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

[jrandom]

I'd also like to add another to that list, Hodo:

4. More wings == more drag. Add just enough wing parts for the correct amount of lift. More than that and you'll just have a drag-happy thing that doesn't go nearly fast enough. Determining this is pretty much trial/error for most of us (at least, I certainly don't know how to calculate it, and don't really want to within the confines of KSP). But trial-and-error is the Kerbal way!

[Alshain]

Yep. That's plane design for you! A plane that is stable at supersonic speeds is going to be super touchy on the pitch at subsonic speeds. This is true of real aircraft. And the opposite: a plane that is stable at subsonic speeds is going to have pretty much zero pitch control when supersonic and will act like a lawn dart instead. The trick is to find that special middle ground where it's juuuust stable enough at subsonic speeds that you can fly it up to thinner atmosphere and higher speeds where the plane will be more stable. I still have to trial-and-error this but I've gotten much better at my CoM/CoL placements to give me something that has a good chance of working right out of the gate.

When I get home after work I'll take screenshots of the aircraft I posted above, but with the CoM/CoL markers activated so you can see the placement of them. That plane is pretty sensitive on the pitch control when climbing up to mach 1, but past that it flies extremely well in straight lines. Turns like a dead fish, though. It's not the best-ever design.

The reason I keep pushing for dry CoM images is that the dry CoM is not as intuitive as it sounds. I had no idea this was the case until I started using RCS BuildAid and could see it. It changed my entire approach to plane design.

Edit: I just saw some images you posted earlier and wanted to chime in on those: A plane that has lifting surfaces in the front is going to make the oversensitive pitch at subsonic speeds worse, since it'll act as a sort of feedback loop: the plane tilts back so the front surfaces generate more lift, pitching the plane up even more which then generates even more lift, etc... You have to be extremely careful with designs that have any kind of lifting surface up near the front of the plane.

Ok, so that's starting to make some sense. Your saying to sacrifice the subsonic because the sonic is more important (it's needed to break out of the atmosphere). But doesn't it have to land at subsonic?

Snip

I appreciate you trying to help, I really do. But I don't see how that relates to my control issues. I'm not having problems with wings ripping off until I lose control. I'm not making sharp turns... on purpose anyway. My goal is to go straight on 90deg without turning. I did turn the yaw off on the elevators and the pitch off of the tail. Is that right? I think I got that from one of the videos on the front page.

Edited July 1, 2014 by Alshain

[jrandom]

Ok, so that's starting to make some sense. Your saying to sacrifice the subsonic because the sonic is more important (it's needed to break out of the atmosphere). But doesn't it have to land at subsonic?

Yes. In fact, you might be able to take advantage of placing the dry CoM farther forward than the wet CoM to improve subsonic behavior during landing when you're nearly out of fuel. I've not actually tried this yet as it 1) just occurred to me, and 2) I've been playing RSS where building an SSTO plane is right up on the edge of impossible (although it's really interesting to try).

Edited July 1, 2014 by jrandom

[Alshain]

Yes. In fact, you might be able to take advantage of placing the dry CoM farther forward than the wet CoM to improve subsonic behavior during landing when you're nearly out of fuel. I've not actually tried this yet as it 1) just occurred to me, and 2) I've been playing RSS where building an SSTO is right up on the edge of impossible (although it's really interesting to try).

Ok, well I've been doing that backward. I was following the wiki that said to get it stable between 0.2-0.8 first.

[ferram4]

@Alshain: I am actually working on a fork of FAR that is a dumbed-down, simplified version of it to see if people are interested in it as I see people constantly stating that FAR is too complicated. Although it's going to be missing quite a few wing interaction effects, as well as the existence of Mach numbers and all the GUIs, it still feels very much the same as FAR, for the most part. This means that even though the tools in FAR would still be useful, they are removed because user feedback has been consistent in that user are not capable of making use of it and that it is confusing, so tools will be removed so as not to confuse the users.

