Friday, November 02, 2007

Faster, Better Open-Loop Gain Measurements

It is often desirable to measure the performance of a feedback network in a particular configuration so as to model the open-loop behavior. But this type of measurement is often challenging due to high gains, small input signals, and long settling times. This article shows you how to characterize the frequency response of the network rapidly and with minimal effort, using available tools and a small amount of special circuitry.

2 Comments:

At 7:40 PM, Anonymous Anonymous said...

Interesting article.

Noise reduction can be accomplished by cross-correlating the input and output signals to determine gain and phase response. That approach is described in an application note on the Syscomp web page (www.syscompdesign.com)
"A Software-Based Network Analyzer"

The theory was also described in the September 2006 issue of Circuit Cellar Magazine.

Peter

 
At 6:47 PM, Anonymous Anonymous said...

Interesting approach.
Sometimes one needs to know the Bode plots, including phase, to determine how stable a loop is with some unknown device. My favourite tool then is the old HP3577A low-frequency network analyzer configured with 1MOhm 'scope probes in place of the S-parameter test set. It can be programmed to calculate the open-loop (Bode plots) of a closed loop as long as it can be made stable. You need to design some simple 'hooks' around one of the op-amps already in the loop, but it's so convenient, and avoids the high-gain problems because the loop is closed. I have used it in AGC loops with log-amps, PLLs and analog adaptive equalizers. As my tech note is not published, a similar reference is:
"Measure PLL Open-Loop Response With Loop Closed", by Kevin Johnson, in Electronic Design, pp. 79-86, November 22 1990.

M.K.

6:51 PM

 

Post a Comment

<< Home