Physics significance = 1: how do these parts affect CoM?

[allmhuran]

I have been playing around with "physicsless" parts. These parts previously contributed no mass to a vessel, my understanding is that they now add their mass to their parent part. However, this does not seem to be taken into account in the construction facilities.

I made a test craft in the VAB composed of a long ibeam, with a large RCS tank at one end. The CoM marker was, as expected, pretty much inside the large RCS tank. When I added my massive (but physicsless) part, weighing about the same as the RCS tank, to the other end of the ibeam, the CoM marker did not shift at all. If the mass of the physicless part was added to the ibeam I would have expected to see the CoM marker shift towards the middle of the Ibeam.

[Farex]

As it is now:

Physicless parts: add mass to the parentbody, CoM does not change

Physic parts: have their own mass, CoM does change.

Pre-1.0 there were parts that didn't have any mass at all (like antennas, or the rcs-engines). This is no more. Now everything has mass. Difference see up above.

[The_Rocketeer]

Physicless parts: add mass to the parentbody, CoM does not change

Physic parts: have their own mass, CoM does change.

Are you saying that the added mass is basically only used for purposes of dV and not for balancing CoM with CoL/CoT? I find this very peculiar! I'd always assumed that the physicsless parts were so lightweight that they didn't impact the CoM noticeably.

Edit: Ah, I think I see. When you say 'parentbody' you mean the mass is added to the centre of the entire stack/assembly that the physicsless part is attached to? I'd always supposed/understood it was added to the specific part to which it was attached.

Edited September 14, 2015 by The_Rocketeer

[Warzouz]

As it is now:

Physicless parts: add mass to the parentbody, CoM does not change

Physic parts: have their own mass, CoM does change.

That's not entirely true. If you attach a physicless part to a physic part which is radially attached, the CoM changes (because the weight of the radial physic part is changed by the physicless part).

Always attach physicless parts to the central stack of you ship, not on radial parts.

That's a major change since 0.9.

I think it's the same for drag.

[Farex]

@Warzouz

I must disagree with you (for now). I will test this once I get home and post here.

I refer to the patch 1.02 (or was it 1.03?), which changed the heatshields from physicless to physic parts, because the heatshields were physicless then meaning adding only mass to the CoM of the capsule -> this resulted in a constant flipover during reentry. Now that they are no longer physicless, they change the CoM so the capsule stays stable.

[Red Iron Crown]

Farex, I think they mean when there is more than one part in a stack. Imagine two identical tanks in a stack together, the CoM would be at the point where they join. Add a physicsless part to the top tank and its mass is added to the top tank, moving the overall CoM upward.

For clarity: Physicsless parts have their mass and drag added to the first physicsfull parent part.

[Alshain]

Warzouz is correct. If I attach a landing gear to the cockpit of a Mk2 plane, that landing gear will add the mass to only the Cockpit, and assuming the cockpit is in a typical location, it will bring the center of mass forward, albeit very little. In your tests, Farex, keep in mind that most physicsless parts have very little mass to begin with, so you may have to add a lot to visually see a difference.

[Red Iron Crown]

In your tests, Farex, keep in mind that most physicsless parts have very little mass to begin with, so you may have to add a lot to visually see a difference.

The 3.75m decoupler is especially good for this, 800kg of physicslessness.

[Farex]

I will add hundreds :-)

Just to make sure, the little Cubic Octagonal Strut is still a physicless part?

[Warzouz]

@Warzouz

I must disagree with you (for now). I will test this once I get home and post here.

I refer to the patch 1.02 (or was it 1.03?), which changed the heatshields from physicless to physic parts, because the heatshields were physicless then meaning adding only mass to the CoM of the capsule -> this resulted in a constant flipover during reentry. Now that they are no longer physicless, they change the CoM so the capsule stays stable.

