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

[Van Disaster]

@Ferram: yeah that sounds a pretty bad idea, specially when using the wildcard version of MM will do the same thing.

I don't know much ( anything particularily ) about subs, but ships have some interesting interactions with the surface... the only thing resembling a hull I've seen is the firesplitter floats, though. Caveat, not checked that carrier mod. I guess if you ever did ship hulls properly ( which would need proper hulls first! ) it wouldn't be a giant stretch to do waveriding aircraft.

Edited August 26, 2013 by Van Disaster

[DuckZero]

I ran into that exact behaviour once before; it was caused when I changed FAR versions from one that did not use ModuleManager to one that did. Could that be your problem as well?

That was it. I looked at the part.cfgs and found that they still had the part modules in them.

The culprit was Taverio’s Pizza and Aerospace not being updated to play with FAR nicely. (Plenty of posts about it in that addon's thread).

Thanks

Edited August 26, 2013 by DuckZero
accidentally a word

[cameroon]

So I've caved and am going to ask

I can't find resources that spell out how I should read the Static Analysis graph (believe it or not, google for '"Static Analysis" aerodynamic graph' you end up at brief mentions within the FAR thread).

While I understand a general "Negative Cm is stable" and that the sharp divergence of the various plots happens at a stall, I can't seem to determine how I should interpret them for the purposes of flight operations. In particular, while the help for the graph states that it plots a positive and negative AoA I don't know if that's the divergence of the lines (and then no idea how to read them if it is). My current "best guess" is that the graph below is telling me that the aircraft would be stable with the nose pitched up to ~16 degrees from the velocity vector, after which it would stall and then not recover until the nose was pitched down to somewhere around -28 degrees? Or do I have that reversed (or, likely, just altogether wrong)?

I've come and gone from FAR a couple times, in part because I feel like I've never had a grasp of the analysis tools, but I'd really like to make it one of my must-have mods because it really improves the experience. So thank you in advance!

[taniwha]

Going by your graph, your vessel should be stable out past 45 degrees. So long as the yellow (Cm) line is negative (for positive AoA), your vessel is stable. Try checking out to 89 degrees (90 crashed FAR last time I tested (a few versions back)). If it's negative all the way (going by the slope at 45 degrees, I'd say it is), you have an always stable vessel.

[cameroon]

Going by your graph, your vessel should be stable out past 45 degrees. So long as the yellow (Cm) line is negative (for positive AoA), your vessel is stable. Try checking out to 89 degrees (90 crashed FAR last time I tested (a few versions back)). If it's negative all the way (going by the slope at 45 degrees, I'd say it is), you have an always stable vessel.

Thank you taniwha, as I understand it the craft would be stable ("controllable") but the stall around 16 degrees would kill my lift. The negative Cm line means I would (in theory) have enough control authority to try and get my nose pointed back at my velocity vector for recovery.

I think one reason I was having trouble was that my aircraft didn't have the pitch authority to hit 16 degrees AoA (it does now, in my quest to figure this out).

That still leaves the splits - I definitely don't know what the split is actually saying.

[Van Disaster]

Er, stall at 28 and recover at 16 seems more likely ( I am rather tired & can't remember those properly ). One wing stalling before the other would end up in a spin, which is not really controlled flight

[taniwha]

Yes, enter stall at 28, leave stall at 16 (or so, I'd use 25 and 15 for safety).

cameroon: with that graph, your vessel should recover from any AoA even with your hands off the controls.

[cameroon]

Er, stall at 28 and recover at 16 seems more likely ( I am rather tired & can't remember those properly ). One wing stalling before the other would end up in a spin, which is not really controlled flight

Ah, ok, that does seem to make more sense - the graph is showing all positive AoAs, just one is post-stall. I do wish the line was a different color or something. I was reading decreasing AoA as "negative", but that doesn't really make sense now that the fog is lifting.

And yes, spin recovery is something I plan on practicing (I have vague notions individually firing one of the engines to just disrupt the spin). Until then, abort systems that recover the cockpit are the order of the day

[cameroon]

Yes, enter stall at 28, leave stall at 16 (or so, I'd use 25 and 15 for safety).

cameroon: with that graph, your vessel should recover from any AoA even with your hands off the controls.

