Where to put Radiator parts?

[Noel32]

My question is, as the title says, where can I put radiators and have them still be effective. Can they be anywhere on the vessel or do they have to be on the part that needs cooling?

[Snark]

Moving to Gameplay Questions.

4 minutes ago, Noel32 said:

Can they be anywhere on the vessel or do they have to be on the part that needs cooling?

Depends on the radiator type.

For the deployable radiators (i.e. the folding ones that rotate to stay edge-on to the sun), no, it doesn't matter where you put them.  At all.  You can have the hot stuff at one end of the ship, and the radiators at the opposite end-- doesn't matter, they pump heat from the whole ship.

For the static radiators, they need to be mounted close to the thing to cool.  Whether they only cool the parent part, or can extend their reach a bit farther, I'm fuzzy on-- I get the impression that KSP has kinda gone back and forth on this somewhat over the past few versions, and in the past (pre-1.2) I observed them cooling stuff from all over the ship even though supposedly they're supposed only to cool locally.  So I'm not completely clear on their exact behavior now (not least because I never use them, ever; the folding panels are just better).  However, broadly speaking, I believe that yes, their placement matters.

[DrLicor]

 

Quoted from @Laie

"

"Radiator panels" (the fixed type) assume the temperature of their parent part without noticable delay. There's also a smaller difference between core and skin temperature, but in my test case that difference wasn't very large even for the parent. In effect, they increase the surface area of the parent. Compared to fuel tanks of 16t and above, even the large panel doesn't provide much extra area, though. And anyway, radiation depends very much on temperature: in order to radiate well, a part has to be very hot. Which in this case means that the panel's parent has to be very hot -- a situation you probably want to avoid from the outset. Still, a panel or two may make the difference between a part going poof, or merely running dangerously hot. But as I said two sentences ago: if a part does not become dangerously hot, the panel will also have little to do.

"Thermal control systems" (the extendable swivel type) gather heat from all over the vessel. They can easily be the hottest part on the ship and thus also radiate disproportional amounts of heat. For purposes of heat managment, it does not matter where you place them. However, they will break if they collide with other parts of your vessel, so you better make sure they have enough elbow room.

So far, it seems as if the fixed radiators were nearly useless. However, when in atmosphere they may also help with shedding heat through convection. Think of a nosecone that gets really hot during reentry and continues to dump heat into a nearby part long after you've been through the worst of it. Or running a nuclear engine on a spaceplane, where nearby parts could conceivably become even hotter than the surrounding atmosphere. I did not test this, however.

"

 

[fourfa]

2 hours ago, Snark said:

For the deployable radiators (i.e. the folding ones that rotate to stay edge-on to the sun), no, it doesn't matter where you put them.  At all.  You can have the hot stuff at one end of the ship, and the radiators at the opposite end-- doesn't matter, they pump heat from the whole ship.

For the static radiators, they need to be mounted close to the thing to cool.  Whether they only cool the parent part, or can extend their reach a bit farther, I'm fuzzy on-- I get the impression that KSP has kinda gone back and forth on this somewhat over the past few versions, and in the past (pre-1.2) I observed them cooling stuff from all over the ship even though supposedly they're supposed only to cool locally.  So I'm not completely clear on their exact behavior now (not least because I never use them, ever; the folding panels are just better).  However, broadly speaking, I believe that yes, their placement matters.

In terms of game lore, the deployable radiators have more mass than the statics for a given cooling capacity because they include a system of internal coolant pipes all over the ship (saw it somewhere in the KSPedia).  In terms of cooling capacity, the range of sizes goes much, much bigger than the static panels as well.

The static panels (AFIAK!) still only draw directly from the part they are attached to (the 'parent' part) and the ones connected to that ('child' parts).  For instance, put down a 2.5m battery or reaction wheel.  Radially attach two large static panels to it.  Radially attach two large drills (which generate lots of heat while mining).  Drop a large ISRU (another large heat source) on top.  Add an ore tank and power anywhere in the ship.  That is a complete mining package that will stay cool in any planet mining situation I've tried (asteroids may vary).

As for the rest of Snark's post - consider that all hot parts naturally cool off over time, all by themselves.  If your SSTO cockpit is glowing red hot when you reach orbit, its heat will partially radiate into space, and partially conduct into the other parts of the ship (which then also radiate heat away). If you add static radiators, they only directly cool the parent and child parts, but they do assist with that conducted heat by making their parts cooler.  Then, heat from other parts has a path to slowly flow to the cooler parts and be radiated away.

