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

[Wanderfound]

The current dev build of Mechjeb appears to have solved the roll issue, BTW. Lovely to have my cruise control back.

http://jenkins.mumech.com/job/MechJeb2/

[Tygroux]

Thanks for the input! Will try to apply that.

I was about to post pics, and found out that the unstability problem was caused by a bug with the B9 large S2 cargohold, wich had an offset COL.

I'm using mostly centered delta-wings, as I find they give the best stability at re-entry speeds, but encounter the same problem with every craft:

Tthe SAS keeps pushing my nose down (or up...), despite the controls being perfectly capable to keep it stable.

And MJ smart A.S.S. was just going crazy on the controls.

I guess I will try the dev build then, thanks Wanderfound!

[Wanderfound]

Thanks for the input! Will try to apply that.

I was about to post pics, and found out that the unstability problem was caused by a bug with the B9 large S2 cargohold, wich had an offset COL.

I'm using mostly centered delta-wings, as I find they give the best stability at re-entry speeds, but encounter the same problem with every craft:

Tthe SAS keeps pushing my nose down (or up...), despite the controls being perfectly capable to keep it stable.

And MJ smart A.S.S. was just going crazy on the controls.

I guess I will try the dev build then, thanks Wanderfound!

You're welcome.

Even when working right, an autopilot will never match a good kerbal pilot, though. Use Smart A.S.S. to hold pitch and heading when you feel like it, but disengage it and fly manually if you want to do any manoeuvres more complicated than tweaking pitch a degree at a time. Even in cruise mode, keep an eye on it; if the plane starts wobbling about, disengage the autopilot and restabilise by hand before turning it back on.

Are you adjusting your control surface tweakables? With mid-mounted wings, the control surfaces there should only be acting as ailerons (i.e. set to influence roll, but not pitch or yaw) or flaps. Pitch needs to be controlled with elevators at the tail and/or canards up front.

[Tygroux]

You're welcome.

Are you adjusting your control surface tweakables? With mid-mounted wings, the control surfaces there should only be acting as ailerons (i.e. set to influence roll, but not pitch or yaw) or flaps. Pitch needs to be controlled with elevators at the tail and/or canards up front.

Not when I'm going for a spaceshuttle style... spaceshuttle (aka delta wing, no elevator), but otherwise I cut pitch on the wings, yes.

I do need to use more the others tweakable though...

[Wanderfound]

Does anyone know where I can find a table of correspondences between altitude, atmospheric density and temperature for FAR?

[Volt]

I took your guys' advice and downloaded the latest dev version, and encountered a FAR FAS UI that disappears after 20 seconds of flight, and the world's most horrific instability issues on an aircraft with derivatives all in the green. Oddly enough, the problems only present on that aircraft, including the vanishing UI. Code is weird, especially where Unity is involved... Bugger it. I'll see you when 14.1.2 comes out.

Edited September 15, 2014 by Volt

[ferram4]

@Wanderfound: You would have to refer to the stock pressure and temperature functions, and then density can be calculated from the ideal gas equation.

@Volt: Reproduction steps and a log please? Bug reports without that basically just tells me that I have to stop continuing development until I can get information to try and fix the issue.

[B787_300]

so ferram what is next on the dev list?

[Volt]

Yeah, sorry, this is what's annoying. I have no idea what's causing this, the logs don't change whatsoever when or before the UI disappears, and I because I can't see any cause for the stability problem with that one plane (which flew near-perfectly with the slightest roll bias to one side in the release version); it's just behaving utterly randomly. At one point it went from being in a slight straight climb to a deep single-wing stall for no apparent reason. Really odd. I might have buggered up the installation, mind, because some weird stuff I didn't expect happened while I was extracting the files. This may all be on my end!

[Wanderfound]

@Wanderfound: You would have to refer to the stock pressure and temperature functions, and then density can be calculated from the ideal gas equation.

Thanks for the answer.

That's a bit over my head, I'm afraid; bioscientist here, mathematics are not the language of my people.

Basically, what I'm after is just something that will tell me what temperature and density values to plug into the data & stability derivatives window when I'm trying to predict high-altitude aerodynamics: i.e. "alright, it looks okay at sea level, will it still work at 25,000m?".

Is there a simple way to know approximate density values for a few typical altitude bands (say, 5,000m increments from 20,000m to 40,000m)? If anyone could toss me a handful of ballpark figures, that'd be great; if not, I'll just have to continue relying on flight testing and the EVA Parachutes mod...

(either that or send up Bob with a thermometer, barometer and notepad)

Edited September 15, 2014 by Wanderfound

[ferram4]

Density falls off roughly exponentially; it's much more affected by pressure than temperature. Since Kerbin has a scale height of 5 km, you can work out the reduction from a SL 1.225 kg/m^3 fairly easily.

