AGC Loop; VGA+ Log Detector

This contains the scottyspectrumanalyzer yahoo group backup
Post Reply
Sam Wetterlin
Posts: 646
Joined: Sat Apr 28, 2018 6:48 pm

AGC Loop; VGA+ Log Detector

Post by Sam Wetterlin » Wed Mar 12, 2008 11:47 am

I recently redesigned the AGC Loop boards (VGA, noise filter, AGC) to

match the SLIM format. Dick Gummer assembled them (very well, I might

add) and sent them to me for tests. The results, which were very

good, are posted at:

http://www.wetterlin.org/sam/SA/AGC/AGC_Loop_SLIM.htm



That page includes photos, graphs, and links to the schematic and PCB

layout. My previous post discussed ringing resulting from narrow RBW

filters; the same applies to a narrow noise filter, and the web page

has some 'scope shots showing ringing.



The function of the AGC Loop is identical to the original. The main

change is that the two separate identical VGA boards are combined into

one board with two AD8330s, and the board sizes match the SLIM format.

I actually anticipate building the IF2 processing chain as a separate

assembly, rather than integrated into the egg-crate. In that

configuration, the board size does not matter.



I am also working on using a single VGA board with a noise

filter--either my switchable wide/narrow board or the filter on

Scotty's control board--as a front end to the Log Detector to expand

its low end range. The AGC loop controls the VGA so as to produce a

precisely fixed output level regardless of input level, and the

control voltage indicates the amount of gain and hence the input

level. In contrast, the VGA + Log Detector, with the log voltage fed

back to control the VGA gain functions as an expanded range log

detector and effectively provides a near-square wave output of fixed

level. It's two ways to get to the same place.



The initial tests of the VGA + noise filter + Log Detector were very

good. The VGA expands the total range and at the same time avoids

using the highest Log Detector gains. This puts more of the gain

before the noise filter, which makes the filter more effective. It

also makes the Log Detector more stable, and "stretches" the ripple in

the Log Detector response to give it a longer period. The more the

gain is shifted to the VGA, the flatter the overall response. I am

going to build one of Cash Olsen's VGA boards to pin down the design.



Sam W.

Post Reply