What does dihedral actually do?

[AeroGav]

I give up ! All I want to do is make an airplane that has a tendency to return to wings - level if disturbed to a few degrees of bank angle,  one that does not constantly need to have uncommanded roll angles taken out by hand,  and can stay on a steady 90 degrees due east heading without continual course corrections.

I build a simple test plane

• big tail fin, yaw stability
• plenty of dihedral

Despite the above, roll angle gets bigger the longer i leave it uncorrected, plane would spiral dive into the ground if i let it.

However, the smallest elevons are having trouble putting any bank on that i do ask for, it's almost impossible to turn this plane at all without the rudder - so maybe it is doing something after all?

 

Edited April 21, 2016 by AeroGav

[regex]

Get FAR if you want realistic aerodynamics.  Vanilla uses a pretty good system for vanilla craft but it's fairly simple in its simulation.

[bewing]

Try using the offset gizmo to raise your wings significantly above your CoM.

[lobe]

While stock aero is simple, it does a good job at making airplanes look like airplanes in the subsonic regime. I think I rebuilt your airplane in my 1.1 sandbox save, and found it had neutral roll stability and the roll rates were pretty good using the Elevon 1 or 4 (I am not sure what you meant by the smallest), no need for rudder. That is, unless you expect the aircraft to complete a 90 degree course change in under 10 seconds. If you want to build a positive stability aircraft, I would suggest experimenting with Learjet- or F-86-alike aircraft. Though I think the best design would be something like a Cessna 152 or 172 with high wing, high dihedral.

[AeroGav]

3 hours ago, lobe said:

While stock aero is simple, it does a good job at making airplanes look like airplanes in the subsonic regime. I think I rebuilt your airplane in my 1.1 sandbox save, and found it had neutral roll stability and the roll rates were pretty good using the Elevon 1 or 4 (I am not sure what you meant by the smallest), no need for rudder. That is, unless you expect the aircraft to complete a 90 degree course change in under 10 seconds. If you want to build a positive stability aircraft, I would suggest experimenting with Learjet- or F-86-alike aircraft. Though I think the best design would be something like a Cessna 152 or 172 with high wing, high dihedral.

 

http://kerbalx.com/AeroGav/Tundra-Goose

This STOL spaceplane I was working on also seems to have negative roll stability, despite dihedral on outboard wing segments at 30 degree angle.    Wonder if you can confirm that's actually the case or it's just my flying that's off.

 

[lobe]

First I need to ask, did you strut this before putting it on kerbalx? If you didn't a lot of your roll (and probably pitch) problems were caused by this. Second, is there a specific regime where the negative roll stability is most pronounced? I had SAS off, torque off, set throttle to about 50% at 2000 meter, my speed was 90-120 m/s, then I entered a 30 degree bank and let it fly itself out, only touching pitch to maintain altitude because the aircraft is nose heavy. So no, your aircraft has positive roll stability. 

It actually looks like kerbalx might have destroyed all of the non-physical connections, no fuel to the nukes (if you forgot, they are stage fed, not even-fed like the jets and RAPIER. Though before I ran out of fuel and air the ship got to 25000 meters at 800 m/s, with only slight yaw deviation, probably the tiny asymmetric thrust or from a slight bank on on take-off.

 

Edited April 21, 2016 by lobe

[AeroGav]

49 minutes ago, lobe said:

First I need to ask, did you strut this before putting it on kerbalx? If you didn't a lot of your roll (and probably pitch) problems were caused by this. Second, is there a specific regime where the negative roll stability is most pronounced? I had SAS off, torque off, set throttle to about 50% at 2000 meter, my speed was 90-120 m/s, then I entered a 30 degree bank and let it fly itself out, only touching pitch to maintain altitude because the aircraft is nose heavy. So no, your aircraft has positive roll stability. 

It actually looks like kerbalx might have destroyed all of the non-physical connections, no fuel to the nukes (if you forgot, they are stage fed, not even-fed like the jets and RAPIER. Though before I ran out of fuel and air the ship got to 25000 meters at 800 m/s, with only slight yaw deviation, probably the tiny asymmetric thrust or from a slight bank on on take-off.