Temperature will be ~300 K at SL, which will steadily decay until it's around 260 K or so at ~10 km, and it stays roughly constant there. IIRC.

[Wanderfound]

Density falls off roughly exponentially; it's much more affected by pressure than temperature. Since Kerbin has a scale height of 5 km, you can work out the reduction from a SL 1.225 kg/m^3 fairly easily.

Well, some of us can; my mathematical abilities exponentially decay as soon as you get past shopping algebra.

The temperature on the analysis panel is in C, not K, yes? It defaults to 20.

(google google... http://wiki.kerbalspaceprogram.com/wiki/Atmosphere ...oh, so that's what scale height means)

If I'm understanding this right, I should just divide the density by 2.718 every time I go up 5,000m?

So:

0m = 1.225 kg per cubic metre.

5,000m = 0.45

10,000m = 0.16

15,000m = 0.06

20,000m = 0.02

25,000m = 0.008

30,000m = 0.002

I've got no idea what's so special about 2.718, but that's okay; all I need is an approximate figure to plug into the analysis systems.

Boil the cauldron, chant the spell, wait for the magic numbers to come out. Maths as usual.

Thanks, Ferram.

Edited September 16, 2014 by Wanderfound

[TheGatesofLogic]

my very large rocket lifters are always very roll unstable, and i don't understand why, they also seem to lack any semblance of control even though they have MASSIVE Control and Stabilizing fins on them (by massive i mean 6mx5m deltas) anyone had this issue before and know how to fix it?

[tetryds]

my very large rocket lifters are always very roll unstable, and i don't understand why, they also seem to lack any semblance of control even though they have MASSIVE Control and Stabilizing fins on them (by massive i mean 6mx5m deltas) anyone had this issue before and know how to fix it?

Make sure you disable rolling control on rudders.

Also make sure your tail fin is reasonably big.

And check the derivates at the speed you get issues.

Posting a pic of one of the planes with the issue also helps.

Because you can be going overkill on dihedral, or smth.

[TheGatesofLogic]

sorry, i was referring to my rockets, i am absolutely lost as to why a perfectly radially symmetrical stick would be ROLL unstable...

[ferram4]

@Wanderfound: e comes from calculus; e^x is its own derivative and the natural log (log base e) is the solution to the integral of 1/x from 1 to any value. But yes, you're doing it correctly.

@TheGatesOfLogic: They are perfectly rotationally symmetric. That is the only symmetry you get from the VAB. That doesn't prevent them from gaining some net amount of twist that would cause roll.

Then again, I'm confused; 0.24.2 has engines thrust vectoring to provide roll control now, so you should have ample roll control unless your design is not radially symmetric to the degree where it should be visible. You will want to check your design for misalignment of parts.

[Wanderfound]

@Wanderfound: e comes from calculus; e^x is its own derivative and the natural log (log base e) is the solution to the integral of 1/x from 1 to any value. But yes, you're doing it correctly.

Calculus, that thing that I did none of in high school and only passed in university because the whole class was taught so badly that my 30% fail got scaled up into a bare pass. Despite having HD averages in both of my majors (psych/neuro, history & philosophy of science).

@TheGatesOfLogic: They are perfectly rotationally symmetric. That is the only symmetry you get from the VAB. That doesn't prevent them from gaining some net amount of twist that would cause roll.

Then again, I'm confused; 0.24.2 has engines thrust vectoring to provide roll control now, so you should have ample roll control unless your design is not radially symmetric to the degree where it should be visible. You will want to check your design for misalignment of parts.

...and make sure that Mechjeb isn't involved. Still not accounting for roll gimballing outside the dev build.

[TheGatesofLogic]

Okay... I think it might have something to do with KW rocketry actually, the engine thrust vectoring seems to be acting weird. My previous designs were not using the KW engines so this would explain a few odd bits.

[Wanderfound]

Any chance of adding something like this image to the guide at https://github.com/ferram4/Ferram-Aerospace-Research/wiki/Notes-on-stability-derivatives ?

I can't speak for anyone else, but I find it a lot easier to wrap my head around that stuff if I'm not distracted by constantly attempting to hold a visualisation of the direction of the assorted variables while I'm reading it.

[rifter]

Okay... I think it might have something to do with KW rocketry actually, the engine thrust vectoring seems to be acting weird. My previous designs were not using the KW engines so this would explain a few odd bits.

