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

[Renegrade]

Here is a <30m/s ultralight

I'll take a peek - thank you!

The UI flickering starts shortly after switching vessels.

Quick question: Are you both updated with KSP 0.24.2 and FAR 0.14.1.1? Prior versions did have a weird flickering effect..

[DaMichel]

@tetryds: Oh this is really nice

I noticed the GUI flickering and Exception spamming, too btw. I think there were some problems already before KSP 0.24.2. But now it seems worse. The GUI completely disappeared during my latest rocket launch. Not sure what conditions trigger it. It was a fairly complex launch with multiple payloads in different fairing compartments. My actions were essentially what steve_v described.

[softweir]

If you're meaning me...

In my case it's pretty straight forward:

Launch a multi (in this case 3) stage rocket.

Seperate final stage from payload once in stable ~100km orbit.

Switch to decoupled stage.

The UI flickering starts shortly after switching vessels.

So far it's repeatable for me, at least as far as the same ascent stages with 3 different payloads.

Just started in on 0.24 so I haven't had a chance to test any other designs yet.

Goes without saying I can't speak for brusura

Does it do it with ALL 3-stage rockets, or just the one design you are working on? It would be most helpful if you upload the .craft file somewhere. Remember to list all the mods you are using! (Better still, reproduce it with all stock.)

Come to that - does it do it with FAR uninstalled?

[peadar1987]

Anybody else having problems with the Rockomax Adaptor causing huuuge drag?

I've built a rocket using it, and it's completely unflyable.

As you can see from the screenshots, removing the adaptor and replacing it with some girders makes a huge difference to the problem (although it doesn't completely solve it). Surely just having a big flat face normal to the airflow should create far more drag than a nice smooth adaptor?

I get the same effect when I recreate the shape of the Rockomax adaptor using procedural fairings, so I think it's more likely that the issue is either with FAR, or with my understanding of aerodynamics!

Anyone have any advice?

For reference, my mod list is:

Blizzy's Toolbar

Environmental Visual Enhancements

Better than Starting Manned

Coherent Contracts

Custom Biomes

Deadly Reentry

Distant Object Enhancement

Kerbal Engineer Redux

Extra Contracts

FAR

Rasterprop Monitor

Kerbal Alarm Clock

Kerbal Construction Time

KerbPaint

KSPX

Hot Rockets

Final Frontier

Precise Node

Procedural Fairings

RealChute

Remote Tech 2

ScanSat

Stage Recovery

[Kalloran]

For the flickers, It might be KER, especially if you're using the release version (1.0.2)

cybutek mentions it.......here:

http://forum.kerbalspaceprogram.com/threads/18230-0-24-2-Kerbal-Engineer-Redux-v0-6-2-9-and-v1-0-2?p=1309987&viewfull=1#post1309987

At least, I'm fairly certain that's referring to the UI flickering. (It's a problem for me too)

Edited July 28, 2014 by Kalloran

[Master Tao]

It is – cybutek was referring to a bug report I'd posted on the previous page.

[ferram4]

@steve_v: I've confirmed that, and I'm testing a fix now.

@peadar1987: The drag of the adapter is correct. The lowered drag using that girder setup is a known exploit that I haven't figure out how to fix without making the game lag every time you stage / lose a part / breaking mods where the attach nodes are inconsistent sizes.

[steve_v]

Good point, will try to whittle down the mods this evening (I do have KER 1.0.2 installed btw)

Damn work interfering with my kerbal time.

Blimey you're quick ferram :-)

Many thanks for all the work, I really can't play without FAR anymore.

Edited July 28, 2014 by steve_v
Ninja'd

[NathanKell]

An even more vivid example of the exploit is to replace all your nosecones with Communotron antennae.

[thorfinn]

An even more vivid example of the exploit is to replace all your nosecones with Communotron antennae.

Trident-style extensible aerospikes!

[peadar1987]

@steve_v: I've confirmed that, and I'm testing a fix now.

@peadar1987: The drag of the adapter is correct. The lowered drag using that girder setup is a known exploit that I haven't figure out how to fix without making the game lag every time you stage / lose a part / breaking mods where the attach nodes are inconsistent sizes.

Thanks for the response Ferram!

So the position of the centre of lift is correct in my screenshots? It seemed a little high to me, but I don't know how it's calculated in the VAB display. If it's calculated for beyond the point when the control surfaces stall out and the entire contraption turns into a flying brick, then it makes a bit more sense to me. I'll just have to be a bit more gentle when I'm manoeuvring in-atmosphere.

[Renegrade]

@peadar1987: The drag of the adapter is correct. The lowered drag using that girder setup is a known exploit that I haven't figure out how to fix without making the game lag every time you stage / lose a part / breaking mods where the attach nodes are inconsistent sizes.

