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

[Lei07]

Still having an issue returning a simple Mk1 pod from LKO Don't get me wrong, I love the challenge FAR provides, and I just love playing with shuttles (which I ironically find easier), but I think this update is beyond my capabilities.

Was awake into the early hours trying to stop going splat over the surface of Kerbin, or squashing myself under g-forces, and I'm back at it today with no success. Does NEAR offer a bit more of a ... "soupy" atmosphere? One that slows craft down more without suffering the g-force that would incur in FAR?

Would be a shame to change over, I'd rather stick with trying to learn this but I truely think it is beyond me. I can honestly say this is the first time FAR has me stumped ^^;

[Master Tao]

Learning new things is only going to strengthen my addiction to KSP. /sigh

And so it begins.

Responding to the rest of your post, there's a key piece with the CoM vs. CoP that needs to be clarified. Along the direction of travel, the CoP needs to be behind the CoM for stability. So when you're launching a rocket, you want the CoP to be more towards the engines than the CoM, and when you're reentering, you want the CoP farther from the heatshield than the CoM. The other trick is that in the VAB, the direction of travel is always modeled toward the roof, even if you flip the whole rocket upside down.

I ran a couple more tests with the pod. Flip it upside down in the VAB to simulate reentry, and everything looks good and stable. Change the orientation by 5-10 degrees in any direction, though, and suddenly the CoP is right on the CoM, so it becomes inherently unstable. That's why it flips, and it's something the Chaka Monkey team will need to take care of unless you go forward with the custom FAR config.

Does NEAR offer a bit more of a ... "soupy" atmosphere? One that slows craft down more without suffering the g-force that would incur in FAR?

Just the opposite. NEAR doesn't include supersonic drag effects, so there's less drag at reentry speeds than with FAR.

How high is your PE? It should be about 20 km for reentry from LKO. With just a Mk. 1 capsule and a Mk. 16 parachute, the pod should slow down to safe parachute deployment speeds by 7-10 km ASL.

[Felbourn]

And so it begins.

That's why it flips, and it's something the Chaka Monkey team will need to take care of unless you go forward with the custom FAR config.

I don't want someone to do this for me. It's a good learning experience. But... do you know of any mods/examples where a FAR config was created that I can examine as part of the learning process? I'm a programmer, so I started with the FAR source and the KSPField entries for the "basic drag" module. I'm not even sure if that's the right one to attach though, and also currently still trying to learn what fields to change (Cm curve, Cl, etc, many fields in this module) even if that IS the right module. I looked at the CFGs that come with FAR but was not enlightened, and I found one config in FASA but it looked like it was just removing drag entirely from a decoupler.

[Lei07]

Hey Master Tao, well, in FAR for KSP 0.24.2 and prior I had the rule of PE 20,000m from a 100,000m LKO, and 30,000m from a return from a moon, and I had no problems at all. I basically got these numbers from a Scott Manley video, and he is a rather good indicator on what and what not to do!

This now leads me to the current version for KSP 0.25. I've restarted my Interstellar mod based save game to incorporate the new KSP features from the start, so currently my tech level is rather basic. After scanning about the last 30 pages of this thread, and a few other info sources, I posted this basic information and my problem: returning a basic Mk1 Pod from a 100km LKO, posting my previous guide rules. I have been told that due to aerodynamic changes in the mod this is now way too steep.

So I was quite literally up long into the night trying out different things, different Pe and AoAs, looking up youtube videos of how Apollo did it, trying to create a asymetric "floating body", trying to find equations, trying to find specifics as regards Kerbin and FAR etc. I'm really not that intelligent when it comes to maths and what have you, I'm more of a "recreate as close as possible then tweak til it works" kind of person.

And after spending today with no success either it has started to get to me somewhat. I did find that, if I keep my upper stage attached, use the engine to help me decelerate through the lower atmosphere, then ditch the thing once I was under 300m/s surface speed, it worked and everyone survived and didnt get squashed or turned into paint. But that isn't a very realistic, long-term solution ^^;

[blowfish]

I looked at the CFGs that come with FAR but was not enlightened, and I found one config in FASA but it looked like it was just removing drag entirely from a decoupler.