I also had some problems like this where with SAS on, the rocket would start rolling more and more. Turned out to be that I had to reverse the controls (a right click tweakable on the engines) before this behavior would correct itself. Only really applied if you are using SAS. Also, with RCS build aid, you can check to see if you have any torque being generated by your engines in the VAB.

[Subasean]

I've been playing with building a Venturestar SSTO more in concept than form, (IE: Flies like, does not look like) and have run into some fascinating lifting body effects.

First off, the ship:

Javascript is disabled. View full album

(That was a quick screenshot taken while toying with wing positions. The final Procedural Wing design and position put the CoL behind the CoM, as it should be)

At low angles of attack, it lawn-darts mercilessly, but once a certain threshold is breached, it tumbles tail first. I pulled the wings off and flew just the body and as I suspected, the CoL is forward of the CoM due to the wide, flat body, however, small stabilizer wings lose their effectiveness once the body presents its lower surface to the air stream, and the body aerodynamics dominate the flight characteristics. I'm trying to find a solution, but I am aware that there's a reason Lockheed Skunk Works were designing the X-33, and my abilities/tools may not be up to the task. This would be a lot simpler if I weren't hung up on a glider return, but that's part of the fun of the challenge.

My gut is telling me the CoL marker isn't going to be a lot of help here, less so than my Shuttle and most of the design is going to have to be sorted with the graphs.

Anyone here pulled off a Venturestar-ish ship with B9 and Ferram? Any advice on taming a lifting body?

**Edit**

I spend a few days fighting this, finally break down and ask for help, then go back to the game and get it pretty much fixed. Argh. Ditched the Procedural wings and fancy designs, grabbed a B9 delta wing part and then just reallocate the mass of the ship when empty and voila! All is solved.

Edited September 16, 2014 by Subasean

[James_Eh]

Welp, I have to say thank you for this mod. I have been KSPing for quite a while now and finally bit the bullet and installed FAR with a new 24.2 install. Wow, I haven't blowed up this many rockets in a LONG time. Courtesy of this forum I knew it was something I had to try, but had not anticipated just what a game-changing experience it would be. I am actually looking forward to the challenge of ascents now, since I often experience the aforementioned "blowed up" phenomenon when I screw up...

[MAKC]

I've been playing with building a Venturestar SSTO more in concept than form, (IE: Flies like, does not look like) and have run into some fascinating lifting body effects.

First off, the ship: http://imgur.com/a/PDqzm (That was a quick screenshot taken while toying with wing positions. The final Procedural Wing design and position put the CoL behind the CoM, as it should be)

At low angles of attack, it lawn-darts mercilessly, but once a certain threshold is breached, it tumbles tail first. I pulled the wings off and flew just the body and as I suspected, the CoL is forward of the CoM due to the wide, flat body, however, small stabilizer wings lose their effectiveness once the body presents its lower surface to the air stream, and the body aerodynamics dominate the flight characteristics. I'm trying to find a solution, but I am aware that there's a reason Lockheed Skunk Works were designing the X-33, and my abilities/tools may not be up to the task. This would be a lot simpler if I weren't hung up on a glider return, but that's part of the fun of the challenge.

My gut is telling me the CoL marker isn't going to be a lot of help here, less so than my Shuttle and most of the design is going to have to be sorted with the graphs.

Anyone here pulled off a Venturestar-ish ship with B9 and Ferram? Any advice on taming a lifting body?

**Edit**

I spend a few days fighting this, finally break down and ask for help, then go back to the game and get it pretty much fixed. Argh. Ditched the Procedural wings and fancy designs, grabbed a B9 delta wing part and then just reallocate the mass of the ship when empty and voila! All is solved.

You can physically rotate (Shift+W/S for pitch) your craft in the SPH to see how the CoL behaves during different angles of attack. You can also use the FAR lift/drag graph for the same purpose but I find just rotating to be a lot quicker and easier to test.

[TheNorthFace]

First of all, a big shoutout to FAR. This mod is a MUST HAVE, and really shines a light on how terrible the current KSP aerodynamics system is.

One thing that bothers me however is how the deltav requirements to get into LKO drops from 4500dv (stock) down to 3100dv or maybe less with FAR. For me it shifts the game balance a bit too much.

I've counterbalanced this deltav benefit a bit by installing deadly reentry and TAC life support.

I have two questions:

1. What is the opinion and design philosophy of the FAR developers about the deltav reduction requirements to reach LKO? Do you plan on trying to match the dv requirements of FAR with stock, or would it negatively affect aerodynamics realism?

2. Any tips on how I could modify my own experience through mods or editing some parameters in order to match dv requirements of FAR with stock?

[Camacha]

Ferram is way ahead of you Try KIDS and find a setting that suits you. If you want to go one bigger you might try Real Solar System.