KSP 1.2.2 - Radiators do not work correctly in atmospheres

[foamyesque]

I came across this because I have some Mk 1 crew parts on a spaceplane I want to avoid getting thermally soaked during the speed run to orbital speeds. I thought i'd try adding radiators to pull out their core heats, since it's so much lower than their skin temperature tolerance. However, the radiators do not appear to actively pump heat until the hard transition to vacuum. I turned on the thermal debugging mode and overlays to show what's happening. First, what it looks like in normal operation, once in space:

 

As you can see, the radiators are bright yellow in the overlay and the tooltip has them with a nice and toasty skin temperature of nearly 1700K, with all the various fields showing valid data, including heat transfer values.

This is the same ship, on the same flight, about fifteen seconds earlier while still, but only just, in the atmosphere:

 

No heat transfer whatsoever -- in fact the values for it are listed as "??", which suggests to me the calculation choking on an undefined value somewhere -- and a skin temperature of ~500K, which is actually less than the skin temperature of any of the parts around it, when it should instead have the active heat pump still pushing it up into the 1700K range just as in space. It's my understanding that the radiators are supposed to be less effective in atmosphere, but I'm pretty sure they're still supposed to do something.

Edited December 8, 2016 by foamyesque

[LatiMacciato]

I can confirm that there is something wrong with radiators not cooling in orbit .. also drills do not get cooled .. maybe there's something borked with the "cooling in vaccum section" of KSP v1.2.2

[Starwaster]

18 hours ago, foamyesque said:

I came across this because I have some Mk 1 crew parts on a spaceplane I want to avoid getting thermally soaked during the speed run to orbital speeds. I thought i'd try adding radiators to pull out their core heats, since it's so much lower than their skin temperature tolerance. However, the radiators do not appear to actively pump heat until the hard transition to vacuum. I turned on the thermal debugging mode and overlays to show what's happening. First, what it looks like in normal operation, once in space:

 

 

As you can see, the radiators are bright yellow in the overlay and the tooltip has them with a nice and toasty skin temperature of nearly 1700K, with all the various fields showing valid data, including heat transfer values.

This is the same ship, on the same flight, about fifteen seconds earlier while still, but only just, in the atmosphere:

 

 

No heat transfer whatsoever -- in fact the values for it are listed as "??", which suggests to me the calculation choking on an undefined value somewhere -- and a skin temperature of ~500K, which is actually less than the skin temperature of any of the parts around it, when it should instead have the active heat pump still pushing it up into the 1700K range just as in space. It's my understanding that the radiators are supposed to be less effective in atmosphere, but I'm pretty sure they're still supposed to do something.

Values of ?? just means that those fields haven't been changed from their default values. No code has run to update them since the debug fields were turned on for them (?? is the default string value). If they're not turned on or the radiators are not currently cooling anything then they don't update.

As to why they aren't cooling it's because they don't cool if they are supersonic to prevent them from being abused in reentry situations. I think that's hard coded too unless there's an option in their config that turns that off.

@LatiMacciato I think you're 'confirming' something different from what foamy is experiencing?

Edited December 9, 2016 by Starwaster

[sal_vager]

Guys, this isn't a hardcoded limitation of the radiators, just take a look at the external temperature in Foamyesque's second pic.

It's six thousand, four hundred and forty four point seventy six degrees kelvin!

6,444.76!

Despite the rarified nature of the atmosphere at those altitudes you're compressing it at 2264m/s,  If the radiators were running they would be pumping this heat into your vessel.

Unlike Earth, Kerbin has an atmosphere that instantly cuts off above 70km, putting you in the slightly cooler vacuum of space at 4 degrees K.

[sal_vager]

More info on KSP's heat system here.

 

[LatiMacciato]

2 hours ago, Starwaster said:

Values of ?? just means that those fields haven't been changed from their default values. No code has run to update them since the debug fields were turned on for them (?? is the default string value). If they're not turned on or the radiators are not currently cooling anything then they don't update.

As to why they aren't cooling it's because they don't cool if they are supersonic to prevent them from being abused in reentry situations. I think that's hard coded too unless there's an option in their config that turns that off.

@LatiMacciato I think you're 'confirming' something different from what foamy is experiencing?

possibly but most defenetly connected to every radiator i tried out in space on an asteroid with any asteroid drill

they just do not cool them formerly

[Starwaster]

1 hour ago, sal_vager said:

Guys, this isn't a hardcoded limitation of the radiators, just take a look at the external temperature in Foamyesque's second pic.

