From: Dave Nelson davefnelson@.......
Subject: Re: nonlinearities


All,
=20
The simple fact remains -- Feedback seismometers are the world standard.  N=
o other technology can come anywhere near their performance and operational=
 flexibility.
Their noise levels and bandwidth of the typical off the shelf instrument fr=
om several sources are such that the only area where significant improvemen=
t is desired relates to extremely long period performance for the study of =
whole earth modes at 1000 seconds and longer.
Most of the development activity is related to making smaller and less expe=
nsive instruments and greater flexibility in installation. One significant =
exception is Metrozet where a new instrument to replace the STS-1 is under =
development.
$$$$ Is that the only area where significant improvements are desired ? WHO=
 sell these LESS expensive instruments, please ? And less expensive than wh=
at ? A new car maybe ?
=20
Non feedback instruments are a relic of the past or short period geophone o=
r geophone-like instruments which have their niche in local event monitorin=
g and the amateur community.
$$$$ The UK is covered by an official network of Willmore vertical seismome=
ters. The sensor band is either variable 1 to 3 seconds or sometimes 20 sec=
onds to 20 Hz, depending on the Model. But we have over 400 horizontal scho=
ol seismometers sensing from 5 Hz to 25 seconds producing signals comparabl=
e to the "professional" seismometers. You simply require reasonably good de=
sign and construction. See http://www.mindsetsonline.co.uk/images/Seismomet=
er.pdf  Long period vertical seismometers are more difficult to make.
=20
 The optical open loop instrument is unlikely to have any success (in my op=
inion) simply because it will never be stable.  When I first read the paper=
 I concluded the project had no chance of becoming an operational instrumen=
t but it was an interesting a academic exercise. Others with the best crede=
ntials in the business have shared that sentiment.
$$$$ If you add "in it's present form", I would agree with you. But go back=
 a few centuries and "Others with the best credentials in the business" sai=
d that the world was flat !
=20
Regarding creep effects --   When a spring is first installed in an instrum=
ent there will be  "pops" related to what is probably dislocation effects i=
n the spring material . Their frequency will gradually reduce in time. The =
solution is too bake the assembly, with the spring at its operational stres=
s, at ~ 160 C for several hours. This will essentially eliminate the effect=
.
$$$$ I wish it were that simple. Heat treating the spring under tension doe=
s greatly reduce the noise, but it doesn't eliminate it. You also have nois=
e from the feedback circuit. Differentiating a signal is a noisy process. I=
t could be a distinct advantage to combine active damping with passive damp=
ing, which is quieter.
See also Wielandt's article quoted below.=20
=20
Dave Nelson=20
=20
 Dave,
    I took some data from one occasion when you had baked a spring showing =
the noise before and after. See:  http://bnordgren.org/seismo/YumaBefore%26=
After.pdf
Brett








Also to All:
      Non linearity (better called  mechanical complexity) is what ultimate=
ly limits, at low frequencies, the performance of every seismograph; but it=
 is not the kind of non linearity that feedback overcomes in remarkable fas=
hion, as is well known.  This kind is not accommodated  and is related to t=
he very reason materials creep =C2=97because of defect structures.  At the =
mesoscopic level, these defect structures cause the potential energy well t=
o be other than smooth.  In other words, the force required to accomplish f=
eedback is not able, at very low levels, to operate on an error signal that=
 is consistent with simple-minded theoretical expectations.   =20
       If the force feedback approach were as perfect as some want to belie=
ve, then there would have been no reason to hold the IRIS sponsored =C2=93b=
roadband conference=C2=94 several years ago, which I attended.   A poster s=
ession that I presented at that conference is online at http://www.iris.edu=
/stations/seisWorkshop04/PDF/tahoeI1.pdf
     Anyone with practical experience in materials science will recognize t=
hat internal friction of the spring in a seismometer has got to have conseq=
uences.  One of the first occasions for the reality of this fact to be note=
d was when Gunar Streckeisen measured the damping as a function of period o=
f a vertical instrument operating with a LaCoste zero-length spring.  What =
he found as a grad student doing this experiment =C2=96 was that the qualit=
y factor was not proportional to the frequency as required by a viscous dam=
ped simple harmonic oscillator theory.  Rather it was measured to be quadra=
tic in the frequency, which is described by the nonlinear damping theory th=
at I developed years ago.  If you want to Google =C2=91nonlinear damping=C2=
=92 and also =C2=91linear damping=C2=92 (without the tick marks of a litera=
l search) you will see that I have devoted many years of intense research t=
o this subject.  It is indeed complex, to the point of bewildering.   =20
      If you Google =C2=93optical seismograph ucsd=C2=94, you will find a p=
aper concerned with a latest generation instrument that is not of force fee=
dback type.  The author list includes Prof. Wielandt.  Much of the work pre=
sented in that paper was performed by PhD student J. Otero, http://www.proq=
uest.com No: AAT3368955.  I encourage folks to take a look at this article,=
 since it describes an instrument that could once and for all settle the de=
bate that has come now to Larry Cochrane=C2=92s list-serve.  =20
    By the way, five years ago Dr. Wielandt wrote a paper that you might al=
so want to look at; it is online at
http://www.docstoc.com/docs/48159410/Hysteresis-Creep-Internal-Friction-and=
-Damping-of-mechanical--dislocation
=20
    Randall
=20













































=
From: Dave Nelson davefnelson@.......

Subject: Re: nonlinearities









All,


 


The simple fact remains -- Feedback seismo=
meters are the world standard.  No other technology can come anywhere =
near their performance and operational flexibility.


Their noise levels and bandwidt=
h of the typical off the shelf instrument from several sources are suc=
h that the only area where significant improvement is desired relates =
to extremely long period performance for the study of whole earth modes at =
1000 seconds and longer.


