Aircraft yaw instability

[Cavscout74]

So I was messing around today with an aircraft I had designed a few days ago.  It wasn't for anything important, but the design seemed like it had potential when I first flew it the other day.  With further testing, however, I've found it has very poor directional stability.  My poor test pilot has either used the pilot ejection system or bailed out on every flight but one.  The first improvement was adding a pair of basic fins underneath to supplement the delta deluxe tail fin, then I switched to the AV-R8 fins with control turned down to 50%, set to yaw only.  Neither made much difference - if I get even slightly off left or right of prograde, it snaps into a spin that is only recoverable if I've got a few thousand meters of altitude.  Speed doesn't seem to matter - mach 1+ or <200 m/s, it spins equally well. 

The only other craft I've had with this much instability were a V-tail jet (which flew perfectly well when switched to a conventional tail) and a spaceplane that just needed a shallower reentry angle since the high AoA reentries I was trying to do blocked the relatively small rudder.  What am I missing on this one? 

This was the cleanest screenshot I have to show the layout.  Approximate locations of CoM & CoL are marked.  That's at full tanks, but I haven't flown it long enough to deplete too much fuel.  The homemade rockets don't effect CoM much - it moves aft a little more with all 6 plus a pair of drop tanks under the fuselage.  But it's still ahead of CoL.  And the spinning problem was noticed before I added anything too it.  It's actually pretty tame flying as long as you don't yaw more than 1 or 2 degrees left or right. 

Second version, loaded, with the basic fins

Clean, version 3 with the AV-R8 fins.  Yes, that is a pretty high angle of incidence on the canards, but the CoL was nearly at the back of the wing before I increased the canards incidence

 

 

[bewing]

This is not with FAR aerodynamics installed, right?

I think the canards are the problem. They need to be built with parts that provide more lift and therefore much less incidence. Or you need to add dedicated lifting winglets to the front end. The instability has nothing to do with CoL. The problem is your Center of Pressure/Drag moving in front of the CoM when your canards deploy upward. Your canards are 100% control surfaces, and therefore have relatively low lift. But very high drag, which increases quadratically with incidence + deployment angle. The drag from the canards, multiplied by the radial arm of the distance to the CoM gives a lot of random torque to the craft. But you have a lot of control in both roll and pitch -- so those don't cause you problems.

 

Edited August 27, 2019 by bewing

[lemon cup]

Are you flying with SAS on? If so, try turning it off and setting pitch trim to about +20 or so (depending on your planned flight profile) to keep the nose up, instead of letting SAS use all of your control surfaces to do it and applying yaw when you don’t need it.

[T-Bouw]

Bewing is right on the money in my opinion and gives you 1 way to look at the problem, so I'll give you the reverse solution since you seem to want to make an agile plane. Make your vertical stabilizer bigger.
Notice how fighter jets with canards almost always have equally impressive vertical and/or horizontal stabilizers. Sometimes the vertical stabilizer(s) has/have almost as much surface area as a single wing! This is so the planes have authority in the relevant axes in almost any situation, sometimes even spins.

You have insane pitching ability because of the canards, ailerons and the big wings (yes, the wings help too, remember the Fokker triplane?).
Rolls might be also fairly high because of the 2 sets of ailerons and possibly the canards if you didn't disable roll for them.
For yaw you only have the single little surface. See the pattern here?

So add a nice big back fin, or two. That should make the jet naturally stable. If it reacts very aggressive when you use yaw, only then dial back the angle of actuation. This last bit is also true for pitch and roll. Dial back and/or limit the number of participating surfaces if the plane is too aggressive handling.

[Cavscout74]

Thanks for the responses everyone.  I switched from the all-moving canard to a two piece stabilizer & control surface like @bewing suggested & it flies pretty normal now.  It's still a little weak in yaw - perhaps a larger tail would help - but even getting 5 degrees or so yawed left or right, it settles back to prograde within 2 seconds instead of going the other direction instantly.  I didn't have to eject once on this morning's test flights. 

 

[Klapaucius]

4 hours ago, Cavscout74 said:

Thanks for the responses everyone.  I switched from the all-moving canard to a two piece stabilizer & control surface like @bewing suggested & it flies pretty normal now.  It's still a little weak in yaw - perhaps a larger tail would help - but even getting 5 degrees or so yawed left or right, it settles back to prograde within 2 seconds instead of going the other direction instantly.  I didn't have to eject once on this morning's test flights. 

 

I use canards on the front all the time. The thing is to play with the authority limiter until you achieve a good balance between stability and maneuverability.  Also I spend a lot of time playing with the initial position and (non-deployed) angle of the canards.  Small changes have a significant impact on performance.

 

In addition,  I have found that tail fin you are using (the Delta Deluxe) is pretty useless. The rudder is far too small to do much of anything.  On one of my craft--which uses a lot of the same parts (KerbalX link here) I use twin C-7s, in which the entire surface moves.

 

Here is my craft in flight, to give you an idea. Very maneuverable but very flyable. In the video I have them set for full movement so I could barnstorm, but I can reconfigure the canards (which are larger than yours) by lowering the authority, and the plane becomes much more forgiving.

 

 

Edited August 27, 2019 by Klapaucius