PSN-L Email List Message

Subject: Re: Vertical design
From: ChrisAtUpw@.......
Date: Fri, 20 Dec 2002 23:03:32 EST


Hi Meredith,

       Short period verticals are not too difficult to make, but they are of 
somewhat limited use. There are a number of practical / engineering 
constraints. I note that at least one design uses the U Alnico magnet as the 
seismic mass on the end of the arm. This is not a good idea and it will react 
to environmental changes in the local magnetic field. In the normal home, you 
are likely to see a very noisy trace. It is quite easy to damp a seismometer 
using readily available NdBFe magnets. Fluid damping is very temperature 
sensitive and can be quite messy. The much higher field supplied by NdBFe 
magnets can also be used to increase the output of coil detector systems and 
improve their range and linearity. A survey of a range of systems can be 
found at http://quake.eas.gatech.edu/Instruments/InstrumentSurvey.htm. 
Prof.Braile's link may be found by removing the seismometer reference.
 
       To get much longer periods, there have been two successful approaches, 
one largely mechanical and the other involving quite a lot of electronics. 
The La Coste mechanical design dates from 1934 and developments on it rely on 
your being able to reduce the effects of temperature on the spring constants, 
on the linear expansion and on any slow 'creep' changes. Springs normally get 
'weaker' as the temperature increases, while the metal of the arm expands and 
the load 'moment' increases. There are no useful materials which shrink, or 
get stiffer, as they get hotter. You can design a temperature compensated 
arm, but getting adequate rigidity may be more difficult. They are used on 
pendulum clocks (eg a gridiron pendulum) to give good temperature 
compensation and high accuracy.
       To wind a zero length spring, you have to use metal which can be cold 
formed and this inevitably has high stresses 'built in'. Leaf springs can be 
made with a higher temper and lower initial stress, but it is desirable to 
use a NiSpan material to reduce the effects of temperature. There probably 
are sources of NiSpanC, but I do not know of any. Leaf springs have less 
problems with stray resonances than coil springs. A seismometer has to be 
highly stable over a range of temperature and over time, or it is not of much 
use. With a purely mechanical system, this is not too easy or cheap to do to 
the required precision. You only stand a reasonable chance of success if you 
design the system to be stable and can also adjust it. 'Cut and try' attempts 
are unlikely to have much success.

       Sean Morrissey took an alternative practical approach in choosing a 
fairly easily made design using a leaf spring (similar in principle to a La 
Coste) and damping it with small magnets. He provided a very sensitive 
distance transducer, a magnet / coil force feedback system and a box of 
electronics to control it all. The electronics damps the system and in doing 
so it is possible to control the frequency response; in doing this, it 
measures the force necessary to keep the mass fixed in relation to the ground 
and outputs any seismic signals; it also separates out the effects due to 
temperature changes and drift and puts these into the 'integrated' feedback, 
which you do not measure. Even so, Sean had to provide an additional 
mechanical weight adjustment with a small electric motor. 

       Unless you can put your vertical seismometer in a sealed pressure 
container, it will react to small naturally occurring changes in air 
pressure. These tend to 'float' the seismic mass and cover a wide band of 
frequencies. This is a major and serious source of noise in vertical sensors. 
It should be possible in principle to significantly reduce this noise by 
using a sealed float mounted on an extension to the arm beyond the hinge. 
This would make the apparatus appreciably longer and Sean was not 
enthusiastic over the idea. 

       Another possible approach is to use a 1 or 2 Hz geophone and fit it 
with a precision distance transducer and a set of electronics similar to 
Sean's. This is rather more like precision engineering, but the mass, the 
spring and the feedback coil are ready made. There are constructional 
adaptations other than the one Aaron Barzilai used, which could be more 
successful. See 
http://micromachine.stanford.edu/smssl/projects/Geophones/  He also used a 
rather bulky square wave excited capacitative distance transducer, which 
seemed to be rather noisy. An LVDT or a magnetic reluctance system might be 
more appropriate for this type of miniaturised application.

