Warning Modders - Parts: Really Low Mass Plus Large Surface Area - 1.0.3 & 1.0.4

[Probus]

I just wanted to warn all the mod developers that if your part has a really low mass that it may cause runaway heating of that part over time in 1.0.3 & 1.0.4.

I am still working on finding the threshold but a really easy way to check for potential problems is using the temperature overlay (F11) key to turn on temperature coloring. If any of your parts are giving the thermal system problems they will flicker rapidly. In certain situations this can cause the part (or other parts by convection) to slowly overheat and explode.

Feel free to update this thread if you have more information on the problem, but I know for sure that if your mass is in the 0.00001 range, problems may ensue.

For now, a module manager fix is in work that catches parts that have unrealistic physics numbers. Sarbian's link is in the thread below.

EDIT: Also some parts (like some of the wheels) that are connected symmetrically to parts with high surface areas, like octagonal struts, can cause this run away heating effect also, yet if you use a cubic octagonal strut the surface area is low enough to dampen the runaway heating. So, If you have any .craft files with runaway heating please post them in this thread to help the devs find the conduction clamping issue.

Edited June 25, 2015 by Probus
Additional Information

[Probus]

The threshold is 12.5 g or mass = 0.0000125. You can also increase the thermal mass modifier to increase the thermal mass above that threshold.

[Beale]

The threshold is 12.5 g or mass = 0.0000125. You can also increase the thermal mass modifier to increase the thermal mass above that threshold.

How does this affect parts that are above this mass, but using physicsignificance = 1 ?

Anyway, helpful information - thanks!

[sarbian]

This MM patch will set part mass under 100g to 100g

https://dl.dropboxusercontent.com/u/30298951/KSP/minMass.cfg

[NathanKell]

Beale: doesn't affect them at all.

Thermal mass don't care if you have a rigidbody or not.

[Babalas]

I'm also finding certain parts stuck together seem to be causing overheating. In my case I've got a ship with cubic octagonal parts holding solar panels that are overheating rapidly. Also the combination of BoxSat frame, hatch, and dipole antennae causes the same rapid overheating. Gave the MM patch a try but no effect.

[hieywiey]

I also think TweakScale *may* scale the heat tolerance too, as I find that scaled-down parts overheat for no reason. But I haven't tested that scientifically.

Also, maybe you should see if that is fixed in 1.0.4, as that's the latest version, although I haven't used it yet. I personally don't really like the heating system, it most of the time makes parts overheat for literally no reason; I would like if it made parts seperate from each other first instead of just overheating, I also think the atmospheric (reentry and mach) effects need to be redone since they're kinda boring.

[Nansuchao]

About the heating, I noticed that with 1.04 radiators are necessary most of times. I made a launch to Minmus, and every time the lander tank inside the Fairings exploded due to overheating.

With some radiators on the transfer stage, no more issues.

[Babalas]

No radiators didn't help in my case (having just rebuilt my probe). I'm going to see what happens with 1.0.4 but otherwise I'm glad I backed up 1.0.2

[Probus]

New information on thermal runaway issue:

Some parts (like some of the wheels) that are connected to parts with high surface areas, like octagonal struts, can cause this run away heating effect also, yet if you use a cubic octagonal strut the surface area is low enough to dampen the runaway heating. So, If you have any .craft files with runaway heating please post them in this bug report or thread to help the devs find the conduction clamping issue.

http://bugs.kerbalspaceprogram.com/issues/5174