The cryogens are held as liquids and kept cold with suitable compressors (such as a G-M tube).

The efficiency of those devices sets the amount of heat that is ejected per unit time.

Let's look at some typical equipment:

Here's an unexceptional G-M tube:

https://www.cryomech.com/products/al330/

Note the compressor dissipates 6 to 7 kW at full power - the beauty is that the compressor can be some distance from the coils (of course!) so there are many options for ejecting heat from the compressor.

See:

https://www.researchgate.net/publication/277598274/figure/fig1/AS:616355128692746@1523961922398/Schematic-of-a-Magnetic-Resonance-Imaging-MRI-scanner-direct-recovery-helium-plant.png

If it's a standard NMR spectrometer the liquid helium is held inside an outer layer of liquid nitrogen. Any cooling (of the room) is down to the boil-off rate (which in modern equipment is very low). Air-conditioning is normally required to keep the electronics cool (the heat given off by the electronics is much more than the cooling effect of the cryogens).

Another consideration is ventilation. In modern installations there should be an emergency fast ventilation system coupled to an oxygen sensor. What can happen (very rarely) is the magnet quenches boiling off all the liquid cryogens quickly displacing all the oxygen in the room and potentially asphyxiating anybody unfortunate to be close by.

Dear Omar,

Interesting question....I never thought about it. If once the cool-down is done (only done once while installation of the machine) the calculation for "regular operation" can probably done like this:

Latent Heat of Evaporation of N2 is 200 kJ/kg. If we assume a loss of 100 l per week (actually it is lower for a 300 MHz machine) we end up if I calculated correctly about "30 W" of "cooling power" 200.000[kJ/kg]*0.8[kg/l]*100[l/week]/(24[h]*7[days]*3600[s])..I'll guess the He contribution can be neglected, but the electronics of the console adds a few hundred watts.

So at the end for a decent air-conditioning no problem!

I hope I calculated half-way correctly

Alfred

Thanks