It is, and the reason is that FAR actually approximates the drag from the shape, except in the case of wings or when various overrides are used.

[Master Tao]

@Felbourn:

blowfish is right – FAR zeroes out stock drag values they don't interfere with it's own drag model, which is largely based on the shapes of the part models. I'm afraid I don't know of any non-wing parts that use custom configs. The FAR wiki would help teach you what those values mean, and ferram has recommended good aerodynamics textbooks at least a few times in this thread.

@Lei07:

That profile (100 km LKO to 20 km PE) works beautifully for me. It burns off less than half the ablative shielding and slows down to < 200 m/s by 5.5 km.

Sounds like FAR isn't running, meaning either you're on Windows x64, it's installed incorrectly, or something such as permissions are interfering with it. Could you post your output log and the other information requested in this post?

[Jovus]

Just to back up Master Tao, I just did a reentry at Mach 10 with a PE at 21km and no heatshield (using an LV-909) and that worked fine. Well, the engine stuck within 10 degrees of exploding for about three minutes straight, but everything survived. So it sounds like something is weird with your install: either FAR, or DR. If the output.log doesn't help, check the DR debug screen (press ALT+D+R while in a flight scene) and make sure you haven't fiddled with your DR numbers.

[Zidane]

*snip*

OK here's how i did it, i lowered my pe to 30km like before but i did a burn at 50km to help things along and managed to land in one orbit.

One thing though, my air intakes show 0 drag. In fact, most parts in my SPH show 0 drag. When I tinted the center of drag with the FAR flight computer it showed all my drag concentrated on the cockpit.

Shouldn't they add drag when open?

[nli2work]

I don't want someone to do this for me. It's a good learning experience. But... do you know of any mods/examples where a FAR config was created that I can examine as part of the learning process? I'm a programmer, so I started with the FAR source and the KSPField entries for the "basic drag" module. I'm not even sure if that's the right one to attach though, and also currently still trying to learn what fields to change (Cm curve, Cl, etc, many fields in this module) even if that IS the right module. I looked at the CFGs that come with FAR but was not enlightened, and I found one config in FASA but it looked like it was just removing drag entirely from a decoupler.

explanation of the values are explained here.

https://github.com/ferram4/Ferram-Aerospace-Research/blob/master/README.md

[Lei07]

@ Master Lao.

Just on Skype with a friend that is quite knowledgeable about this kinda stuff, I've gotten him to install FAR into an otherwise stock KSP 0.25. We made a practically identical craft (ie, me missing out some stuff from my extra mods) and got into a similar orbit. He had no issue and gave similar paramaters for a 20km PE to what I had in 0.24.2 and before.

It was at this stage I realised I didn't have a heatshield on, and with DRE installed, I was interested to see how many sparks I would be making. However, although I entered slightly faster than he did, I did infact slow down within much more acceptable parameters and much more like how the mod was before 0.25.

So that means that the drag model (I guess?) for the small Mk1 Pod heatshield is incorrect, be it due to Ferram or Starwaster, or both, of course. I will make a quick post over there also, after scanning the thread to see if anyone else has had this issue.

At the moment not outrulling a dodgy install, or some kind of imcompatibility issue, but I run 0.25 in 32 bit, as the 64bit is worse than before and many mods do not support it. I have also done many installs of the game in the past, testing mods, adding mods, always making a fresh install with the smallest of changes. It's not impossible I've made an error, but I do doubt it. All functionaility of FARs GUIs and toolbar stuff etc work.

EDIT: My friend thinks that the drag vector on the MK1 Pod heathshield is inverted, so I'll mention that in the DRE thread, if no one else has reported it.

I'm happy to give a logfile, I don't however know how to do one! Will figure it out and do so if still needed

Edited November 5, 2014 by Lei07

[indylead]

Right, Lei's so-called knowledgeable friend here, just to add a few more details. I haven't played this version of KSP until today, and being an engineer wanted to be methodical!