 

Thanks for checking.   I'm coming to the same idea.   It starts to wander off heading by the time it's flown the length of the KSP runway, couple of degrees out, then as the nose rises to 30 or 40 degrees a wing begins to drop.

Yes it needs a fuel balancer mod to keep the feeder tanks for the NERV topped off. 

I've not used fuel lines or struts because I heard they had VERY high drag levels - couple times more than the cockpit.

I'll probably have to take down that version of the ship on KerbalX - it's too unstable in high dynamic pressure situations.  Got to about mach 2 @ 16km or so then when I brought the nose up 5 degrees off the prograde it suddenly snaps in to an 18g pull up and a dramatic reconfiguration of the airframe 

So, it turns out you got that thing closer to space than i ever did.

I've tried shifting the CoL further aft and making the canards smaller, trouble is the max AoA you can now achieve is 6deg, costs us most of our STOL capability....

[lobe]

It didn't even take that many struts. I used 8, but 6 should be fine. Honestly, unless you're going for a challenge in efficiency, struts and fuel lines aren't that bad as long as they aren't over done. The amount of fuel that is wasted through flappy wings or wobbly rockets is probably an order of magnitude greater than what drag the struts add.  Your SSTO had tons of fuel left in it when it quit, probably 80% or greater, and that was the strutted version.

[I_Killed_Jeb]

In 1.0.5 I found that dihedral wings actually decreased stability, esp. roll stability, at any trans-sonic speed. Dunno yet about 1.1, waiting for the next hotfix before jumping in

[AeroGav]

1 hour ago, lobe said:

It didn't even take that many struts. I used 8, but 6 should be fine. Honestly, unless you're going for a challenge in efficiency, struts and fuel lines aren't that bad as long as they aren't over done. The amount of fuel that is wasted through flappy wings or wobbly rockets is probably an order of magnitude greater than what drag the struts add.  Your SSTO had tons of fuel left in it when it quit, probably 80% or greater, and that was the strutted version.

I'll keep on iterating the design till the coffee runs out.     Mark II  has the following changes from the one you flew

• two standard canards replaced by three advanced canards, clipped to make a blended wing body effect in the nose
• two big-s strakes removed, replaced by the small strakes to seal up the gaps.  reduces lift at the front end a bit
• i've been repeatedly urged to add incidence to my wings. After opening the aerodynamic debug i was shocked to see how much drag the body was generating, so i added a few degrees by hand.  Hard to do on a complex design like this.   Body drag is indeed lower, but not sure overall picture is better.  I've always stayed away from this for fear of making stall/departure characteristics worse on a canard design.  Still only getting Lift drag ratio of 10 (subsonic) to  3 (high speed).  But I must admit, it's nice being able to fly much of the mission with SAS set to Prograde and forget about it.

Next thing to try is struts , as you say.

BTW I am using GPOSpeedFuelPump to keep the engine pre-coolers, which serve as feeder tanks for the NERV, topped off.  It was the first fuel balancer to become available for 1.1.0  and I like the way you can set it up in the SPH/VAB and the settings are kept.  TacFuelBalancer had to be switched on every time you loaded a quicksave.

First orbit in 1.1.0

[pincushionman]

Keep in mind, lift and drag forces are applied AT THE WING PART ROOT.

NOT at the 3d center of pressure of the wing segment.

So, unless your wing is made from multiple segments along the beam (a la the Albatross), your moment arm is effectively only half the width of the fuselage, and dihedral will be FAR less effective than in reality for that reason. God forbid you're trying to make big wings with the FAT-455's and hoping for good dihedral.

Squad really should fix this. 

[AeroGav]

43 minutes ago, pincushionman said:

Keep in mind, lift and drag forces are applied AT THE WING PART ROOT.

NOT at the 3d center of pressure of the wing segment.

So, unless your wing is made from multiple segments along the beam (a la the Albatross), your moment arm is effectively only half the width of the fuselage, and dihedral will be FAR less effective than in reality for that reason. God forbid you're trying to make big wings with the FAT-455's and hoping for good dihedral.

