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

[gallexme]

One small question is it normal that when u turn on SAS all "movable wings" they wobbling like hell? and it causes some fps drops which flips your plane instantly?

[Tharios]

I went and changed some things, and when version 0.12.4 comes out body lift will be more viable than it is currently. I did some math and it turns out that body lift at low angles of attack isn't quite high enough, so I'm boosting it back up a bit; pure structural part lifting bodies won't be quite as effective, but one with a pair of tiny wings attached should be doable. Pure structural part lifting bodies will need some thrust behind them to counteract the large amount of drag on them (which is what kills them now), but with tiny wings a glider will be possible.

Other than that, the solution to making a lifting body is always to just go faster. At high enough speeds, even barn doors make significant amounts of lift.

Also, I think I narrowed down and squished the "supersonic roll twitch" bug. So that's something nice.

The emphasized portion is what I was about to post a question about. I finally got my planes stable without becoming twitchier than a meth addict after a month of sobriety...and then when I get up to 15km altitude and about mach 3+, it rolls inevitably to one side or the other with no way to roll back with the little control surfaces I use for roll (usually to the left).

Glad to hear that may be fixed.

Never messed with lifting bodies...after the update, maybe I'll try them out.

[ferram4]

@Zyglrox: The control surface time constant defines how long it will take for a control surface to properly deflect, modelling the servo controlling it as a first-order system where the time constant is used to determine how much to change the control deflection every frame. If the control input changes from one constant to another, the system will reach 67% of the commanded deflection 1 time constant after the command is initially issued, reaching ~98% of the commanded deflection 3 time constants after the deflection occurs. It's selected not because the modelling is simple, but because it makes modelling the control surface response in a more advanced control system (say, a more intelligent SAS) far, far simpler.

Every time I tried increasing the time constant, SAS became far, far worse at controlling my planes in every case. Every time I tried decreasing it SAS has had an easier time handling the issue. The problem is when you have a not perfectly balanced plane, and SAS commands a deflection that should be enough to bring things under control, but the control surface hasn't reached that deflection yet. So it commands more deflection, and when the control surface blows past the balance point the plane starts heading in the other direction.

The reason it's hard mode is that it makes SAS far less forgiving. And since SAS got fixed in 0.21 and "fixed" further in 0.22, people have assumed that SAS will be perfectly fine with whatever they throw at it. And that will never be the case.

@gallexme: If you're having fps drops caused by turning on SAS then that sounds like you either have background applications interfering with KSP or something else. I seriously doubt FAR is causing the fps issues if they're that intermittent.

@Tharios: No, that's not what I'm talking about. Your issue is caused by uneven flexing in the plane / a lack of sufficient yaw stability and is a perfectly realistic and intended behavior. What I'm talking about is getting a plane up to supersonic speeds and the computation freaks out and causes a single frame of larger-than-realistic lifting force on one wing, causing the plane to suddenly start rolling, but the force causing it disappears the next frame.

[Hodo]

Not sure exactly what is causing this issue but I have narrowed it down to FAR. When I undock something it catapults away at the speed of plaid, which is quite concerning as you can imagine. Last night I docked a ship to my low kerbin station at 100km orbit, undocked it and it launched through the station and away at over 1000m/s away from the station. It killed the crew but did no damage to the ship itself, it was a G-Force overload on the crew. I checked the logs and there was no faults in the output log other than the RCS related ones in the SPH and VAB, but nothing to do with the docking or undocking.

My list of mods are below but I have tested this with only FAR installed and it repeats itself. It is like the decoupling bug that happened when .23 first came out. Seeing as I haven't tested anything with decouplers I can't say if it is the same.

I am using the latest FAR, and my list of current mods are in my sig. I have tested this without the mods and with the mods, FAR is the only mod that I can repeat this with.

[ferram4]

Are you sure you're using the most recent version of FAR? Because I just played around with docking and undocking and i couldn't reproduce the issue. Could you provide more detail about how to cause the issue?

[Hodo]

Are you sure you're using the most recent version of FAR? Because I just played around with docking and undocking and i couldn't reproduce the issue. Could you provide more detail about how to cause the issue?

Just checked again, my FAR was saying 12.2, so I guess I wasn't using the latest, even though the one I downloaded was from the spaceport. I just changed over to the other download link and it gave me 12.3. So I guess that fixed it, I will test later, Merry Christmas and happy holidays!

[RTK]

One small question is it normal that when u turn on SAS all "movable wings" they wobbling like hell? and it causes some fps drops which flips your plane instantly?

@Gallexme - in relation to the wings going crazy it could be your control surfaces. in SPH right click on relevant flaps and reduce the Control Maxi, also set your wings flaps to be inactive for Yaw, also set your rudder to be inactive for Pitch and Roll, SAS is highly likely not the cause of the wings going crazy.

I found on my test plane which was really light and quick build (so not really well designed on purpose) I had to reduce control maxi almost right down.

in relation to the Frame rate drops that is something else causing the problem, I'm only running FAR mod and no others with no performance issues in KSP .23

I hope the above helps, it is only a suggestion I have as I recently experienced the wings going crazy once I got into flight.

