In a message dated 19/04/2009, rog@.......... writes:
I have a question regarding the lower pivot. I read the material you
provided previously. It seems a SS plate on the horizontal arm working against a
SS ball bearing on the vertical is a good choice. Is my understanding
correct?
Hi Rob,
That is quite correct. One system that I use has this. I usually work
at 20 seconds to minimise tilt drifts and optimise surface wave pickup, but
I can set it up for a stable 60 second swing. The SEPUK1 which has crossed
WC needle rollers can also be set up for periods much greater than 20
seconds.
I am planning a crossed carbide drill shafts for the upper pivot.
Mount the balls or sloping vertical needle rollers on the upright
stand and the plates or horizontal rollers on the arm and the diagonal
suspension. This minimises angular contact errors and lessens significantly the
chance of the bearings slipping in operation.
You can also use a flexing wire in tension for either or both
suspensions. See _http://jclahr.com/science/psn/gate.html_
(http://jclahr.com/science/psn/gate.html)
You could use crossed shanks for both bearings? This may be the
simplest solution, but not the cheapest.
Or use SS ball bearings for both? Drill out a larger SS bolt to take
say a 5/16" bearing and turn / file off the top threads to leave a short
thin tube with a V bottom. File off the top of the tube at the angle of the
top suspension and stick a ball to the V bottom. Mount the bolt vertically on
an L bracket fitted to the upright. Bend a U of brass strip and solder a
flat brass plate across the end. Stick a rectangle cut from a scalpel blade
to the inside of the flat plate with two part acrylic glue. Epoxy is too
brittle for this job. Two part polyurethane should also be OK. Neither of
these set 'brittle hard'. This U fitting fits over the vertically mounted SS
bearing.
Consider an Al base 30" long. The expansion of Al is 23 ppm / C deg.
So a 10 C change of temperature results in a 30x10x23^-6 = 7 thou" change in
length! This will make the support legs to slide over their mounting
plates and will move any play / slop in the threads. Commercial seismometers may
use fine thread support fittings over 1/2" long, which are slotted
parallel to the axis. The slot is compressed with a bolt to clamp the outer thread
to the threaded shaft after the position has been set.
Alternatively, you can use a wavy washer and a lock nut on top of the
frame to clamp the thread in place. I use a 3" x 1" x 1/8" Al U channel
frame fitted with 6" corner plates. I drill 1/2" OD Al alloy rod 3/4" long
with a 6 mm drill and turn off the ends square. I drill a 6 mm hole in the
frame. I fit a 6 mm bolt + locknut through the rod, put acrylic adhesive on the
free end of the rod and on a spare nut, put the bolt through the frame,
thread on the nut + adhesive and clamp the fitting firmly with the locknut
until the adhesive is set. Lightly coat the bolt thread with Vaseline and
then wipe it to prevent adhesion.
The ground will likely expand by a smaller amount and more slowly, so
the three mountings are trying to slide about most of the time!!! I tip my
6 mm SS mounting bolts with 5 mm SS bearings and provide 2" square x 1/8"
flat SS support plates stuck to the concrete floor with pool cement. I don't
have any significant tilt drift problems. You are likely to see tilt
drifts over time if you don't use a hard flat stable base / flat ground plates
and don't clamp the threads.
You need to make the arm quite rigid. Amateur designs seem often to
have no damping / mechanical constraint to counter rotational motion of the
seismometer arm about it's long axis. If the centre of the damping force is
either above or below the line joining the centre of mass to the lower
bearing, any quake signal will try to rotate the arm as well as move it side to
side. Since there is no or low damping on this motion, the movement can be
quite large and this may show up as a large background peak at a few Hz.
This may be opposed by fitting a rigid top suspension tube similar to the
arm, or by using a V cable suspension from the top of the support column to a
6" to 8" long crossbar bolted to the arm near the mass. I use 7 core nylon
coated SS 30 lb fishing trace for this. The cable ends are loops fitted in
round V grooves on the crossbar and are crimped with the tubes provided.
This seems to work well and it is not expensive. The top fitting is a 1.5"
OD SS mudguard washer with a V groove turned / filed around it's edge. This
is stuck onto the top bearing frame / wire clamp.
You should choose a site with the minimum daily temperature variations
around the seismometer. It MUST be shielded from direct sunlight. You can
buy LCD digital thermometers which run off a 1.5V battery, indicate to 0.1
C Deg and record minimum and maximum readings. Covering the floor with
insulating material around the seismometer can also help. You definitely do
need an airtight shielding case to prevent drafts and to minimise any air
convection, preferably made out of Celotex sheet foam or similar. A large
folded cover sheet of bubble wrap with the bubbles inside can help.
You can also get reverse convection noise late at night / about dawn
if the case temperature falls below the ground temperature. Rolls of air
peel off the inside of the case, fall to the ground and push the arm about.
This can be cured by mounting a small heater ~10W inside the top of the
seismometer case to keep a +ve air temperature gradient inside at all times.
Have a look at the damping and sensor designs at
_http://jclahr.com/science/psn/chapman/index.html_
(http://jclahr.com/science/psn/chapman/index.html)
I hope that these comments may be of help. The Lehman designs on psn
use out of date / poor techniques which need to be completely revised. I
advise against trying to use oil damping; point or knife blade bearings;
separated suspension, damping and sensor modules; Alnico U magnet and coil
sensors; magnets and ferromagnetic components on the arm and single wire
suspensions. These are ALL best avoided!
Regards,
Chris Chapman
In a message dated 19/04/2009, rog@.......... writes:
I have a
question regarding the lower pivot. I read the material you provide=
d
previously. It seems a SS plate on the horizontal arm working against a=
SS
ball bearing on the vertical is a good choice. Is my understanding=
correct?