It's six thousand, four hundred and forty four point seventy six degrees kelvin!

6,444.76!

Despite the rarified nature of the atmosphere at those altitudes you're compressing it at 2264m/s,  If the radiators were running they would be pumping this heat into your vessel.

Unlike Earth, Kerbin has an atmosphere that instantly cuts off above 70km, putting you in the slightly cooler vacuum of space at 4 degrees K.

Active radiators only 'pump' heat from a part's internals to the radiator's internals. NEVER the reverse. The only way external (environmental) temperatures would be an issue would be as a hard coded limitation. Even if the part's internal temperature exceeds the radiator's the radiator still stops pumping in that situation. And it's for exactly the reason that I indicated. To prevent OP radiators in reentry situations.

 

[foamyesque]

5 hours ago, sal_vager said:

Guys, this isn't a hardcoded limitation of the radiators, just take a look at the external temperature in Foamyesque's second pic.

It's six thousand, four hundred and forty four point seventy six degrees kelvin!

6,444.76!

Despite the rarified nature of the atmosphere at those altitudes you're compressing it at 2264m/s,  If the radiators were running they would be pumping this heat into your vessel.

Unlike Earth, Kerbin has an atmosphere that instantly cuts off above 70km, putting you in the slightly cooler vacuum of space at 4 degrees K.

 

I'm aware of the external temperature, but this explanation isn't.

 

Radiators, by definition, radiate. Radiation is governed by the skin temperature of a radiator. This is independent of the external temperature. Even if the external temperature was sufficient to cause a net flux into the plane, an active radiator should have a skin temperature up in its operational ranges. Moreover, you're wrong about the external temperature being so high the radiator wouldn't work. The temperature is very high, yes, but it's attenuated and doesn't have much effect actual heat fluxes involved. You can tell because the radiator, and everything else on the plane, have skin temperatures well under a radiator's operating temperature, and that's largely due to the fact that it's still shedding heat from flight in the lower atmosphere.

EDIT: @sal_vager: Further proof of this can be seen in the action menu for the radiator of the atmospheric screenshot, where the convective flux (i.e. atmospheric heating) is displayed as nil. On the other hand, you can see the huge jump in conductive flux between the atmospheric and vacuum situations (~80kW delta) once the heat pumps engage.

I strongly suspect I could build an ore refinery, put it in a 69x71km orbit, and watch my radiators toggle on and off as I cross the invisible boundary, while maintaining an external skin temperature under 500K.

 

6 hours ago, Starwaster said:

Values of ?? just means that those fields haven't been changed from their default values. No code has run to update them since the debug fields were turned on for them (?? is the default string value). If they're not turned on or the radiators are not currently cooling anything then they don't update.

As to why they aren't cooling it's because they don't cool if they are supersonic to prevent them from being abused in reentry situations. I think that's hard coded too unless there's an option in their config that turns that off.

@LatiMacciato I think you're 'confirming' something different from what foamy is experiencing?

That's correct, this is about cooling in atmosphere. Cooling in vacuum seems to work correctly, or at least the debug tooltips say it is. :v

And, just to be clear, yes, my radiators were engaged from the moment I turned my engines on; the activation (not toggle!) was mapped to the stage action group.

Edited December 9, 2016 by foamyesque

[LatiMacciato]

possibly that is still connected tho..

...tried a stock game (only hyperedit), drill is not cooling, radiator works on Kerbin ground to cool perfectly, stops working while attached on an asteroid.
There MUST be something wrong with the cooling stuff inside KSP at all.

screen:

regards

Edited December 10, 2016 by LatiMacciato

[RoverDude]

20 hours ago, LatiMacciato said:

I can confirm that there is something wrong with radiators not cooling in orbit .. also drills do not get cooled .. maybe there's something borked with the "cooling in vaccum section" of KSP v1.2.2

Ok.  There are apparently two different things being discussed in this thread.  The first is skin/internal temp and atmospheres.  In your pic you are referring to core heat.  Apples and tractors.  Are you on a modded install?

[foamyesque]

@RoverDude:I dunno what @LatiMacciato's problem is; it may or may not be related to this, but I don't think so. I believe they've posted about it elsewhere before, though I don't have a link handy.

