Why is the TR-38-D "massless"?

[Galane]

What is PhysicsSignificance = 1 doing in the cfg for the TR-38-D? That's supposed to be a large and fairly heavy part.

With each release from .24 I have commented that line out. Why? Because with it in there, it causes a heap of problems. I can build an all stock parts rocket that demonstrates how very bad it can get, then simply change that 1 to a 0 or comment out or delete the line and no problem.

It took me a while to find out what exactly was causing the problem in the rocket I was building. Then I built some test rockets of all stock parts to see if the problem only manifested through scaling up some stock parts. Nope, same exact problem all stock. That's what led me to examine the cfg files for all the stack decouplers and separators.

Then I saw it, the only thing the TR-38-D cfg file had which none of the others had. PhysicsSignificance = 1 Changed 1 to 0 and all my test rockets worked as built.

Besides that, it's quite cheaty having a 0.8t part effectively weightless in flight.

So please, Squad, change this in the next release and put out a notice advising people to edit the part.cfg in .24 through .90. It will only cure a bunch of problems people may experience using the TR-38-D in anything but a straight up single stack, or single asparagus stacks. If a TVR-2160C Mk2 is involved, it has its own little bug that can combine with the TR-38-D's as-shipped part.cfg to do some very strange things.

*Going to rebuild that test craft since I've altered it significantly due to it turning out to be a good base for a heavy launcher.

[CaptainKipard]

WHat's the problem exactly?

[Red Iron Crown]

I don't think we've ever heard an official explanation of the reasoning behind many parts being physicsless. Some of the theories I've seen floating around:

- Makes craft easier to balance (though this applies more to radial parts than an inline decoupler like the TR-38D)

- Better performance through reduced physics calculations (massless parts are inertialess)

- Workaround for Unity disliking parts of greatly disparate mass directly connected to each other.

I'm not sure any of those are accurate, or if more than one apply. I've also read some suggestions for how to handle the missing mass, my favorite is to add the mass of the massless part to the first massful parent part.

[Overfloater]

O_O

it's massless? Then this whole time i've been building my Eve landers wrong. I used only the "tiny" size decouplers thinking they're the lightest, when instead I should have used the giant decouplers to make it more efficient.

(xD not really, it would look ugly)

[Galane]

Demonstration of "massless" TR-38-D interacting badly with TVR-2160C Mk2. It has its own curious bug, which seems to cascade down to parts connected below it. Note that when staging the tanks off the top of the Kerbodyne side boosters, only one goes instead of all four and a lot of force is projected downward, destroying the tank below. The TR-38-D's are mounted upside down so their weight goes with the fuel tanks - if the physics line is changed to 0 or commented out or deleted. Without being "massless" the TR-38-D works in either orientation, just like all the other stack decouplers.

http://pastebin.com/xPHeUM64

Demonstration of the same craft, with the TR-38-D's right side up, which without the Physics Significance line would leave an extra 3.2t weight on the rocket. It stages with or without the physics line.

http://pastebin.com/M2H5kzSs

To demonstrate another problem, again an interaction between the "massless" TR-38-D and TVR-2160C Mk2, requires using the Bigger, Beastlier Quad Coupler upscaled version in ReStock, then attaching four Kerbodyne tanks below it with four KR-2L. Put four TR-38-D under those, four more Kerbodyne tanks and four more KR-2L. Strut as needed.

With the physics line in the TR-38-D cfg file the TR-38-D's won't all trigger together. One pops then a short bit later the exhaust of the other engines explodes the others. Take out the physics line and they all decouple properly.

Here's a test craft that demonstrates the TVR-2160C Mk2 bug http://pastebin.com/dA1Prjny Toss it onto the pad, throttle up then start popping the space bar. Note how the 2nd probe engine fails to activate. Revert to VAB then grab *any* one of the stages in the stack and move it anywhere else in the order. Don't do anything to the craft, just move a stage up or down.

Launch again and amazingly, all four probes will separate and ignite correctly. It's one crazy weird bug in that part and it's been there since at least .21 Want weirder? Sepratrons are not affected by the bug. Could be because unlike the liquid fuel engines, they don't need an active command/control part on the probe they're attached to in order to ignite.

That's the top end of my Armed Camp kinetic kill vehicle launcher (with KKVs simplified to demonstrate the bug). IIRC when I first built it, two of them wouldn't separate and activate but while re-arranging part to make it fail again all I could get was the one failure.

Edited January 1, 2015 by Galane

[thereaverofdarkness]

http://imgur.com/a/5wy9b#0

Thank you, OP, for this beautiful event.

[King Arthur]

^ Magnificent.

[Galane]

See? This is why we can have crazy things.

[thereaverofdarkness]

http://forum.kerbalspaceprogram.com/threads/105674-Stock-0-90-TR-38-D-Decoupler-Propulsion

^ ^ ^ my new challenge ^ ^ ^

http://imgur.com/a/vhzNW

^ Kerbin escape ^