Stock Radiator physics help?

[RoverDude]

Hopping in late - but (a) show a pic of the craft, there's something going on.  and (b) if you tweakscale those radiators, I would not be surprised if truly horrible things happen because you are dramatically changing their radiative area, and their balance.

 

 

[FreeThinker]

On 10-12-2015 at 8:33 PM, Snark said:

 

Can someone please clear this up for me?

I see statements like the above from Claw.  And I see that if I go and look at the static radiator panels' .cfg file, it does indeed have a "maxLinksAway" property on its ModuleActiveRadiator, which is set to 2.  It's clear that they're meant to be cooling only the two-links-away parts.

However, when I use panels in-game, I swear I see them cooling the whole ship.  I'll launch a ship with some static panels up top, and SRBs down at the bottom, and they're nowhere near "two links away"... and yet the radiators are glowing cherry-red very quickly after activating the SRBs, it's clear that they're pulling heat from them.  As far as I can tell, the static panels are cooling the whole ship, same as the active panels.  What gives?

(And I don't think it's just me. I'm pretty sure I've seen someone else in the forums mention observing the same thing, at some point.)

Is this a bug?

Seems to me maxLinksAway  has no effect , at least not anymore in 1.1.2 because it has no backing field in the class

[bewing]

23 hours ago, FreeThinker said:

Seems to me maxLinksAway  has no effect , at least not anymore in 1.1.2 because it has no backing field in the class

I just tested it in the game, in 1.1.2 -- and it works. So they must have done some coding trick or other.

[FreeThinker]

1 minute ago, bewing said:

I just tested it in the game, in 1.1.2 -- and it works. So they must have done some coding trick or other.

They do!? how have you tested?

[bewing]

1 hour ago, FreeThinker said:

They do!? how have you tested?

I put a couple drills on a lf fuselage, another fuselage on top of that, an ore tank on top of that, with 2 gigantors and 2 large radiators. Started the drills. Drill temp went up and up and up.

[FreeThinker]

I still don't understand. Also why isn't this information documented anywhere. I realy hate the guessing game.

[Plusck]

 

47 minutes ago, FreeThinker said:

I still don't understand. Also why isn't this information documented anywhere. I realy hate the guessing game.

Because with the changes it's relatively new, and there's always a ton of things that has to be tested before it can be really documented somewhere. The interactions are quite complex.

So, without further ado, here are some tests:

I tried to test three types of heating: skin heating, "internal" heating for an engine, and core heating (drill, ISRU), and how fixed radiators dealt with that. On the left, radiator on the part next to the tested part (i.e. distance of 2 parts). On the right, radiator on the next part along (i.e. distance of 3 parts)

Spoiler

Unfortunately, the LV-N just doesn't get hot enough in the atmosphere to give any conclusive results. And its exhaust isn't hot enough either. So I had to use the Vector to heat a structural panel.

However, the findings from this test were:

• 1.1.2 fixed radiators most definitely do have an effective distance of 2 parts. They don't in 1.0.5. (screenshots not included since they show only that radiator placement has only a marginal effect)
• Radiators have a specific function when working with ISRU / drills: when the slightest bit warm, each part will "use" a certain percentage of the cooling power of the radiator. This only applies within the 2-part distance (as advertised).
• A single large radiator is enough to keep a drill at 500°K, but only if it is not trying to cool anything else down. If any other part tries to claim part of the radiator's function, efficiency drops to 86% (on Kerbin, YMMV).
• A single large radiator can only keep an ISRU - running a single process - at about 90% efficiency.
• Radiators are a bit slow to get working - they start cooling parts within range immediately, but the biggest effect only occurs after a few minutes as they warm up
• Parts conduct heat better the warmer they are: while heating the structural panel without a radiator in range, its heat peaked quite quickly, but as the part it was attached to heated up (and was therefore cooled by the radiator), the panel started cooling down despite being under constant heating from the engine exhaust.

And finally:

Spoiler

• if you feed an engine using an ISRU (i.e. no other fuel available), it produces a constant spark show and a beatbox-style rhythm

 

Edited May 6, 2016 by Plusck

[FreeThinker]

28 minutes ago, Plusck said:

  Reveal hidden contents

 

• 1.1.2 fixed radiators most definitely do have an effective distance of 2 parts. They don't in 1.0.5. (screenshots not included since they show only that radiator placement has only a marginal effect

 

 

Well this can be explained by the property parentCoolingOnly , which does have a backing field in the ModuleActiveRadiator class and which is set to true on all Radiator Panels. Seems to me this is the only relevant  field. The maxLinksAway does not have any backing field and therefore cannot have any effect.

Edited May 6, 2016 by FreeThinker

[Plusck]

24 minutes ago, FreeThinker said:

Well this can be explained by the property parentCoolingOnly , which does have a backing field in the ModuleActiveRadiator class and which is set to true on all Radiator Panels. Seems to me this is the only relevant  field. The maxLinksAway does not have any backing field and therefore cannot have any effect.

I haven't looked at the code, but "parentCoolingOnly" sounds like it would only affect the part that the radiator is attached to, and that is most definitely not what is happening here.

In the third screenshot, the structural panel was a child to the fuel tank to which the radiator was attached. The radiator registered the hot structural panel as soon as it was heated, and you can see that the radiators have specific fields (HeadRoom, ExcessHeat etc.) that the more distant radiators don't.

[FreeThinker]

2 hours ago, Plusck said:

I haven't looked at the code, but "parentCoolingOnly" sounds like it would only affect the part that the radiator is attached to, and that is most definitely not what is happening here.

In the third screenshot, the structural panel was a child to the fuel tank to which the radiator was attached. The radiator registered the hot structural panel as soon as it was heated, and you can see that the radiators have specific fields (HeadRoom, ExcessHeat etc.) that the more distant radiators don't.

no I read that parentCoolingOnly  mean it parent and those arround it.

[Sereneti]

On 7.1.2016 at 2:22 PM, Nightshift83 said:

Hey there, been playing KSP since 0.7 or something like that, but never understood how radiators exactly work since they were introduced.

I mean if static ones can only cool up to two parts away, why should it matter ? Since as stated above temperature should distribute itself evenly across the whole ship.

Do they have any effect during atmospheric entry ? If yes how does it work.

And how does it work in real life btw ? Since they operate in (almost) vacuum how do they get rid of the heat they acumulated before ?

 

Oh and great community for a great game! Keep it up!

I will try to leave my sleeper role and try to be more active from now on

Yes, Radiators can help during reentry and launch...

My vehicle blow up during reentry, then i attached 2 small(!) radiators to it, and it dosnt exlode so easy...

if you put it in the slipstream of other parts , it wont be touched by the plasma, and cann cool down the attached parts...