No thermal conduction

[Arthropode]

KSP version : 1.0.4 Linux 64 bit without Steam

Detailed explanation : There is no heat conduction inside the craft, so all engines I have tested heat very quickly. This problem happens even with a fresh installation.

I tested in sandbox and career, and in both the engines overheated too quickly.

To show the bug I used the mod HotSpot, but the problem is the same without the mod.

System specs :

[TABLE]

[TR]

[TD=class: field]Kernel[/TD]

[TD=class: value]Linux 4.0.5-1-MANJARO (x86_64)[/TD]

[/TR]

[TR]

[TD=class: field]C Library[/TD]

[TD=class: value]GNU C Library version 2.21 (stable)

[/TD]

[/TR]

[TR]

[TD=class: field]Distribution[/TD]

[TD=class: value]Manjaro Linux[/TD]

[/TR]

[TR]

[TD=class: field]CPU[/TD]

[TD=class: value]4x Intel® Core i7-3520M CPU @ 2.90GHz

[/TD]

[/TR]

[TR]

[TD=class: field]GPU[/TD]

[TD=class: value]NVIDIA GeForce GTX 660M[/TD]

[/TR]

[TR]

[TD=class: field]Memory[/TD]

[TD=class: value]8069MB[/TD]

[/TR]

[/TABLE]

[Arthropode]

I don't understand : I had the problem 2 days ago and yesterday there was no issue : the thermal conduction worked fine, but today the problem reappears.

I restarted my computer and downloaded a new fresh installation, but it didn't anything.

Why the conduction works every other day ?

[Finndibaenn]

I have this issue too, with atmospheric planes.

The jet engines do not seem to cool down at all. I have a contract to do crew report on the other side of the planet, and engine (1 ram jet) reach the extreme overheat at maybe 75% distance.

I tried adding passive radiators as close as possible to the engine, but they don't seem to change a thing.

Active radiators are out of the equation as you need to deploy them, and they get ripped off if you do so inside the atmosphere

[sal_vager]

If your Physics.cfg is old you can get a new one from the game zip or from here.

That should solve the issue.

[Arthropode]

No, it isn't a problem of Physics.cfg. I changed it and I ran the latest zip, and I still have no thermal conduction

[sal_vager]

Well it should have worked, all the variables the heat system uses are in there, can you upload your save and current physics.cfg please.

[fbfree]

KSP 1.0.4 Linux x64

I've encountered the same problem. I cannot use a Basic Jet Engine for more than 25 minutes without it exploding and it does not cool off after shutdown. After 25 minutes at cruise 80% throttle, 11,000m 180m/s cruise, Cd = 0kW, internal temperature 1900 K, External 238 K, Cv=-0.01 kW, R=0.20 kW, I=1200kV with engine on, 0kW when shut down.

Kernel: Linux 3.16.0-4 64-bit

C Library: GNU C library 2.19-18

Distribution: Debian Stretch (testing)

CPU: 8x Intel Core i7 - 4700MQ CPU @ 2.40GHz

Intel integrated graphics controller

8GB RAM

[Arthropode]

Here is my save : [snipped]

I only added HotSpot to display thermal data and made a test with Kerbal X in a creative game.

Edited July 4, 2015 by sal_vager

[sal_vager]

Please try uploading again Arthropode, all I need is your save file and physics.cfg

fbfree, that's a long time to run any engine, all of the engines will overheat eventually, please use the alt+f12 debug menu to turn on heat data in the right click actions menus, you will see there if parts are cooling down.

Update:

Arthropode, please try the Physics.cfg I provided again, but this time set the file to read-only before you start KSP.

See here for more info.

Edited July 7, 2015 by sal_vager

[Arthropode]

Sorry for the first uploading.

I tried to set Physics.cfg read-only, but I still have the problem even if I have the correct file (md5sum confirmed). I put here a save, my logs and the Physics.cfg : https://www.dropbox.com/s/nt5fty7q43eczjm/KSP_thermal_bug.zip?dl=0

[sal_vager]

Well, the save is fine, but I see you're still loading addons.

[LOG 00:24:18.833] PhysicsGlobals: Loading database

[LOG 00:24:19.139] Load(Assembly): /ModuleManager.2.6.6

[LOG 00:24:19.141] AssemblyLoader: Loading assembly at /home/yann/Data/bin/KSP_linux/GameData/ModuleManager.2.6.6.dll

[LOG 00:24:19.191] AssemblyLoader: KSPAssembly 'ModuleManager' V2.5

[LOG 00:24:19.191] Load(Assembly): HotSpot/Plugins/HotSpot

[LOG 00:24:19.191] AssemblyLoader: Loading assembly at /home/yann/Data/bin/KSP_linux/GameData/HotSpot/Plugins/HotSpot.dll

[LOG 00:24:19.193] AssemblyLoader: KSPAssembly 'HotSpot' V0.4

Please retest this issue with an unmodded copy of KSP, and run the KSP.x86_64 directly, not via a shortcut.

[Arthropode]

I run a 5 minutes old installation directly (but with primusrun) and still the problem : https://www.dropbox.com/s/24d5410wno3n1qt/KSP_thermal_bug_2.zip?dl=0

What I really don't understand is that from time to time, the conduction works. But when I restart KSP, it disappears again. Ans I don't understand why sometimes it is OK. It's not just after a restart, but at random.

[sal_vager]

I am having no luck reproducing this, your files all seem fine, do you start KSP in the same way every time or does it vary?

[Arthropode]

Well, after a lot more test, I discovered that there is the bug when I launch KSP from the command line (directly or through primusrun/optirun). But when I launch KSP from my file manager (thunar), it is OK. I tested with a .desktop file and launching it whith my file manager and it is also OK, both directly or with primusrun. I didn't tested to launch the desktop file from the command line because I didn't find how.

Apparently the bug can be bypassed with the desktop file. But I don't know why there is this bug from the command line. A problem of environment variable ?

[sal_vager]

Arthropode I think I know what is going on, depending on how you start KSP it is either able to find its config files because it is starting from the directory that contains them, or is unable and is looking for the files in the location where you started it, such as from a desktop link.

Please ensure you cd to the KSP folder before starting the game, either from the command line, a desktop file, a script or even from primusrun, and not by using "./PathToKSP/KSp.x86" KSP like many other games will look local to where the binary was started, not where the binary actually is.