# Talk:Noise temperature

WikiProject Physics (Rated Start-class, Mid-importance)
This article is within the scope of WikiProject Physics, a collaborative effort to improve the coverage of Physics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
Start  This article has been rated as Start-Class on the project's quality scale.
Mid  This article has been rated as Mid-importance on the project's importance scale.

## Duplication

There are already articles on Noise Figure and cascade equations. Pleas consider using them as sources for references instead of duplicating the information here. Thbusch (talk) 19:53, 28 August 2008 (UTC)

In the last section of Applications, I feel the claim that the temp. of the antenna has no influence on the noise temperature of the antenna is incorrect. One could say it has a *negligible* effect, but in reality I would expect heating to change things to some degree, right? —Preceding unsigned comment added by 132.162.77.5 (talk) 01:01, 26 February 2010 (UTC)

## Why so many subscripts?

I don't really see the need for all the subscripts in this equation.

${\displaystyle P_{RL}=k_{B}T_{s}B_{n}}$

Much easier to read is simply P = k T B. Just to compound the problem, part of the article uses k with a subscript and part does not. Drkirkby (talk) 23:30, 1 May 2011 (UTC)

It's because someone copied it from a book. I'm fixing it now.... Interferometrist (talk) 15:51, 5 May 2011 (UTC)

## NF and source temperature

Hi Interferometrist, you're doing some nice work on the article! I'm not sure about a couple of points though. The article currently states:

"The noise figure can also be seen as the decrease in signal to noise ratio (SNR) caused by passing a signal through a system if the original signal had a noise temperature of 290 K."

Isn't NF independent of the temperature of the original signal? I thought that was the point. I'm also not convinced the source temperature is often 290K - doesn't it depend on the antenna (or whatever the source) and the matching network? But I'm not certain, so I thought I'd ask. GyroMagician (talk) 13:52, 6 May 2011 (UTC)

Hi, thanks for the feedback. Well it isn't the "temperature of the signal" (or perhaps that isn't what you meant to say) but the noise accompanying the signal. In the extreme case, there would be a signal that had no noise whatsoever (thus SNR=infinity) but then you amplify it using an amplifier with a noise temperature of 290 K. Now the SNR has gone from infinity to something finite so it's been degraded my much more than the NF. It's just that the definition of NF is with respect to an arbitrary temperature 290 K (didn't have to be) which makes it directly useful in cases where the source happens to have a noise temperature of 290 K which they often do. That was the point I was making in that section. But if it isn't clear to YOU, then perhaps it needs to be reworded.
Also you're right that this depends on the matching network doing its job. If there is a mismatch then the noise can only be worse (since you're not coupling the full power of the source into the amplifier) but I don't think you can easily quantify how much worse because that depends on how the amplifiers noise temperature varies according to source impedance. For instance, one amplifier might have current noise intentionally introduced (which would be foolish, but to make the point.....) at its input terminals. Shorting out the input will make the noise disappear. A different amplifier might have noise intentionally introduced on the output terminal of the amplifier, so that shorting out the input wouldn't change the output noise level at all. Clearly these amplifiers will respond differently to a mismatch which causes the input impedance to differ from its specification. So I don't think you can generalize about the effect of a mismatch except to say that it can only hurt the SNR (assuming that the amplifier actually DOES have the best NF when hooked up to the specified input impedance, which itself might not be true!).
I'm also not convinced the source temperature is often 290K
Well I think it "often" is, but of course that depends. I gave explicit examples where it would be higher or lower using an antenna. I believe (haven't measured it though) that a 600 ohm microphone would have a thermal noise given by the formula using T=290K, and the noise from a photodetector in the extreme low-light case (where there is no shot noise) can be calculated using the room temperature and the shunt conductance of the detector modelled as a current source. And an antenna at VHF or microwave frequencies pointed horizontally should be looking mainly at room temperature black body radiation and thus be 290 K (I believe that's where they started using the term noise figure defined in that manner). So the stated NF is convenient in those cases (where the source has 290 K noise). But in other cases NF (if that is the specification you received) really needs to be converted back into noise temperature and then you add that noise temperature of the amplifier to the noise temp of the source to get the resulting noise temp (and from which you can deduce the decrease in SNR which will be different from the NF as stated). Does that make sense? And did I answer your questions? And does the article need rewording? (which you can do as well of course ;-) -- Interferometrist (talk) 16:00, 6 May 2011 (UTC)
I prefer not to use noise figure for temperature noise calculations for cascading components. I think it much better to use the noise temperature for each component and its gain not in dB so the gain greater than zero. Letting the gain be less than one allows one to put in the losses in a straight forward manner. When you do this, then the arbitrary 270 K term goes away that is used for Earth. Once done then the simple summation of the noise temperature can be put into the link budget to calculate SNR or Eb/N0. — Preceding unsigned comment added by 128.156.10.80 (talk) 19:58, 1 May 2014 (UTC)