[1.9.x] RCS Build Aid Continued - New Dependencies

[MaverickSawyer]

Wait, found the problem. Remember how I said I grabbed the file off of Github?

The .dll file isn't in the download from the master branch, for whatever reason.

EDIT: Hell, the entire Plugins folder was missing from that. Huh.

Edited May 21, 2020 by MaverickSawyer

[linuxgurugamer]

7 minutes ago, MaverickSawyer said:

Wait, found the problem. Remember how I said I grabbed the file off of Github?

The .dll file isn't in the download from the master branch, for whatever reason.

EDIT: Hell, the entire Plugins folder was missing from that. Huh.

Ummm, I don't put dll files into the repo on Github.  That's what the Releases is for:

https://github.com/linuxgurugamer/RCSBuildAid/releases

NONE of my mods have any DLL files in the source directory.  Only in the Releases

[DStaal]

8 minutes ago, MaverickSawyer said:

Wait, found the problem. Remember how I said I grabbed the file off of Github?

The .dll file isn't in the download from the master branch, for whatever reason.

EDIT: Hell, the entire Plugins folder was missing from that. Huh.

That's because the download from the master branch isn't a release download - it's a *code* download.  As in all the code to develop the mod, but it's not what you install.

You need to grab the release.

[Superfluous J]

9 minutes ago, MaverickSawyer said:

Wait, found the problem. Remember how I said I grabbed the file off of Github?

The .dll file isn't in the download from the master branch, for whatever reason.

EDIT: Hell, the entire Plugins folder was missing from that. Huh.

This is why SpaceDock exists.

[linuxgurugamer]

This is why I prefer CKAN.  It knows what to install, where to get it from and where to install it.

[RaccoonTOF]

I realize this has not been updated to 1.10 yet, so if that is what is causing this issue, please just ignore

Using version 0.10.0, when looking at ENGINES they are always displaying tiny torque values if located far enough away from the CoM/have enough thrust.  This occurs even with perfectly centered engines (unless EVERY LF engine is somehow offset slightly, I've tried about 40 different engines from stock and a huge variety of mods, same issues with all of them if appropriately far from CoM for their thrust), in a stack consisting of nothing but fuel tank(s) and the engine on the bottom of the stack.  It is definitely not an issue with the CRAFT CoM being off-center, as I can take the exact same test "craft", slap enough symmetrical RCS thrusters on it at the very bottom to have anywhere from 1kn to over 800kn thrust, and there is never any TRANSLATION torque shown (for the forward thrust direction - obviously there is torque in sideways directions due to them being mounted so far from CoM).  The problem is exacerbated by having either higher thrust engines, or further from CoM, as would generally be expected from an off-center mount or thrust vector.

EDIT: After further testing, it appears that it is registering a very, very slight off-centeredness in the thrust along the yaw axis that does not actually exist.  When on the ENGINES setting, I can adjust gimbals in pitch and roll directions, and opposing directions are always the same torque value, but in the yaw direction, they are off from each other by exactly double the torque value of the "centered" position - again as would be expected by a slightly off-center thrust vector.  However, this appears to be happening with ALL engines, and I kinda doubt that all of their thrust vectors are off-center by the same amount relative to their thrust...(could be wrong though...)

Edited July 11, 2020 by RaccoonTOF

[linuxgurugamer]

2 hours ago, RaccoonTOF said:

I realize this has not been updated to 1.10 yet, so if that is what is causing this issue, please just ignore

Using version 0.10.0, when looking at ENGINES they are always displaying tiny torque values if located far enough away from the CoM/have enough thrust.  This occurs even with perfectly centered engines (unless EVERY LF engine is somehow offset slightly, I've tried about 40 different engines from stock and a huge variety of mods, same issues with all of them if appropriately far from CoM for their thrust), in a stack consisting of nothing but fuel tank(s) and the engine on the bottom of the stack.  It is definitely not an issue with the CRAFT CoM being off-center, as I can take the exact same test "craft", slap enough symmetrical RCS thrusters on it at the very bottom to have anywhere from 1kn to over 800kn thrust, and there is never any TRANSLATION torque shown (for the forward thrust direction - obviously there is torque in sideways directions due to them being mounted so far from CoM).  The problem is exacerbated by having either higher thrust engines, or further from CoM, as would generally be expected from an off-center mount or thrust vector.

