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

[ferram4]

@AncientGammoner: I would consider a "good" method of adding difficulty to be dropping Isps; while the dV necessary to get anywhere might be small that amount of mass necessary to do it will end up being either where they are now or requiring launch vehicles more like what we have in real life. The dV really doesn't matter; it's the mass necessary to achieve it that matters in the long run. I actually think that idea has come up a few times and I'd be very happy to see someone take on the "lower Isps to keep payload ratios low with FAR"-type of project, but it's not something I want to do myself. Too many part packs that would need to be taken into consideration.

I dislike the "make the atmosphere out of soup" method of adding difficulty since it adds very little to any aspect of gameplay; for launches it just adds a slow, 10km vertical ascent and for planes it adds a nearly vertical ascent to the same altitude. It requires that jet engines to have too much thrust to make flight at sea level possible, which makes reaching top speed at maximum altitude in a plane almost laughably easy; it also makes jets a viable almost-SSTO-stage for rockets when they should be no more than extra boosters in that role, if they should be able to do anything there at all. It also makes exploring the surface of a planet tedious; you're limited to very low speeds by the large drag forces or you have to burn tons of fuel to move at relatively slow speeds. Finally, there's almost no challenge landing on a planet with a half-decent atmosphere once you get past reentry; just let the atmosphere slow you down to ~100 m/s and then land from there; no danger of slamming into the ground with a vertical lander or trying to touch down to fast with a plane, since the atmosphere will slow you down more than enough.

My problem with the drag losses isn't that they're there, it's that they're way too high compared to all the other loses. There's almost no fun launching a rocket since so much time is spent waiting for it to be high enough to actually do something.

As for the rocket, I see what the problem is there: FAR uses the size of the attachnodes to estimate some drag parameters and the top of the MkII Lander Can is a size 1 when it should be a size 2 based on the part diameter. There are a lot of stock parts that just don't use the proper attachnode size. There's really nothing I can do there since I think using ModuleManager to try and fix that will just cause craft files to break.

@egreSS: That is an issue I've noticed myself (but probably not to the same extent that you're noticing it). Hopefully a fix will be in the next revision.

Edited August 27, 2013 by ferram4

[Torminator]

I've had that "skip" issue with and without FAR. I think it's just a KSP thing. Drives me insane.

How detailed does FAR's model get? For instance, if I have a secondary wing above a large delta, will the delta affect the characteristics of the secondary wing above it?

[a.g.]

As for the rocket, I see what the problem is there: FAR uses the size of the attachnodes to estimate some drag parameters and the top of the MkII Lander Can is a size 1 when it should be a size 2 based on the part diameter. There are a lot of stock parts that just don't use the proper attachnode size.

Does it work correctly with the new scalable procedural fairing base? The attachment node indicator size doesn't appear to change in VAB as you scale it.

[jrandom]

To weigh in here: I don't want "easy" or "hard", I want realistic -- as close as can be had within the constraints of the game engine. Making the atmosphere soupy for no other reason than to increase to-orbit delta-v is just bad, bad thinking. I don't want the game to be my enemy, I want physics to be my enemy. If certain tasks are easier, fine. If others wind up being harder, that's fine too. Just give me an atmosphere that behaves like an atmosphere -- that is the joy of using this mod.

[KerbMav]

I just upgraded from 9.5.5 to the latest version and ran into an issue only appearing in the VAB/SPH in which the game ran extremely choppy. specifically - not a normal video/FPS lag but alternating smooth FPS and short pause. like if someone was pressing pause and unpause key really fast or a background application that used alot of harddisk IO(I checked, only KSP).

Ah, so this is what ... at least I can reinstall my mods now ...

[Van Disaster]

I've had the micropause issue too - I traced mine back to Extraplanetary Launchpads doing something silly ( and found why & pointed that out, and apparently it's being fixed ); it wasn't the full story as something else was magnifying the issue, so I guess that thing was FAR - but removing EL made my SPH smooth again, so it can't really have been that bad.

