[Elecraft] Re: Help with distorted transmit audio

[Don Wilhelm]

Steve,

If you attempt to use AGC to change the noise level, you will also be 
potentially pushing all the signals down to that level as well (how much 
so depends on the amount of AGC slope you have dialed in.

A receiver's sensitivity is measured at the antenna input end, and if I 
understand correctly. what you are proposing can only change the level 
at the output (speaker or headphones).
The receiver sensitivity can be related to its internal noise level (for 
any given gain, more internal noise reduces the sensitivity at the 
input) - and that is a very different thing than band noise.  The amount 
of noise picked up by the antenna and fed to the receiver is effectively 
'just another signal'.

If the band noise is a -90 dBm, one would ideally want the AGC threshold 
to be at something above that - say -85 dBm for best reception (no 
matter what the meters are indicating).  Set that way, signals will 
activate the AGC and not the noise - the human ear/brain can perceive 
signals down to (and below) the noise floor with a little practice.  
With band noise at -90 dBm, the K3 receiver sensitivity is still -140 
dBm, but the band noise covers up actual signals until their level is 
greater than -90 dBm, no receiver can change that situation.

If band noise activates the AGC, the receiver will not hear as well 
because the AGC has reduced its gain. 

Band noise is just another 'signal', but unlike real signals it is 
broadband and unwanted.

73,
Don W3FPR

Steven.Zabarnick at notes.udayton.edu wrote:
> Jack's measurements of the K3 AGC on the link listed below are quite
> interesting. As my noise level on the low bands is usually at least S3 or
> 4, his plots indicate that the noise (S3 is approx -90 dBm) is either in
> the linear range of the AGC, or just barely in the software AGC range.
> Don't we really want the band noise in the  flat part of the AGC curve,
> down near -130 dBm? Would operating with the ATT on and/or the RF gain
> turned down help move the band noise to this part of the curve, resulting
> in a higher S/N ratio for signals just above the background noise?
>
> Steve N9SZ
>
>   
>