EDIT: After further testing, it appears that it is registering a very, very slight off-centeredness in the thrust along the yaw axis that does not actually exist.  When on the ENGINES setting, I can adjust gimbals in pitch and roll directions, and opposing directions are always the same torque value, but in the yaw direction, they are off from each other by exactly double the torque value of the "centered" position - again as would be expected by a slightly off-center thrust vector.  However, this appears to be happening with ALL engines, and I kinda doubt that all of their thrust vectors are off-center by the same amount relative to their thrust...(could be wrong though...)

Can you please provide a craft file and a picture showing the problems?

And, did you actually test the rocket to see if it may be the game causing the problem?  In other words, launch it and see what happens

Following up, I've tested this on 1.9.1 and 1.10, and in both cases with identical vessels, the torque shows was the same.  Very small, and essentially meaningless.

 

Edited July 12, 2020 by linuxgurugamer

[RaccoonTOF]

Yup, what you are seeing is exactly what I was talking about, as shown in your image there.  And for MOST things (sane rockets, with reasonable amounts of control wheels/gimbals) it doesn't actually cause a major issue.  However, it DOES become an issue when trying to build large, multi-part/multi-launch ships where small errors can really add up - and where getting the CoM exactly balanced for each part can be very important to the final stability of the completed multi-part vessel.  Also very important for certain VTOL designs, or "modular" designs which have a common core but you want to be able to have various options for payloads AFTER the initial build in the VAB.

Of course, you can get around this by just building everything in perfect symmetry...but my biggest use of the mod in the first place is for balancing crafts that are NOT perfectly symmetrical  

EDIT: With a gimbal-less engine centered on a single fuel tank with a probe core and nose cone on top, disabling the reaction wheels, and launching straight up...it appears to not waver on the nav ball at all, even with the off-balance torque.  However, if you adjust the craft so that the DISPLAYED torque from the mod is 0 (tested by clipping a test weight into the center of the fuel tank, perfectly centering it with EEX, then using fine adjustment to offset it until the readout was 0kn) it WILL arc the rocket slightly...so definitely an error in reporting rather than an imbalance in the craft itself in the original case.

 

Edited July 13, 2020 by RaccoonTOF

[linuxgurugamer]

It's called floating point error, and is so small that it's not worth the effort.  I hear what you are saying, but this issue is not a small fix, but rather a much larger issue that actually concerns the whole game

Sorry to disappoint, but you are literally the first person who has mentioned this in the entire time that I've been maintaining the mod

[m4v]

The transforms that give the direction of thrust and position are regularly off by  a magnitude of the order of 1e-6 and 1e-7, I don't know if that's a floating point error or if the models ended up that way when Squad made them, but that's how it is and RCSBA can't workaround that, those are the transforms were thrust is applied when the ship fires its engines.

It doesn't matter anyway,

23 hours ago, RaccoonTOF said:

However, it DOES become an issue when trying to build large, multi-part/multi-launch ships where small errors can really add up

Large ships also have a big moment of inertia, a minuscule torque will not rotate them, that's why the plugin doesn't display the circular arrow,  the torque is too low for move a ship that big. Sure, you might see some rotation if you burn for 10 minutes but that's why you have attitude actuators. RCSBA is for minimize imbalances but within reason.

Anyhow how do you know that tiny imbalance is really the issue you're having with your design? At this level of precision I can't even be sure that what you see in the editor will be the same in flight when all the physics are enabled, part' mass are changing and joints are flexing by tiny amounts.

Edited July 14, 2020 by m4v

[RaccoonTOF]

Oh I'm well aware that it isn't usually an issue - I've been using RCSBA for something like 3 years now, and I've been happy with imbalances of .01 or less.  It wasn't until I started recently building complex, multi-launch, relatively lightweight but large ships and using EEX to try to get them as perfect as possible that I even noticed the issue.  But yes, it IS definitely the issue as reported - if you read the edit of my last post, you can see that if I build a craft that is theoretically perfectly centered but RCSBA shows as off-balance, it is indeed centered in flight.  However, if you actually DO shift the CoM (without changing anything else, done with a part-clipped test-weight in that case) so that it zeroes out according to RCSBA, then it DOES get noticable pitchover (actually, yaw-over?) in flight.