[Donziboy2]

@Gallexme - in relation to the wings going crazy it could be your control surfaces. in SPH right click on relevant flaps and reduce the Control Maxi, also set your wings flaps to be inactive for Yaw, also set your rudder to be inactive for Pitch and Roll, SAS is highly likely not the cause of the wings going crazy.

I found on my test plane which was really light and quick build (so not really well designed on purpose) I had to reduce control maxi almost right down.

in relation to the Frame rate drops that is something else causing the problem, I'm only running FAR mod and no others with no performance issues in KSP .23

I hope the above helps, it is only a suggestion I have as I recently experienced the wings going crazy once I got into flight.

The aircrafts noodle effect does not help, if its long and flexible that tends to screw up SAS. So a long craft that bends alot will have control issues, a few Struts usually solve the problem.

[Hodo]

Are you sure you're using the most recent version of FAR? Because I just played around with docking and undocking and i couldn't reproduce the issue. Could you provide more detail about how to cause the issue?

Ok installed the latest patch 12.3 and it still shows 12.2 on the GUI so not sure if it took.

But tested the docking again, I am getting this when it undocks along with catapulting it away at a high rate of speed.

This is the error that is showing up in the output log RIGHT after it undocks.

Joint::setBreakable: maxForce should be nonnegative!

(Filename: ..\..\Physics\src\NpJoint.cpp Line: 294)

Not sure what that means.

[t3hk0d3]

How to fix mirrored wing "bending"?

Primary (placed) one acts normaly. But mirrored wing bending on slightest wingload. That cause plane being extremenly wobbly on maneuvers.

Lastest version.

[softweir]

Struts. I'm afraid that's the only answer, at least until Squad fix that bug!

[AbeS]

Or maybe KJR

[ferram4]

@Hodo: That sounds like some other mod is causing the bug; I know you have a list of mods in your sig, but would you mind listing them out, including version numbers and what version of KSP they're guaranteed to be compatible with? Perhaps it's a known issue with one of them or exacerbated by something.

@t3hk0d3: Struts. This is a stock bug, nothing I can do about it.

[Chris_W]

Using the latest version, I'm getting a lot of errors when I use any winglets on the bottom of a rocket and disable all control directions via tweakables. It keeps saying lift is NaN in flight for every winglet over and over, happens every time.

I usually do this since my designs are stable enough to be controlled through a single gimbaling inboard engine, enabling the winglets to move would just make them less controlable.

Example design (Parachute test craft)

[ferram4]

Does it do that in flight, or are those errors from using the CoL in the editor?

I'm looking at the code and I'm not sure how a NaN error could occur... Right-clicking on the winglets confirms that they're not producing any lift or drag, correct?

[Peppe]

Ok installed the latest patch 12.3 and it still shows 12.2 on the GUI so not sure if it took.

But tested the docking again, I am getting this when it undocks along with catapulting it away at a high rate of speed.

This is the error that is showing up in the output log RIGHT after it undocks.

Joint::setBreakable: maxForce should be nonnegative!

(Filename: ..\..\Physics\src\NpJoint.cpp Line: 294)

Not sure what that means.

I had that isssue in stock with kerbals in the external seat in .22. Think that was on docking. Do you have kerbals in external seats or maybe the command pod you are using is doing something funky?

[Chris_W]

@Ferram: It occurs in flight, and indeed the crafts behave as if there were no wings attached.

I don't know what causes it, but in ... .2 this did not happen, only just recently with the .3 at the end version.

Maybe it's something about the tweakables returning a zero value? Divide by zero usually gives NaN afaik..

The VAB log seems clear of errors.

Flight Scene

Edited December 27, 2013 by Chris_W

[Hodo]

Found and fixed the error, it was a combination of problems.

FAR12.3 had a fault in the download somehow, and then there was a error in the B9 file. Found and fixed both.

[Shad0wCatcher]

Ferram, long time lurker, first time poster. Thank you for all you've done so far. Really appreciate everything you've offered the community.

On to my quandary now. Currently running FAR with a bunch of mods, including B9 and KW Rocketry. I'm not sure if my issue I'm running into is something for yourself or kickasskyle (i.e. model issue or aerodynamics not taking the shape of the part into account properly); but every time I use either a structural size adapter from KW rocketry (inverted or not) or a LFO size adapter (inverted or not) the drag per part in an aerodynamic profile seems ludicrously high (anywhere from 30-100+ kilonewtons of drag). I'm not sure if you use KW at all or can do a test on them really quickly or not (I'm using 2.5.5 of KW Rocketry and 12.3 of FAR).

To reproduce take any KW rocketry 2.5 or 3.75 fairing plate/decoupler combo (or hell; use the 3.75-1.25 LFO adapter and try and make a "needle" styled rocket with some type of nose-cone at the top; I can guarantee 100 KN of drag); attach a 3.75-2.5 structural adapter inverted (to reduce the lifter assembly of course and maintain an aerodynamic profile as the adapter will be on the trailing edge of the fairings) or a 2.5-1.25 structural adapter plate inverted under the plate/decoupler combo. Build your lifter as usual afterwards.

