alt.engineering.electrical

In article <47cb8dd0$0$16687$4c368faf@>,
"Tim Perry" wrote:

> I'll give this some thought although for equipment I deal with it simplest
> explainations are usully the culpret.
> In one case the theory was advanced that group delay of ganged passband
> cavities before the receiver was contrubuting to a decoder problem.
> By using a triple conversion receiver with a passband optimized for digitaly
> encoded I was able to forgo the cavities in one instance. another receiver
> fed by the same antenna required them to make it work. Different frequency,
> more RFI. This is at a location that has problems similar to LA's Mount
> Wilson or Black Mountin in Las Vegas. In short these are RF jungles where
> there are so many signals comming in to and going out of that no practical
> method of predicting all the combinations of mixing products is feasible.
> It's just RF 'soup'.

To summarize simply, normal distributions arise when fluctuations arise
from a sum of separate independent fluctuations (the central limit
theorem). For turbulence, transmission goes through a lot of non-lossy
blobs. Each blob multiplies the signal by a factor close to one. The log
of this product is a sum of the logs of the individual transmission
factors. The central limit theorem applies to this sum. Thus, the log of
the transmission is the random sum of the individual logs. That is, the
log of the transmission is normally distributed.

At this point, that is more than I know.

Bill