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

[Hodo]

Ok I'm having a really head scratcher of a problem with my large SSTOs.

According to the "Center Of Lift Behind Center Of Mass" Rule, they are stable, and indeed in the lower atmosphere it is indeed stable. However as I climb above 10km and over 300km, the aircraft begins to pitch and roll widely, going into a series of violent nose-up maneuvers, basically going belly-first into the airstream, then dropping the nose back down into the airstream, before repeating the process over and over again.

Can someone explain why exactly this is happening? The airplane is stable, and at no time does the CoL go ahead of the CoM.

Ok several questions in response.

1- Do you have a lot of reaction wheels on the craft?

2- Do you have your control surfaces set to specific tasks?

3- Do you have a lot of flex in the aircraft, does it wiggle like a worm through the air?

4- Is your empty fuel tank CoM behind your loaded CoM?

If you answered yes to 1, then remove some of the reaction wheels.

if you answered yes to two, set your control surfaces to some assigned tasks, like Roll, Pitch, and Yaw, but not all three.

If you answered yes to 3, then you should strut the body more, and/or download Kerbal Joint Reinforcement.

If you answered yes to 4 or I don't know, then I suggest getting RCS build aide and TAC fuel balancer, or PWB Fuel pumps.

And pictures would be helpful.

[Scootaloo]

1 - No. Just the S2 one near the front. It's B9.

2 - No. But how do I do that?

3. I think thats the problem. Even with struts it wiggles though. I gotta check and see if Joint Reinforcement is installed.

3. I'm flying with fuel balancer. So no.

[ferram4]

I'm going to assume that you meant 300 m/s, not 300 km, since there's no air up that high. In that case, congrats, you've found transonic instability; your plane is actually unstable at that Mach number and what's happening is it pitches up, the additional drag slow it down, and then the slowdown drops it into a stable regime, so you end up with this oscillatory motion. Weird behaviors like this are why no one ever wants to fly at the speed of sound, only faster or slower. What you're going to want to do is go through that regime with the lowest AoA you can manage and just try to get supersonic quickly, where your plane should be a lot more stable. As for the roll problems, that hints at needing a larger vertical tail.

As to your other questions:

Right click the control surfaces in the editor to assign control axes, just like you would have done in the stock game.

For flexing, use your struts intelligently. Don't strut two parts that are already attached together, strut together parts that aren't attached yet.

[AndreyATGB]

Is the way FAR handles this region around Mach 1 modeled correctly? Do planes really instantly get most of their control authority back at Mach 0.99-98 when dropping from 1.00? That's my experience with this at least, the second I go supersonic unless the plane is made for it, usually it's uncontrollable and starts pitching down. However the second it drops from 1.00 it is completely fine again.

Another thing, I noticed on your git that the dumb down version of FAR is called NEAR. I have to admit, it didn't cross my mind that the name is just the opposite of far (away) until in the shower last night. Have you thought of its long name too or just the acronym?

[BudgetHedgehog]

Have you thought of its long name too or just the acronym?

It better be Not Exactly Aerodynamic Research...

(I think 5thHorseman came up with that)

Edited July 4, 2014 by ObsessedWithKSP

[Roaken Corporation]

Hello, I just want to comment a few things about transonic instability.

In fact, in real life you will never want to fly at transonic speeds, but in the game there are ways to do that, as well as many reasons why you would want to fly at such speed.

After much research, we at Roaken Corporation found ways to keep the aircraft steady and fully controlable (keyboard piloting without fine controls) even when wildly breaking soundspeed and going back so subsonic over and over.

We also did manage to do that while with some big AoA, a few side effects happen though, you just will not want to break sound barrier while turning, it can destroy your aircraft. But most of them are negligible with proper polishing on the design.

Going from supersonic to subsonic is not much of a bad issue though, as seen at this point of this video (http://youtu.be/MWdYnl3hbHY?t=2m8s) you only lose a bit of pitch control, the difference on pitch is what causes it to oscilate, on a straight flight it would be steady.

There are many things (other than trial and error) that helps on keeping the aircraft controlable at such speed range (Mach 0.95~1.05), and we are interested on creating tutorials to show how to do it, or at least to teach how to not fully lose control when passing on that region.

We are still unsure if the tutorials are going to be in the form of text or videos, or text with few example videos and pics.

That could come together with a series of FAR tutorials, from basic to advanced, we want more and more people to lose the fear of this awesome mod.

One important thing that we should mention is that stability derivates, even though they are extremelly important, are not rules, and knowing which one to care about the most can be very helpful.

We have never seen a single aircraft which is "all green" at any Mach 1.0 setting, the thing is to know which ones will destroy your craft and which can be fixed by hand when piloting (as well as which matter at that speed at all).

Also, always check the stability derivates at Mach 0.8, given they are fine on Mach 0.2 and 0.8 it's very likely that it will be stable on that range.

Thanks for the attention, if you want us to help with your aircraft, suggest/request any vehicle or try some of ours, check our topic: http://forum.kerbalspaceprogram.com/threads/84738-Roaken-Corporation-High-end-FAR-planes

[Scootaloo]

I'm going to assume that you meant 300 m/s, not 300 km, since there's no air up that high. In that case, congrats, you've found transonic instability; your plane is actually unstable at that Mach number and what's happening is it pitches up, the additional drag slow it down, and then the slowdown drops it into a stable regime, so you end up with this oscillatory motion. Weird behaviors like this are why no one ever wants to fly at the speed of sound, only faster or slower. What you're going to want to do is go through that regime with the lowest AoA you can manage and just try to get supersonic quickly, where your plane should be a lot more stable. As for the roll problems, that hints at needing a larger vertical tail.

As to your other questions:

Right click the control surfaces in the editor to assign control axes, just like you would have done in the stock game.

For flexing, use your struts intelligently. Don't strut two parts that are already attached together, strut together parts that aren't attached yet.

Sounds good. It does accelerate pretty slowly, so it does tend to loiter at that transonic speed for a bit. Guess I just need to give it a little boost with the rockets to get it past the sound barrier.

What is the physics significance of struts? Do exposed struts contribute much to drag at the moment in KSP?

[R0cketC0der]

Sounds good. It does accelerate pretty slowly, so it does tend to loiter at that transonic speed for a bit. Guess I just need to give it a little boost with the rockets to get it past the sound barrier.

What is the physics significance of struts? Do exposed struts contribute much to drag at the moment in KSP?

I think struts have no drag at all.

[ferram4]

@AndreyATGB: While the suddenness of the change near Mach 1 isn't exactly accurate (there's a change in the latest dev build that makes it a bit gentler on the supersonic end), planes are much less stable at transonic speeds M < 1 and are much more stable at supersonic speeds, so if the plane is lacking control authority on either side that implies that it needs more control surfaces or that it needs fuel pumped around to shift the CoM (which real supersonic planes do). I have been making some tweaks to that, since I managed to find a nice detailed section in the 1978 Stability and Control DATCOM on what's happening there (which has led to another change in the dev build, which is that swept wings are much less affected by that effect).

@ObsessedWithKSP: The acronym is good, but I dislike "Not Exactly Aerodynamic Research", particularly since it's not research in any way. For FAR it's technically accurate, as it's trying to solve the problem of "given a player that can through whatever they want together, what is the best way to accurately and efficiently model aerodynamic effects around that so that the game can still run in real time?" So for NEAR, the two possibilities I've been looking at are "Not Elaborate Aerodynamics Replacement" as a descriptive one and "Newbie's Easymode Aerodynamics Replacement" for a snarky one.

@Scootaloo: Try going up higher and breaking Mach 1 in a short dive. That's how the F-104 got to supersonic speeds as quickly as possible (although real jet engines vary in thrust with altitude, unlike KSP engines, so that was easier at the low end of the dive).

Also, struts create no drag. It's possible to figure out where they are, but so many end up being used to hold large, elaborate wing designs together (inside the wing) that punishing them would be silly.

[jrandom]

"Newbie's Easymode Aerodynamics Replacement"

This one gets my vote! (Or possibly shorten to "Newbie's Easy Aerodynamic Replacement".)

Hrm...

Newbie's Enhanced Aerodynamic Replacement? That gets rid of the word "easy" which a lot of people will stumble over the first time they use non-stock aerodynamics, even if it's easier than FAR.

[Roaken Corporation]

Hello ferram, you mentioned some changes on the dev build.

How big would these differences be?

And how much time would such modifications take to be implemented?

[AndreyATGB]

Hello ferram, you mentioned some changes on the dev build.

How big would these differences be?

And how much time would such modifications take to be implemented?

Are you talking about FAR or PANELFAR? Because FAR is always even already compiled on his git so the stuff he mentioned about transonic is downloadable if you want to try it. The other thing I don't understand because it's well, aerodynamics.

