PSN-L Email List Message
Subject: Re: feedback semantics
From: Brett Nordgren Brett3mr@.............
Date: Fri, 15 Feb 2008 10:24:19 -0500
No problem. My only issue had been that I was interpreting 'soft' as
meaning 'low loop gain' and because of that had privately dismissed the
concept without pursuing it. In fact, you are describing a feedback
approach which has low/no loop gain at the mid-band frequencies you are
hoping to observe, though the integral feedback should become quite strong
and effective at sufficiently low frequencies.
Regarding the centering forces required, it's my feeling that they are not
all that small. For example, in the STM-8 example, 1 degC changes the
spring force by about 10 dynes, which I believe is quite large compared to
the forces you might be wanting to measure.
I reread your June 2005 paper which referenced the Sprengnether setup and
now understand much better that you were describing 'integral only'
feedback. I was hoping to use that as a sample case in my analysis
programs to better understand how it performs.
In particular, the program would need to know the main instrument
parameters most of which I can find, assuming it was an S-5100-V, except
I'd need to know what natural frequency you had set it up for?. And do I
understand correctly in Figure 3 that "Sensor output" is where you are
obtaining your instrument output, with a sensitivity of about 2000V/m?
At 06:18 PM 2/11/2008 -0500, you wrote:
> I am not very familiar with the terminology used in the world of
> engineering controls.
>My term 'soft' for the feedback scheme that I have used seemed reasonable
>to me for the
>following reasons: (i) it is as you noted, happerning at a much lower
>frequency than most of
>the signals of interest (say teleseisms at 20 s period, and (ii) it is
>nowhere near as
>powerful as the forces required to do force-balance; i.e., keep the mass
> You mention the matter of using feedback with a pendulum. I can't
> imagine a reasonably
>simple pendulum for seismic purposes where feedback would ever be
>necessary. The primary
>source of motion at very low frequencies is the change in shape of the
>earth. Every mass
>part of our planet contributes to the local field, and so a plumb bob
>concerning eigenmodes and tides (as the VolksMeter has demonstrated).
> Your question about my use of the word 'noise' relative to force balance
>systems--anything that works with the derivative of the position of the
>inertial mass (flat
>to velocity sensor) will fail to see earth motions at really long periods
>(starting around a
>few thousand seconds or even less). Even though the period of the
>pendulum is only about 1 s, it is well suited to the study of earth
>changes happening over
>days, months, and even years.
My e-mail address above should be working, but if not
you can always use my mail form at: http://bnordgren.org/contactB.html
using your Web browser.
Public Seismic Network Mailing List (PSN-L)
[ Top ]
[ Back ]
[ Home Page ]