Step 1: Completely new clean install using KSP v0.25 32bit with FAR 14.3.2. Basic rocket built, small capsule etc. Ap. 102km, Pe 20km. Everything flew as expected - i.e. safe landing.

Whilst chatting, Lei had built the same rocket as me, and then commented that he;d left out a heat shield and might die due to deadly re-entry. The re-entry was slower than he'd been seeing, although a little faster then my version.

Step 2: Added Deadly Reentry 6.2.1, and using the same rocket, added only an appropriate heat sheild, then went into same orbit (Ap 100, Pe 20). This time, on starting reentry effects at 45km the speed started to increase from 2369m/s @45km to 2390m/s @37km - i.e. I was accelerating! Reentry effects did not go until 5km. At 10km the capsule was still travelling at 1,450m/s, impact speed at 0 was 302m/s, unfortunately with the loss of Bill Kermin.

So it seems either the small heat shield drag values are screwed, or there is some sort of clash between FAR and DRE.

Edited to add: video showing reentry with the Deadly Reentry Heatshield in place here: linky

Thanks for the help, folks!

Edited November 5, 2014 by indylead

[ferram4]

If you were using the small capsule, then the heat shield was unnecessary; the 1-man capsule has a built-in heat shield. Your problems are consistent with clipping the heat shield into the pod, which results in extremely low-drag behavior due to confusing the blunt-body drag system.

[indylead]

An update:

We've been posting on the Deadly Re-entry thread here. The developer of DRE seems pretty sure the bug is not with his heat shield. It does seem that the problem is with the aerodynamic properties of the part as modeled in FAR. Looking at the speeds involved on the video, it looks like the atmosphere is having zero effect on the speed (graphical effects aside) and the acceleration is not the drag vector being 'in the wrong direction' but a result of the acceleration as the craft moves towards periapsis.

If the dev wants any more info, please let me or lei know.

[NathanKell]

I believe you were just ninja'd

[indylead]

It wouldn't be the first time!

I was just saying on the other thread that as this was the first time I've posted here, it was good to see two mod devs respond quickly. Now going to have to fight the urge to get back into the game

[Van Disaster]

I haven't built anything in about a year - now any sizeable spaceplane I build seems to be stalling out/getting into an uncorrectable undemanded pitch-up situation in the ~30km altitude range no matter what wing design I try, so I suspect something fundamental has changed while I've been away. Smaller spaceplanes seem fine, but they could just be disguising a problem with extra thrust or something. I lean towards canard/tandem wing arrangements usually with fairly short chord, maybe I'm just not providing enough wing area? it's not one design in particular, which is a bit confusing.

Any thoughts?

Edited November 6, 2014 by Van Disaster

[ferram4]

Well, I'm not seeing that behavior. Planes at 30 km show the same stall characteristics that they show at all other altitudes, and anything above about 15 degrees AoA is getting into stalling. Stall behavior is as it should be.

[Van Disaster]

Yes - not saying there's a bug, just wondering if anything changed that'd have affected on-limit areas like that. It's wierd, planes will be happily climbing at ~10deg and up goes the nose and there's no way of pulling it back. I'll make an concious effort to use more wing area.

[blowfish]

Might it have to do with fuel drain? If tanks near the front of the plane are draining it could be bringing the CoM back and causing pitch instability.

[Felbourn]

@NathanKell; @Master Tao; @blowfish; @nli2work

Yes. Yes. Yes! Yes! Yes!

I needed to create a custom drag module, and I needed to rename one of the parts from "Cargo Area" to "Service Compartment" to get it to work.

But it was glorious. Perfect reentry now.

[Van Disaster]

Nah, it's all balanced ok - I tend to either use fuel burn as a design feature or shift fuel around to adjust pitch in that part of the ascent anyway. I just tried a rather large delta on the same fuselage which as expected was a real pig once supersonic, but once I'd wrestled it to 20km it went up to 150km without even a twitch, so it appears to be my wing design. I've been basing them on old successful designs but I guess there's a combination of differences which make them unviable.