Time to climb to sufficient altitude to check that out

[cameroon]

Yes, enter stall at 28, leave stall at 16 (or so, I'd use 25 and 15 for safety).

cameroon: with that graph, your vessel should recover from any AoA even with your hands off the controls.

So some testing seems to indicate that it does recover, unless it enters a spin - which it has an alarming tendency to do. Fortunately I wouldn't actually fly this aircraft with this much pitch authority anyway. With this much pitch, it's just as likely to snap in half as it is to actually get where it's going (plus it is super twitchy).

Van Disaster and Taniwha, thanks for your explanations and help - I think I now grasp what that graph is telling me so I should have less frustrating builds

[Van Disaster]

Opposite controls and dive to get out of a spin is the basic idea, not much you can usually do about a flat spin but try and power out of it though. Have you tried the graph with no fuel yet? ( click the Full button ).

[Torminator]

I use TAC Fuel balancer, and if I get into a flat spin, I just dump all my fuel into the front-most tank. Doesn't work on ascents into orbit, but coming back down it makes the plane extremely stable. (So long as down is where you want to go.)

Also works wonders for avoiding re-entering tail-first.

[ferram4]

@Torminator: So while taniwha and Van Disaster should have covered most of it, keep in mind what happens to your Cm (yellow line) as stall occurs. From what it states, your pitch moment becomes more positive as stalling begins; this indicates that your aircraft begins to stall from the rear. Also, the fact that Cm slopes upward very during the stall-transition region basically means that your plane is pitch unstable in that region where the amount of stall is changing; in a fully-stalled or fully-unstalled state the plane is stable though. What this means is that once your plane stalls it will be highly resistant to coming out of that stall. A good way to try and fight this is to build your plane with the forward lifting surface having a larger aspect ratio than the aft lifting surface; alternatively, build in positive AoA to the forward surface or negative AoA to the aft surface to try and force the front surfaces to stall first; this will make the plane stable in stall.

As a heads-up to the B9 air intake bug: I think I've got it solved, just some more testing on the way. It does seem like the drag from the large SABRE intakes is accurate though; they're just too blunt to be all that aerodynamically effective.

[AncientGammoner]

I like the concept of this mod, but the impression I've gotten the couple of times I've tried it is that I don't think it fits well into the Kerbal universe even if it is "realistic". First off, I launch rockets only, so maybe it's great for planes, but for rockets all I see is that drag loss is essentially eliminated for launches. I launched a totally stock, very simple two stage 2.5m vehicle, once with FAR and a flattop, once with FAR and a nosecone, and once without FAR and a flattop, and the results were 2725 m/s, 2635 m/s, and 1659 m/s, respectively, delta-v remaining after burnout (yes the nosecone actually hurt, probably because the 2.5m nosecone is so heavy). Now maybe eliminating virtually all the drag losses is "realistic", but it changes the game far too much and makes getting to orbit (which is already easy) even easier, not sure why that would be a desirable change at all but that's just my opinion. Also, trying to impose realism on Kerbin might seem like a noble goal, but Kerbin is far far from realistic intrinsically. I mean the planet's density is 3 times that of Uranium

[Torminator]

Yes, now try launching a large rocket with an awkward payload. It makes simple rockets easier, and big rockets a nightmare (for me). I've actually taken to shuttling station modules with my heavy SSTO spaceplane.

Ferram, while I'm here, I notice several of my planes , past a certain angle of attack, will pitch up violently, like if a gust of wind catches a sheet of plywood (probably exactly like). My best guess is a combination of factors:

1) I use delta wings. All the time. This presents a huge surface area to the air, and that drag intensifies the upwards pitch

2) I also frequently use large canard surfaces to move my CoL forward. While I'm assuming that rear horizontal fins naturally bring the plane back to 0 AoA, forward surfaces want to intensify any deviation. Same thing with having a vertical stabilizer on your nose. This is because, if a rear stabilizer moves out of the velocity vector, the drag force naturally pushes it back, whereas if it's in the front of the CoM (CoL?), the drag force wants to make it continue around until it is in back. Like a caster wheel? Maybe? I'm rambling.

Is this an accurate (or at least passable) assessment? On a lark, I made a plane shaped similarly to an MD-80, incredibly long fuselage, long, narrow wings (compared to deltas, at least), and engines way in the back near the tail. It ended up being the easiest to fly plane I've ever made, even though my sense of aesthetics was furious.