The only way to stabilize the rocket is using N sets of stabilizer fins; where N is a set for each stage until reaching complete vacuum. (I still had 0.4 KN of drag at 65 KM -- well beyond where anything else was producing drag; no RSS or anything resized or any universe replacer; a stock Kerbin). If this isn't clear enough I'll see about loading up the game and taking a couple screenshots to post on imgur.

My mods are all completely up-to-date as of yesterday(apart from the two aforementioned parts mods; though I've applied the 0.23 fixes for B9s SABRES) and include the following; FAR, Deadly Re-entry, KSPX, KSPI, RealChute, B9, KW Rocketry, KAS, KJR, Near Future Propulsion, Romfarer Lazors, Mechjeb (2.1.1 stable), VOID, TAC Fuel Balancer, TAC Life Support, Firespitter's DLL, PartCatalogue, Visual Enhancements (6.4; in the process of updating to 6.6), Kethane, ScanSat, 6S Compartment Tubes, The Texture Reduction packs for Squad, B9, and KW Rocketry, Toolbar, Kerbal Crew Manifest, Biomass's science dev version, Davon TC Systems, FTmN NTRs, Kerbal Alarm Clock, Atmospheric Sound Enhancement, Station Science, SDHI Service Module, rbray's Active Memory Reduction, Haystack, Extra-Planetary Launchpads, Sam Hall's Mk2 Cockpit internals, NavyFish's Docking Port Alignment Indicator, and Enhanced Navball....Holy hell that's a lot of mods now that I'm looking at the list.

[ferram4]

@Chris_W: I'll look into it, but the right-clicking on the part does state that it's making 0kN of lift and drag, correct?

@Hodo: Good to hear.

@Shad0wCatcher: That's correct. Changes in part diameter add lots of drag to the shape; you should try putting a standard fuel tank on its side at the same dynamic pressure and see how much drag it makes. The reason that the straight fuel tank makes so much less drag is that there is only skin friction drag acting on it, which is a drag coefficient of ~0.005. A tapered tank will also have pressure drag acting on it, which can go up to a drag coefficient of ~1.86 for a purely flat face. Use a part with a less-sudden tapering.

[Shad0wCatcher]

Thanks for the response ferram and I can completely understand that as a non-inverted part that's facing atmosphere. I was less concerned about a needle-shaped rocket than I was about using the same tanks as sort-of trailing edges to taper between a larger payload fairing and smaller upper-stage motor. I can understand some sort of vortex effect of wind coming around the large fairing and pushing back onto the tapered tank; I'll see about getting a screenshot (in the VAB right now) to try to explain more clearly. It could very well be my limited understanding of how trailing edge drag functions; but I would think that the larger fairing would create more of a vacuum effect behind it (underneath from a vertical standpoint) as opposed to massive amounts of drag from what is otherwise a very aerodynamic body.

EDIT: Image follows; fair warning it was taken at 1920x1080 so is too large to preview on imgur. The part in question is just below the fairing adapter/decoupler and has (at least to my eyes) the same taper as it, but it has orders of magnitude more drag than the fairing plate just above it. (Hash marks are where the fairing adapter starts if you aren't familiar though you probably are)

Image link

Edited December 28, 2013 by Shad0wCatcher

[ferram4]

The bottom of the fairing tapers less than the tank below; it's the relative change in diameter that matters, not the absolute change diameter, and the fairing adapter has a smaller relative change in diameter than the fuel adapter, so that would result in lower drag on the fairing adapter. I'll admit that I don't quite understand the point of this design; wouldn't it make more sense to use a wider tank under the fairing and then use the fuel adapter at the bottom of it?

I can look into it to see if anything truly messed up is going on, but I doubt that anything really is going wrong.

[Hodo]

Ferram4 I have a request for the next version of FAR. Can you add a settings option for the flight information window, so we can turn off some of the information and reduce the size of the window.

While I love all the info for test craft, I also use it for normal use craft and I no longer need to know the c/l/d information or the AoA, really I want is the intake % required and a few other things. It would be great to be able to turn off what we don't want without having to go in the config.

[BC000001]

Just a question about drag in real life; if you have a flat object and are moving it through the wind, other than the air you are pushing but is there more drag because you are pulling the air behind you into the vacume you may be creating? I only ask this because I noticed nose cones on the ends of some fuselages and wondered why they were there.

[Hodo]

Just a question about drag in real life; if you have a flat object and are moving it through the wind, other than the air you are pushing but is there more drag because you are pulling the air behind you into the vacume you may be creating? I only ask this because I noticed nose cones on the ends of some fuselages and wondered why they were there.

Not sure about planes, but I know in cars if you have a flat rear the car creates a "wake" or a vortex swirl behind it, which can create lift off of the rear wheels. Which as you would imagine is bad for a rear wheel drive car. But this only happens at high speeds in cars, more than 100mph, and rarely has any affect on daily drivers.

I imagine something like that happens with aircraft.