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

[capi3101]

On 11/11/2016 at 8:11 PM, StellarumSectatio said:

Haven't ever used Ferram but i'm looking to get into it (I need more realistic aerodynamics to advance any further in my glider development)  Does anyone know if Ferram aerospace was/is/will be compatible with B9 Procedural Wings?  They're pretty essential for my purposes.

Absolutely. All of my designs use B9 Procedural Wings and I fly far - for an example, here's a screenie of the Auk II.

This is from 1.1.3, of course; I haven't switched to 1.2.1 pending FAR.

[DoctorDavinci]

11 hours ago, ITman496 said:

So I've been reading through the thread but I haven't quite figured this out yet...  When KSP updated to 1.2, what broke in FAR that prevented it from working?  Just curious, mostly. I like hearing about this kind of thing from a programmers point of view.

It broke quite a bit as many of the procedure calls within the code don't work no more (they've been changed in U5) ... for example, one of the .dll's from pre 1.2 has been merged with another one which means any snippet of code that references that pickle will end up null spamming and such

There is a fair bit that has changed which makes updating highly complex mods more of a challenge ... just think of how long it took for us on the BDAc team to get a working release out (there was quite a bit of work involved and FAR is way more complex)

Hope this helps you understand 

[ITman496]

Oh!  I forgot about the engine change.  That is a mess.  I wish everyone the best of luck in finding the time to de-pickle everything and get it all working again!

[FunnelVortex]

2 hours ago, DoctorDavinci said:

It broke quite a bit as many of the procedure calls within the code don't work no more (they've been changed in U5) ... for example, one of the .dll's from pre 1.2 has been merged with another one which means any snippet of code that references that pickle will end up null spamming and such

There is a fair bit that has changed which makes updating highly complex mods more of a challenge ... just think of how long it took for us on the BDAc team to get a working release out (there was quite a bit of work involved and FAR is way more complex)

Hope this helps you understand 

 

Didn't the change to U5 happen in 1.1 though? Don't see how that's the issue. Something else probably changed between 1.1 and 1.2.

[sarbian]

1 hour ago, FunnelVortex said:

Didn't the change to U5 happen in 1.1 though? Don't see how that's the issue. Something else probably changed between 1.1 and 1.2.

Unity was upgraded to 5.4 and that leads to some legacy Unity code not working anymore (shader loading from string for example). And there was some quite large code change on KSP side in 1.2. 

[ss8913]

6 hours ago, sarbian said:

Unity was upgraded to 5.4 and that leads to some legacy Unity code not working anymore (shader loading from string for example). And there was some quite large code change on KSP side in 1.2. 

anything I can do to help?  can I pull the 1.2 branch, install it, and submit bug reports/send logs?  Or are we not at that point yet?  @ferram4 please do let me know when that kind of help is desired, I will provide.

[Maeyanie]

On 2016-11-12 at 0:54 PM, stellargeli said:

Is that correct?  Kerbin has the same gravity as Earth.  Is the difference the density of the atmosphere?  Or is it just that Deadly Reentry isn't realistic to start with?  Size of the planet should not matter.  Gravity and atmosphere would be what matters.

Sorry for posting off-topic one last time, but I felt it important to explain conceptually why the formulae people listed are true.

Being in orbit isn't really being in zero gravity... in fact, gravity is still pretty close to the same as on the surface. Being in orbit is, in a highly simplified way, that you are moving sideways so fast, that by the time that gravity pulls you in, you have gone far enough sideways to miss the ground and stay in space. So if the planet is larger, even with the same gravity and altitude, you have to be going faster to be able to miss it.

It's kind of like the old Douglas Adams joke, "There is an art to flying, or rather a knack. The knack lies in learning how to throw yourself at the ground and miss."

Edited November 15, 2016 by Maeyanie

[MaxRebo]

13 hours ago, ss8913 said:

anything I can do to help?  can I pull the 1.2 branch, install it, and submit bug reports/send logs?

You can do exactly that. In fact, I've been playing with it for 2 days now. So far I experienced nothing out of the ordinary to report though. In other words, it seems to be extremely stable already.

Edited November 15, 2016 by MaxRebo

[FunnelVortex]

8 hours ago, MaxRebo said:

You can do exactly that. In fact, I've been playing with it for 2 days now. So far I experienced nothing out of the ordinary to report though. In other words, it seems to be extremely stable already.

There's a dev build out already? Where can I download it?

 

I'll be sure to report on bugs, of course  

Edited November 15, 2016 by FunnelVortex

[Akira_R]

8 hours ago, MaxRebo said:

You can do exactly that. In fact, I've been playing with it for 2 days now. So far I experienced nothing out of the ordinary to report though. In other words, it seems to be extremely stable already.

 Dev build!! Sweetness why did I not think to check the branches on Git, @ferram4 sorry to poke you but does the current dev build have the pull request submitted by Shadowmage to help SSTU parts play nice with FAR merged in it?