 

I can also inform you that the heat-pumps-not-on issue occurs with both the fixed radiators (as above) and the TCS systems; I've tried both and they both stay inoperative and at unheated skin temperatures until crossing into vacuum. Next trial, I suppose, is to see how they work landed; I might heat something with a rocket engine and infinite fuel and see if the radiator engages or something.

[LatiMacciato]

I filed a bug report about my issue, it might or might not be connected .. just thaught it is because this thread is about radiators beeing used and do not work correctly in KSP (even beeing in atmosphere or not) and reporting anything related or similar can be helpful aswell.

I'm sorry if that does not interact or match this topic here.

Regards

Edited December 10, 2016 by LatiMacciato

[sal_vager]

Radiators will automatically switch off when the external temperature is higher than the hottest part that the radiator is serving, I don't think this is new to 1.2.2 but I'll have to test some more to be sure.

Core heat is something else again, and completely separate.

[foamyesque]

2 hours ago, sal_vager said:

Radiators will automatically switch off when the external temperature is higher than the hottest part that the radiator is serving, I don't think this is new to 1.2.2 but I'll have to test some more to be sure.

That's physically stupid behaviour, then. It should only disengage if the radiator is being heated beyond its operational temperatures.

[sal_vager]

Didn't say it wasn't but I spent hours testing this only for Imgur to chew up my gallery, mess up the order of images, lose text then remove my edit permissions so I can't fix it.

[Starwaster]

5 hours ago, sal_vager said:

Radiators will automatically switch off when the external temperature is higher than the hottest part that the radiator is serving, I don't think this is new to 1.2.2 but I'll have to test some more to be sure.

So in other words, hard coded behavior specific to the radiator and separate from any of Flight Integrator thermals.

Edited December 11, 2016 by Starwaster

[sal_vager]

Seeing as it is triggered by flight integrator thermals I'd hardly call this separate, and the radiators are the only parts that actively move heat so it's hardly surprising that this is specific to the radiator, but if you look closely the radiators aren't shutting off, my assumption that they were shutting down was wrong.

The important part here is the second line from the top.

AR:CoolingParts: 0/0

It's not because you're supersonic, if your parts were hotter than the shockwave the radiators would cool them, if they hadn't exploded.

I'm doing some more testing here but this is in 1.1.3

And in 1.0.5.

[foamyesque]

1 hour ago, sal_vager said:

Seeing as it is triggered by flight integrator thermals I'd hardly call this separate, and the radiators are the only parts that actively move heat so it's hardly surprising that this is specific to the radiator, but if you look closely the radiators aren't shutting off, my assumption that they were shutting down was wrong.

The important part here is the second line from the top.

AR:CoolingParts: 0/0

It's not because you're supersonic, if your parts were hotter than the shockwave the radiators would cool them, if they hadn't exploded.

I'm doing some more testing here but this is in 1.1.3

And in 1.0.5.

 

I have parts that are hotter than the radiator, and the radiator's not cooling them; the logic being used here does not make sense to me.

[sal_vager]

2 minutes ago, foamyesque said:

 

I have parts that are hotter than the radiator, and the radiator's not cooling them; the logic being used here does not make sense to me.

Even if the radiators were cooling them it is possible to get into a situation where the parts are hotter than the radiators, the radiatorHeadroom value in the cfg file controls the percentage of the radiators maxTemp they can handle before they are saturated (cooling 99.99%) and cannot draw in any more heat.

Most parts have a maxTemp that is higher than the radiators saturation temperature.

[foamyesque]

3 hours ago, sal_vager said:

Even if the radiators were cooling them it is possible to get into a situation where the parts are hotter than the radiators, the radiatorHeadroom value in the cfg file controls the percentage of the radiators maxTemp they can handle before they are saturated (cooling 99.99%) and cannot draw in any more heat.

Most parts have a maxTemp that is higher than the radiators saturation temperature.

 

I am 100% certain that temperature is higher than 500K. Like, that check makes sense, but it's not applicable here.

Edited December 11, 2016 by foamyesque

[sal_vager]

2 minutes ago, foamyesque said:

 

I am 100% certain that temperature is higher than 500K. Like, that check makes sense, but it's not applicable here.

It is applicable here, because that's how they work when the outside temperature is below the temp of the part being cooled.

