In a message dated 23/01/2007, rsparks@.......... writes:

Randle,   Thanks for the thought provoking response.

I  can understand what you are saying about direct coupled sensing of 
pendulums,  but I think I am correct in saying that your comments are limited to 
sensors  that actually extract some power from the pendulum motion.  A sensing  
system that (for instance) that counted area sweeping past a camera in a time  
period should not count as a direct coupled sensor.
Hi Roger,
 
    It should apply to any sensor, but if it exchanges  energy there will be 
a perturbation.
    If it senses displacement you will sense  position.
    If it is a velocity sensor you will sense  velocity.
    If it is a force sensor you will sense  acceleration.

I am  using a magnetic/coil sensing system with a heavy damper. The system is 
 near or even more than critically damped. My sensing magnet and coil are  
not providing much load on the pendulum, most of the energy captured by the  
pendulum is dissipated in the damper. I certainly agree that my sensing system  
is really acceleration based as it does load and reduce the pendulum  speed. 
The sensor also acts to accelerate the pendulum from a standing  start when a 
wave arrives. These are two reasons to describe the system  as an acceleration 
sensor.  

Despite these acceleration events, I  have described my sensor as a velocity 
device. I do this because maximum  output of the sensor occurs when the 
pendulum is moving at maximum  velocity. Or at least that is what I believe is 
happening. 
    It IS a velocity sensor in this case. For short  periods, the mass is 
'stationary' and the relative linear motion = the  earth motion. You have an 
inductive  / magnetic sensor, which responds to  the differential of the field 
value = velocity. If instead you had an LVDT  position sensor, you would read 
linear distance = quake amplitude for the same  signal, but at a different time.

I think  of my pendulum as being set in motion by my direct coupled 
damper, direct  coupled sensor, supporting spring, hinge, and surrounding 
air.    The sequence of events is: (1)earth moves; (2)damper, sensor, 
spring,  hinge and air move; (3)pendulum responds to forces of 
acceleration from  damper, sensor, spring, hinge, and air.   Once moving, 
the  pendulum begins integrating all the instantaneous forces which will  
increase at different rates, and will not stop relative to earth until  
all the absorbed energy has been dissipated, which will not occur until  
some time period after the earth motion has stopped.   This time  delay 
is dramatic if one watches an in-car pendulum while stopping the  car.

There seems to be some logic in classifying a sensor by when peak  output 
is reached relative to pendulum velocity as measured against  earth.   A 
strain gage mounted on the side of a flexible  pendulum arm should be a 
pure acceleration measure.   A magnetic  system can be closely coupled an 
be an acceleration system, or could be  lightly coupled with very little 
effect on the velocity of the  pendulum.   A capacity measuring position 
indicator would have a  negligible force couple to the pendulum but not a 
zero force because a  conductor or dielectric will be pulled into the 
space between the fixed  plates proportional to the applied voltage. 
    Depends on the type of capacity detector. The  forces are usually small 
compared to mg.

With  negligible coupling, capacity sensors would be position  indicators.   
Direct coupled sensors appear to cover nearly the  entire range between 
sensing acceleration and sensing  displacement.
    They don't have to, but all the capacity detectors  that I know of sense 
position.

Subject:  Re: Digest from 01/21/2007 00:00:40
From:    Randall Peters  
Date:    Mon, 22 Jan 2007 09:47:19  -0500


Roger, I have respnded to your request for help;  i.e.,
"This brings up a very important point - what is the pendulum  sensor
reading?  Is it acceleration, velocity, or  displacement? I need help here. 

In practical terms,  this equivalence of inertial reference frames means that 
it is impossible to  detect uniform motion on the basis of measurements 
conducted inside a box,  such as a seismometer. Thus the only feature of motion 
having  any
importance whatsoever to a seismometer is acceleration of the case that  
supports its inertial mass M. It is very common to erroneously believe that  any 
type motion of the case will be met with displacement of M relative to the  
case, because of the inertia of M. 
    I am trying to think of an occasion when this is  not true? If the 
inertia is large compared to any force, the mass will be  'stationary' / moving with 
a fixed relative velocity.

