Randall Pratt rpratt@.............
Sent: Fri, 3 Jun 2011 17:20
Subject: Seismograph pictures



Here are links to a couple pictures of the latest incarnation of my continu=
ing experiment in recycling.  http://mit.midco.net/rpratt/images/Vertical.J=
PG    I started with optical position sensing and moved to different coil a=
nd damping configurations.  The upper spring attachment is adjustable over =
about 4" in out up down. The attachment to the carbon fiber boom was also i=
nitially adjustable but is now in place for the current 2.68 second period.=
  I have experimented with periods of 1 to 4.5 seconds.  I cast the lead ma=
ss which has the vertical damping vane integral and a brass bolt threaded i=
nto it for fine balance adjustments. http://mit.midco.net/rpratt/images/Coi=
lDamp.JPG  Damping is adjustable by turning the damping magnets on threaded=
 attachment bolts to set the gap between the 2 sides.  I tried a knife edge=
 pivot and found it basically worthless so now have a flex pivot in compres=
sion.  The main frame is 3/16" by 1 1/2" steel angle brazed together.  The =
magnet has an extension to the center pole into the coil and the outer pole=
 face is removed to prevent shortcut of field below the coil. =20
=20
I like the steel angle construction for simplicity and quick assembly.  I l=
ike the damping system as the force lines are not offset and damping is eas=
ily tuned to a fine level and not easily bumped out of adjustment.  The dam=
ping system is also fixed to the frame so it does not pick up separate vibr=
ations.  The solenoid type coil set up is very sensitive.  The change from =
knife edge to flex is a must for anyone contemplating construction.  In the=
 future I may try the optical position system as a feedback loop but it did=
 not compare in sensitivity to the coil.=20
=20
 Hi Randall,
=20
    Congratulations on completing a successful project!=20
=20
    To eliminate any sensitivity to horizontal motion, the centre of gravit=
y of the arm needs to be at the same vertical position as the flex suspensi=
on. The arm needs to be ~horizontal.
=20
    You can also make a very satisfactory frame by BOLTING L section mild s=
teel together! For a Lehman I used a 12" cross bar, 1" x 1 3/8" L section, =
with two level adjust bolts at the ends. Two 30" horizontal L bars with a s=
ingle 5" cross bar on the end with a third level adjust bolt at it's centre=
, to hold the magnet blocks etc. Two 18" vertical L bars with a 4.75" top b=
ar and a 4.5" lower cross bar to support the boom bearing. Connect two 13" =
diagonal 1" x 1" L barsat 45 degrees between the upright and the horizontal=
 frames to cross brace them.
=20
    I agree with your finding that a real knife edge does NOT give a satisf=
actory bearing. Neither does the point in a cup arrangement. I wonder if so=
me guy years ago didn't realise that the "knife edge bearings" used in chem=
ical balances are in reality tiny cylinders?
=20
What material / dimensions did you use for the flexures under compression, =
please?=20
Have you tried an old fashioned twin edged safety razor blade? You clamp on=
e blade to the arm and the other to the upright. The two U loops at the end=
s of the blade do the flexing. They are about 3.7 thou" thick, 0.14" wide a=
nd 0.28" long. You can easily narrow the width using a small carbide cuttin=
g disk in a drill.=20
    Two other high load bearing types which DO work well are a plane rollin=
g on a sphere and crossed circular section rods. You can buy tungsten carbi=
de balls (www.smallparts.com) and solid triangular lathe tool flats, but a =
stainless steel scalpel blade rolling on a stainless steel ball bearing is =
much cheaper < $2! The SEPUK uses crossed tungsten carbide needle rollers S=
ee http://www.mindsetsonline.co.uk/images/Seismometer.pdf=20
=20
    The way you can increase the period of a vertical seismometer is to red=
uce the angle the spring makes with the horizontal arm and mount the top at=
tachment FURTHER  BEHIND the vertical pillar. This increases the tension in=
 the spring. Extension springs have a natural zero tension length. This may=
 be measured by measuring the tension in the spring at two lengths, plottin=
g them on a graph and measuring the zero intercept. If you run the system w=
ith a "zero length" spring, theoretically you get an infinite period, but w=
ith a steel spring the system will be unstable and it may collapse as the t=
emperature increases. An 8 second period is about the best that you can hop=
e for. Amateur constructors can use a period compensation amplifier to reco=
ver signals from up to 10x the natural period.
=20
    There are two problems with using ferrite magnets. The saturation field=
 is relatively low and it is also quite strongly temperature dependant. Thi=
s requires the use of rather large damping plates and sensor coils. Efficie=
nt dampers and sensors using NdFeB magnet quads are described at www.jclahr=
..com/science/psn/chapman The very sensitive rectangular coils use only abou=
t 2,000 turns! Using pairs of flat rectangular magnets removes the requirem=
ent for having a magnetic return path through the mounting frame.
=20
    You can use optical sensing quite easily, but the system needs to be co=
rrectly designed using large area photo diodes and a very stable light sour=
ce. Tungsten filament bulbs run at about 3/4 their rated voltage, are excel=
lent. You can use Infra Red LEDs with metal cases, but you do need to stabi=
lise their temperature, since the photo output is strongly temperature depe=
ndant.=20
    I have used BPW34 photo diodes for a +/-1 mm range and my photo amplifi=
er circuit is shown on the above website. The noise level was about 14 nano=
 metres. You need to run the photo diodes at about 50 to 100 micro amps. Th=
e practical limit is provided by the photon noise. For a total sample charg=
e of N electrons, the noise is SQRT(N).  The charge on a single electron is=
 1.602^-19 coulombs and 1 coulomb / second =3D 10 amps. You can also buy lo=
ng narrow photo diodes more suitable for a Lehman.
=20
    Have fun!
=20
    Regards,=20
=20
    Chris Chapman  =20











































