Hi Rob,
That is quite correct. One system that I use=
has
this. I usually work at 20 seconds to minimise tilt drifts and optimise su=
rface
wave pickup, but I can set it up for a stable 60 second swing. The SEPUK1=
which
has crossed WC needle rollers can also be set up for periods much greater=
than
20 seconds.
I am
planning a crossed carbide drill shafts for the upper pivot.
Mount the balls or sloping vertical needle ro=
llers
on the upright stand and the plates or horizontal rollers on the arm and=
the
diagonal suspension. This minimises angular contact errors and lessens
significantly the chance of the bearings slipping in operation.
You can also use a flexing wire in tension fo=
r
either or both suspensions. See http://jclahr.com/science=
/psn/gate.html
You could use crossed shanks for both bearing=
s?
This may be the simplest solution, but not the cheapest.
Or use SS ball bearings for both? Drill out=
a
larger SS bolt to take say a 5/16" bearing and turn / file off the top thr=
eads
to leave a short thin tube with a V bottom. File off the top of the tube=
at the
angle of the top suspension and stick a ball to the V bottom. Mount the bo=
lt
vertically on an L bracket fitted to the upright. Bend a U of br=
ass
strip and solder a flat brass plate across the end. Stick a rectangle =
;cut
from a scalpel blade to the inside of the flat plate with two part acrylic=
glue.
Epoxy is too brittle for this job. Two part polyurethane should also be OK=
..
Neither of these set 'brittle hard'. This U fitting fits over the vertical=
ly
mounted SS bearing.
Consider an Al base 30" long. The expans=
ion of
Al is 23 ppm / C deg. So a 10 C change of temperature results in=
a
30x10x23^-6 =3D 7 thou" change in length! This will make the support=
legs to
slide over their mounting plates and will move any play / slop in the thre=
ads.
Commercial seismometers may use fine thread support fittings over 1/2=
"
long, which are slotted parallel to the axis. The slot is compressed with=
a bolt
to clamp the outer thread to the threaded shaft after the position has bee=
n set.
Alternatively, you can use a wavy washer and=
a lock
nut on top of the frame to clamp the thread in place. I use a 3" x 1" x 1/=
8" Al
U channel frame fitted with 6" corner plates. I drill 1/2" OD Al alloy rod=
3/4"
long with a 6 mm drill and turn off the ends square. I drill a 6 mm hole=
in the
frame. I fit a 6 mm bolt + locknut through the rod, put acrylic adhesive=
on the
free end of the rod and on a spare nut, put the bolt through the frame, th=
read
on the nut + adhesive and clamp the fitting firmly with the locknut=
until
the adhesive is set. Lightly coat the bolt thread with Vaseline and then=
wipe
it to prevent adhesion.
The ground will likely expand by a
smaller amount and more slowly, so the three mountings are tryin=
g to
slide about most of the time!!! I tip my 6 mm SS mounting bolts with 5 mm=
SS
bearings and provide 2" square x 1/8" flat SS support plates stuck to
the concrete floor with pool cement. I don't have any significan=
t tilt
drift problems. You are likely to see tilt drifts over time if you don't=
use a
hard flat stable base / flat ground plates and don't clamp the
threads.
You need to make the arm quite rigid. Amateur=
designs seem often to have no damping / mechanical constraint to counter=
rotational motion of the seismometer arm about it's long axis. If the=
centre of the damping force is either above or below the line joining the=
centre
of mass to the lower bearing, any quake signal will try to rotate the arm=
as
well as move it side to side. Since there is no or low damping on this mot=
ion,
the movement can be quite large and this may show up as a large backg=
round
peak at a few Hz. This may be opposed by fitting a rigid top suspension tu=
be
similar to the arm, or by using a V cable suspension from the top of the=
support
column to a 6" to 8" long crossbar bolted to the arm near the ma=
ss. I
use 7 core nylon coated SS 30 lb fishing trace for this. The cable ends ar=
e
loops fitted in round V grooves on the crossbar and are crimped with=
the
tubes provided. This seems to work well and it is not expensive. The top=
fitting
is a 1.5" OD SS mudguard washer with a V groove turned / filed around it's=
edge.
This is stuck onto the top bearing frame / wire clamp.
You should choose a site with the minimum dai=
ly
temperature variations around the seismometer. It MUST be shielded from di=
rect
sunlight. You can buy LCD digital thermometers which run off a 1.5V batter=
y,
indicate to 0.1 C Deg and record minimum and maximum readings. Covering&nb=
sp;the
floor with insulating material around the seismometer can also help. You=
definitely do need an airtight shielding case to prevent drafts and=
to
minimise any air convection, preferably made out of Celotex sheet foam or=
similar. A large folded cover sheet of bubble wrap with the bubbles=
inside
can help.
You can also get reverse convection noise lat=
e at
night / about dawn if the case temperature falls below the ground temperat=
ure.
Rolls of air peel off the inside of the case, fall to the ground and push=
the
arm about. This can be cured by mounting a small heater ~10W inside the to=
p of
the seismometer case to keep a +ve air temperature gradient inside at all=
times.
Have a look at the damping and sensor designs=
at http://jclahr.co=
m/science/psn/chapman/index.html
I hope that these comments may be of help. Th=
e
Lehman designs on psn use out of date / poor techniques which need to=
be
completely revised. I advise against trying to use oil damping;=
point
or knife blade bearings; separated suspension, damping and sensor modules;=
Alnico U magnet and coil sensors; magnets and ferromagnetic components on=
the
arm and single wire suspensions. These are ALL best avoided!
Regards,
Chris Chapman