Be sure  to understand that the only property of the motion that is 
``resisted'' by M  is the acceleration. Thus the acceleration is the only thing that 
can be  directly measured!!
OK

Velocity  and position, the other kinematic variables so frequently discussed 
in  seismology, can only be inferred from the acceleration measurement. 
Unlike  the
quintessential acceleration, they cannot be directly measured, even  though 
they are frequently specified.
OK

The output from a seismometer is directly  proportional to acceleration, as 
long as the acceleration takes place at a  frequency lower than the natural 
(eigen) frequency of the instrument, and  additionally, it is operating with 
damping that is near critical. 
--> Very long period signals > resonant period

When the  frequency of the drive is higher than the natural frequency of the 
instrument,  the response of the instrument is attenuated by the ratio of the 
square of the  drive frequency to the square of the eigenfrequency. If one is 
talking about  the ground displacement, as opposed to the acceleration, just 
the  opposite
behavior is found. 
--> NORMAL operation < resonant period.

For  those who want to believe that a seismometer responds directly to
ground  displacement, complete confusion results.
--> Not for this bunny.

It is also important to note that the horizontal  seismometer, such as a 
pendulum,
responds to more than one type of  acceleration. From ``inside the box'' of 
the instrument there is no way to  distinguish between these two forms of 
acceleration, which are (i) horizontal  acceleration of the instrument, and (ii) 
changes in orientation of the box  (tilt) relative to the direction of the local 
field of the earth g of the  earth, having the magnitude of 9.8 m/s2.
--> It can get confusing when you ALSO rotate the seismometer case, like  in 
a real quake, rather than just moving it laterally. 
 
    Regards,
 
    Chris Chapman





In a message dated 23/01/2007, rsparks@.......... writes:
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000=20
  size=3D2>Randle,   Thanks for the thought provoking response.
I=20
  can understand what you are saying about direct coupled sensing of pendulu=
ms,=20
  but I think I am correct in saying that your comments are limited to senso=
rs=20
  that actually extract some power from the pendulum motion.  A sensing=
=20
  system that (for instance) that counted area sweeping past a camera in a t=
ime=20
  period should not count as a direct coupled sensor.
Hi Roger,
 
    It should apply to any sensor, but if it exchan=
ges=20
energy there will be a perturbation.
    If it senses displacement you will sense=20
position.
    If it is a velocity sensor you will sense=20
velocity.
    If it is a force sensor you will sense=20
acceleration.
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>I am=20
  using a magnetic/coil sensing system with a heavy damper. The system=20=
is=20
  near or even more than critically damped. My sensing magnet and coil=20=
are=20
  not providing much load on the pendulum, most of the energy captured by th=
e=20
  pendulum is dissipated in the damper. I certainly agree that my sensing sy=
stem=20
  is really acceleration based as it does load and reduce the pendulum=20
  speed. The sensor also acts to accelerate the pendulum from a standin=
g=20
  start when a wave arrives. These are two reasons to describe the syst=
em=20
  as an acceleration sensor.  

Despite these acceleration events=
, I=20
  have described my sensor as a velocity device. I do this because maximum=20
  output of the sensor occurs when the pendulum is moving at maximum=20
  velocity. Or at least that is what I believe is happening.=20

    It IS a velocity sensor in this case. For short=
=20
periods, the mass is 'stationary' and the relative linear motion =3D th=
e=20
earth motion. You have an inductive  / magnetic sensor, which responds=20=
to=20
the differential of the field value =3D velocity. If instead you had an LVDT=
=20
position sensor, you would read linear distance =3D quake amplitude for the=20=
same=20
signal, but at a different time.
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>I think=20
  of my pendulum as being set in motion by my direct coupled 
damper, dir=
ect=20
  coupled sensor, supporting spring, hinge, and surrounding 
air. &n=
bsp;=20
  The sequence of events is: (1)earth moves; (2)damper, sensor, 
spring,=20
  hinge and air move; (3)pendulum responds to forces of 
acceleration fro=
m=20
  damper, sensor, spring, hinge, and air.   Once moving, 
the=20
  pendulum begins integrating all the instantaneous forces which will=20
  
increase at different rates, and will not stop relative to earth until=
=20
  
all the absorbed energy has been dissipated, which will not occur unti=
l=20
  
some time period after the earth motion has stopped.   This=20=
time=20
  delay 
is dramatic if one watches an in-car pendulum while stopping the=
=20
  car.