=
Randall Pratt rpratt@.............

Sent: Fri, 3 Jun 2011 17:20

Subject: Seismograph pictures











Here are links to=
 a couple pictures of the latest incarnation of my continuing experiment in=
 recycling.  http://mit.midco.net/rpratt/images/Vertical.JPG &nbs=
p;  I started with optical position sensing and moved to differen=
t coil and damping configurations.  The upper spring attachment is adj=
ustable over about 4" in out up down. The attachment to the carbon fiber bo=
om was also initially adjustable but is now in place for the current 2.68 s=
econd period.  I have experimented with periods of 1 to 4.5 seconds.&n=
bsp; I cast the lead mass which has the vertical damping vane integral and =
a brass bolt threaded into it for fine balance adjustments. http://mit.mid=
co.net/rpratt/images/CoilDamp.JPG  Damping is adjustable by turnin=
g the damping magnets on threaded attachment bolts to set the gap between t=
he 2 sides.  I tried a knife edge pivot and found it basically worthle=
ss so now have a flex pivot in compression.  The main frame is 3/16" b=
y 1 1/2" steel angle brazed together.  The magnet has an extension to =
the center pole into the coil and the outer pole face is removed to prevent=
 shortcut of field below the coil.  


 



I like the steel =
angle construction for simplicity and quick assembly.  I like the damp=
ing system as the force lines are not offset and damping is easily tuned to=
 a fine level and not easily bumped out of adjustment.  The damping sy=
stem is also fixed to the frame so it does not pick up separate vibrations.=
  The solenoid type coil set up is very sensitive.  The change fr=
om knife edge to flex is a must for anyone contemplating construction. =
; In the future I may try the optical position system as a feedback loop bu=
t it did not compare in sensitivity to the coil. 


 



&nb=
sp;Hi Randall,


 


    Congra=
tulations on completing a successful project! 


 


    To eli=
minate any sensitivity to horizontal motion, the centre of gravity of the a=
rm needs to be at the same vertical position as the flex suspension. The ar=
m needs to be ~horizontal.


