Incorrectly angled thrust with small radial (Twitch) engines attached with sym in SPH

[Wintersdark]

See very simple craft here: https://dl.dropboxusercontent.com/u/28016101/WHY.craft

I originally discovered this bug while attempting to get nice centered thrust for a large rover using 4 Twitch engines to have it land on it's wheels. Placing them so the thrust should vector directly through the CoM while being placed vertically, I discovered the direction of thrust was always angled a few degrees to one side.

This was in a modded install, so in an effort to troubleshoot, I removed all my mods and started a new sandbox game.

I then took, in the SPH, a command pod, fuel tank, and 2 Twitch engines mounted with symmetry and angle snap along the center line (can be seen in the attached craft file) and sure enough, the thrust is still inexplicably offset a few degrees to one side.

What's going on? Is there anything I can do about this?

[sal_vager]

Hi Wintersdark, I'm not sure I understand, are you saying the thrust is angled or that the thrust is detached from the engines?

It should look like this.

If it doesn't, then one of the addons you removed has already modified your parts files, and you will have to replace them.

[Wintersdark]

This. This is in the modded install, but the result is exactly the same with all mods removed. Said modded install is very minor (KER, StageRecovery, etc, and all mods that add parts do so via modulemanager. I install things carefully (that is, I always add mod subfolders directly, never just overwriting GameData), and already removed and used Steam to rebuild GameData/Squad to be sure.

- - - Updated - - -

But even if the engines where somehow modified, and not projecting thrust straight.... as they're applied with angle snap and symmetry along the midline of the craft, no matter how wonky they were they should "cancel each other out" so to speak.

- - - Updated - - -

Further investigation.

If I add a single Twitch engine w/o symmety, the thrust is angled correctly slightly off center but exactly along the angle of the engine.

See:

So it's angling outboard correctly, I assume, here. However, when a second engine is added on the other side in the opposite position (without symmetry, so I'm not getting 3 engines total), the thrust vector doesn't center but leans off as in my first pic above.

[sal_vager]

Looks like you found a bug in how KSP handles the angled engines, whether they are placed in symmetry or not, this also affects the o-10 puff monoprop engine and any other engines place at an angle.

To fix this with the Twitch, turn on angle snap, hold left-shift and tap W or use the rotate tool to rotate the engines 5 degrees, they will look like this.

Twitch engines are 5° off from vertical.

[Wintersdark]

This *may* be a vector display issue in the VAB/SPH (the same thing happens in both); as my tests with small vessels off the launchpad seem like they launch straight up. Hard to tell, though, if that's because of aero or not. I'm not sure how to test that.

[sal_vager]

They will launch straight up yes, the two engines combined thrust vector is zero despite what the thrust indicator claims.

[Wintersdark]

Yes, but with them placed in symmetry but 5 degrees off vertical in opposing directlyions, the end result should be a straight thrust vector, right? I'm *reasonably* sure this is display only, but I'm not sure how to test this, as it's a slight angle and may get "eaten" by drag...?

- - - Updated - - -

Ok, testing with a tiny little probe:

Launching straight up w/o SAS, it goes directly up until around 300m/s then flips sharply east.

Reverting, rotating the whole vessel 180 degrees in the VAB (so the thrust is angled in the opposite direction) and it still flies straight up then flips sharply east at 300m/s.

So... I'm not sure how to better test that (yeah, I should have done the tests unmodded given the forum, but I get the same display results from unmodded)... But as it stands, it seems the displayed thrust vector is just screwed up with angled engines. Note that this happens whether or not symmetry is used.

- - - Updated - - -

Confirmation with manually rotated Reliant engine:

[basic.syntax]

I just tried this with a simple "24-77"-engines-in-symmetry-around-tank design in both 1.0.2 and v.90.

This is -at least- a bug with the display of the thrust vector.

It is shown going straight down in .90, with the same symmetrical design, and off at an angle in 1.0.2

edit:

Oops, read right over sal_vager post #4 writing the same thing.

Edited May 11, 2015 by basic.syntax

[JoCRaM]

This is -at least- a bug with the display of the thrust vector.

.

Put a "mammoth" under a probe core - thrust vector is down. Put an ant on top of the the core, pointing down, and it overpowers the mammoth...

However, remove the mammoth and replace it... and thrust is upwards again.

Prettu much completely useless for balancing thrust - although with thrust varying with ISP and velocity balancing thrust is no longer as trivial as it was.

[Claw]

[quote name='Fwiffo']I think this still looks like an issue with 1.0.5? I noticed my Twitch thrust vectors are "misaligned" (using RCS Build Aid to view). Having trouble getting 3 radial mounted ones to cancel out and give me a torqueless vector.[/QUOTE]

It should be fixed. Can you post a pic of exactly what you're talking about?

Cheers,
~Claw

[Fwiffo]

Hi Claw,

Nevermind, my bad. I didn't realize I had a "turn" key active in RCS Build Aid at the time. Sorry to resurrect an old post for a red herring. Tried to delete it but you beat me to action ;-)

Thanks for the offer to help; s'all good now.

-Fwiffo

[Claw]

No worries, I'll just lock this one to help avoid confusion. :)