Squad really should fix this. 

Yeah I'd heard of that - in all my spaceplanes, the inboard wing sections have 0 dihedral, but the outboard has 30 degrees.   The test craft i pictured also had outboard wing dihedral (gull wing style), but that V shaped section under the main fuselage was below the CG, making it a low wing aircraft which would counteract the effect.

 

As regards the spaceplane,  2 pairs of struts seems best.  I am tieing together the two wing segments furthest apart on the craft file tree structure, but the really key strut is the one that ties the two tailcones to the cockpit.   The problem was that at high speed the lift forces were torquing these upwards,  and the forward 2 wing segments are mounted to this part - the two front wing segments therefore get increased AoA and more lift than the rears, and act as giant canards.  

I've noticed that the reluctance to respond to roll input may be down to the trailing edge wing segments torquing up when using ailerons to try pick the wing up.   This results in the wing as a whole angling downwards generating less lift, counteracting the effect of the aileron. Or it's just an adverse yaw resulting from aileron drag, not sure which yet.

Edited April 22, 2016 by AeroGav

[lobe]

42 minutes ago, pincushionman said:

Keep in mind, lift and drag forces are applied AT THE WING PART ROOT

(Since the editor won't let me post outside of this box, lobe's reply will be in the box. Wouldn't this mean to have dihedral effect you would need 2 wing segments, the inside one attached to the flat plane on an angle, with the outside one providing the root lift force?)

FREE MEEEEE!!!!!!!!!!!!!!!!

[Boris-Barboris]

5 hours ago, pincushionman said:

Keep in mind, lift and drag forces are applied AT THE WING PART ROOT.

Simply not true.

13 hours ago, AeroGav said:

All I want to do is make an airplane that has a tendency to return to wings - level if disturbed to a few degrees of bank angle,  one that does not constantly need to have uncommanded roll angles taken out by hand,  and can stay on a steady 90 degrees due east heading without continual course corrections.

Then you're in need of shoulder or high wing with very slight (1-2 degrees) or no dihedral. Wing angle defines your craft's reaction to non-zero AoA and sideslip, it's airflow frame, if you want to change behaviour in planet reference frame, you need to play with CoM\CoL balancing = wing positioning.

[AeroGav]

8 hours ago, Boris-Barboris said:

Simply not true.

Then you're in need of shoulder or high wing with very slight (1-2 degrees) or no dihedral. Wing angle defines your craft's reaction to non-zero AoA and sideslip, it's airflow frame, if you want to change behaviour in planet reference frame, you need to play with CoM\CoL balancing = wing positioning.

The only trouble with high wing positioning, is that moving the mass of the wing upwards also raises the centre of gravity of the airplane.  Which means that engines may now be below the centre of gravity,  causing a pitch up tendency when power is applied.  This is especially noticeable if it's a wing that contains fuel, since it has a greater mass,  and especially noticeable when operating in close cycle mode - engine thrust remains high as the air gets thinner, but the control surfaces have less and less ability to fight the pitch up tendency. 

Ah well,  this is where Jebediah earns his pay.

[pincushionman]

19 hours ago, Boris-Barboris said:

Simply not true.

Okay, I'm wrong. It used to be that way, but they've changed it. Disregard my post.

[Wanderfound]

I've been getting into scratchbuilding RC planes lately. A lot of the simpler ones are three-channel designs: elevator, throttle and rudder only, no ailerons. They rely entirely upon dihedral or polyhedral wings for roll control. Like this one:

 

As you give it rudder input, the plane begins to sideslip. This pushes the "outside" wingtip into the air, generating extra lift, which banks you into the turn. Once you stop sideslipping, the inside wingtip now generates more lift than the outside wingtip, rolling you back to level.

I've no idea about stock, but the same concept does work in FAR. With enough dihedral, ailerons are unnecessary.

They're still nice to have, though; dihedral forces often don't act as quickly as you like, and on a high-dihedral airframe there is frequently a tendency to oscillate on the roll axis.