Mind you, as I said, I've been working without that level of precision for years now, as until I started messing with the really fine-adjust tools in EEX I had no choice but to do so.  And if it all comes down to a floating point error in display, then yes, it's probably not worth the effort to correct it - simple enough to use reaction wheels to counter the effect when under engine thrust.  I'm rather curious as to why the offsets for RCS ARE more accurate than for engines though, even at higher levels of TWR/distance offset (the above test vehicle with 800kn of RCS thrust all arranged around the bottom of the tank still showed 0.0 torque in the "forward" direction, without the 0.001 error that the 206kn liquid engine gave.  Then again, I suppose that is likely because I usually have RCS placed in symmetry that would tend to cancel out any floating point errors, while I often don't with engines...might need to do more clustered engine designs for the finicky designs mentioned above.)

[m4v]

On 7/14/2020 at 10:27 PM, RaccoonTOF said:

However, if you actually DO shift the CoM (without changing anything else, done with a part-clipped test-weight in that case) so that it zeroes out according to RCSBA, then it DOES get noticable pitchover (actually, yaw-over?) in flight.

Probably drag, clipping doesn't shield parts from aero forces. Disable aero forces and test it again or put the test weight in a service bay.

On 7/14/2020 at 10:27 PM, RaccoonTOF said:

I suppose that is likely because I usually have RCS placed in symmetry that would tend to cancel out any floating point errors

That's what's happening.

[katateochi]

Some things are heavier than others! 
ok that was badly put, what I mean is; some parts seem to alter the COM of a craft more than other parts of the same mass.
I'm working on a tiny science probe and I'm building it so its visually asymmetrical but is still well balanced, so it's quite sensitive to small masses.  And I noticed that some parts with a mass of say 0.01t have a big impact to it's COM (and the Torque value shown in the build aid window), while others with the same mass have no effect. For example;
If I attach a single KAL-1000 controller (which has a mass of 0.01t) then it doesn't alter the Torque value at all (even if I use the offset tool and drag it way off center).  But the Micro Goo Containment Pod (from DMagic Orbital Science, which has the same mass of 0.01t) placed in the same position will alter the Torque value quite a bit (and offsetting it has a big impact).
So I'm wondering why this is. Is it something to do with physics-less parts?

[swjr-swis]

2 minutes ago, katateochi said:

Is it something to do with physics-less parts?

It used to be that 'physicsless' parts applied their drag and mass at the CoM of the parent part. I'm not sure if that is still the case.

[linuxgurugamer]

On 7/16/2020 at 1:31 PM, swjr-swis said:

It used to be that 'physicsless' parts applied their drag and mass at the CoM of the parent part. I'm not sure if that is still the case.

It is

[darthgently]

A question for the RCS Build Aid community.  There is one aspect of RCS balancing that defies my understanding so far.  When balancing yaw/pitch attitude on a well balanced craft (wet/dry/avg COM line up), I still (apparently) get a thrust in some translational direction sometimes from equally spaced attitude only thrusters (I separate out my yaw/pitch thrusters on the ends and put another dedicated set for roll/translate at COM).  Or maybe I'm not reading RCSBA correctly.  What does the "thrust" value and the green vector arrow from COM mean when adjusting for pitch/yaw?

Nevermind, I had a hidden thruster that got buried/clipped into a procedural tank when I increased its diameter. <facepalm>

Edited August 23, 2020 by darthgently

[darthgently]

So I'm in the VAB and hovering the pointer over a radial decoupler and thinking wouldn't it be so excellent if when I did this if a dry weight COM marker appeared in the subassembly downtree from that decoupler.  It would make placing separatrons and stuff so much easier to prevent collisions in flight

[linuxgurugamer]

Hi All,

@m4v has returned and is picking up development of RCSBuildAid.

So for now, I'll only maintain this since it is somewhat incompatible with his version, there are a number of significant changes which did not get migrated back into his version.

I will be looking at his version and will see if I can migrate some of his changes over to this, but for the foreseeable future, this will be a background effort.

[Xyro]

This is one of my essential addons and have been using it for the longest time. Thanks for all your work to those maintaining it over the years. I know it's somewhat in limbo, but I'd like to bring to light an issue. Maybe it's hidden and I can't find it, but this bad boy REALLY needs a GUI scaler. I recently got a 4k "monitor" (really a 55" QLED lol) and the window and text are EXTREMELY small. Some other mods seem to work off of KSPs scaling, while other mods have their own scale.

Thanks again!

[linuxgurugamer]

@Xyro in the post immediately above yours is a very relevant post.

The original author has picked it up again.

The current thread for this mod is here

 

[Geonovast]

Thread locked at OP's request.