[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[SnappingTurtle]

How should I balance gravity and aerodynamic losses for a rocket launch? Should I limit Q to some number?

[ferram4]

@NathanKell & Dragon01: Took a look at it; I know what's wrong. FAR's screwing up on the decoupler calculation, since it's a very large change in diameter from top to bottom on that part. The best solution would be to manually define a FARBasicDragModel to handle it; look in the FAR readme for what to put in the config file. Give it very low drag, but a non-zero area so that things don't break. Since the strap decoupler is a very strange case, I don't think it deserves me trying to code in some type of special exception.

Also, some of the models didn't resize with that config, or resized too much. The shroud around the Mercury retropack was too big and one of the side tanks on the Atlas was much taller than the other. I don't know if that's the way it's supposed to be, but it looks wrong.

@azza276: You need to define a surface attach for both wings, so that the CoL is placed correctly. Then, if the CoL is offset with both attached, you need to "mirror" the surface attach on one of them so that the CoL is shifted properly. What it sounds like is that the right wing places the CoL properly, but the left wing places it in the proper place if the wing was clipping through the fuselage. Does that make sense? So if you flip the direction of the surface attach FAR will know to fix that.

@SnappingTurtle: If you think that you can find some way to reduce gravity losses from 1 km/s and raise drag losses above 100 m/s, more power to you. The thing is though, you won't be able to reach terminal velocity without risking a loss of control on your rocket by going ridiculously fast. There is no reason to reduce throttle except for control purposes, the atmosphere does not eat all of your dV anymore.

[NathanKell]

ferram: thanks! Will do.

Yes, the model is correct: the retropack is tiny but the spacecraft adapter goes between booster and the max width of the capsule, and the Atlas did have different-length sides.

[ferram4]

Well, I will say that it looked like the longer side tank went up to almost the top of the tank. Maybe it's just my perception, but that screams wrong.

What I meant was that the shroud didn't fit to the mercury capsule or the adapter below in the config you posted. So there was this magical floating ring around the mercury-atlas connection.

[NathanKell]

Ok, *that's* weird. I will check. Last I saw on mine it worked, but maybe I changed something else somewhere.

[Tharios]

So...I build planes, that are mostly very stable on their own. They have one vertical fin with one control surface, and upward-canted wings with one control surface each, and a single jet engine. The CoL is always centered just at the edge of the CoM sphere. I do not usually add a stand-alone SAS or ASAS module as they don't seem to be necessary on such small, nimble craft.

I am using a fresh install of the stock game except for FAR, but even without FAR, I have exactly the same issue. As much as I can tell, I've installed it correctly, as everything about it seems to function as intended despite the problem noted below.

Even so...they always roll and pitch up, very slowly to the left. That wouldn't be a big deal, except attempting to correct it in-flight is problematic at best. Even with "soft" controls active, it jumps a full 15 to 20 degrees in angle no matter how gingerly and quickly I tap the keys. There seems to be no difference at all between normal and "soft" controls. One is a nail bat, and the other is a nail bat with one less nail in it.

Activating the built-in SAS makes things worse. The gimbal, the control surfaces, and the reaction wheels all go berserk as soon as I turn it on, flinging it back and forth nauseatingly. Locking the gimbal does nothing. It all ceases the instant I deactivate the SAS. If I do add a stand-alone SAS or ASAS module, nothing changes at all in the craft's behavior at activation. It still thrashes wildly.

If I take too long to lift-off from the runway, these tendencies make taking-off impossible, as one side lifts while the other stays in contact with the ground, causing a catastrophic rollover.

The same thing happens with rockets, but is somewhat easier to deal with since the lift comes mostly from the engine and not the winglets.

This all pretty much makes the game totally unplayable, though.

I realize it's not your mod doing it, since the stock game does it just as badly. But install after install, modded or unmodded, it doesn't seem to change anything. What else could be responsible for this?