[tetryds]

If nothing has been announced in the thread it makes sense that it should only be used by people with deeper knowledge about the KSP code and coding itself, who can report nasty bugs and help improving it.

As always, the best practice is that you should not be using it if you don't already know how to do so.

IMO it's also a disservice for the mod development to help other people do that.

It's important to mention that since a development version was not even mentioned you should expect absolutely no support from it, only report issues that you have tested and confirmed and report them on github, never here.

Otherwise people will think there is a version they can use to play and everything turns into chaos

Edited November 15, 2016 by tetryds

[MaxRebo]

1 hour ago, FunnelVortex said:

There's a dev build out already? Where can I download it?

If you stare at Github long enough it will eventually reveal itself Just remember that you also still need MFI, which obviously isn't included in the FAR repository. (I hope that's not helping too much @tetryds)

But as I said, it seems to be so stable already that I doubt there's much left we can help with from our end at this point. An actual release is probably imminent.

[ss8913]

So I've installed the test build... It does seem to be functional, my planes no longer crash immediately on takeoff as they did with stock aero, so that's a marked improvement

Few things I've noticed, and I'll file these as bugs/issues in github as soon as someone confirms that this isn't expected behavior:

1. SPH - transonic design, all lines are pink.  guides used to be black, curves used to be yellow, blue, green.. and lines do not show through parts like they used to.
2. In-flight - seems like stall speeds are generally higher, but I also notice that stock Mk3 fuselage, cockpit, etc parts are not producing any lift as per the aero overlay (F12)

Are these functioning as intended or should I file bugs?

 

update: after a few flights I noticed that in the FAR window the mach numbers are stuck at 0.00... indicating that, perhaps, FAR had stopped working; KSP.log has no direct evidence of this (is the log in the KSP_x64_data/ folder preferable to the one at the top level?)
update2: yes, I do have modular flight integrator installed

 

side note: All that work for 1.2 and the wheels are still terrible.  Planes still pull off the runway and still require parachutes to stop any plane over 200t or so.. wheel brakes won't cut it.  Unless this behavior is correct and 1.0.5's behavior was unrealistic(?)  but that's not a FAR issue, I digress.

Edited November 16, 2016 by ss8913

[MaxRebo]

2 hours ago, ss8913 said:

1. SPH - transonic design, all lines are pink.  guides used to be black, curves used to be yellow, blue, green.. and lines do not show through parts like they used to.
2. In-flight - seems like stall speeds are generally higher, but I also notice that stock Mk3 fuselage, cockpit, etc parts are not producing any lift as per the aero overlay (F12)

[...]

update: after a few flights I noticed that in the FAR window the mach numbers are stuck at 0.00... indicating that, perhaps, FAR had stopped working; KSP.log has no direct evidence of this (is the log in the KSP_x64_data/ folder preferable to the one at the top level?)

I experience none of the problems other than (1) - but I currently use OpenGL so I blamed that. Mach is displayed correctly no matter how often I revert or start new flights and all parts seem to have proper body lift. Are you sure you grabbed the right branch?

[ss8913]

44 minutes ago, MaxRebo said:

I experience none of the problems other than (1) - but I currently use OpenGL so I blamed that. Mach is displayed correctly no matter how often I revert or start new flights and all parts seem to have proper body lift. Are you sure you grabbed the right branch?

maybe I didn't?  I grabbed the "KSP_Update" branch which last had a commit a few days ago.. that one seemed correct, is it not?

[MaxRebo]

@ss8913 Yeah, that's the one I have... I guess your best bet is to compile reproduction steps and file a github issue for real. If the bug is down to the execution environment (64bit KSP on Windows 7 + lots of mods for me) that will hopefully become clear that way.

[ss8913]

@MaxRebo yeah I'm on 64-bit KSP windows 10, plus lots of mods... I need to try to track down the "mach showing 0.00" as I think that's the most serious error.  Things are definitely behaving differently for a lot of my designs though; my basic design shape for spaceplanes no longer seems to be aerodynamically stable with this version of FAR, and it was fine in 1.1.3; however a "traditional"-style airplane seems to work just fine, so either the old version of FAR was letting me get away with things that it shouldn't have, or the new version of FAR isn't calculating lift correctly.  How are you determining that your fuselages are generating lift?  there's a blue lift line when you push F12 coming from the body?  I'm only getting those from my wings and control surfaces, although I'm not sure if that was or was not happening in 1.1.3 at this point.

