02:03:38	Jon:	So Gray, an amplifier at a fixed Tsys will add a fixed amount of noise, but you can have a choice about how much noise is added as imprecision (flat distribtution) and what fraction is added as back action (lorentz distribution). With typical noise matching, they add equally on resonance. But if you change the matching, you will lower the imprecision noise and raise the back action noise. Overall, you will have lower SNR over the full bandwidth, but that is made up for by having a much wider bandwidth.
02:13:28	Gray Rybka:	Does the amplifier refrigerate the resonator in that case?
02:16:08	Marianna Safronova:	Just curious - where does DM density 0.45 GeV/cm^3 value is from?
02:21:58	Jon:	@Gray I am not sure I understand your question. Why would the amplifier refrigerate the resonator?
02:22:40	Gray Rybka:	It's historical so all the plots agree.  Check out Cirian's talk from Monday, it gives justification for both 0.3 and 0.45, many sigma apart.
02:23:05	Marianna Safronova:	Thanks!
02:24:28	Gray Rybka:	@Jon, I like to analyze my understanding of thermal noise from the point of view of "can I build a refrigerator out of this".  (If I can and its efficiency is better than carnot, then I know I have to redo my calculations)
02:25:28	Gray Rybka:	So if I don't understand which way heat is flowing, I don't really understand what is going on.  So I must still not really understand what is going on.
02:30:47	Chiara Salemi:	I’m not sure if this is exactly what you are asking, but the resonator’s thermal noise is flowing into the amplifier, and showing up in the data with that line shape that I showed in red