
View attachment 244114
View attachment 244115

Ok, now I understand what you did.
Take a look at the latest video on Flight Test's Youtube site dealing with the "Will this 3D printed Horten fly?" (put up one day ago)
Very interesting.

 

[telnar1236]

Watched the video. Looks like another great design from Eclipson. Not sure how much it applies to what I'm trying to do though. It looks like it has a much higher aspect ratio wing which is a lot easier to get stable without a tail, and, crucially, a clear vertical stabilizer. I know the Prandtl v2 from Eclipson does completely do away with a vertical stabilizer, just like its namesake, but it's still also a very high aspect ratio design.

 

[L Edge]

All I am trying to say is, when you add "EDFs" to your model from motor/props, you need to add taken a look at summation of forces besides just air flow study. When they added the 2 EDF's, they had to resort to the vertical component to make it stable. Again, the roll/yaw exist to a degree in any plane. In fact, if the Prandtl v2 was changed to an EDF, it would have stability problems. You need to look at all forces.

Another is the SR-71, plane works good with props/motors, not so good with EDF's. Same with A-10. Need to solve roll/yaw into the summation.
HH solved the roll/yaw and lessoned wobble on the SR-71 by using differential thrust, that why you don't have any rudder.

This I did years ago, if a flap, not rudder(double deflection) could be printed and more reliable, could be the answer by opening both to control yaw/roll and closed together to work as one like a flap.

For rudderless NYGAD single or dual EDF's, problem become much harder because of forces generated by stealth EDF exit not being symmetrical. Look at a Harber's YF-23, it pitches up.

 

[L Edge]

Here is another single EDF wing, certainly has yaw/roll problem. From NACA, designed a staiblizer, took away vertical components, added a 64mm 5 bladed 1300mah battery and with adjustments to the EDF, my summation of forces (don't even use reflex at all) overcame the roll/yaw and reduced the drag so I can get a 10 plus minute flight(ESC cut it off) and now it's capable of stable flight even when gusts hit it in enclosed airfield.

So now I found 2 ways to solve roll/yaw except that on a NGAD plane, stabilizer not allowed. Even the TV can't be used with the odd shaped exhaust.

 

[telnar1236]

In terms of adding EDFs I think it normally actually makes the plane behave much closer to the simulation. Props create complex flows behind them that are much more difficult to deal with. However, props do keep air flowing over the wing which delays stall and makes the plane fly better normally. In terms of pitch-up, that's generally caused by non-linear aerodynamics which are also much harder to predict. It's why I've done so much testing with chuck gliders for this, only using CFD to point me in the right direction. It is easier to mess up thrust line with an EDF, but if you make sure your thrust line goes through the CG you're mostly good

 

[telnar1236]

Here is version 2 of my tailless plane. I have been using the acronym TSTB (tailless stability test bed) and it has been dubbed the Testy Bee based on that and a tendency of the version 1 plane to buzz when given throttle (I think the skin somewhere vibrated at a similar frequency to the EDF). It uses the new wing design which should be much more stable. I have also increased the size to a 64mm EDF and moved the inlets to the bottom to help give it more power. And the exhaust is now more similar to that in a lot of the concept art floating around. I would like to go back to a thinner wing in the final version, but for now this is what seems most likely to get something flying.

I also hot glued the version 1 back together last weekend, and a friend hand-launched it. This made a pretty big difference. One of the wings had warped from the sun which resulted in a strong right roll tendency I couldn't, but it didn't spiral at all with the faster launch, so this new version will also have fixed gear to let it get off the ground at a higher speed.

 

[Piotrsko]

not using the control surfaces for augmented stability or are you satisfied with the concept as is?

 

[telnar1236]

not using the control surfaces for augmented stability or are you satisfied with the concept as is?

Not totally sure I understand what you're asking. But I think I am not. My goal is to try and get something stable purely through aerodynamics (if possible) without a vertical stabilizer. I definitely could add in additional surfaces, or a gyro and drag rudders or thrust vectoring, but that would somewhat go against what I want to achieve. But if this version doesn't work, then time for the gyro it is

 

[L Edge]

not using the control surfaces for augmented stability or are you satisfied with the concept as is?

In his design above, he just added a new force problem to the equation., He is going to have a pitch-up as soon as he is flying.
The reason no one has designed NGAD planes is the two problems, doesn't have any rudders(roll/yaw needs to be solved) and designs of exit EDF exhaust of weird shapes cause additional forces that need to be balanced out..
One sample of exhaust shape testing that I did on my testbed plane has the plane doing a loop at full throttle.

This is why I picked this model.

The center segment between the edf's may just become a pitch control flap(gyro?) and the rectangular spline exhaust I am exploring to see what forces exist and see if I can redirect/null out. Otherwise, it is going to pitch up.

 

[telnar1236]

In his design above, he just added a new force problem to the equation., He is going to have a pitch-up as soon as he is flying.
The reason no one has designed NGAD planes is the two problems, doesn't have any rudders(roll/yaw needs to be solved) and designs of exit EDF exhaust of weird shapes cause additional forces that need to be balanced out..
One sample of exhaust shape testing that I did on my testbed plane has the plane doing a loop at full throttle.

