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

[ferram4]

Alright, so v0.15.5 "Haack" is out, with a bunch of optimizations and bugfixes. Also some changes to control surface behavior to make them somewhat more useful.

[Drew Kerman]

my first reaction of course to seeing the release title made me think it was some quick dirty fix for a serious issue

[Major SNAFU]

I have a couple of quick questions that I hope someone can provide some insight upon.

1) I installed (using CKAN) FAR onto and existing career and when I went into the SPH to build a new plane I noticed that a) the center of lift ball is now much smaller and has no arrows. and that no matter where I put the wings, move them to extreme ends of plane, etc. the center of lift ball remains in a single location. It doesn't shift at all that I can see. This made designing planes virtually impossible.

2) So I started a new career giving myself a ton of science points to see if I could replicate the issue in a new career. I went into the SPH and now the center of lift appears to shift when I move the wings around, etc. However, it is still very small and still doesn't have the vector arrows.

So I can now build and fly planes again, but I will have to start my career completely over it appears.

Any observations of suggestions?

[cantab]

I don't believe FAR accurately adjusts the CoL marker. Ignore it and instead use the analysis tools. Run the Sweep AoA on the first page and you want to see the yellow line sloping slightly downwards, that means you're stable in pitch. And remember FAR will make your plane behave differently at different Mach numbers and altitudes, so you need an idea of your flight envelope.

[ferram4]

Well, for one, CKAN installs aren't supported; they always seem to have something wrong and certainly have issues providing correctly functioning and up-to-date FAR installs. Any issues with it are for CKAN's authors to deal with, not me.

For two, while FAR attempts to set the CoL to the correct location (actually placing it correctly is surprisingly difficult to do, because aerodynamics are complicated), and the result is usable, the arrow provided by the stock game makes no sense considering that the CoL is simply a position with no vector associated with it and is absurdly confusing when it comes to rockets. Lots of people claim that FAR is screwing up when they see a rocket with a CoL that has an arrow coming out the side, but that's what should happen if it's supposed to be consistent with the system on planes.

[Major SNAFU]

Well, for one, CKAN installs aren't supported; they always seem to have something wrong and certainly have issues providing correctly functioning and up-to-date FAR installs. Any issues with it are for CKAN's authors to deal with, not me.

For two, while FAR attempts to set the CoL to the correct location (actually placing it correctly is surprisingly difficult to do, because aerodynamics are complicated), and the result is usable, the arrow provided by the stock game makes no sense considering that the CoL is simply a position with no vector associated with it and is absurdly confusing when it comes to rockets. Lots of people claim that FAR is screwing up when they see a rocket with a CoL that has an arrow coming out the side, but that's what should happen if it's supposed to be consistent with the system on planes.

Many thanks.

I am relatively new to KSP and I have only been using CKAN because is seemed the easiest way to add/remove mods. I didn't know that CKAN wasn't supported.

Now knowing this, I will begin the process of learning how to add all of the mods I have loaded manually and see how it goes from there.

[DaMichel]

Ferram4, there is a minor issue with the new control surface gui. The "show" state is cleared when i use undo in the editor.

Perhaps it would be better to make the show state persistent. I think this should fix it. It would be nice also because the state would persist over saves and reloads in flight, and between editor and flight(?).

[Torih]

Hi, I've started to get a lot of FAR errors since this new update. Planes are flying OK so not sure if its an issue or not?

FAR Error: Aerodynamic force = NaN AC Loc = 2614.30385346932 AoA = 0

MAC = 4.92000007629395 B_2 = 0.980000019073486 sweepAngle = 0

MidChordSweep = 0 MidChordSweepSideways = 0

at B9.Aero.Wing.Procedural.TypeB

[ferram4]

@DaMichel: I suppose, but the entire point of hiding it is supposed to make things less cluttered when you go to tweak a part, I kinda think that would defeat the purpose.

@Torih: Need full reproduction steps and the conditions that this is happening under. I cannot reproduce any issues with that.

[DaMichel]

I tend to hide the gui manually. But nevermind. Do as you please. The new gui is a nice improvement over previous version regardless. I very much like it.

[BahamutoD]

@Torih: Need full reproduction steps and the conditions that this is happening under. I cannot reproduce any issues with that.

Not exactly repro steps, but if it helps, I don't get those errors on a plane I built from scratch with Haack, but I do get those errors on an existing plane built during Goldstein.

Edit: You can probably ignore this. It didn't come up on every plane I built in Goldstein.

Edited August 25, 2015 by BahamutoD

[TheHengeProphet]

Control surfaces seem somewhat less capable in this update, but I guess that can be attributed to the adjustment in aerodynamics. I am, however, struggling with that many of my planes now have drastically increased drag, and almost 30% their original control, while also having only about 60% the lift. After wondering what was going on, I started wondering if it was the B9 Procedural wings, so I tested a craft that had stock wings and control surfaces. It worked better than ever... Then I took a craft that was made not long after that (though prior to the update), which was essentially the same, but with procedural control surfaces, and it was nearly un-flyable. I then decided to attach stock control surfaces to one of my B9-winged planes, and it handled fairly well compared to its original arrangement.