I've always been a bit mystified about how you actually detect changes in size.. I don't think parts have any real description of their size aside from their actual graphical mesh...?

An even more vivid example of the exploit is to replace all your nosecones with Communotron antennae.

I have to admit.. I've done this in the past. And.. not so past. Besides, I often need Communotrons anyhow so...

(Actually I've discovered that procedural fairings can be used to enclose various tidbits like antenna/batteries/science/probe cores, plus making a sleek nosecone-y looking nose, so I'll be doing that a bit less in the future )

[DaMichel]

@Renegade: I have seen some code in FAR which looks at the top and bottom most vertices of the graphical mesh and computes an approximate radius for each side. Not sure if this is actually used though.

@Ferram: Does NEAR also have a FARControlSys instance? Where is the NEAR source code? I assume a switch determines if the code ends up as NEAR or FAR, but i could not find it.

[ferram4]

The graphical mesh method is for determining the properties of a given part, but it's highly approximate. Attach nodes are used to quickly determine if something is really blunt and deserves special drag treatment for that; in theory, I could store the end data from the mesh method and use that, but I suspect that would have issues involving some mod parts until I sort out all the details there.

There isn't a switch in the code to change between FAR and NEAR; that would result in a large amount of crap in the NEAR code that would be unnecessary, as well as adding some (arguably negligible) slowdowns to check against that flag. The NEAR source code exists in another branch of the FAR git repo, with unnecessary stuff removed completely. As a note, FARControlSys doesn't exist in NEAR, as user feedback indicated a desire to fly blind and that information readouts were confusing / scary. I don't understand the logic, but that's the gist of what I got from all of that.

[acc]

I get some weird behavior on .24.2. If I use radial decouplers they almost ever instant explode while decoupling in low and high atmosphere, even if I add powerfull seperatrons and all is exactly centered. the top of the external tank/booster always crashes into the midle section of the rocket. but it also happens if there is nothing attached to the seperators. Duno if it's a stock issue with the ejection force part module or if the latest FAR version is much more agressive than in past.

Another thing is KW rocketry. the pack contains now fins and a (apollo like) petal adapter. Guess all that stuff has actualy no FAR integration and ends with extremly high drag / no shielding.

And at least: for testing parts in atmospheric flight I use small rockets. And to give the rocket the ability to fall back down without crazy turning and flipping I wanted to add 4 symmetry air brakes on the top. But they don't work the way I would expect:

(no fairings just for ground-testing. It's all covered until reaching AP)

So what I would like to have is:

- the possibility to use air brakes on rockets too

- integration for the KWR parts

- the source of the strange behavior of radial decouplers

Thanks

Edited July 29, 2014 by acc

[Renegrade]

As a note, FARControlSys doesn't exist in NEAR, as user feedback indicated a desire to fly blind and that information readouts were confusing / scary. I don't understand the logic, but that's the gist of what I got from all of that.

I found the readouts confusing, but I never wanted them to go away. Since it's a highly technical, field-specific interface, us non-aero-engineers could use some slightly less ..jargony.. documentation (the wiki is actually a good step in the right direction there)... but never for it to go away..

I get some weird behavior on .24.2. If I use radial decouplers they almost ever instant explode while decoupling in low and high atmosphere, even if I add powerfull seperatrons and all is exactly centered. the top of the external tank/booster always crashes into the midle section of the rocket. but it also happens if there is nothing attached to the seperators. Duno if it's a stock issue with the ejection force part module or if the latest FAR version is much more agressive than in past.

I've had that in stock - there's a thread for it in the general forum. Plus a fix, although I don't recommend using that fix in FAR as it uses those rectangular struts.

Another thing is KW rocketry. the pack contains now fins and a (apollo like) petal adapter. Guess all that stuff has actualy no FAR integration and ends with extremly high drag / no shielding.

Mods have to specifically support FAR for things like wings (ooh rhymes). You could create aero entries for the KW parts by copying the equivalent out of FerramAerospaceResearch.cfg. That would give the winglets the same performance as stock winglets at least... There are entries in the configuration window of FAR for cargo bays/interstage/fairings etc, you might be able to set up the petal adapter by putting a piece of it's name in there...

[DaMichel]

Ferram4, thanks for the information.

I asked because i want to include some support for NEAR in my FlightData plugin. A stall warning comes to my mind. I haven't actually tried NEAR yet but as i understand planes in NEAR can stall, right? I assumed maybe some variation of FARControlSys is still around invisibly. So the data could be pulled from there, the same way i do it with FAR.

