This is why I will not start use cooling on
my 47 GHz EME LNA

I have seen several Ham's trying to cool the LNA using a
Peltier element

I have been considering to do this also but I first wanted to
try and calculate the effects of such cooling

At first I did not understand how I could make calculations
for this

When I was trying to work out the losses involved using a bad
image rejection I suddenly
found the way to calculate the effects of LNA cooling

As starting point I used my pre amp with nF of 2,45 dB
Now I want to know how much I will gain by cooling the pre amp.
The result is as follow, calculations are done with ambient temperature of 290 K
For 10 degrees cooling nF = 2,38 dB
For 20 degrees cooling nF = 2,32 dB
For 30 degrees cooling nF = 2,25 dB
For 40 degrees cooling nF = 2,18 dB
For 50 degrees cooling nF = 2,11 dB

When Using the peltier cooling we might expect a cooling of about 20 at most 30
degrees below ambient temperature

This will give a reduction in nF of about 0,2 dB, in my
opinion to little gained and a lot of effort needed
with many new problems like water vapour in the LNA

For now, my conclusion is that cooling is not worth the effort.
If cooling is considered then you should follow a more serious approach in using
liquid nitrogen cooling
and bringing temp down as far as possible.
Sergei RW3BP has done this with good success,
take a
look at his report