[R-390] Spectrum Analyzer

[2002tii]

Barry wrote:

>The above is the part I don't quite understand.  I thought that the 
>AGC would attempt to keep the IF output as constant as possible.

Correct.

>When the sweep signal is outside of the bandpass, then the reduced 
>AGC voltage would result in an increase in the gain in the IF stages.

Correct.  That is, it would try as hard as it could to raise a signal 
down on the skirt of the IF response toward full output.

>When the signal is inside the bandpass, then the AGC voltage 
>increases and thus decreases the gain.

Correct.  That is, as the input signal moves up the skirt of the IF 
response, the AGC adjusts the IF gain so that the IF output remains 
substantially constant.

Said differently, the AGC adjusts the IF gain so that the IF output 
remains substantially constant regardless of signal level.  This is 
true whether the difference in signal level is due to a change in the 
antenna input signal level OR (with the input signal held constant) 
the movement of the signal up the skirt of the IF response to the 
peak response as it changes frequency.

>If that's correct, then why would the IF output pop to full 
>output?  It would seem the opposite would be true.

For the reason given just above.  (Forget I said "pop."  The signal 
would BE at full, or near-full, output as the frequency entered the 
passband, even though the IF response was attenuating it by many tens 
of dB, and would remain substantially constant even as it moved 
toward the peak response and was attenuated less and less by the IF 
response.)  Both the signal down on the skirt of the IF response and 
the signal at the peak of the IF response would produce the same IF 
output, so it would appear to you (looking at the IF output) that the 
IF response did not drop off (or did not drop off as much) at 
frequencies away from the peak of the IF response.  Thus, you would 
not get an accurate picture of the IF response.

>BTW, when you say "perfect AGC with infinite slope", do you mean an 
>AGC with an instantaneous response time?

No, I mean that it could hold the output perfectly constant over a 
wide range of input levels.

Best regards,

Don


Copyright (c) 2011.  Not for redistribution