Arbitrary changes in drag? who decides how much drag a nosecone removes and based on what criteria? the stupidity of using arbitrary rules like that is you've got a perfectly good, well understood, tested and balacned set of rules in place already, called real life physics. We definitely need some real-life tuned rocket components.

[KerbMav]

I do not use EL, but get the lags in VAB.

[asmi]

Just found this video: http://www.youtube.com/watch?v=SnNNa_kmKyU

Flat spin is obviously not a problem for this plane! Just WOW!

[Van Disaster]

Woah. Sukhoi have some crazy talented designers... I wonder what the bounds for pulling that sort of thing off is. Unfortunately for this sort of craziness I think the future is drones, and then you can abandon a lot of the rules for conventional design anyway.

Ferram, are there any gotchas for swept-forward wings, as far as FAR is concerned? I have this:

which as it looks is pretty docile - however when I've deliberately completely stalled it has been totally unrecoverable.

[ferram4]

Add a larger vertical tail and rudder to that to try and make it more resistant to asymmetric stalls. Then make that canard have a higher aspect ratio so that it stalls before the main wing.

I've fixed the B9 intake bug and I think I've fixed the VAB and SPH lag; now I just have to look into the MOI errors with procedural wings and then the revision will be released.

[BubbaWilkins]

Add a larger vertical tail and rudder to that to try and make it more resistant to asymmetric stalls. Then make that canard have a higher aspect ratio so that it stalls before the main wing.

I've fixed the B9 intake bug and I think I've fixed the VAB and SPH lag; now I just have to look into the MOI errors with procedural wings and then the revision will be released.

How about the propeller drag issue?

[Van Disaster]

Hm, larger stabilizer, yes. Oddly when I've been watching the wings in flight, the canard is 100% stalled before the main wing despite what the graph's been saying.

[ferram4]

@BubbaWilkins: Just remembered that one; I think I can work something out, although it will probably require propellers be powered by the Firespitter plugin.

[BubbaWilkins]

@BubbaWilkins: Just remembered that one; I think I can work something out, although it will probably require propellers be powered by the Firespitter plugin.

I'm not aware of any that exist that aren't, so that would probably work just fine.

[DaRocketCat]

What exactly do I do with modulemanager.dll?

[KerbMav]

What exactly do I do with modulemanager.dll?

Put it directly into the gamedata folder, you only need one (it comes with serveral mods or as a standalone download), use the youngest.

[egreSS]

On Rockets;

What are the current thoughts on, and common solution(loaded question, I know) to uncontrollable spin issues about the lateral axis(pitch & yaw) of large >= 3.75m rockets during the gravity turn ascent phase?

While building the rocket I used a very heavy but perfectly symmetrical(x1) test payload which worked fine, I've only been getting issue's after changing the payload.

• Exactly which metrics should I be looking at to troubleshoot loss of attitude control (pitch & yaw) during initial ascent turn?

• Is there a recommended ascent profile for FAR rockets? Or what have you had success with on your larger rockets?

• Are lift-producing components(B9 winglets) an absolute no-no on the bottom of the first stage?

• Also, do proc. fairings, etc. al. negate drag effects of the contained payload?

I realize a picture would help but I'm asking in a general sense, looking ahead and want to try to fix it myself - for didactic purposes.

My theory is the problem is caused by the payload producing too much drag or the bottom stage winglets producing too much lift(or too little drag). How can I prove which one?

[NathanKell]

I can answer 2-4. Recommended ascent profile is start turn early--anywhere from 200m+, certainly no more than 1km or so--and end it by, say, 45km.

2. (For a 2-stage rocket with TWR->MAX of 1.5->2.5 each, I did .7km start, 38km end, shape 38, and did 3140m/s to 80x80km. Can be optimized further.)

3. Which stage? They're very important for your lowest stage. If you add them to upper stages, make sure you add even more to your lower stage, such that CoL is lower than CoM in the VAB. You WANT lots of lift at the bottom! That's how you get stability.

4. Yup!

[NathanKell]

Now, my own issue. For my mission log (and for early MCE missions) I've been launching lots of 0.625m rockets. And I find they suffer FAR (heh) higher drag losses than the 1.25m stack size rockets, for me at least.