[FunnelVortex]

17 minutes ago, ss8913 said:

@MaxRebo yeah I'm on 64-bit KSP windows 10, plus lots of mods... I need to try to track down the "mach showing 0.00" as I think that's the most serious error.  Things are definitely behaving differently for a lot of my designs though; my basic design shape for spaceplanes no longer seems to be aerodynamically stable with this version of FAR, and it was fine in 1.1.3; however a "traditional"-style airplane seems to work just fine, so either the old version of FAR was letting me get away with things that it shouldn't have, or the new version of FAR isn't calculating lift correctly.  How are you determining that your fuselages are generating lift?  there's a blue lift line when you push F12 coming from the body?  I'm only getting those from my wings and control surfaces, although I'm not sure if that was or was not happening in 1.1.3 at this point.

 

You can determine if parts are generating lift in the SPH. Place parts and see where the blue dot goes or if it responds to you placing the part.

 

And also, it's a dev build. And dev builds usually have problems and bugs. I gave the dev build a spin last night, and no, it is not yet calculating everything properly. But @ferram4 has been having to rewrite a lot of code, so any issue is most likely because he hasn't gotten around to coding that yet. 

 

I'd just wait for an official stable release before seriously playing with any new build of FAR.

Edited November 16, 2016 by FunnelVortex

[DoctorDavinci]

28 minutes ago, ss8913 said:

my basic design shape for spaceplanes no longer seems to be aerodynamically stable with this version of FAR, and it was fine in 1.1.3; however a "traditional"-style airplane seems to work just fine, so either the old version of FAR was letting me get away with things that it shouldn't have, or the new version of FAR isn't calculating lift correctly

You do realize that there is no new version of FAR yet ... What you are reporting on is a development build that has not been finished

Referring to it as the 'new' version is misleading and could incite people to flood this thread with requests for a download link for the new version when there is no new version

[MaxRebo]

4 hours ago, ss8913 said:

How are you determining that your fuselages are generating lift?  there's a blue lift line when you push F12 coming from the body?

Yeah, like that. It's not much (no idea if it's supposed to be higher), but there's definitely lift from the cockpit and fuselage:

Spoiler

Mk2:

Mk3:

^ Note lifting force pointing downward for the Mk3 stuff. But then again, they look to me like that would actually be correct.

I didn't create any really exotic designs yet (everything I made so far relies on dedicated lifting surfaces), so that might be why I didn't experience anything craft-breaking. But regardless of how correct or incorrect the calculations are in this state of development, it hasn't just "stopped working" for me yet, not that I noticed anyway.

Edited November 16, 2016 by MaxRebo

[ferram4]

On 11/15/2016 at 2:29 PM, tetryds said:

It's important to mention that since a development version was not even mentioned you should expect absolutely no support from it, only report issues that you have tested and confirmed and report them on github, never here.

@MaxRebo and @ss8913, you are both absolutely terrible at reading. I appreciate the effort, but take it to github where I can track things, not here where you end up burying @tetryds' relevant post.

Also, since it seems necessary, I fully endorse all of tetryds' post, as well as @DoctorDavinci's.  Follow their example, let's not have a repeat of the leadup to FAR for 1.0.

[MaxRebo]

Alright, I'll just be quiet now. For the record tho, I want to strongly point out that I was neither expecting (or even requesting) support nor reporting an issue (as I had none). Just felt responsible to help ss8913 figure out if it's even something he should report, since it was me who said that everything seemed to work fine (maybe PM would have been more appropriate for that) - I wouldn't have replied at all otherwise. Anyway, happy and fruitful bug hunting @ferram4 and Co.

/edit And just so it doesn't get buried again:

On 15.11.2016 at 8:29 PM, tetryds said:

It's important to mention that since a development version was not even mentioned you should expect absolutely no support from it, only report issues that you have tested and confirmed and report them on github, never here.

 

Edited November 17, 2016 by MaxRebo

[ss8913]

22 hours ago, DoctorDavinci said:

You do realize that there is no new version of FAR yet ... What you are reporting on is a development build that has not been finished

Referring to it as the 'new' version is misleading and could incite people to flood this thread with requests for a download link for the new version when there is no new version

good point.  was just trying to get some extra eyes on the dev build, as apparently so has MaxRebo, in order to help Ferram4 with the process.  That was the intent.  Your point is valid, however, that this isn't even an 'official' dev build and people shouldn't be asking for help/etc.  I'll submit all issues found via github, unless they're already reported

Edited November 17, 2016 by ss8913

[Shnyrik]

On 16.11.2016 at 11:40 AM, ss8913 said:

All that work for 1.2 and the wheels are still terrible.  Planes still pull off the runway and still require parachutes to stop

Well, there is something strange with wheels. For example, with stock aerodynamics stock Gull can roll at speeds 50 m/s and higher, with FAR it overturns randomly. The same with braking: it suddenly slips, turns and crashes. That problem existed in stock, but FAR seems to multiply it.

Edited November 18, 2016 by Shnyrik

[GBry]

Is it possible to add limits to a control surface, for example to make spoilerons so that it can go between 0 and +45, so that it doesn't go under the wing? Sorta infernal robotic's controls...