Open NirViaje opened 4 years ago
NirViaje
commented
4 years ago 
NirViaje
commented
2 years ago A lot depends on the temperature requirements for your superconductors. That in turn depends on how much magnetic field you need, your mass budget for the coils, and how exotic a material you can afford.
There's a nice summary of (mostly active) space cryogenics here. Passive cooling to 50K has already flown on Planck (they were planning 60K, but did better), and the JWST is planning 40K.
The whole telescope and both instruments are first cooled to 50 K (-223oC) by taking advantage of the cold temperatures of deep space by radiative cooling. This is achieved by the three large “petals” that can be seen in the picture. These have a large surface area that maximizes the cooling power. The Low Frequency Instrument (LFI) is then cooled to 20 K (-253oC) by a “sorption cooler”. This works by using materials (metal hydrides) that absorb hydrogen at low temperature and evolve it at high temperature. This material is used in beds that are sequenced so that when one bed is absorbing another is being heated to give out hydrogen. In this way, the high pressure gas is pre-cooled to the temperature of the radiative cooler then expanded through a narrow orifice (a process called Joule-Kelvin expansion). Under these conditions the hydrogen cools to around 20 K. The use of the metal hydride beds to provide compression of the gas means that there are no moving parts. The cooler therefore has inherently low vibration that could upset the measurements.
The environment matters a lot. "Cooling" comes from radiating to space at ~5k (averaging over star flux in the local area of the Galaxy); heating comes from any local object. The more and larger shields you need for Sun, Earth, Moon, etc, the less solid angle you have to radiate to the rest of the sky.
JWST's 5-layer shield was designed to meet the requirement of no more than 2W passed flux, which in turn could be radiated to space for a telescope instrument temperature of 40K. Because flux goes like T^4, keeping all things the same, you'd need to drop the flux by a factor of 4^4=256 to get to 10K. If everything was perfect, this would just require 2 (maybe 3) more layers.