[ferram4]

The rolling issues sound like a slight asymmetry in the design, normally caused by the vehicle flexing unevenly. Add more struts to fix it.

If your planes are to twitchy under control inputs, either reduce the number of control surfaces, make the plane more stable, or use FAR's control surface GUI to set a lower deflection angle for the control surfaces.

The SAS jitters are caused by the same issue; your plane is too responsive, and the solution is to reduce the amount of control surfaces. If the SAS jitters are primarily in roll, remove the number of roll control surfaces.

All of the issues sound like design issues mostly.

[Tharios]

The rolling issues sound like a slight asymmetry in the design, normally caused by the vehicle flexing unevenly. Add more struts to fix it.

If your planes are to twitchy under control inputs, either reduce the number of control surfaces, make the plane more stable, or use FAR's control surface GUI to set a lower deflection angle for the control surfaces.

The SAS jitters are caused by the same issue; your plane is too responsive, and the solution is to reduce the amount of control surfaces. If the SAS jitters are primarily in roll, remove the number of roll control surfaces.

All of the issues sound like design issues mostly.

I...don't know how to reduce the number of control surfaces to less than 3. I need two wings to fly at all...one surface each. I need at least one vertical tail fin to keep it stable...that's another single control surface. Is it even possible to build a craft with less than three flaps?

The strutting may be an issue, but they don't otherwise seem to be flexing much if at all, but I'll try it.

[ferram4]

What kind of craft are you flying? It sounds like you just need to shift the wings back a little bit and set up separate, smaller control surfaces for roll so that it isn't capable of doing 10 ailerons rolls per second. If you're building your plane out of nothing but control surfaces, you're going to have a bad time; use unmoving wings more then.

[SnappingTurtle]

I have the same twitch controls for my tailed planes. Reducing the maximum control surface deflection isn't ideal because it reduces the control you need for low-speed landing. So for flying tailed planes I pretty much always use trim (alt + w/a/s/d/q/e, and alt+x to zero trim).

[Guest]

Perhaps you should limit the control surface deflection? That usually dampens out the input quite a bit.

Also, Ferram, could you provide me with a sample drag config for a 1.8m by about 0.1m diameter disk (in other words, that retropack decoupler)? The FAR drag config has a lot of values, none of which seem to directly relate to part radius.

[azza276]

Ok, getting there. Got the CoL of the wings sorted, (orientation of the attach nodes on the +/- x plane. That being said, should the x-plane direction be so that the wing is physically correct in orientation IRT the surface.

e.g. left wing X plane is - toward the tip from the root. set node orientation to x+ )to toward the surface to attach to and the wing is in the correct physical orientation. Obviously when orientation is x- the wing the upside down (but the orientation is pointing toward the tip) Just want to see what is the best (correct) orientation for FAR.

But, I have tried with both and it still want to fly backwards. I have confirmed no unattached stack nodes, all stack nodes names correctly, CoL is behind CoM (maybe a little too far, but it flips when pitching up, not when flying and nosing down). Its like the nose has too much drag but there are no unassigned stack (or attach) nodes. Drag settings seem nominal. Is there a FAR model I can use for fuselage parts?

Also is FAR affected by scaling?

Edited December 14, 2013 by azza276

[ferram4]

@Dragon01: Well, ignoring the end points, you'd just set it to have an area of ~0.254 (since it's just the outer area of the cylinder), which is the only thing affected by radius. The rest is taper-based stuff. So, since there should be negligible pressure drag (which is what is handled by the drag curves) set that to 0, ranging from -1 to 1 so that it gets all directions. Lift and moment curves should also be near 0, since it's really not that large a part, so just set those to zero as well. That should handle it fine.

@azza276: Have you set up the control surfaces properly? How about the vertical tail? If those aren't set up properly there might still be issues.

FAR is only affected by scaling on the fuselage parts, but those you shouldn't need to mess with.

[NathanKell]

Fixed, thanks!