Now this delta craft doesn't want to re-enter, that's definitely harder than it used to be

[Gfurst]

Nah, it's all balanced ok - I tend to either use fuel burn as a design feature or shift fuel around to adjust pitch in that part of the ascent anyway. I just tried a rather large delta on the same fuselage which as expected was a real pig once supersonic, but once I'd wrestled it to 20km it went up to 150km without even a twitch, so it appears to be my wing design. I've been basing them on old successful designs but I guess there's a combination of differences which make them unviable.

Now this delta craft doesn't want to re-enter, that's definitely harder than it used to be

There are some new features as wing mass, I've been wondering this too, having to redesign my craft as the default mass seems heavier than before.

Just note that, more mass means harder structure resistance. Something interesting that wings now tend to bend just like in real aircrafts.

Other thing though to get lift on higher altitudes, where air density is considerably lower, the craft usually needs much higher speed.

Thats how it happens in reallife too where crafts simply don't enough thrust to keep high velocities to maintain stable flight, what we call operational ceiling.

[tetryds]

Nah, it's all balanced ok - I tend to either use fuel burn as a design feature or shift fuel around to adjust pitch in that part of the ascent anyway. I just tried a rather large delta on the same fuselage which as expected was a real pig once supersonic, but once I'd wrestled it to 20km it went up to 150km without even a twitch, so it appears to be my wing design. I've been basing them on old successful designs but I guess there's a combination of differences which make them unviable.

Now this delta craft doesn't want to re-enter, that's definitely harder than it used to be

Can you post a pic of your aircraft?

Because deltas tend to be roll-happy.

[Wanderfound]

Yes - not saying there's a bug, just wondering if anything changed that'd have affected on-limit areas like that. It's wierd, planes will be happily climbing at ~10deg and up goes the nose and there's no way of pulling it back. I'll make an concious effort to use more wing area.

It's entirely possible that it's actually too much wing. Have you accounted for the increased weight (and strength) of the wings? Use the tweakables to wind the mass down to 0.4 if you haven't already.

What you actually need to do is run an AoA sweep at the speed it tends to happen at and check that your Cm (the yellow line) is still sloping downwards at the AoA you want to be able to fly at. If it's aiming down, the plane is stable; if it's sloping up, it ain't.

If you're looking at the stability derivatives screen, you'd be using -20/.01/5 as your input data. Wiggle your wings about and see if you can get more green happening, particularly on the pitch numbers. A bit more space between CoM and CoL is likely to help.

[Van Disaster]

There are some new features as wing mass, I've been wondering this too, having to redesign my craft as the default mass seems heavier than before.

Just note that, more mass means harder structure resistance. Something interesting that wings now tend to bend just like in real aircrafts.

Other thing though to get lift on higher altitudes, where air density is considerably lower, the craft usually needs much higher speed.

Thats how it happens in reallife too where crafts simply don't enough thrust to keep high velocities to maintain stable flight, what we call operational ceiling.

I was actually wondering if it was a combo of the two, although I remember the nose being rather glowy on at least two runs. However short-chord wings are not terribly rigid ( specially p-wings ), so perhaps there was some extra twist involved. I've been trying to reduce strut count given recent developments ( and Ferram's joint strength mod ) but I've rapidly discovered you can't go totally without

Graphs all looked tolerable, and I usually make sure derivatives are green even if some instability in one direction doesn't really matter, because it just makes planes annoying to fly unfortunately I don't really know how to apply the derivative results other than guessing that I don't have enough surface in one axis or another. It looks like Down vel Mw is positive at very low air density/mach 5 for the troublesome craft which goes some way to explains the sudden backflips ( although not the complete lack of warning ) but other than brute-force adding more pitch control surface I wouldn't know what to do about that. Derivatives are fine in denser air.

This was my first attempt at something actually useful this time round:

screenshot0

It goes up & down quite happily. Next one is one of the troublesome ones, and I guess it probably just doesn't have enough control authority ( it's a bit underpowered ):

screenshot1

Pitch control is only from the canard surfaces, the rear wing has flaperons & brakes.

And this flies beautifully, but I wish I knew how to solve the roll-yaw couple thing - having the engines all the way out there does make it yaw happy.

screenshot3

screenshot2