From: "tchannel" tchannel@..............

Date: Mon, 4 Jun 2007 18:34:27 -0600

Ben and Chris, This is great information, Thanks, for your response. ----- Original Message -----=20 From: ChrisAtUpw@.......... To: psn-l@................. Sent: Monday, June 04, 2007 6:25 PM Subject: Re: Springs for Verticals In a message dated 2007/06/04, tchannel@.............. writes: Hi Everyone, I wanted to ask some questions about choosing a spring = for a vertical spring sensor. Pictures three situations, three different springs being pulled = straight down by a mass. No triangular modifications, just straight = down. 1 One is a strong spring like a screen door spring, pulled down by a = large mass. 2 One is a weak spring being pulled down by very little mass. 3 One is a rubber band being pulled down by a small mass. Just for comparisons, say they all had the same period of 1.5 = seconds. Is one of the three better for recording earthquakes? Two = questions here: Is there an advantage in using a weak spring, or strong = spring, if the resulting period is the same. =20 You will only get this sort of period with quite large extensions. = You need a certain mass for the motion not to be swamped by thermal or = environmental agitation.=20 It is the product M x T x Q, which is important for noise = considerations, not just the mass.=20 For a simple spring, the extension E =3D g x T^2 / (2 x Pi)^2, = where T is the period. Thus to get a period of say 10 sec, you need an = extension of ~25 metres.... And Secondly is a rubber band spring every used? No. Rubber is highly temperature sensitive and it also has a = lot of loss (hysteresis) associated with the extension. What prompted the question is seeing how small the spring in a = geophones is. Is the idea to have the smallest spring which would move = under the smaller stimulus. The coil springs used for seismometers have very specific = characteristics. They are designed to have a zero or negative length = when extended. The wire is twisted as it is wound and this makes the = coils clamp tightly together. If you plot the length versus the load, = the length stays constant up to a certain load and you then get an = extension proportional to the load. A plot of the load / length graph = can be extended back to zero load and the 'zero load extension' can be = negative. You need a net zero length to get an 'infinite period' on a = vertical seismometer. The changes in metal properties with temperature give a = practical limit of about 5 seconds for ordinary steel springs and you = may only have ~a 5 C Deg temperature range before re-balancing becomes = necessary. The 'easy' way around this is to extend the period = electronically, but the noise performance then becomes critical. A = maximum period extension of about x10 max is practicable. It is not very = difficult to extend a 2.5 second period sensor to 25 seconds. The use of = Sm-Co or NdFeB magnets has enabled the output to be considerably = increased over Alnico magnet systems, reducing problems with noise. Regards, Chris ChapmanBen and Chris, This = is great=20 information, Thanks, for your response.----- Original Message -----From:=20 ChrisAtUpw@.......Sent:Monday, June 04, 2007 = 6:25 PMSubject:Re: Springs for = VerticalsIn a=20 message dated 2007/06/04, tchannel@..............=20 writes:

Hi Everyone, I wanted to ask some questions about = choosing a=20 spring for a vertical spring sensor.

= Pictures three=20 situations, three different springs being pulled straight down by a = mass. No=20 triangular modifications, just straight down.

1 One is a = strong spring like=20 a screen door spring, pulled down by a large mass.

2 One is a weak = spring being=20 pulled down by very little mass.

3 One is a = rubber band being=20 pulled down by a small mass.

Just for = comparisons, say they=20 all had the same period of 1.5 seconds. Is one of the three = better for=20 recording earthquakes? Two questions here: Is there an = advantage in=20 using a weak spring, or strong spring, if the resulting period is = the=20 same.

You will only = get this=20 sort of period with quite large extensions. You need a certain mass = for the=20 motion not to be swamped by thermal or environmental agitation.=20

It is the product M x T x Q, which is = important for=20 noise considerations, not just the mass.

= For a=20 simple spring, the extension E =3D g x T^2 / (2 x Pi)^2, where T is = the period.=20 Thus to get a period of say 10 sec, you need an extension of ~25=20 metres....

And=20 Secondly is a rubber band spring every=20 used?

No. Rubber is highly = temperature sensitive and it also has a lot of loss (hysteresis) = associated=20 with the extension.

What prompted the question is seeing how small the = spring in a=20 geophones is. Is the idea to have the smallest spring which would = move under=20 the smaller stimulus.

The=20 coil springs used for seismometers have very specific characteristics. = They=20 are designed to have a zero or negative length when extended. The wire = is=20 twisted as it is wound and this makes the coils clamp tightly = together. If you=20 plot the length versus the load, the length stays constant up to a = certain=20 load and you then get an extension proportional to the load. A plot of = the=20 load / length graph can be extended back to zero load and the 'zero = load=20 extension' can be negative. You need a net zero length to get an = 'infinite=20 period' on a vertical = seismometer.

=20 The changes in metal properties with temperature give a practical = limit of=20 about 5 seconds for ordinary steel springs and you may only have ~a 5 = C Deg=20 temperature range before re-balancing becomes necessary. The 'easy' = way around=20 this is to extend the period electronically, but the noise performance = then=20 becomes critical. A maximum period extension of about x10 max is = practicable.=20 It is not very difficult to extend a 2.5 second period sensor to 25 = seconds.=20 The use of Sm-Co or NdFeB magnets has enabled the output to be = considerably=20 increased over Alnico magnet systems, reducing problems with=20 noise.

=20 Regards,

Chris=20 Chapman

=20