There seems to be some logic in classifying a sensor by when p=
eak=20
  output 
is reached relative to pendulum velocity as measured against=20
  earth.   A 
strain gage mounted on the side of a flexible=20
  pendulum arm should be a 
pure acceleration measure.   A magn=
etic=20
  system can be closely coupled an 
be an acceleration system, or could b=
e=20
  lightly coupled with very little 
effect on the velocity of the=20
  pendulum.   A capacity measuring position 
indicator would ha=
ve a=20
  negligible force couple to the pendulum but not a 
zero force because a=
=20
  conductor or dielectric will be pulled into the 
space between the fixe=
d=20
  plates proportional to the applied voltage. 
    Depends on the type of capacity detector. The=20
forces are usually small compared to mg.
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>With=20
  negligible coupling, capacity sensors would be position=20
  indicators.   
Direct coupled sensors appear to cover nearly=20=
the=20
  entire range between 
sensing acceleration and sensing=20
displacement.
    They don't have to, but all the capacity detect=
ors=20
that I know of sense position.
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>Subject:=20
  Re: Digest from 01/21/2007 00:00:40
From:    Randall Peters=20
  <PETERS_RD@..........>
Date:    Mon, 22 Jan 2007 09:47:=
19=20
  -0500


Roger, I have respnded to your request for help;=20
  i.e.,
"This brings up a very important point - what is the pendulum=20
  sensor
reading?  Is it acceleration, velocity, or=20
  displacement? I need help here. 

 In practical terms=
,=20
  this equivalence of inertial reference frames means that it is impossible=20=
to=20
  detect uniform motion on the basis of measurements conducted inside a box,=
=20
  such as a seismometer. Thus the only feature of motion having=20
  any
importance whatsoever to a seismometer is acceleration of the case=20=
that=20
  supports its inertial mass M. It is very common to erroneously believe tha=
t=20
  any type motion of the case will be met with displacement of M relative to=
 the=20
  case, because of the inertia of M. 
    I am trying to think of an occasion when this i=
s=20
not true? If the inertia is large compared to any force, the mass will be=20
'stationary' / moving with a fixed relative velocity.
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>Be sure=20
  to understand that the only property of the motion that is ``resisted'' by=
 M=20
  is the acceleration. Thus the acceleration is the only thing that can be=20
  directly measured!!
OK
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>Velocity=20
  and position, the other kinematic variables so frequently discussed in=20
  seismology, can only be inferred from the acceleration measurement. Unlike=
=20
  the
quintessential acceleration, they cannot be directly measured, even=
=20
  though they are frequently specified.
OK
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000=20
  size=3D2>    The output from a seismometer is directly=20
  proportional to acceleration, as long as the acceleration takes place at a=
=20
  frequency lower than the natural (eigen) frequency of the instrument, and=20
  additionally, it is operating with damping that is near critical.=20

--> Very long period signals > resonant period
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>When the=20
  frequency of the drive is higher than the natural frequency of the instrum=
ent,=20
  the response of the instrument is attenuated by the ratio of the square of=
 the=20
  drive frequency to the square of the eigenfrequency. If one is talking abo=
ut=20
  the ground displacement, as opposed to the acceleration, just the=20
  opposite
behavior is found. 
--> NORMAL operation < resonant period.
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000 size=
=3D2>For=20
  those who want to believe that a seismometer responds directly to
groun=
d=20
  displacement, complete confusion results.
--> Not for this bunny.
<=
FONT=20
  style=3D"BACKGROUND-COLOR: transparent" face=3DArial color=3D#000000=20
  size=3D2>    It is also important to note that the horizont=
al=20
  seismometer, such as a pendulum,
responds to more than one type of=20
  acceleration. From ``inside the box'' of the instrument there is no way to=
=20
  distinguish between these two forms of acceleration, which are (i) horizon=
tal=20
  acceleration of the instrument, and (ii) changes in orientation of the box=
=20
  (tilt) relative to the direction of the local field of the earth g of the=20
  earth, having the magnitude of 9.8 m/s2.
--> It can get confusing when you ALSO rotate the seismometer case,=20=
like=20
in a real quake, rather than just moving it laterally. 
 
    Regards,
 
    Chris Chapman