[Roaken Corporation]

Thanks a lot!

We haven't heard of PANELFAR before, that sounds awesome!

Edit: also "Nearly Easy Aerodynamics Reworked" (or Replacement).

Edited July 5, 2014 by Roaken Corporation

[Hodo]

Ferram got a question about how flaps are modeled and operate in FAR.

I understand how flaps work in reality, and why they work. But I don't get how they work in FAR. For some reason they have a tendancy to pitch the nose down on some aircraft, and work fine on others. And is it possible to add leading edge slats?

EDIT- and one last thing, the concept and function of swing wings. I have built a successful swing wing aircraft that works quite well in RO, but would there be any benifits of a foward swing wing design.

Edited July 5, 2014 by Hodo

[ferram4]

Well, that's because flaps cause a pitch-down moment about the aerodynamic center of the wing they're attached to, which is a simple consequence of how they create lift: they increase the camber of a wing, and cambered wings have a pitch-down tendency. Now, if they're attached to a wing that has an AC that's ahead of the CoM of the plane, then the additional lift will probably counteract the additional pitch-down moment. If they're attached to a wing high above the CoM of the plane, the extra drag might be able to counteract the pitch-down moment. If the design has the horizontal tail set up to add some additional deflection when the flaps are deployed automatically, that would hide the effect of the pitch-down moment. In any case, that's proper behavior and it's a problem to be solved through designing the plane properly.

In general though, delta wing vehicles don't have flaps on them. Or they're the Tu-144 and they deploy canards when the flaps are deployed in order to try and counter the pitch-down moment.

Leading edge slats are also correctly modeled, but given that you don't often run into stall during launch due to the incredibly short landing gear that you have to work with, there's not much use for it's max-AoA-increasing abilities. But if you want them, go ahead.

The main benefit of a forward-swept wing would be that they shouldn't have as much of a rearward shift of the AC when the wings are swept, since more of the wing would be forward, which would help counter the rearward AC shift due to the flow changing for supersonic flight. On the other hand, it would be more prone to be ripped off. Either way, it'll look cool.

[Melisandre]

Just created this account to thank you (and on my way to other few mods ) for your awesome job and keeping it updated.

[Hodo]

Well, that's because flaps cause a pitch-down moment about the aerodynamic center of the wing they're attached to, which is a simple consequence of how they create lift: they increase the camber of a wing, and cambered wings have a pitch-down tendency. Now, if they're attached to a wing that has an AC that's ahead of the CoM of the plane, then the additional lift will probably counteract the additional pitch-down moment. If they're attached to a wing high above the CoM of the plane, the extra drag might be able to counteract the pitch-down moment. If the design has the horizontal tail set up to add some additional deflection when the flaps are deployed automatically, that would hide the effect of the pitch-down moment. In any case, that's proper behavior and it's a problem to be solved through designing the plane properly.

In general though, delta wing vehicles don't have flaps on them. Or they're the Tu-144 and they deploy canards when the flaps are deployed in order to try and counter the pitch-down moment.

Leading edge slats are also correctly modeled, but given that you don't often run into stall during launch due to the incredibly short landing gear that you have to work with, there's not much use for it's max-AoA-increasing abilities. But if you want them, go ahead.

The main benefit of a forward-swept wing would be that they shouldn't have as much of a rearward shift of the AC when the wings are swept, since more of the wing would be forward, which would help counter the rearward AC shift due to the flow changing for supersonic flight. On the other hand, it would be more prone to be ripped off. Either way, it'll look cool.

Ok learned something new about delta wing aircraft, and that is a good thing. Seeing as I still am not really good at designing true delta wings aircraft. Most of mine come from the Mig-21 school of design.

I may have to do some experiments on the forward swept wing design.

[Jed]

Hello,

I have a strange problem:

- I have FAR 0.13.3 and I've created a space plane with stock and space plane plus parts.

- I take off from KSC, dock with a space station, refuel, quick save, undock, crash during reentry, quick load

- Now my ailerons controlling pitch and roll don't work (reaction wheels and RCS work fine, yaw ailerons too).

- If I launch another identical spaceplace, pitch and roll ailerons work fine.

Does anyone know what might be causing this?