Most of the development activity is relate=
d to making smaller and less expensive instruments and greater flexibility =
in installation. One significant exception is Metrozet where a new instrume=
nt to replace the STS-1 is under development.


$$$$ Is that the only area where signi=
ficant improvements are desired ? WHO sell these LESS expensive instruments=
, please ? And less expensive than what ? A new car maybe ?


 


Non feedback instruments are=
 a relic of the past or short period geophone or geophone-like instrume=
nts which have their niche in local event monitoring and the amat=
eur community.


$$$$ The UK is covered by an official&=
nbsp;network of Willmore vertical seismometers. The sensor band is either v=
ariable 1 to 3 seconds or sometimes 20 seconds to 20 Hz, depending on =
the Model. But we have over 400 horizontal school seismometers sensing=
 from 5 Hz to 25 seconds producing signals comparable to the "pro=
fessional" seismometers. You simply require reasonably good design and=
 construction. See http://www.mindsetsonline.=
co.uk/images/Seismometer.pdf  Long period ve=
rtical seismometers are more difficult to make.


 


 The optical open loop instrument is =
unlikely to have any success (in my opinion) simply because it will never b=
e stable.  When I first read the paper I concluded the project had no =
chance of becoming an operational instrument but it was an interesting a ac=
ademic exercise. Others with the best credentials in the business have shar=
ed that sentiment.


$$$$ If you add "in it's present form"=
, I would agree with you. But go back a few centuries and "Others with the =
best credentials in the business" said that the world was flat !



 


Regarding creep effects --   Whe=
n a spring is first installed in an instrument there will be  "pops" r=
elated to what is probably dislocation effects in the spring material =
.. Their frequency will gradually reduce in time. The solution is too b=
ake the assembly, with the spring at its operational stress, at ~ 160 C for=
 several hours. This will essentially eliminate the effect.


$$$$ I wish =
it were that simple. Heat treating the spring under tension does great=
ly reduce the noise, but it doesn't eliminate it. You also have noise from =
the feedback circuit. Differentiating a signal is a noisy process. It =
could be a distinct advantage to combine active damping with passive dampin=
g, which is quieter.


See also Wielandt's article quoted below. <=
/div>


 


Dave Nelson 


 


 Dave,

    I took some data from one occasion when you had baked a =
spring showing the noise before and after. See:  http://bnordgren.org/seismo/YumaBefore%26After.pdf

Brett









=
=








Also to All:


      Non linearity (better called  mechanical complex=
ity) is what ultimately limits, at low frequencies, the performanc=
e of every seismograph; but it is not th=
e kind of non linearity that feedback =
overcomes in remarkable fash=
ion, as is well known=
.  This kind is not accommodated  and is related to=
 the very reason ma=
terials creep =C2=
=97because of defect structu=
res.  At the mesoscopic level, these defect structures cause =
the potential energy well to be other than smooth.  In=
 other words, the force required to accomplish =
feedback is =
not able, at very low levels, to operate <=
FONT face=3D"Arial, Helvetica, sans-serif">on an error signal that is consistent with =
simple-minded theoretical ex=
pectations.    


      &=
nbsp;If the force feedback approach were as perfe=
ct as some want to believe, then there wo=
uld have been no reas=
on to hold the IRIS<=
/FONT> sponsored =C2=93<=
FONT face=3D"Arial, Helvetica, sans-serif">broadband conference=C2=94 several years ago, =
which I attended.   A poster session that<=
/FONT> I presented at that c=
onference is online at http://www.iri=
s.edu/stations/seisWorkshop04/PDF/tahoeI1.pdf<=
/div>


     Anyone with pr=
actical experience in materials science will recognize that internal friction of th=
e spring in a seismometer has got to have =
consequences.  One<=
/FONT> of the first occasion=
s for the reality of=
 this fact to be noted was when Gunar Streckeisen measured the damping as a function of period of a vertical instrument operating a LaCoste =
zero-length spring.  Wh=
at he found as a gra=
d student doing this =
experiment =C2=96 w=
as that the quality f=
actor was not proportional to the frequency as required by =
a viscous damped simple harm=
onic oscillator <=
FONT face=3D"Arial, Helvetica, sans-serif">theory.  Rather it was<=
/FONT> measured to be quadra=
tic in the frequency, which is described by nonlinear <=
FONT face=3D"Arial, Helvetica, sans-serif">damping theory that I develop=
ed years ago.  If want to Google =C2=91nonlinear damping<=
/FONT>=C2=92 and also =C2=91linear damping=C2=92 (wi=
thout the tick marks =
of a literal search) you will see that I have devoted many =
years of intense research to this subject.  =
It is indeed complex, the point of bew=
ildering.    


      <=
FONT face=3D"Arial, Helvetica, sans-serif">If you optical seismograph ucsd=C2=94, will find a paper concerned with=
 a latest generation instrument that is not of fo=
rce feedback type.  The author list includes Prof. Wielandt.  Much of <=
FONT face=3D"Arial, Helvetica, sans-serif">the work presented in =
that paper was performed by PhD student J. O=
tero, http://www.proquest.com No: =
AAT3368955.  I enco=
urage folks to take a=
 look at this article, it describes an instrument that could once and for =
all settle the debate=
 that has come now to Larry Cochrane=C2=92s list<=
/FONT>-serve.  &nbs=
p;


    By the way=
, five years ago Dr. =
Wielandt wrote a paper that<=
/FONT> you might also want to look at; it =
is online at



http<=
/FONT>://www.docstoc.com/docs/48159410/Hysteresis-Creep-Internal-<=
FONT face=3D"Arial, Helvetica, sans-serif">Friction-and-Damping-of-mechanical--dislocation



 


    Randall