174 noise
Debeli said:
Thank you again, but once more I am not sure if you got actualy my question.
I do not want to measure exact noise in a system, I was just wandering if a filter can suppress the out-of-band noise bellow kT. Therefore I choose hypotetical Bs>Bf for the experiment. If the noise can not be suppressed bellow kT than result is kTBs, otherwise it should be kTBf.
Regards
D.
Of course everyone may have it's own (and different) opinion regarding your question.
i may write my opinion, but tell me the possibility to write that is the usual, canonical and well accepted in my field of operation (radioastronomy).
I'll call G the available gain. I'll call T the Operative Temperaure (Agilent AN57-1).
Noise power = K ∫ T(f) G(f) df
What about T(f)?
1) inside the flat zone of Bf, supponing the reflection coefficient very close to 0, T(f) is supposed to be constant and is the sum of Noise Temperature of the receiver + The noise generated by the filter "Bf". The second is term is numerically same to the actual ambient temperature.
2) Very far outside Bf but inside Bs, The noise power coming from the filter is very low (supponing |Γ|≈1- see note) but the noise temperature of the receiver change because Te is dependent on Γsource.
3) on the the rising and falling zone close to the flat zone, the situation is more complicated.
The result is the following:
1) Noise power at the output of a network may change vs. T and Gain, it's impossible to discriminate the cause.
2) In order to sicriminate the cause of the noise power, a 2 port network should be examinated.
3) In any case the thermal equilibrium must exist. It is impossible speak or think that a receiver may see a target with a noise temperature less than it's phisycal temperature.
Note: the available noise power still KTB, but if the |Γsouce| is close to 1, the power delivered to a matched load is very low.