I will also admit that I find the wiki's page on stability derivatives lacking. However, given that you have not once told us what stability derivatives you are have problems with, nor have you specified the flight condition that has them being problematic, no one has been able to help you, and I can't make changes to the wiki to improve it. Since you asked some to explain why the numbers are green / red, I'll give you an answer: when the derivative is green, that means that it will promote static / dynamic stability in whatever range of motions it controls; red would indicate that it discourages static / dynamic stability. For example, Mw being green would mean that the plane is pitch-stable, while Mq being green would indicate that pitching motions dampen out over time (red would indicate that they grow over time, and combined with a green Mw would indicate that the plane is statically stable, but dynamically unstable. This is rare behavior, but would not be revealed by the simplistic CoL indicator).

I'll also note that you haven't described how you lost control; the vehicle sideslipping and being destroyed requires a different solution to a vehicle that suddenly pitches up and flies apart. But if you're losing control as you run out of intake air, then your problem isn't with FAR; that's stock behavior messing you up, and you would be served well by keeping track of how much thrust your engines are making while running near the edge of the breathable atmosphere.

Edit: another point: most supersonic vehicles pump fuel forwards and back to maintain balance as they switch from subsonic to supersonic flight. You will need to do this if you want ideal flight characteristics in both regimes.

@Starwaster: You have no means to get the changes you want to happen through FAR's existing code, and I don't plan on supporting all-in-one wing/flap parts, as I think they discourage design choices and would either result in many, many calculations on the fly to get things done properly or allow balancing issues to crop up due to FAR. You would be best served by finding another plugin to apply the force for you.

[Alshain]

@ferram4 Thanks I'll look forward to that. Specifically the ones I've got now are Zu (Mach 1), Xw, and something that looks like L-beta @ Mach 2+. I have seen others go red but this is as refined as I've been able to get them. I'm at my computer now so screenshots are coming... oh here they are

I do use TAC Fuel Balancer already to prevent that issue.

Here is the plane, DCoM is not an issue apparently, though that was completely by accident. I'm not certain about the ideal TWR on a plane so tips there would be help.

EDIT: OK here they aren't. Give me a sec to figure out the spoiler tag.

EDIT2: Well since that isn't working here is the album.

Ok, I was wrong earlier recounting from memory. The instability comes AFTER the rapiers switch modes, I try to pitch up and it almost immediately starts going every which way.

Edited July 1, 2014 by Alshain

[ThorBeorn]

@Alshain: I am actually working on a fork of FAR that is a dumbed-down, simplified version of it to see if people are interested in it as I see people constantly stating that FAR is too complicated. Although it's going to be missing quite a few wing interaction effects, as well as the existence of Mach numbers and all the GUIs, it still feels very much the same as FAR, for the most part. This means that even though the tools in FAR would still be useful, they are removed because user feedback has been consistent in that user are not capable of making use of it and that it is confusing, so tools will be removed so as not to confuse the users.

I hope the "complicated" not dumbed-down version will still be available. I find the derivatives very helpful for designing purpose built aircraft. There's even explanations of many derivatives in FAR if you click the help button. Then there's google and the forums. I've learnt a lot about aerodynamics thanks to FAR (and it's complexity) and I hope others will be able to do so too with future updates.

[regex]

I hope the "complicated" not dumbed-down version will still be available. I find the derivatives very helpful for designing purpose built aircraft. There's even explanations of many derivatives in FAR if you click the help button. Then there's google and the forums. I've learnt a lot about aerodynamics thanks to FAR (and it's complexity) and I hope others will be able to do so too with future updates.

A "fork" implies that the original will remain available and, likely, worked on. Anyway, I highly doubt ferram4 would stop working on FAR proper in order to focus on a dumbed-down version...

[ferram4]

There's no hard and fast rule for TWRs, but I generally go with a minimum of 0.3, since that's about the lowest you can get on takeoff without it having terrible climb performance.

