[TowerTalk] Noise figure in cascaded networks

[TexasRF@aol.com]

There is a minimum noise level form galctic sources that varies inversely 
with frequency squared. We have lots of published data documenting this for 
135Mhz and higher. The data has been gathered primarily to allow satelite 
link signal to noise calculations to be done with some degree of accuracy. 
This allows system engineers to provide a minimum cost solution to space 
communications.

Other services benefit from the data: radio telescope operations and amateur 
moon bounce (my passion) operations to name a couple. We know from studying 
the data that the noise level from thses galactic sources varies greatly with 
what part of the sky the antenna is looking toward. for example, at 144 Mhz, 
the very lowest reading is toward the northern and southern galactic poles 
where the noise level is about half that produced by a 50 ohm dummy load 
connected to the receiver. The highest  level is toward the center of the 
Milky Way where the constellation called Sagitarious is located and the noise 
level at 144Mhz is about 10 times the level produced by the 50 ohm dummy 
load. You can see from this that the signal to noise level can vary over a 
range of about 20 to 1 (13 db) depending on where the antenna is pointed.

If the antenna is fixed at zero derees elevation, both the lowest level and 
the highest level noise sources, as well as every thing in between are 
encountered every day at any location in the USA and most of Europe. The time 
that a given level is encountered varies with the season. Remember how we see 
the Milky Way at night in the summer and don't see it in the winter?

At 28 Mhz, the noise levels are approximately 25 times stronger (freq 
squared) than at 144 Mhz. This is equivalent to a level of about 12 times 
that of a 50 ohm dummy load at a minimum and a much higher level as a 
maximum. The actual amount for the higher level depends on the beamwidth of 
the antenna as the Milky Way occupies only part of the sky and is probably 
not filling the entire antenna patern with noise at h.f.. This is a whole 
'nuther subject to be discussed seperately.

The bottom line of all this is that this galactic noise causes a noise floor 
to exist in the receiver that degrades the signal to noise ratio of the 
system. At 28Mhz this floor is roughly equivalent to a 12 db noise figure. If 
you add a 6 db attenuator to the system, the effective noise figure increases 
to about 18 db. These figures are the very best you can do while the antenna 
is pointed toward a low noise part of the sky. It gets much worse as the 
antenna "sees" higher noise parts of the sky. The signal to noise test 
results reported here will vary depending on just when you did them and which 
direction your antenna was pointed! 

The next step is to extrapolate the total noise performance of the system 
using these noise figures and you can calculate down to .1 db what the effect 
of adding coax loss will be to your signal to noise ratio. Perhaps others 
will take the time to expound on this but we are getting way off topic for 
tower related issues!

Tnx/73 de Gerald Williamson, K5GW
Owner/General Manager, Texas Towers

--
FAQ on WWW:               http://www.contesting.com/towertalkfaq.html
Submissions:              towertalk@contesting.com
Administrative requests:  towertalk-REQUEST@contesting.com
Problems:                 owner-towertalk@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm