[Elecraft] Noisy K3 receiver

[Geoffrey Mackenzie-Kennedy]

Wes Stewart, N7WS, wrote on Friday December 18, 2009 at 4:43 PM:

<snip>

> We have DXpeditions operating on some chunk of coral that are running 
> wireless networks between stations and uploading their logbooks in near 
> real-time to the Internet via satellite yet their operators  > are still 
> "riding the RF gain control". Something is wrong with this picture.

Wes,

What is wrong IMHO is that many if not most commercial receivers designed 
for the amateur market are designed with too much gain ahead of their 
roofing filter, followed by a small dynamic range IF "system" - not a good 
mix which requires several roofing filters of various bandwidths, and in 
some cases multiple loop AGC systems. A multi loop AGC scheme can cause 
"noise problems" if any "see-sawing" between the AGC loops takes place as 
the result of overshoot in any one loop, for example, certainly if the loop 
time constants are incorrect. No wonder then that "riding the RF gain" is 
required in some cases.

One approach used in the design of high performance downconversion superhets 
is always to use negative gain ahead of the roofing filter, followed by a 
low noise large IMDDR3 IF system, with minimnal required AGC applied using 
"electronic attenuators", not current starvation of an amplifier, to protect 
the "weakest link", be it the DSP's ADC if used or an audio amp. The HF 
noise figure of such a receiver without preamp is typically 9 to 10 db.

When using these high performance receivers it is not necessary to "ride the 
RF gain control" when digging out weak signals in the presence of QRM.

73,
Geoff
GM4ESD