As for the stability derivatives, Zu and Xw can be wrong and everything will still be fine, it's just odd. What Zu being wrong is telling you that going faster means you make less lift, which kind of makes sense near Mach 1; it's undesirable, but the only fix is to go faster. Xw being wrong is stranger, because that implies you get less drag from increasing your angle of attack, which doesn't make much sense at first glance. Then the fact that you're checking at Mach 2 flight at SL (make sure to set your density correctly, it changes the results a lot, you probably want 0.2-0.4 for this test) makes it very clear: you're at a low AoA already, so a small increase isn't going to cause much additional drag due to lift, however, it might lead to the vertical tails becoming swept enough that they make less drag, and that ends up being a larger change than the induced drag. In both cases, these don't matter too much and are just flukes.

Lb being wrong is the tricky one, and I suspect it has more to do with the low AoA resulting from the messed up density conditions than anything else. Generally the solution for that would be to add some dihedral to your wings, but I don't think that's what you want to do here, since delta wing craft normally have proper Lb values in standard flight conditions. Now, if you want to be bold, leaving this as it is might be a good idea; I'll explain why:

There's a flight motion that planes with highly-swept wings can experience at high AoA called "wing rock," where the plane rolls back and forth uncontrollably. This is because highly-swept planforms have really strong Lb values at high AoA (in the correct direction, mind you, but while it does make the plane want to roll itself to the "correct" orientation, it does so too... vigorously; Lb is strong, but it is too strong compared to Nb and Lp, which would counteract sideslip (the initial cause of the motion) and rolling motions (the problem itself), respectively). So having a highly-swept planform with improper values at this flight condition (Mach 2, SL flight) might give you much better Lb values at a different flight condition (say, Mach 3.75, 35 km) that will make ascent and reentry easier. It's a balancing act, since planes are pretty much impossible to make perfect in ALL conditions.

[Alshain]

Ok, good info. I didn't know about the density thing. Without changes to the craft, it didn't help with Mach 2, but Mach 3 and 4 it goes green on Lb at 0.4 density leaving only Xw red. This is actually closer to my speed at the time of instability. But if that's the case I can't figure out why I can't control it.

EDIT: Oh, look at that.

I moved the canards forward a smidge and added the control surfaces to them (an earlier suggestion). That is definately capable of making it to my station. It wiggled a bit at the engine switch but stabilized rather quickly.

Edited July 1, 2014 by Alshain

[jrandom]

Heh. I just came in here to recommend you add canards to the front and also move your center of lift closer to the CoM. Which is precisely what you did! When you have rear-mounted wings, it really helps to have front control surfaces to help complete the pitching lever.

[Hodo]

@ferram4 Thanks I'll look forward to that. Specifically the ones I've got now are Zu (Mach 1), Xw, and something that looks like L-beta @ Mach 2+. I have seen others go red but this is as refined as I've been able to get them. I'm at my computer now so screenshots are coming... oh here they are

I do use TAC Fuel Balancer already to prevent that issue.

Here is the plane, DCoM is not an issue apparently, though that was completely by accident. I'm not certain about the ideal TWR on a plane so tips there would be help.

EDIT: OK here they aren't. Give me a sec to figure out the spoiler tag.

EDIT2: Well since that isn't working here is the album.

Ok, I was wrong earlier recounting from memory. The instability comes AFTER the rapiers switch modes, I try to pitch up and it almost immediately starts going every which way.

After looking at your craft I haft to suggest moving your CoL, closer to your CoM, right now you will suffer from a very heavy nose, or a pitch down situation. Perhaps extending the wing roots a bit more forward and the use of canards near the front of the craft.

On a side note suggestion I cannot stress enough how useful Procedural Wings are and HIGHLY suggest anyone using FAR use P-Wings.

[Alshain]

I didn't care for the procedural wings to be honest. But that's just preference I think. In any case my Kerbal Station Crew thanks you all. They can now eat... and you know, continue to breath.

And I'm just going to leave this here too cause well I thought it was cool looking with the Mun and all.

[jrandom]

Glad you made it! (Now you can continue to use RealFAR, despite Ferram's promise to introduce FARLite. )

[Alshain]

We'll see. I'm going to try building one from an earlier career point for 0.42 practice using something other than the RAPIER. I may give pWing another second chance too.

