Kerbal Space Program

Good find.
This also happens with the XL3 wheels. I haven't been able to find any other part combinations that produce this effect so far.

A minimal assembly to reproduce is an Octagonal Strut with something attached to the two nodes, and either XL3 or TR-2L rover wheels surface attached at the sides.

The devs are investigating this phenomenon.

Here is another example .craft file with run away heating. Press F11 to watch one of the struts flicker then explode.

Physics Warp will cause the Rockomax brand Adapter to experience overheating and combustion in this rocket under 2x or greater physics warping or normal speed while in space flight. There is a FL-R25 RCS fuel tank and Z-1k Rechargeable batter Bank embedded between the adapter and service module.

You can also see the thermal heating extends to the Mk-16 XL parachute.

rakutenshi wrote:

Additionally I've done some poking about, discovered that the overheating went away almost entirely when I made sure no parts were clipping, leading me to believe that this is related to the skin heat transfer. had a craft which was basically the octo core with an aerodynamic nosecone on top, a fuel tank below, clipped it so the nosecone and the fuel tank touched with the probe core hidden inside, leading to rapid overheating after launch (small radiators on side of fuel tank had no effect). Rebuilt, making sure none of the parts were clipping into each other more than necessary and it experienced no thermal issues at all.

I have added a minimal parts variation. We have identified that both nodes of the strut must be occupied as you have noted, and that TR-2L or XL3 wheels attached to the sides will trigger this effect. Clipping does not seem to be a factor, but can't be ruled out.

This particular assembly does not have the same orientation dependency that assemblies in #5181 are showing, but the effects are too similar to dismiss. We need to identify and isolate which part combinations this affects. Please attach minimal examples of craft files for analysis.

All of these are pretty much the same 3 part craft, save for the connecting piece between the Short Mk3 Cargo Bay and the structural fuselage (the fuselage can be replaced with anything I believe)
#1 Explodes in about 12 seconds with no acceleration.
#2 explodes in about 3 seconds with no acceleration.
#3 explodes in about 25 seconds with no acceleration.
#4 and #5 only explode under time acceleration.
#3b is facing south from the cargobay but the cargobay is still E-W, explodes in about 1:40 under time acceleration, heat buildup did not seem to be an issue under real time.

Orientation continues to be a deciding factor in this bug, none of the tests I've done work while the cargobay is oriented perpendicular to the runway (N-S)
Having a part on both nodes of the piece being tested is also critical to this bug happening.
Cargobay doors must be closed in these examples.

Regarding Thermal Runaway.craft, as noted by squelch Orientation does not matter, but still exploded for me even when the attach nodes on the core Octagonal strut were removed (1x Oct. Strut and the two Ruggedized Wheels only) the core oct. strut exploded.

I do not know for sure if it is a factor, but I have noticed that the temperature of the pieces that explode tend to fluctuate rapidly between current high temperature and 4.0K (possibly several hundred times a second or more, it's impossible for me to tell). The conductive flux also rapidly shifts between a very large negative number at 4.0 temp and a very large positive number at the hot temperature. When it gets to critical temperatures, you can see the gauge on the part in question begin to flicker rapidly as well.

Regarding Minimal Test 4 item from #5181, the position of the docking port is very specific, it must be located directly over the window on the Hitchhiker, and creates the heat bug when the Clampotron is oriented between 005 and 055 degrees(Just east of north to a bit east of northeast), and starts blowing up again around 105. additionally the service bay has to be slightly misaligned on top of the hitchhiker or the effect doesn't seem to happen.

Thanks for reporting the resolution.