This is why I picked this model.
View attachment 244169


The center segment between the edf's may just become a pitch control flap(gyro?) and the rectangular spline exhaust I am exploring to see what forces exist and see if I can redirect/null out. Otherwise, it is going to pitch up.

Out of curiosity, what will cause the plane to pitch up? I've been trying to puzzle through, and I don't think I'm following your reasoning.

 

[telnar1236]

I think it's time for a bit of a redesign. I was playing around the wing a bit more in XFLR5 and came up with something roughly 3 times better than the earlier design that uses a thinner wing and doesn't require any anhedral. It's finally to the point that the version with the vertical stabilizer is not a vertical line anymore. The design I tried to fly is the nearly horizontal pink line, the old new design is the blue lines, the new new design is the gray line, and the design with a vertical stabilizer is the orange lines.

The biggest change was sweeping the trailing edge, but I also tweaked the geometry of the longer central section.

It's starting to look like some of the art released by Northrop Grumman, which is what I eventually want to build and was part of what inspired me to play with the aft swept trailing edge.

 

[telnar1236]

The more I look into designing this plane, the more I realize it will need a gyro/augmented stability. Doing some trim analysis it looks like most of the stable designs I can come up with go unstable at low or negative angles of attack. Most of my analysis has been performed for a trimmed steady flight condition (the arrow) at 40 or 50 mph or for higher angles of attack like during take-off or landing. From the analysis the plane should still be stable without a gyro to its top speed, but I can't guarantee I won't unload the wing during flight resulting in a loss of control.

So, I need to decide between trying to design some drag rudders or thrust vectoring. I think drag rudders are ultimately the best option since one of the most likely times for me to unload the wing would be during landing with low throttle and so not much vectoring authority to correct for the reduced stability.

 

[L Edge]

Are your drag rudders going to be automated or manual?

 

[telnar1236]

Are your drag rudders going to be automated or manual?

The plan is to make them automated, probably just attached to a cheap gyro to start with. I don't think my thumbs are fast enough to keep up with this design

 

[telnar1236]

Here is the work in progress newest design. It's mostly similar in concept to the previous one that I didn't end up building but incorporates the new wing geometry. I also switched the inlets back to the top of the plane and added pockets for drag rudders.

To avoid any pitch up from the location of the EDF exhaust the inside is carefully shaped to direct the air straight out the back. I'll lose a bit of thrust, but should still have a TWR of about 1 with a 64mm EDF on 4s.

 

[telnar1236]

In the interest of not having my plane smash into a million pieces each time it crashes, I decided to try and make a foam board version. Still no luck getting it to fly, so I started to suspect that the simulation tool I was using wasn't capturing the whole picture and decided to run a simulation of the whole aircraft using Simflow instead of relying on the simplified XFLR5 model. As I had realized, the XFLR5 model is simply not adequate for this kind of work, which means all of my previous results are at least somewhat suspect. My current design, that I was in the process of 3D printing is in fact very slightly unstable in the yaw direction, and many of my previous designs were fairly unstable. Here are a couple of graphs, one of the current 64mm version, and one of a 70mm version I have been working on.

The old 70mm version is about 3 times as unstable but mostly sees this instability past 4 degrees sideslip. These results also very clearly show that I will need to use augmented stability of some sort or other. That immediately means that building the avionics has now become the most difficult part of the project and that I won't be able to fit the electronics and sensors into any plane smaller than a 6s 70mm jet, and that I might even need to go to an 80mm design to carry everything. The good news here is that neither one is that wildly unstable, so my next step is going to be sizing drag rudders so that they can easily counteract the instability.

Having more complete CFD, also showed where my earlier approach was going wrong. Broadly, flow separates from the cockpit canopy even at smaller sideslip angles and disrupts airflow on the downstream wing. Having the inlets above the wing somewhat reduces this since some of that separated flow is redirected into the inlet, but it still isn't great.

It also looks like having the sharper corners on the cross-section aft of the CG doesn't really help and may even have been hurting.

However, it does look like I was right about the leading-edge vortex coming from where the wing sweep changes having a stabilizing effect, although this is mostly in the roll direction. The vortex coming from the leading edge on the outside of the turn is strengthened and the fuselage keeps it from spilling over as much onto the inside wing.

The good news here, is I now have a much better idea of how much work will actually be needed and how long a road this process could be. The bad news, of course, is that it is enormously longer than I was hoping.
One other cool thing that came out of this CFD is one of the clearest streamline images showing vortex lift generation I have ever seen. When a delta gets to a high angle of attack, it generates a powerful vortex from its leading edge that helps keep the flow attached, which is why delta wing planes are so good at high alpha. Here a couple of plots showing the vortices at 30 degrees angle of attack.

Vs. the same view but at 7 degrees angle of attack. There is still a small vortex at the smaller angle of attack, but it is far weaker and is not the primary contributor to lift generation.