       Summing up, it is more difficult to make a vertical seismometer than 
to make a horizontal one due to difficulties in making the suspension, in 
balancing and in compensating the system. Short period vertical sensors are 
not too difficult to make, but are of limited use. Sean's design seems to 
offer the amateur the best chance of success for a broad band instrument, but 
a fair amount of constructional skill is required.
       
       Regards,

       Chris Chapman
Hi Meredith,


      Short period verticals are not too difficult to make, but they are of somewhat limited use. There are a number of practical / engineering constraints. I note that at least one design uses the U Alnico magnet as the seismic mass on the end of the arm. This is not a good idea and it will react to environmental changes in the local magnetic field. In the normal home, you are likely to see a very noisy trace. It is quite easy to damp a seismometer using readily available NdBFe magnets. Fluid damping is very temperature sensitive and can be quite messy. The much higher field supplied by NdBFe magnets can also be used to increase the output of coil detector systems and improve their range and linearity. A survey of a range of systems can be found at http://quake.eas.gatech.edu/Instruments/InstrumentSurvey.htm. Prof.Braile's link may be found by removing the seismometer reference.

      To get much longer periods, there have been two successful approaches, one largely mechanical and the other involving quite a lot of electronics. The La Coste mechanical design dates from 1934 and developments on it rely on your being able to reduce the effects of temperature on the spring constants, on the linear expansion and on any slow 'creep' changes. Springs normally get 'weaker' as the temperature increases, while the metal of the arm expands and the load 'moment' increases. There are no useful materials which shrink, or get stiffer, as they get hotter. You can design a temperature compensated arm, but getting adequate rigidity may be more difficult. They are used on pendulum clocks (eg a gridiron pendulum) to give good temperature compensation and high accuracy.
      To wind a zero length spring, you have to use metal which can be cold formed and this inevitably has high stresses 'built in'. Leaf springs can be made with a higher temper and lower initial stress, but it is desirable to use a NiSpan material to reduce the effects of temperature. There probably are sources of NiSpanC, but I do not know of any. Leaf springs have less problems with stray resonances than coil springs. A seismometer has to be highly stable over a range of temperature and over time, or it is not of much use. With a purely mechanical system, this is not too easy or cheap to do to the required precision. You only stand a reasonable chance of success if you design the system to be stable and can also adjust it. 'Cut and try' attempts are unlikely to have much success.

      Sean Morrissey took an alternative practical approach in choosing a fairly easily made design using a leaf spring (similar in principle to a La Coste) and damping it with small magnets. He provided a very sensitive distance transducer, a magnet / coil force feedback system and a box of electronics to control it all. The electronics damps the system and in doing so it is possible to control the frequency response; in doing this, it measures the force necessary to keep the mass fixed in relation to the ground and outputs any seismic signals; it also separates out the effects due to temperature changes and drift and puts these into the 'integrated' feedback, which you do not measure. Even so, Sean had to provide an additional mechanical weight adjustment with a small electric motor.

      Unless you can put your vertical seismometer in a sealed pressure container, it will react to small naturally occurring changes in air pressure. These tend to 'float' the seismic mass and cover a wide band of frequencies. This is a major and serious source of noise in vertical sensors. It should be possible in principle to significantly reduce this noise by using a sealed float mounted on an extension to the arm beyond the hinge. This would make the apparatus appreciably longer and Sean was not enthusiastic over the idea.

      Another possible approach is to use a 1 or 2 Hz geophone and fit it with a precision distance transducer and a set of electronics similar to Sean's. This is rather more like precision engineering, but the mass, the spring and the feedback coil are ready made. There are constructional adaptations other than the one Aaron Barzilai used, which could be more successful. See
http://micromachine.stanford.edu/smssl/projects/Geophones/  He also used a rather bulky square wave excited capacitative distance transducer, which seemed to be rather noisy. An LVDT or a magnetic reluctance system might be more appropriate for this type of miniaturised application.

      Summing up, it is more difficult to make a vertical seismometer than to make a horizontal one due to difficulties in making the suspension, in balancing and in compensating the system. Short period vertical sensors are not too difficult to make, but are of limited use. Sean's design seems to offer the amateur the best chance of success for a broad band instrument, but a fair amount of constructional skill is required.
      
      Regards,

      Chris Chapman

[ Top ] [ Back ] [ Home Page ]