[ferram4]

@AncientGammoner: I would argue that the huge drag losses to orbit (which are often equal to gravity losses) are an incredibly poor method of adding difficulty; it's not something to be overcome with skill, it's just a problem to be handled by throwing more rocket at it. Further, anyone coming into KSP with some knowledge of actual flight mechanics (say, someone like me) will end up being incredibly frustrated when they've gone an entire 5km up before starting their gravity turn and the rocket still falls into the ocean due to the large atmospheric drag forces; real rockets start their gravity turn at ~100m and that kind of disconnect between reality and game (for no apparent reason, mind you) is strange at best.

I also don't see the reasoning to get rid of all realism for aerodynamics; we have fairly realistic gravity, we have fairly realistic part strengths, both of which are weighted towards being easier; why should proper aerodynamics be thrown out the window in favor of an atmosphere of pudding so that the difficulty can be bumped back up?

I also have to assume that your rocket designs are already pretty aerodynamic if you actually consider launching with FAR to be "easy" considering the difficulties most people seem to have with constructing aerodynamically stable rockets. Alternatively, you're good at building rockets that don't have to undergo staging events in the lower atmosphere and so you avoid any stability-based failures there.

I will look into the differences between drag with and without a 2.5m nosecone on the top, since that doesn't quite sound right; would you mind posting a picture of the rocket you used, just so that I can remove any other variables from the situation?

@Torminator: Yeah, you're pretty much right. I think I'm not quite modelling the aeroydnamics of delta wings properly, so they might be more prone to being unstable in stall than they should be.

[cameroon]

@Torminator: So while taniwha and Van Disaster should have covered most of it, keep in mind what happens to your Cm (yellow line) as stall occurs. From what it states, your pitch moment becomes more positive as stalling begins; this indicates that your aircraft begins to stall from the rear. Also, the fact that Cm slopes upward very during the stall-transition region basically means that your plane is pitch unstable in that region where the amount of stall is changing; in a fully-stalled or fully-unstalled state the plane is stable though. What this means is that once your plane stalls it will be highly resistant to coming out of that stall. A good way to try and fight this is to build your plane with the forward lifting surface having a larger aspect ratio than the aft lifting surface; alternatively, build in positive AoA to the forward surface or negative AoA to the aft surface to try and force the front surfaces to stall first; this will make the plane stable in stall.

I think this was directed at me actually, thank you for describing how to read the split Cm - that's immensely helpful! I have noticed that once in a stall it's far easier to get that particular craft into a spin then it is to exit the stall. I'll definitely try those suggestions out, I think the lift is substantially in the mid-body, but I think they're all 0 AoA. I'd picked up that some positive AoA at the nose or negative AoA at the tail was a good idea, but had never heard why that was.

Unrelated, but I was testing out the FAR Hypersonic craft that comes with the mod using my newly gained understanding of the static analysis tool and noticed that the fuel tanks drain in a very unbalancing manner. I didn't start editing to fix it, just told TAC to balance them all, but I thought I'd mention it.

[Van Disaster]

I'd rather have rockets behaving like you'd expect rockets to behave based on, you know, actual predictable physics you can look up in books, and adjust the default engines and so on ( this comes up regularily in this thread ). Why is adding artificial and illogical "difficulty" at all desireable? if you want to make things actually hard, start paying for your launch vehicles. Without a cap on resource usage the only challenge is going to be making whatever construction you throw together stay in one piece. Planet densities may be crazily high, but the rest of the universe works just like ours does - except the default atmosphere. And well, the sea, but hey at least that isn't just blue land.

Your test case is also totally artificial - it just shows the change in drag loss without exploring any other differences, like the need for aerodynamic design.

Good news about the intakes ( just after I rebuilt my shuttles that used them, heh - at least they behave better with empty tanks ). I hadn't actually noticed any extra drag from the SABRE M intake anyway.

[cameroon]

Opposite controls and dive to get out of a spin is the basic idea, not much you can usually do about a flat spin but try and power out of it though. Have you tried the graph with no fuel yet? ( click the Full button ).

Absolutely, I build with shifting CoM in mind constantly. I usually give myself a decent margin between CoM and CoL, pushing it forward only to the point where control starts to get "twitchy" rather instead of smooth. It's the only way something like this works

Building VTOLs (with Infernal Robotics) is one of the reasons I decided to dive into FAR completely. Stock KSP can't really handle a VTOL if it has lifting surfaces anywhere - all sorts of bad things happen. As ungainly as that ship looks, it flies reasonably well.

