Hello engineers
I've heard some talk that the default radiator setup seen on many stations (set of three radiators per tile connected into a long line with manifolds) is one of the worst setups in terms of cooling power and flow.
Anyone got some pointers on how to improve the setup? I've heard that using filters with manifolds as gas reservoirs can improve both the cooling and the flow 
.
That works pretty well, yeah! Just as long as you're not putting the radiators on a series and not creating packets of pipes in between them with no devices to enforce flow, it should be pretty fine. Flow gets marginally worse the more radiators there are in a consecutive line, in exchange for a negligible amount of cooling.
There's not much you can really do past enforcing the flow, as creating a truly efficient cooling loop for the SM (which I am assuming what it is being used for) requires an absurd amount of space, time and resources, that it's better to just rely on simple setups like what you've been shown.
In other cases, have you ever tried using a turbine for cooling instead of radiators? It's pretty good!
Fushirimaya wrote: Fri Mar 13, 2026 2:20 pmIn other cases, have you ever tried using a turbine for cooling instead of radiators? It's pretty good!
Yeah, I decided to give it a go on the ol' Roid and seems to work pretty well! 
Things get pretty scary considering you need keep it above room temp to get the cooling effects, but it was able to handle a 90% Trit / 10% Hyper-Nob mix running with no delams or fires!
In a real shift (assuming you even get a turbine to use) you would do a mix of CO2 and NO2.
Fushirimaya wrote: Fri Mar 13, 2026 2:20 pmThat works pretty well, yeah! Just as long as you're not putting the radiators on a series and not creating packets of pipes in between them with no devices to enforce flow, it should be pretty fine. Flow gets marginally worse the more radiators there are in a consecutive line, in exchange for a negligible amount of cooling.
There's not much you can really do past enforcing the flow, as creating a truly efficient cooling loop for the SM (which I am assuming what it is being used for) requires an absurd amount of space, time and resources, that it's better to just rely on simple setups like what you've been shown.
In other cases, have you ever tried using a turbine for cooling instead of radiators? It's pretty good!
Flow does not get marginally worse - it gets 50% worse with each consecutive radiator. Is the maths! I can link the document I wrote on the subject
For the more mathematically inclined, I have also created a Desmos graph which is a 1 for 1 transcription of the Space cooling code for radiators - all you do is input the total heat capacity of the gas packet being cooled, and it spits out the loss in both temperature and kilojoules.
Note that past a certain heatmass (a very low mass, relatively), the kilojoule gains slow down at lower temperatures ; that is, higher heatmasses only matter at really high temperatures.
For the purposes of something like a Supermatter, radiators will only ever do you 2kj or less depending on what your static temperature is. N2O can raise the safety temp of the SM which makes radiators significantly more efficient for keeping it cold; but N2O also reduces SM output significantly so be careful.
For something like Trit cooling, just a single radiator with a decent amount of gas passing through it regularly can pretty much keep your trit chamber at stable operating temperature forever. Maybe 2.
The real funny becomes when you need to cool something ludicrously hot - such as the HyperTorus. When you get to the millions of degrees, Radiators are cooling about a gigajoule of thermal energy every tick; and the HFR can generate dozens or even hundreds of gigawatts depending on the setup. That gigajoule figure is with Frezon as the coolant, by the way - with Hypernoblium it quadruples to 4 gigajoules. That heatcap density matters at these ludicrously high temperatures
https://www.desmos.com/calculator/kit1z4k1ja
Anyways heres the graph have fun