Example: with roughly equivalent TWRs, I get about 3x the drag loss for the 0.625m rocket, which means >500m/s more cost to orbit.

Pics of said rockets: (and note that while the example here uses an SRB first stage, I see the same thing with a liquid first stage--I know the SRB has a shorter burn time, which contributes, but it's true of the upper stage too.)

In flight:

Also, a weird problem with the upper stage of the mini rocket:

With no engine (looks normal for CoL)

With engine (WEIRD):

[taniwha]

egreSS: So long as your CoL (really CoD, but only CoL is shown) is below your CoM, you will be stable. Fins help and are often required. KSP lift (even with FAR) relies on angle of attack, so no need to worry about undesired lift.

[Duxwing]

On Rockets;

What are the current thoughts on, and common solution(loaded question, I know) to uncontrollable spin issues about the lateral axis(pitch & yaw) of large >= 3.75m rockets during the gravity turn ascent phase?

While building the rocket I used a very heavy but perfectly symmetrical(x1) test payload which worked fine, I've only been getting issue's after changing the payload.

• Exactly which metrics should I be looking at to troubleshoot loss of attitude control (pitch & yaw) during initial ascent turn?

• Is there a recommended ascent profile for FAR rockets? Or what have you had success with on your larger rockets?

• Are lift-producing components(B9 winglets) an absolute no-no on the bottom of the first stage?

• Also, do proc. fairings, etc. al. negate drag effects of the contained payload?

I realize a picture would help but I'm asking in a general sense, looking ahead and want to try to fix it myself - for didactic purposes.

My theory is the problem is caused by the payload producing too much drag or the bottom stage winglets producing too much lift(or too little drag). How can I prove which one?

How:

For a stable launch in FAR, keep your rocket vertical until your atmosphere indicator goes to the middle bar, which it should do around 12km, and then slowly turn your rocket to the horizontal while keeping an eye on your apoapsis: if it rises near 70km, and you're not pointing into the brown, then deactivate your engine and turn until you are. Once you so are, burn in a direction that expands your orbit expands while maintaining your apoapsis until your apoapsis turns too far away. At that point, coast to a comfortable distance to apoapsis and circularize.

Modifications:

Put fins on the bottom stage of your rocket or replace a non-gimablling lower stage motor with a gimballing one in order to give yourself the control authority that you'll need to keep your rocket straight despite occasional errors of the PID.

Why:

FAR simulates air as air, and objects experience apparent wind while travelling through air. If you've ever stuck your head out the window of a moving car or run at a full sprint, then you've experienced apparent wind: it's the apparent motion of air relative to you. Apparent wind applies to rocketry because rockets upwards hurtle through the air at breakneck speed and thereby generate equally roaring gales. When the rocket is pointed straight upward, these gales act on the nose-cone of the rocket and are simply interesting; however, when the rocket turns, they act on the exposed side of the rocket and press the turn outward. Long, skinny rockets are particularly susceptible to this push because their entire lengths act as levers.

Yet the force of the apparent wind is proportional not only to the airspeed, but the density of the medium. Again, we experience this phenomenon firsthand: we hardly mind when a column of air the size of a refrigerator rushes at us at a hundred kilometers per hour, but I doubt that anyone here would stand tall in the face of a column of water that's likewise hurtling. I've studied my Kerbal Engineer readouts and observed that the fluid medium through which we move on Kerbin--its atmosphere--becomes exponentially less dense as we rise past 12km. In this case, the practical implication of this rapid decrease in density is that gravity turns are much easier and safer above 12km than below 12km; the natural conclusion to draw from this implication is to take the shortest route to 12km (a straight line) and then slowly turn over for orbital maneuvers.

Fairings: Yes! Use them to cover up anything that looks gnarly.

-Duxwing

Edited August 31, 2013 by Duxwing

[egreSS]

Thanks Dux, I'll try that out as soon as I get a chance.

As a side note, I just had an amazing idea - Using part highlighting to color a part with dynamically updateable values for user defined variables.

e.g. a tail fin that turns from blue to red depending on it's Cl or other specified value.