[Starwaster]

@Starwaster: You have no means to get the changes you want to happen through FAR's existing code, and I don't plan on supporting all-in-one wing/flap parts, as I think they discourage design choices and would either result in many, many calculations on the fly to get things done properly or allow balancing issues to crop up due to FAR. You would be best served by finding another plugin to apply the force for you.

WOW. Well at least I'll know where to go next time I need crucifying, so, thanks for that

[jrandom]

We'll see. I'm going to try building one from an earlier career point for 0.42 practice using something other than the RAPIER. I may give pWing another second chance too.

pWing can come in pretty darned handy at times.

It's a little late for this, but here's the CoL on that crazy lifting-body jet I made. (The CoL is up slightly because of the little wingtips I have that are bent upwards a bit to help with roll stability.)

[NathanKell]

jrandom: ferram pointed out that actually dihedral and anhedral reverse affect in the supersonic regime. You'll get better roll stability subsonic and roll *instability* supersonic with those.

It explains wings like the F-104's, which I'd wondered about.

[ferram4]

@Starwaster: I'm sorry, it's just not possible for me to put in without making FAR more hacky and complicated than it is right now. If I try to calculate the flap effect in code from the animation directly, that first requires figuring out what vertices make up the flap to begin with, which isn't necessarily an easy task. It would be easily broken by anyone that did anything more complicated than plain flaps or split flaps. The alternative is to have it as a thing in the config file, but then that requires the hassle where everyone needs configs to make it work, which is something I've been trying very hard to get away from. But then that has the standard problems of someone messing up the input values and all of a sudden weird things are happening, especially since for flaps it can be a little more tricky than just figuring out wing geometry.

Ultimately, the FARWingAerodynamicModel and FARControllablerSurface classes are already so bloated that I don't want to add more to them if I can avoid it.

Also, if this is wrt the KSO, those flaps are in a terrible position. The only thing you'd get out of them is a pitch-down moment to put the plane into the ground; delta wings do not do well with flaps unless there are canards to balance them, which the KSO does not have.

[Guest]

Would that be possible to introduce another FAR module to handle animated wings? KSO flaps aside (they are rather poorly placed, flaps need to be near the COM), there are a few variable geometry wings, wings with leading edge slats and so on that could use FAR support. Say, you'd enter the wing's state with animation "on" and "off" (much like with the current modules), then assume transition between those states is linear (it usually is).

[jrandom]

jrandom: ferram pointed out that actually dihedral and anhedral reverse affect in the supersonic regime. You'll get better roll stability subsonic and roll *instability* supersonic with those.

It explains wings like the F-104's, which I'd wondered about.

Well, that explains that instability I kept running into.

[Alshain]

Would it be possible to make that density entry calculate from elevation? I don't know the density but I do know what elevation I hit at certain speeds.

EDIT: Oh nevermind, after submitting I realized that would assume Kerbin, which may not be where you are.

[RoverDude]

So I am in a quandry.

I recently created an escape pod - it consists of two parts. One has a heat shield, and the other has a parachute. As a result, with FAR, both of them have zero drag, resulting in much sadness (I expect this is because of how FAR auto-adds attributes).

Does anyone have an example of, either, how to add my own FAR attributes to a capsule, or how to make FAR recognize these parts?

TIA!

[ferram4]

FAR should take care of the heat shield, unless you have neglected to place an attach node on the blunt bottom of it. You should look at how DRE's heat shields are set up as an example. As for the pod, you do NOT want FAR to handle that if it has a parachute; all the drag calculations will go to pieces when the chute is deployed. You will have to separate that into two parts or you will have to live with it.

[Scootaloo]

Ok I'm having a really head scratcher of a problem with my large SSTOs.

According to the "Center Of Lift Behind Center Of Mass" Rule, they are stable, and indeed in the lower atmosphere it is indeed stable. However as I climb above 10km and over 300km, the aircraft begins to pitch and roll widely, going into a series of violent nose-up maneuvers, basically going belly-first into the airstream, then dropping the nose back down into the airstream, before repeating the process over and over again.

Can someone explain why exactly this is happening? The airplane is stable, and at no time does the CoL go ahead of the CoM.