## PSN-L Email List Message

Subject: Re: Spring Constant and Temperature
From: Barry Lotz barry_lotz@.............
Date: Tue, 16 Dec 2008 06:38:45 -0800 (PST)

```Chris
I added a bymetal spiral spring from a dial thermometer to my Sean Morresse=
y vertical to help compensate for the temperature. It is mounted off to the=
side and rests on the boom in a fishing pole configuration. I move it up a=
nd down the boom until I get the proper compensation
Barry

--- On Fri, 12/12/08, ChrisAtUpw@.......  wrote:
From: ChrisAtUpw@.......
Subject: Re: Spring Constant and Temperature
To: psn-l@..............
Date: Friday, December 12, 2008, 7:20 PM

=20

In a message dated 13/12/2008, gel@................. writes:
The=20
spring constant changes with temperature for the typical materials used f=
or=20
springs. For instance, for carbon steel, the modulus of elasticity is 29.=
5E6=20
psi at 70 degrees F and then changes to 28.8E6 psi at 200 degrees F. This=
=20

means that the spring constant changes with temperature. Is this a proble=
m=20
that I should be concerned about. My vertical seismometer is located in t=
he=20
wine cellar and so the temperature does not vary during the day, but the=
=20
temperature may change 10 degrees F with the seasons.=20

Hi Gary,
=A0
=A0=A0=A0=A0This normally limits the maximum period that you=20
can set a steel spring vertical to about 6 seconds. The Tc of the modulus i=
s=20
about =A0-2.4x10^-4 / C Deg, which is large. If you try to make a vertical=
=20
with a very long natural period, it will collapse as the room temperature=
=20
rises.
=A0
=A0=A0=A0=A0The 'work around' for a simple amateur system is to=20
make a vertical sensor with a 1.5 to 2 second period and then fit a=20
1/f^2=A0low frequency boost amplifier=A0to recover the long period=20
signals. You can get at least a factor of x10 this way before you run into =
noise=20
problems, more if you use quad NdFeB magnet + coil sensor. I run a sensor o=
f=20
this general type and it picks up the Rayleigh waves very nicely.=20
=A0
=A0=A0=A0=A0You can get much greater period extensions if you=20
use a position sensor and force feedback - see the SG type=A0pendulum=20
seismometers on psn. You don't have to use the RF oscillator + tuned sensor=
=20
circuit. Large area Si photodiodes, 10 to 20 sq mm, and a tungsten filament=
lamp=20
will work OK. Small photodiodes and photo transistors both have too high no=
ise=20
levels.=20
=A0
=A0=A0=A0=A0The professionals now=A0tend to use N-SpanC=20
springs and force feedback systems to get periods to 360 seconds. The old=
=20
Sprengnether verticals used an Elinvar spring and got=A0stable operation to=
=20
30 seconds.=A0
=A0
=A0=A0=A0=A0Sean Morrissey developed a force feedback vertical=20
with a steel leaf spring which was OK for=A0about=A0+/-10 deg.=A0It is=20
all linked on psn, if you look!
=A0
=A0=A0=A0=A0Regards,
=A0
ChrisI added a bymetal spiral spring from a dial thermometer to my Sean Morressey vertical to help compensate for the temperature. It is mounted off to the side and rests on the boom in a fishing pole configuration. I move it up and down the boom until I get the proper compensationBarry--- On Fri, 12/12/08, ChrisAtUpw@....... <ChrisAtUpw@.......> wrote:From: ChrisAtUpw@....... <ChrisAtUpw@.......>Subject: Re: Spring Constant and TemperatureTo: psn-l@..............Date: Friday, December 12, 2008, 7:20 PM

In a message dated 13/12/2008, gel@................. writes:
The
spring constant changes with temperature for the typical materials used for
springs. For instance, for carbon steel, the modulus of elasticity is 29.5E6
psi at 70 degrees F and then changes to 28.8E6 psi at 200 degrees F. This

means that the spring constant changes with temperature. Is this a problem
that I should be concerned about. My vertical seismometer is located in the
wine cellar and so the temperature does not vary during the day, but the
temperature may change 10 degrees F with the seasons.

Hi Gary,

This normally limits the maximum period that you
can set a steel spring vertical to about 6 seconds. The Tc of the modulus is
about  -2.4x10^-4 / C Deg, which is large. If you try to make a vertical
with a very long natural period, it will collapse as the room temperature
rises.

The 'work around' for a simple amateur system is to
make a vertical sensor with a 1.5 to 2 second period and then fit a
1/f^2 low frequency boost amplifier to recover the long period
signals. You can get at least a factor of x10 this way before you run into noise
problems, more if you use quad NdFeB magnet + coil sensor. I run a sensor of
this general type and it picks up the Rayleigh waves very nicely.

You can get much greater period extensions if you
use a position sensor and force feedback - see the SG type pendulum
seismometers on psn. You don't have to use the RF oscillator + tuned sensor
circuit. Large area Si photodiodes, 10 to 20 sq mm, and a tungsten filament lamp
will work OK. Small photodiodes and photo transistors both have too high noise
levels.

The professionals now tend to use N-SpanC
springs and force feedback systems to get periods to 360 seconds. The old
Sprengnether verticals used an Elinvar spring and got stable operation to
30 seconds.

Sean Morrissey developed a force feedback vertical
with a steel leaf spring which was OK for about +/-10 deg. It is
all linked on psn, if you look!

Regards,

[ Top ]
[ Back ]