This same plane went from having a mach wave drag area of ~0.29 in Goldstein to having one of ~.72 in Haack. However, if I adjust the wing down a couple degrees in Haack, the wave drag drops down to ~.27 in Haack, and up to ~.32 in Goldstein. Yet, if I adjust the wing down a touch further, it goes back up into the .7 region in Haack and ~.36 in Goldstein.

Sorry, I'm taking you on a journey here.

Anyways, I built two planes that are nearly identical, using B9 wings and control surfaces, and they operate... nearly identically. I've been trying to hunt down what is going on here, and I just can't pinpoint it, which is incredibly frustrating. May be multiple things going on here...

[ferram4]

Well, for one thing, the stock wings are overperforming greatly. Procedural wings are much more likely to be correct now, and that should be fixed in the wing overhaul that should be in the next update (unless 1.1 comes out first, in which case, compatibility update first).

I want the craft with the touchy wave drag though. Things should not be that twitchy at all.

[Nothalogh]

I have a stupid question, is it possible to temporarily turn off FAR?

The reason is for testing a craft under stock aero easily

[TheHengeProphet]

I have a number of mods installed. I removed all modded parts excluding B9 Pwings, meaning it has an ugly placeholder that brings its wave-drag down to around .7, but it's the wing arrangement which causes the twitchy nature anyways. I will PM you links to download both the stripped and original.

[ferram4]

@Nothalogh: No, because that would be horribly complicated to do and very likely to introduce bugs, and it simply doesn't have that much use anyway.

@TheHengeProphet: So after messing with the craft, I can confirm changes, though not on the order of 3x change in wave drag; in addition, I can't achieve a wave drag of 0.29 with this at all, mine are all reaching a minimum of 0.79, mostly being around 1.0.

Further testing indicates that large variation only occurs with wing parts, so that means all of this waits until the wing overhaul. No point in chasing after a bug there if I'm going to replace it in the next update anyway.

[cantab]

KSP-AVC still isn't picking up FAR 0.15.5, 2 days after release. Is that much delay normal, or has something gone wrong with either KSP-AVC or on ferram's end of things?

(And yes I know I can manually update anyway, but it would still be nice if AVC did its job.)

[TheHengeProphet]

The low drag was specifically attributed to the original Viper1 cockpit in the Mk2 Stockalike Expansion, due to its sharp profile. That stock cockpit I attached to reduce mod count drastically altered its low drag profile. I might be able to approximate it with stock parts if I use one of the Mk2 tail sections and finagle on some air intakes, but... ugh.

I figured this might be the case. Oh well, thanks for looking into it!

[ferram4]

@cantab: Well, I don't know. I updated the version file as I was supposed to, is KSP-AVC now doing CKAN-style screw-ups too?

It amazes me that automated systems that are supposed to be more reliable than humans always end up being less reliable.

[Svm420]

@cantab: Well, I don't know. I updated the version file as I was supposed to, is KSP-AVC now doing CKAN-style screw-ups too?

It amazes me that automated systems that are supposed to be more reliable than humans always end up being less reliable.

Well maybe not always , but it is amazing the work people go through "to save time". Some spend more time not using the mod waiting for it to be on ckan or fixing the automation own glitches they never would have via manual install. LOL

[StevenRS11]

So true.

[cantab]

It amazes me that automated systems that are supposed to be more reliable than humans always end up being less reliable.

I've found apt-get and dpkg pretty reliable and they've been around for years. Even Android auto-updates work well enough. It disappoints me that when it comes to KSP we've still got no good alternative to doing everything manually like we're stuck in the dark ages. (Not really your fault of course.)

I did manual mod management before, in the .90 days and earlier, and it ended up being simply that I'd install the mods and never update. Not what I want to repeat.

[taniwha]

To err is human, to really foul up requires a computer. -- Seen on a sign on the wall at the university bookstore.

[steve_v]

I've found apt-get and dpkg pretty reliable and they've been around for years.

But they're not entirely automated... the factor that CKAN lacks, IMO, is the extensive testing team behind the Debian/apt system. CKAN is a nice idea, but it's just pulling upstream updates, there's no real human oversight and no experimental -> unstable -> testing -> stable transition cycle.

Result: Latest stuff right away, bugs and incompatibilities included. By the time issues have surfaced (usually on a forum the ckan/netkan bot don't read ) the problem has already been distributed. All it takes is a minor packaging change to foul everything up.

That said, CKAN certainly still has it's fair share of outright bugs. I would have thought KSP-AVC pretty safe by now though.

[Kussris]

I used the FAR mod and made a chart of the drag force. I am not sure if the drag is right; it increases after the speed of 300 m/s (?)