[Jovus]

If I should put this somewhere else, please let me know.

I seem to be having serious trouble with FAR. I'm running the latest release (14.1.1) on KSP 0.24.2. I'm also running KW Rocketry 2.6 for the fairings (though I have all the parts) and some other mods like Interstellar, Docking Port Alignment, Kerbal Alarm Clock, TAC Life Support and the Enhanced NavBall. I am not running MechJeb at all.

I understand that in order to have a stable rocket you need to get your center of lift below your center of mass so that the rocket will naturally swing itself back to point toward the prograde vector while in atmosphere. However, I seem entirely unable to drop the center of lift below the center of mass, even with rocket designs that are imminently sensible. Here's an example, built in three steps, showing the CoM and CoL at each step: http://imgur.com/3wDRNG4,XAbURb3,XFYHgCg#0

I have also tried exact copies of other rockets I've seen in Youtube videos as examples of good rockets with FAR, and no luck.

I suspect this means my installation is borked somehow, but I've un-installed and re-installed KSP, FAR, and all my other mods multiple times. Can anyone offer some advice?

Thank you.

[Jovus]

I suspect this means my installation is borked somehow, but I've un-installed and re-installed KSP, FAR, and all my other mods multiple times. Can anyone offer some advice?

Yep. I found the problem; for some reason the FAR stuff wasn't overwriting the stock drag model for Squad-released parts. Figures I would 'give up' and post half an hour too soon.

[peadar1987]

Yep. I found the problem; for some reason the FAR stuff wasn't overwriting the stock drag model for Squad-released parts. Figures I would 'give up' and post half an hour too soon.

Did you find a solution? I've had similar things happen with various mods in the past, so a quick fix would be brilliant!

[tetryds]

Hey guys, since you liked my ultralight you may also like this:

It's the latest revision of E36, from Roaken Corp.

Not yet finished, optimal stability over Mach 1.4 at any altitude.

Works with realism overhaul, you just need to replace the fuel tanks.

Download: https://www.dropbox.com/s/366p08ufw17n3a3/E36%20revision%201.craft

Considering making some guide on how to beef up the max wingload, if anyone is interested.

[ferram4]

@acc: In answer to your questions:

1. They're not air brakes, they're spoilers. Their main purpose is to reduce the lift of the vehicle (and produce some additional drag as well) on planes primarily. Dunno why you'd do that, even if you got symmetrical deflections those brakes would do effectively nothing to slow it down; you'd be better of with a chute.

2. I don't write up configs for mod parts, simply because if I did then I'd quickly find more of my time going into adding compatibility for other people's stuff than fixing any issues / adding features in my stuff. There are tutorials on how to do that, and I accept pull requests on data for that stuff.

3. Radial decoupler behavior seems to be the result of whatever was changed to try and fix the 0.24 x64 windows decoupler bug. Nothing I can do about that. Continue to build as if the bug was still there, I guess.

@DaMichel: You'd have to do it all manually. That said, I'd urge you not to, simply because user feedback seems to indicate that one of the nice features of NEAR is the lack of any GUIs to complicate things, like in FAR.

@Jovus: Without an output_log.txt there is absolutely no way to figure out what happened. However, if FAR stuff isn't being calculated, that indicates that either the installation is very wrong (you did copy the folder in the correct directory, and you do have ModuleManager sitting in the GameData root, right?) or that something you've done in-game breaks FAR horribly.

[Black-Talon]

I understand that in order to have a stable rocket you need to get your center of lift below your center of mass so that the rocket will naturally swing itself back to point toward the prograde vector while in atmosphere. However, I seem entirely unable to drop the center of lift below the center of mass, even with rocket designs that are imminently sensible. Here's an example, built in three steps, showing the CoM and CoL at each step: http://imgur.com/3wDRNG4,XAbURb3,XFYHgCg#0

I suspect this means my installation is borked somehow, but I've un-installed and re-installed KSP, FAR, and all my other mods multiple times. Can anyone offer some advice?

Yep. I found the problem; for some reason the FAR stuff wasn't overwriting the stock drag model for Squad-released parts. Figures I would 'give up' and post half an hour too soon.

Did you find a solution? I've had similar things happen with various mods in the past, so a quick fix would be brilliant!

Just a shot in the dark, but you have ModuleManager.dll installed in the GameData folder?

[peadar1987]

Just a shot in the dark, but you have ModuleManager.dll installed in the GameData folder?

Yep, latest version, so it's not that.

[TheKutKu]

One question: is it normal that i have a big framerate drop (15fps) while i fly a rocket/plane with FAR? It happens with every craft with 100+ parts. Does this mod require a good computer? I only have a celeron 1.8 GHZ.