Would make troubleshooting designs across dynamic environments a lot easier.

[ferram4]

@egreSS: Most people have answered the questions pretty well, but I'll give you my opinions too:

1. One thing to look at is how the CoM shifts during the ascent; it is always possible that it is shifting back enough for the rocket to become unstable, particularly if the rocket has a first stage containing multiple fuel tank parts with any of them above the CoM, though I don't think that's the problem with these rockets. Another thing to look at is the amount of control authority you get from gimbal-equipped rocket engines; if you are too reliant on reaction wheel torque you will probably lose control. A final thing to check is what dynamic pressure you detach the lowest core booster; if the dynamic pressure is high when you do that you might lose control momentarily while waiting for the second stage engine to ignite and give you control.

2. If you have a TWR ~1.4 I've found the best option to be to go straight up until you're going 60 m/s and then angle the rocket about 1-2 degrees off the flight path in the direction you want to turn. After that, try to keep the rocket pointing surface-prograde, and don't drift outside of the prograde ring during the ascent unless you know the rocket can take it.

3. Putting lifting surfaces on the bottom of your first stage is generally a good idea, since it will drag the CoL further behind the CoM.

4. Yes they do.

As for the differences between your rockets, it is always possible that the drag of the fairing for the smaller rocket is worse than for the larger rocket. Another possibility is that the parts (which are mod parts in both rockets) are not set up with the correct size attachnodes to handle drag properly; 0.625m uses size 0, 1.25m uses size 1, 2.5m uses size 2, and so on.

I'm not sure what's causing the CoL bug, but I'll look into it. I suspect I might just have to scrap my current system and rebuild it.

taniwha: The "CoL" indicator is actually the Aerodynamic Center in this, which accounts for where lift and drag is applied. I honestly hate the term "Center of Lift" since it implies that the drag of a vehicle either doesn't affect it's stability or can (somehow?) be considered separately to figure out stability; it's just an ill-conceived term and I've only heard of it after it was implemented in KSP. I'd prefer referring to it as the Aerodynamic Center, but I'd lose everyone if I did, so I'm stuck with CoL, along with the other Kerbalisms like "Asparagus staging" and "Onion staging."

Just to be completely explicit, the Aerodynamic Center is defined as the point on the vehicle where all of the aerodynamic forces can be modeled as separate lift and drag forces (which vary with angle of attack) and a constant pitching moment (which doesn't vary with angle of attack, and so doesn't affect stability).

Edited August 31, 2013 by ferram4
Browser messed up formatting; post changed to include paragraphs

[taniwha]

Ah, so FAR makes it the aerodynamic center? (new term for me, but cool!)

The reason I never completely trusted the CoL indicator is because some of my planes were stable out to about 60 degrees, then suddenly became quite unstable (long nose, wings far back, with a high(?) aspect ratio (like a glider)).

Anyway, thanks for that explanation!

[egreSS]

Thanks I have a much better understanding of what to do. After compiling everything I'm fairly certain the problem is, at least in part related to the rapid angle change during the ascent, the other being the altitude. NathanKell was the one with the possible attachNode issue.

The only thing I'm still fuzzy on are the metrics in the stability analysis screen to check if the other points are met. It may not be needed as I stated above as I'm sure it's my ascent profile - both angle and possibly altitude. I was planning on setting the speed between 60-200m/s and paying close attention to Cl and Cd in the analysis screen along with adjusting my ascent as you mentioned.

Interesting note - The issue happens during the first stage (right before separation of some insignificantly small SRB's) but also before any massive change in CoM occurs due to fuel loss. Once the AoA nears a 90° difference to prograde a truly spectacular display occurs of hundreds of parts both small and large flying around, colliding with each other. I've actually managed to get the payload into orbit once after a spin and subsequent loss of a bunch of different parts, and except for 1 of the 2 oversized, seeming important but apparently not, SRB's, the greater silhouette remained largely intact - Wish I recorded it. The greatest difficulties were ~200m/s and ~4-7km. - I'll post pictures after I get a chance to test things out, and hopefully one with the solution.