PSN-L Email List Message

Subject: Re: Wielandt's "inverse filter" or integrator (long)
From: Brett Nordgren Brett3kg@.............
Date: Wed, 15 Dec 2004 17:54:23 -0500


Dave,

Sorry, I don't have a copy of the "Noise" paper.  If Chris Chapman is still 
monitoring the list, he might be able to help you find it.  I know he'd 
been working to acquire many of the old journal articles.

Enlightenment is good.  Feedback systems are some of the more non-intuitive 
things you're likely to come across.  To tell the truth, when I was first 
doing feedback design, I think it was a couple of years before I could 
honestly claim that I had a "feel" for what was going on.

One general comment on why feedback seismos exist....  Starting with a 
spring-mass which has a position sensor and some means for applying 
force-feedback, the feedback is used to *reduce* and shape the frequency 
response to predictably match the shape you want.  There is a fairly 
standard set of seismo. frequency-response curves that seem to be in common 
use by seismologists for different purposes.  In most cases, around here, 
the Velocity Broad Band response seems to be the most common.  The purpose 
of having feedback is to tightly control the instrument response (at the 
expense of sensitivity).  However in electronic seismo. systems, high 
sensitivity is quite easy to get.  The more fundamental issue will usually 
be instrument noise.

The "inverse filter" you refer to was one of many terms coined in the early 
days of feedback seismometers as designers were first putting together and 
publishing the concept.  A more accurate name might be "one of those things 
you need in order to make the frequency response what you want", though I 
admit, it is a little lacking in elegance and brevity.  His choice for the 
term probably comes from the fact that when *any* element is put in the 
feedback path, the overall instrument response that results will tend to be 
the inverse of that element's frequency response.  That fact is fundamental 
to *all* feedback designs.

If you haven't spent time with them, I would very much recommend Wielandt's 
articles  "Seismic Sensors and their Calibration " online at
http://www.geophys.uni-stuttgart.de/seismometry/man_html/index.html
and "Seismometry", located at
http://www.geophys.uni-stuttgart.de/seismometry/hbk_html/index.html
In particular, the sections on force balance seismometers should be 
helpful.  These are much more understandable and technically refined than 
the early journal articles, which often were not easy to follow.

And you can try my Web site downloads index at
http://bnordgren.org/seismo/index.html
The files "feedback", "complex" and particularly "loop3" might be of 
interest.  Also, today, I put a draft copy of "loop4" on the site, but 
haven't put it in the index yet.  You can access it directly at
http://bnordgren.org/seismo/loop4.pdf
It is considerably improved from "loop3" (and somewhat larger).  These 
relate to an analysis of the feedback loop of S-T Morrissey's STM-8 leaf 
spring vertical.  Its loop design is very similar to those of the 
commercial VBB instruments.  If you can follow the math, great, but there's 
still a lot there that should be helpful even neglecting the math details.

Hope these help.

Brett

At 07:11 AM 12/13/2004 -0500, you wrote:
>Good morning all,
>
>I am looking for some enlightenment on the subject circuit. I know just 
>enough about it to be dangerous and not enough to understand what exactly 
>it does and how to select the corner frequency of the high pass filter. I 
>have not been able to obtain a copy of Wielandt's paper "Noise in 
>Electronic Seismograph Systems", so if anyone has an electronic copy (in 
>english) I would appreciate a copy. Short of that, can anyone provide an 
>explanation of how this thing works, and what the alternatives to it might be.
>
>I understand that the output of the high pass filter is fed back to the 
>input amplifier and subtracted from the input signal, but doesn't the 
>phase shift in the filter prevent this from working as Allan Coleman 
>describes in his papers?
>
>Thanks for any help you may provide
>
>Dave Youden



__________________________________________________________

Public Seismic Network Mailing List (PSN-L)


[ Top ] [ Back ] [ Home Page ]