 


    You ca=
n also make a very satisfactory frame by BOLTING L section mild steel =
together! For a Lehman I used a 12" cross bar, 1" x 1 3/8" L section, =
with two level adjust bolts at the ends. Two 30" horizontal L bars with a s=
ingle 5" cross bar on the end with a third level adjust bolt at it's centre=
, to hold the magnet blocks etc. Two 18" vertical L bars with a 4.75" top b=
ar and a 4.5" lower cross bar to support the boom bearing. C=
onnect two 13" diagonal 1" x 1" L barsat 45 degrees between the uprigh=
t and the horizontal frames to cross brace them.


 


    I=
 agree with your finding that a real knife edge does NOT give a satisfactor=
y bearing. Neither does the point in a cup arrangement. I wonder if some&nb=
sp;guy years ago didn't realise that the "knife edge bearings" used in chem=
ical balances are in reality tiny cylinders?


 


What material / dimension=
s did you use for the flexures under compression, please? 


Have you tried an old fas=
hioned twin edged safety razor blade? You clamp one blade to the arm and th=
e other to the upright. The two U loops at the ends of the blade do the fle=
xing. They are about 3.7 thou" thick, 0.14" wide and 0.28" long. You can ea=
sily narrow the width using a small carbide cutting disk in a drill. 


    Two ot=
her high load bearing types which DO work well are a plane rolling on =
a sphere and crossed circular section rods. You can buy tungsten carbide ba=
lls (www.smallparts.com) and sol=
id triangular lathe tool flats, but a stainless steel scalpel blade ro=
lling on a stainless steel ball bearing is much cheaper < $2! The SEPUK =
uses crossed tungsten carbide needle rollers See http://www.mindsetsonline.co.uk/im=
ages/Seismometer.pdf 


 


    The wa=
y you can increase the period of a vertical seismometer is to reduce the an=
gle the spring makes with the horizontal arm and mount the top attachment F=
URTHER  BEHIND the vertical pillar. This increases the tension in the =
spring. Extension springs have a natural zero tension length. This may be m=
easured by measuring the tension in the spring at two lengths, plotting the=
m on a graph and measuring the zero intercept. If you run the sys=
tem with a "zero length" spring, theoretically you get an infinite period, =
but with a steel spring the system will be unstable and it may collapse as =
the temperature increases. An 8 second period is about the best that you ca=
n hope for. Amateur constructors can use a period compensation amplifier to=
 recover signals from up to 10x the natural period.


 


    There =
are two problems with using ferrite magnets. The saturation field is relati=
vely low and it is also quite strongly temperature dependant. This req=
uires the use of rather large damping plates and sensor coils. Efficie=
nt dampers and sensors using NdFeB magnet quads are described at www.jclahr.com/science/psn/chap=
man The very sensitive rectangular coils use only about 2,000=
 turns! Using pairs of flat rectangular magnets removes the requirement for=
 having a magnetic return path through the mounting frame.


 


    You ca=
n use optical sensing quite easily, but the system needs to be correctly de=
signed using large area photo diodes and a very stable light source. T=
ungsten filament bulbs run at about 3/4 their rated voltage, are excellent.=
 You can use Infra Red LEDs with metal cases, but you do need to stabilise =
their temperature, since the photo output is strongly temperature depe=
ndant. 


    I have=
 used BPW34 photo diodes for a +/-1 mm range and my photo amplifier ci=
rcuit is shown on the above website. The noise level was about 14 nano metr=
es. You need to run the photo diodes at about 50 to 100 micro amps. The pra=
ctical limit is provided by the photon noise. For a total sample charge of =
N electrons, the noise is SQRT(N).  The charge on a single electron is=
 1.602^-19 coulombs and 1 coulomb / second =3D 10 amps. You can also buy lo=
ng narrow photo diodes more suitable for a Lehman.


 


    Have f=
un!


 


    Regards, 


 


    Chris =
Chapman