In a message dated 2008/09/02, gel@................. writes:

> You and others have mentioned temperature affects. What is it that we need 
> to worry about? Is it the spring constant or just simple 
> expansion/contraction that we need to worry about. Are there ways we can use that will nullify 
> the affects, maybe bimetal springs? 

Hi Gary,

       It is the Temperature Coefficient of Young's Modulus of Steel which is 
the problem. It is a whacking -2.4 x 10^-4 per C Degree. This large value 
makes it very difficult indeed to compensate for it's effects, when you need ppm 
accuracy.
       However alloys like Elinvar and Ni-SpanC were developed from the mid 
1930's onwards which have near zero temperature coefficients. So seismometer 
manufacturers just use these alloys. However, since they are relatively 
expensive, they are not widely available for DIY constructors.

       Regards,

       Chris Chapman.   
In a me=
ssage dated 2008/09/02, gel@................. writes:



You and others have mentioned t=
emperature affects. What is it that we need to worry about? Is it the spring=
 constant or just simple expansion/contraction that we need to worry about.=20=
Are there ways we can use that will nullify the affects, maybe bimetal sprin=
gs? 




Hi Gary,



       It is the Temperature Coefficient of Yo=
ung's Modulus of Steel which is the problem. It is a whacking -2.4 x 10^-4 p=
er C Degree. This large value makes it very difficult indeed to compensate f=
or it's effects, when you need ppm accuracy.

       However alloys like Elinvar and Ni-Span=
C were developed from the mid 1930's onwards which have near zero temperatur=
e coefficients. So seismometer manufacturers just use these alloys. However,=
 since they are relatively expensive, they are not widely available for DIY=20=
constructors.



       Regards,



       Chris Chapman.