[StevenJ]

I just wanted to say I've been playing with your mod for a month now and I love it. It does make it harder but far more rewarding to build and fly air planes and rockets with your mod, mainly air planes. Also I saw you wrote not to use flaps on delta wing craft unless you have forward canards. I found a different solution as I don't usually make craft with forward canards. I set the control surface for the flaps also as elevators so if it pitches down too much I'll pull up and it will decrease the angle of deflection as the flaps are set as elevators as well. Seems to work for me. No KSp game is complete without random aerodynamic failure! I'm still trying to build a good jet powered stunt plane. I find if I build one with lighter fuel loads it tends to work better. Though that transitional speed of mach 0.7 to 0.9 is usually where the aero frame fails, as to be expected. That is the most dangerous point. I have found with your mod the only kind of air planes, no matter how hard I try, to get them to fail aero dynamically are the small fire spitter powered acrobatic propeller driven planes I've made. Cause you know, there's nothing like celebrating a hard fought mission to Minmus and back by buzzing the tower in my Extra 300 clone. Now if only I could do sky writing and smoke trails in game!

Edited July 7, 2014 by StevenJ

[mk-fg]

Hello,

I have a strange problem:

- I have FAR 0.13.3 and I've created a space plane with stock and space plane plus parts.

- I take off from KSC, dock with a space station, refuel, quick save, undock, crash during reentry, quick load

- Now my ailerons controlling pitch and roll don't work (reaction wheels and RCS work fine, yaw ailerons too).

- If I launch another identical spaceplace, pitch and roll ailerons work fine.

Does anyone know what might be causing this?

I came here to look for a solution to a similar issue, as a few threads (all seemingly without any resolution) seem to mention that FAR is most likely suspect for this, but removing "FerramAerospaceResearch/Plugins" directory with FAR dll's doesn't seem to fix the issue here, so I strongly suspect it might be unrelated to FAR, at least in my case.

[Hodo]

Hello,

I have a strange problem:

- I have FAR 0.13.3 and I've created a space plane with stock and space plane plus parts.

- I take off from KSC, dock with a space station, refuel, quick save, undock, crash during reentry, quick load

- Now my ailerons controlling pitch and roll don't work (reaction wheels and RCS work fine, yaw ailerons too).

- If I launch another identical spaceplace, pitch and roll ailerons work fine.

Does anyone know what might be causing this?

I am going to say what has been said to me before.

Can't help without an output log. It can be found in your KSP_Data folder.

[dlrk]

Has anyone taken a look at the lack of drag produced by the WIP dtobi inflatable heatshield?

[mk-fg]

I came here to look for a solution to a similar issue, as a few threads (all seemingly without any resolution) seem to mention that FAR is most likely suspect for this, but removing "FerramAerospaceResearch/Plugins" directory with FAR dll's doesn't seem to fix the issue here, so I strongly suspect it might be unrelated to FAR, at least in my case.

I was wrong, FAR definitely affects the issue for me, just needed to remove whole FAR dir, not just dll files.

With FAR removed, all surfaces seem to work in orbit with the same savegame file.

[ferram4]

That issue has been fixed (a while ago) in the dev builds on github. A link to the repo (not the release link) can be found in the OP, where you will be able to download updated dlls with the fixes.

[mk-fg]

That issue has been fixed (a while ago) in the dev builds on github. A link to the repo (not the release link) can be found in the OP, where you will be able to download updated dlls with the fixes.

Thanks.

Should it only apply to newly-launched crafts though?

Loading current save with FerramAerospaceResearch dir from the current master (3c7c50d879) doesn't seem to fix already-orbiting vessel, producing quite a few errors like this one in the log (they weren't there with stable 0.13.3) when I try to steer it:

[EXC 02:11:10.021] ArgumentException: Value does not fall within the expected range.

Unity's Player.log has more info on these events:

(Filename:  Line: -1)

Event Type Ignored: EnterNotify
ArgumentException: Value does not fall within the expected range.
  at ferram4.FARControlSys.StabilityAugmentationUpdate (.Vessel vesselToChangeTo, .Vessel vesselToChangeFrom) [0x00000] in <filename unknown>:0 
  at ferram4.FARGlobalControlFlightObject.LateUpdate () [0x00000] in <filename unknown>:0

Full log with 3c7c50d879 (1.8M): http://fraggod.net/tmp/KSP.far_control_surfaces_issue.log

Player.log from the same run (~5M): http://fraggod.net/tmp/Player.far_control_surfaces_issue.log

Aforementioned errors should be at the very end of these.