Erh, no, you don't disagree with me : I didn't spoke of stacked physicless parts, but radially attached physicless parts which are attached to a already radially attached part (which is not physicless). I should add a screenshot but I'm at work

Physicless parts add their mass (and drag ?) to their parent part. In that case the parent part is not the central stack, but the radial part. As the radial part mass increases, CoM moves toward it. If your test doesn't show any deviation, that mean the definition is wrong : the physicless parts don't add mass to their parent, but elsewhere.

So beware when adding those nice little non-retractable ladders to your side tank/engine ; you should attach them on the central stack to avoid moving the CoM

Heat shields are not radially attached.

[Jouni]

Physicsless parts add their mass to their parent part, unless the parent part is also physicsless. In that case, the mass of the physicsless part vanishes.

The center of mass indicator in the editor doesn't know that. It just assumes that physicsless parts are massless.

If you put the example structure on the launchpad, it tips over slowly. (The fuel tank is empty.)

If you attach the physicsless decouplers to a physicsless cubic octagonal strut, the structure is stable.

Notice how nobody told the graphics engine about physicsless parts. The view is centered on the "real" center of mass.

Edit: The situation has apparently changed since I tested this in 1.0.2. Physicsless parts now add their mass to their first non-physicsless parent, so the rocket in the first picture no longer works. The graphics engine doesn't center the view on the real or the effective center of mass, but to something else.

Edited September 14, 2015 by Jouni

[Red Iron Crown]

Physicsless parts add their mass to their parent part, unless the parent part is also physicsless. In that case, the mass of the physicsless part vanishes.

Are these pics from 1.0.2 or earlier? As I understood it that got fixed in 1.0.3/4, it now moves to the first physicsful parent part.

[Warzouz]

I understand why small science parts are physicless (more should be...) , but I don't understand why the big decoupler aren't...

[Jouni]

Are these pics from 1.0.2 or earlier? As I understood it that got fixed in 1.0.3/4, it now moves to the first physicsful parent part.

The first picture was from 1.0.2, while the rest are from 1.0.4.

[Starhawk]

The first picture was from 1.0.2, while the rest are from 1.0.4.

The second picture must also be from 1.0.2.

It doesn't work like that in 1.0.4.

Happy landings!

[Jouni]

The second picture must also be from 1.0.2.

It doesn't work like that in 1.0.4.

Pictures 2-4 were taken today in 1.0.4 specifically for this discussion. They demonstrate that the visual indicators for the center of mass are all wrong, when it comes to physicsless parts.

[Starhawk]

Pictures 2-4 were taken today in 1.0.4 specifically for this discussion. They demonstrate that the visual indicators for the center of mass are all wrong, when it comes to physicsless parts.

Sorry, my mistake. I quickly tested it in my career save, but that has RCS build aid installed which affects the CoM marker.

You are quite correct. In the stock editor, the incorrect CoM is shown. If you install RCS build aid, you will see the CoM marker correctly placed.

Happy landings!

[Farex]

I'm done testing. Look at what I discovered:

Beware of 1080p pictures

It might not be the best example, sure, but you can see that the CoM shifts towards the Parentpart, which is heavier.

[allmhuran]

Cheers Jouni et al.

From my own testing it seem that, similar to what Jouni said, the CoM of the craft for actual physics purposes does change such that the physicsless mass is added to the first physics parent, but the CoM indicator in the VAB/SPH does not show the same thing (it doesn't move at all). To test this, don't rely on tweakscale or whatever. Just go copy any part config file, alter its mass to, say, 5 tons, and set physicssignificance to 1. You won't have to rely on pixel sized "CoM might have changed slightly maybe?" indications. It will jump around very obviously, or not at all (not at all is the answer).

Edited September 15, 2015 by allmhuran

[Teilnehmer]

Why the 3.75m decoupler is marked physicsless? It is neither a light part nor attached radially. Why?

[Red Iron Crown]

Why the 3.75m decoupler is marked physicsless? It is neither a light part nor attached radially. Why?

As I understand it a large disparity in masses of stacked parts makes the joint strength go a bit wonky, an effect which becomes fairly pronounced in the 3.75m parts. You could try deleting the line in the config and seeing how it behaves, I must confess I haven't tested it myself.