[cameroon]

I think this was directed at me actually, thank you for describing how to read the split Cm - that's immensely helpful! I have noticed that once in a stall it's far easier to get that particular craft into a spin then it is to exit the stall. I'll definitely try those suggestions out, I think the lift is substantially in the mid-body, but I think they're all 0 AoA. I'd picked up that some positive AoA at the nose or negative AoA at the tail was a good idea, but had never heard why that was.

Sorry for all the pictures/posts, but learning what the analysis is telling me has really made for a fun evening - thanks to all who have helped me figure it out

I was wrong - this has a lot of lift from the tail to pull the CoL behind the CoM, but giving the horizontal stabilizers a bit of negative AoA seems to have fixed the stall characteristics. It's been recovering from very off-axis stalls, which it certainly couldn't do prior to the rear AoA change.

[AncientGammoner]

@AncientGammoner: I would argue that the huge drag losses to orbit (which are often equal to gravity losses) are an incredibly poor method of adding difficulty; it's not something to be overcome with skill, it's just a problem to be handled by throwing more rocket at it.

Not sure it why it's 'an incredibly poor method of adding difficulty', it's just a loss that you have to overcome, if you didn't have any losses there would be no real game. What do you think is a good method of adding difficulty? I mean you could make the same argument about the gravity, they scaled down the planet to a 600 km radius, if they kept the same density as an average rocky planet (ala the Earth), the gravity would be tiny, like 0.1g. Do you think someone should make a mod that does this, it would be 'realistic', but if it takes only 1 km/s to get to orbit how boring and unchallenging would the game become.

I will look into the differences between drag with and without a 2.5m nosecone on the top, since that doesn't quite sound right; would you mind posting a picture of the rocket you used, just so that I can remove any other variables from the situation

Like I said I think it's because the mass of the nose cone is so huge (0.4 tons) that any aerodynamic advantage is lost.

[cameroon]

@AncientGammoner - I seriously don't get why you're commenting in such a (seemingly) hostile manner. If it isn't your cup of tea, fine but why bother commenting at all if you don't like the premise.

Edited August 27, 2013 by cameroon
fixed name

[taniwha]

AncientGammoner: that rocket will have very low drag (like a canoe).

As for your gravity argument: Minmus, Ike, Gilly. The sizes of planets isn't to make the game easy, but to help with float precision issues and keeping travel times playable. Stock's drag losses do not make things difficult. They only make for monstrous behemoths and bad flying.

Your test rocket is really pretty useless beyond testing (and maybe delivering kerbals to orbit). Try getting a 100t "wide load" into orbit. Doing so is actually more difficult in FAR. Yes, it takes less delta-v, but unlike stock, it's not just a matter of more boosters and more struts. You have to watch your CoM, CoL and CoD all the way up past 20-30km (I've had rockets flip at around 40km (rare)).

KSP is about having fun while learning a bit about physics. Stock aerodynamics throws that out the window.

[egreSS]

Okay wtf, if this is a double post...

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). I re installed 9.5.5 and the issue went away

also, I loaded multiple crafts(SPH & VAB) and removed previous version first(just gamedata/far) - Wondering if I just missed something?

[AncientGammoner]

@AncientGammoner - I seriously don't get why you're commenting in such a (seemingly) hostile manner. If it isn't your cup of tea, fine but why bother commenting at all if you don't like the premise.

I apologize if I sound hostile that's not my intention, I'm just stating my opinion. I said in my first post that I liked the concept, but the mod claims to "fix aerodynamics" and I don't think lopping 1 km/s off the delta-v to orbit is fixing anything, any more than, say, changing the wet to dry mass ratio of all tanks from 9:1 to 100:1 because that's the ratio in real life. I think mainly I'm wondering why we can't have 'semi-realistic' aerodynamics, where using a nosecone is better than not using one but the level of drag loss is still relatively close to what it is now (which is what I had imagined this mod would be like, not like 100x less drag loss). So if the loss is like 1000m/s now, have using a nosecone drop it to 750m/s while not using one raises it to 1250m/s, but keep it within the ballpark at least.