It's basically academic though.  We run into heat from atmospheric flight, where are deployable radiators are uselessly draggy and fragile.  Static panels can help cool parts a bit in flight, but it's a tiny bandaid where the real cure is part selection (parts have different temp limits), cargo bays (where parts are shielded from aero friction), and flight path (flying too fast too low blows up parts radiators or no, and doesn't get you to space any better if that's where you're headed).  The nuclear engines used to get hot enough to blow up parts, but they don't now in ordinary use - radiators just aren't required.  Mining - this matters.  There is no effective mining without effective cooling, but it is 100% possible with statics (as described above).  That leaves sundiving - flying so close to the sun that parts overheat.  This, finally, is where you will have to do a lot of thermal engineering to see how close you can get before exploding.

For all other situations, radiators are essentially just roleplay.  My Apollo LM replica has one, because the real one does.  The panels also make nice fashion fairings and covers for ugly parts, edges, corners, lines, etc.

Edited January 13, 2017 by fourfa

[Sharpy]

Re: confusion regarding static radiators: They cool only the parent part. BUT all parts conduct heat to "adjacent" (parent/child), and the speed is dependent on temperature gradient - so heat from hot parts goes to cool ones pretty fast. (the exception being radiators which draw heat out of parent, getting hot themselves).

That means: if you can't attach radiators directly to the "hot part" (coincidentally, all "interested": nuclear engine, drills, ISRUs and heatshields don't accept surface attachment) attach them to a parent or child of the "hot part". The same fuselage a drill is attached to, a common battery you slap on top of ISRU, etc, etc. (and while for nukes and heatshields you only care about not exploding, which by itself is not likely, ISRU and drill thermal efficiency drops like a rock when they are too hot. Seriously, adding a radiator can improve your drill/ISRU output a good 15 times over!)

Bad caveat: they don't cool own children, but instead heat them badly. The child part would need to conduct heat to grandparent through the radiator - and radiator "pumping heat" from its parent doesn't conduct heat TO parent, and will most likely be much hotter than the child part. So: do NOT attach drills to radiator panels. This looks like a great solution, but in reality it works really bad.

Some extra thingies:

- Static radiators actively draw heat and use electricity when you activate them, but they are essentially large surfaces with good thermal conductivity and emissivity meaning they work pretty well when switched off.

- attaching radiator to radiator is not that bad, though not great.

- placement/geometry of where you put the radiators is totally moot. You can attach them then use the offset tool to move them to the opposite end of the craft, or clip 50 radiator panels into each other, and they will work just fine.

- they don't work well when they are over the capacity. Actually, then they heat the parent part. You won't get very good results on reentry, near the Sun, etc as radiators heated externally will soon saturate and the active cooling will falter.

- active radiators are good in that they draw heat from the hot parts directly, and multiple to that, regardless of distance. You can have a base mining module and a base cooling module, no need to mix these - and a pair of big radiators will cool quite a few drills. But they are ridiculously fragile. Way more so than solar panels.

 

[Streetwind]

9 minutes ago, Sharpy said:

Re: confusion regarding static radiators: They cool only the parent part. BUT all parts conduct heat to "adjacent" (parent/child), and the speed is dependent on temperature gradient - so heat from hot parts goes to cool ones pretty fast. (the exception being radiators which draw heat out of parent, getting hot themselves).

Was this silently changed? Last time I checked, they cooled the part they were attached to, and one hop out from that. Look into the part config, it should say "maxHops = 2" or something similar, somewhere in the radiator Module.

[Sharpy]

Just now, Streetwind said:

Was this silently changed? Last time I checked, they cooled the part they were attached to, and one hop out from that. Look into the part config, it should say "maxHops = 2" or something similar, somewhere in the radiator Module.

*shrug* then maybe I'm wrong. Regardless, don't attach drills to radiators. I'm not sure how the hops are counted but apparently a drill right on a radiator is not within 2 hops

[Streetwind]

Well the thermal control systems have it set to 0, and that apparently means vessel-wide. One hop cannot mean cooling only itself, because that is both redundant and impossible (a radiator literally cannot actively cool itself, because any heat removed from itself would be added right back to itself), so it must mean the part it is attached to. Two hops can therefore only mean another hop outwards from that part.

And as for attaching things to radiators... yeah, that won't work because of the direction of the Unity part tree. The hops only look up the tree, not down. I'm honestly surprised why surface attachment is even allowed on radiators in the first place

Edited January 16, 2017 by Streetwind

[Aegolius13]

7 hours ago, Sharpy said:

*shrug* then maybe I'm wrong. Regardless, don't attach drills to radiators. I'm not sure how the hops are counted but apparently a drill right on a radiator is not within 2 hops

I usually put my static radiators on fuel tanks, then attach drills and ISRUs to those tanks.  Works like a charm as far as heat control, so I'm pretty sure the statics can cool two parts away.