~500k is the point which radiators try to keep a part cooled at, which they can do while not saturated as any ground test will show.

[foamyesque]

9 minutes ago, sal_vager said:

It is applicable here, because that's how they work when the outside temperature is below the temp of the part being cooled.

~500k is the point which radiators try to keep a part cooled at, which they can do while not saturated as any ground test will show.

500K is the temperature of the radiator, not that of the parts it is trying to cool. I have internal temperatures several hundred degrees above the radiator's skin temperature and as soon as the atmosphere is exited the radiators notice this and turn on full blast. Therefore, that check is not what is blocking radiator function. So it's not applicable here.

For example, the small fixed radiators in this example have a maxTemp of 2500K and a radiator headroom multiplier of 0.75, for a saturation temperature of 1875K. That's certainly above 500K...

Edited December 11, 2016 by foamyesque

[sal_vager]

3 minutes ago, foamyesque said:

500K is the temperature of the radiator, not that of the parts it is trying to cool. I have internal temperatures several hundred degrees above the radiator's skin temperature and as soon as the atmosphere is exited the radiators notice this and turn on full blast. Therefore, that check is not what is blocking radiator function. So it's not applicable here.

For example, the small fixed radiators in this example have a maxTemp of 2500K and a radiator headroom multiplier of 0.75, for a saturation temperature of 1875K. That's certainly above 500K...

Then your radiator has been heated to 500k, but the radiators heat isn't the factor here, it's the external heat and the heat of the parts the radiator can reach.

Also, radiators are just heat exchangers, they can only exchange heat when there is a difference in heat with their surroundings, and heat always flows from the hottest to the coldest until equilibrium is reached, unless work is added such as with a refrigeration system, that's how heat can be moved from a cooler part into a hotter radiator but there's no such system between the radiator, or heat exchanger, and the atmosphere.

It also takes time for heat outside the vessel to conduct inward, that's why your craft temperature increases instead of instantly becoming red hot.

[foamyesque]

10 minutes ago, sal_vager said:

Then your radiator has been heated to 500k, but the radiators heat isn't the factor here, it's the external heat and the heat of the parts the radiator can reach.

Also, radiators are just heat exchangers, they can only exchange heat when there is a difference in heat with their surroundings, and heat always flows from the hottest to the coldest until equilibrium is reached, unless work is added such as with a refrigeration system, that's how heat can be moved from a cooler part into a hotter radiator but there's no such system between the radiator, or heat exchanger, and the atmosphere.

It also takes time for heat outside the vessel to conduct inward, that's why your craft temperature increases instead of instantly becoming red hot.

 

The external heat is not relevant from a physics standpoint and the heat-pump argument is a red herring (and the "it takes time for heat outside to conduct inwards" even more so) because 1: it's immediately obvious that a heat pump mechanism exists given how radiators operate in space and 2. the thing it's supposed to be cooling is 200K hotter than the radiator so it isn't even trying to push heat uphill.

What is relevant are heat fluxes, and the tooltip very clearly shows that the convection heat loading is nil. Therefore the radiator is not being heated by its passage through the air whatever the external temperature is claimed to be. The air has a high average energy (hence the temperature) but it's so attenuated it can't actually transfer any of it, hence the no-convection-flux reading. This is correct behaviour and why the ISS can operate radiators despite slamming through air at 7km/s.

Edited December 11, 2016 by foamyesque

[sal_vager]

5 minutes ago, foamyesque said:

 

The external heat is not relevant from a physics standpoint and the heat-pump argument is a red herring (and the "it takes time for heat outside to conduct inwards" even more so) because 1: it's immediately obvious that a heat pump mechanism exists given how radiators operate in space and 2. the thing it's supposed to be cooling is 200K hotter than the radiator so it isn't even trying to push heat uphill.

What is relevant are heat fluxes, and the tooltip very clearly shows that the convection heat loading is nil. Therefore the radiator is not being heated by its passage through the air whatever the external temperature is claimed to be. The air has a high average energy (hence the temperature) but it's so attenuated it can't actually transfer any of it, hence the no-convection-flux reading. This is correct behaviour and why the ISS can operate radiators despite slamming through air at 7km/s.

None of that actually matters though, as I said the radiators stop cooling parts when the outside air is hotter than the part, that's why you're seeing the effect that you are.