ModuleEngineFX with no thrustTransform?

[Diazo]

Good day all:

I am actually a plugin dev, but as this question pertains to the thrustTransform object on engines, I figure here was the place to post this.

I'm trying to refer to the thrustTransform object on an engine made by someone else and it is coming back as not exiting.

This confuses me as I though that without a thrustTransform, a part would not produce thrust.

So, as far as you all know:

1) Can an engine, specifically the moduleEnginesFX partModule, produce thrust without a thrustTransform?

2) Can an engine have more then one thrustTransform? (Not RCS blocks, main booster type engines such as the Skipper).

Thanks,

D.

[nli2work]

1. doesn't have to be "thrustTransform", that's just a convention people tend to follow. any transform's +Z axis can work for thrust. check the part config to see what transform the engine module is using for thrust. for ModuleEnginesFX, FX and thrust do not have to use the same transform. FX transform is defined in the EFFECT{}; thrust is defined in Engine{}.

2. yes, all transform with the same name given to Engine{} will be used. thrust is the averaged position (and possibly orientation) of all the transforms.

[sarbian]

2. yes, all transform with the same name given to Engine{} will be used. thrust is the averaged position (and possibly orientation) of all the transforms.

It's not an average. For n transform each transform generate a trust of 1/n in its orientation. If all transform are parallel it is equal to the average but if orientation change it's not.

[nli2work]

It's not an average. For n transform each transform generate a trust of 1/n in its orientation. If all transform are parallel it is equal to the average but if orientation change it's not.

so 1 transform at 30 degrees X and 2nd transform at -60 degrees X would yield thrust at -30 degrees X? the xyz components would average out in the end for the final thrust vector no?

[NathanKell]

Consider the case of 2 transforms, one going left and one right. If you "average it out" you get a thrust vector of 0,0,0, which is nonsense; in fact what happens is each one fires at half thrust, leading to net 0 acceleration but not a zero-vector single thrust direction...

[Diazo]

Thank you for the help all, I've figured out where I went wrong.

This does raise an interesting question though, when would you have a part with multiple thrust transforms that pointed in different directions? Because it is a moduleEngines, you can't throttle the transforms individually, so there is no control advantage.

Maybe some sort of helicopter rotor with rockets to spin it?

Interesting puzzle.

D.

[NecroBones]

Thank you for the help all, I've figured out where I went wrong.

This does raise an interesting question though, when would you have a part with multiple thrust transforms that pointed in different directions? Because it is a moduleEngines, you can't throttle the transforms individually, so there is no control advantage.

Maybe some sort of helicopter rotor with rockets to spin it?

Interesting puzzle.

D.

An example:

The "launch escape system" (LES) in the ARM update is an example of using different angle thrust transforms. It has four of them that angle slightly outward rather than straight down, so that the thrust bypasses the capsule underneath. And then there's a fifth transform that comes down on one side only, to provide rotation in one direction, so the capsule gets flung to the side.

[NathanKell]

Or consider a 'quad engine' part where the engines all angle 'outwards' slightly.