Yes, I see what you mean: I guess frizzank decided to extend the side tunnels' lengths. Not getting what you mean about the decoupler though.

[azza276]

Well I gave it a good red hot go, but I still cannot get the TiberDyne shuttle flying forward... I don't like giving up and the challenge is no longer enjoyable.

Thanks for all your help Ferram. I made some positive steps, just could not get over the last hurdle I think. My lack of experience in these kinds of things and lack of knowledge of how FAR works was working against me. If I get bored again, I might look into it.

Cheers

Azza

[localSol]

For people with lots of control power who don't want to also lose surgical flippability without going back to the hangar, I was just trying FAR's built-in dynamic control adjustment feature (DCA). It works well. You can try setting it to 70 for a test, after you've left the runway. Or lower for stronger limiting. The default is 150, I don't know how to interpret the value other than in a practical way yet.

Careful to set it back if you need it or you may wonder why controls are deflecting less if you close the window. It can severely limit your control surfaces temporarily. The effect depends on the type, placement and amount of control surfaces on your craft. In other words, I have been needing to get a feel for a good setting for each craft with it.

[likke_A_boss]

Hey farram, did you hear that DYJ is back? He is updateing his procedural wings with what looks like procedural flaps. Do you think FAR will support it?

[SFJackBauer]

Ferram, does FAR read ctrlSurfaceArea from the CFG? If not, how does I should go about taking an existing control surface and doubling its size? Geometrically I know how the rescaleFactor works, but aerodynamically, which settings should I also rescale? Just double the semi-span?

Also for a fuselage, once I double its size with rescaleFactor, is there anything else I should tweak so that FAR works correctly like node sizes or its just that?

[ferram4]

@likke_A_boss: I don't see why it wouldn't work, but if it doesn't work I'll work with him to make sure it works fine. Odds are it shouldn't be that difficult.

@SFJackBauer: You'll want to double b_2 and MAC; just doubling b_2 would account for increasing the wingspan, but wouldn't account for increasing the chord length.

For fuselages you don't need to change anything.

[SFJackBauer]

Thanks ferram.

Oh man I feel such an idiot! I have overlooked the most useful feature in flying an airplane: TRIM!

I didn't knew KSP had built-in trim controls with ALT+WASD and trim reset with ALT+X. This alone makes landings so smooth I think I am back in Flight Simulator! Haha!

[Beetlecat]

Thanks ferram.

Oh man I feel such an idiot! I have overlooked the most useful feature in flying an airplane: TRIM!

I didn't knew KSP had built-in trim controls with ALT+WASD and trim reset with ALT+X. This alone makes landings so smooth I think I am back in Flight Simulator! Haha!

Ugh. You're not kidding.

That's the sign of a good game --that you can just jump right in and play without having to look up all the controls, but maybe looking at them once would have been a good idea a few space plane crashes ago...

[Goozeman]

How is wing sweep accounted for? I looked through a plane with pWings and couldn't figure it out. The problem that I'm having is that no matter how much I sweep the wing, the NP remains pretty much in the same place. Stability derivatives seem to update correctly, but that may be simply due to the control surfaces hanging way out back.

[ferram4]

Wing sweep is supposed to handle changing the placement of the aerodynamic center, zero lift drag of wings and the lift slope. Currently, I'm absolutely sure it does the latter two correctly, but I'm not convinced with the first one now that you've brought it up; it looks like it's not shifting enough with wing sweep. I was trying to shift it based on the average chord and wing sweep, but I figure it really makes more sense to use span and wing sweep now that I think about it. Should be fixed in the next release.

[ÄlveKatt]

I managed to make an aeroplane that was quite easy to fly. It's still hard to land, but that's mainly because of lack of air brakes. I did managed to land it on the ground, although not the runway ones. My speed was too high though, the engines fell off, but the rest made it. Then it wouldn't stop, I tried to apply down force, but that resulted in a crash. Gonna try adding a parachute.

[Camacha]

Have you tried simply (wheel)braking?