Here is an update on the fastening of the hinges for the STM-8
leaf spring vertical seis:

Epoxying the hinges to the frames is not reliable in the long run.

The seis here went bonk in the night and I assumed that it was simply
out of the capture range of the feedback because of the seasonal
temperature changes. But the remote zero did nothing, and investigation
showed that two of the flexure strips on one side of the hinge had 
become unglued.

In the initial ideas and design of the instrument, I thought that 
gluing the flexures would be a reliable simplification of the design,
and it has survived for several years even on the original prototype.
The epoxy surface preparation was limited to sanding with 400 grit paper.
But any failure in something that has to run unattended for many years is
not acceptable, so epoxying the hinge flexures is probably not a good
idea in the long run; it still might be if proper surface preparations 
are made (McMaster has pages of adhesive primers). And indeed, the 
flexures that remained glued had to be forcibly peeled from the frame 
members.  But one of the major VBB sensor companies has also had problems 
with adhesives peeling loose after several years that jammed the mass. 
So the ultimate long term reliability seems to favor clamped flexures, 
even though this involves many additional parts with the increased
contribution of micropositioning noise from thermal and mechanical stress.
It is possible that using a slow set acrylic adhesive or even Locktite
under the clamps might reduce these effects.

To replace the flexures on the STM vertical (as shown in the web site
drawing) I used the same 0.005" by 3/4" wide phosphor bronze strapping,
but with the full 3/4" width on each side of the 3/8" center section that 
is the actual hinge area (the epoxied flexures were simply straight 3/8"
wide strips). I used a standard paper punch to make 1/4" large clearance
holes for the 6-32 fastening screws; a good punch will make a clean,
minimally stressed cut, and is also used to shape the 3/8" wide center
flexing section. The aluminum brackets (1 1/2" x 1 1/2" x 1/8" thick 
angles 4 1/2" long as in the drawing) are drilled and tapped (with light
oil) for the 6-32 screws which are 3/8" and 1" back from the hinge corner.

The clamping bars are made from 1/2" wide by 1/16" thick brass bar 
(ACE hardware) strips cut 1 1/4" long, and filed to a shallow angle at 
the inside end. The 3/16" screw holes are 5/16" and 15/16" back from the
inside end.  After cutting and drilling, these are carefully deburred 
and rounded at the edges to provide a uniform clamping surface.

To assemble the hinge, the two brackets are clamped in the X shape 
using an L bracket in the center; the hinge corners are spaced 0.005" 
apart each way using the hinge stock under each side of the L bracket
and crossing through the center of the assembly (where the hinges themselves 
cross, the bracket corners are flattened by 1/8" so that the actual 
flexing area is untouched).  The 4 flexures, 8 clamp bars, 16 SS screws 
and lockwashers are all loosely assembled and squared up before being 
gradually tightened so as not to torque the flexures. The inside angled 
edge of the brass clamps is aligned with the outside edge of the flattened 
section of the bracket corner in the flexure area, with the slightly longer
side against the flexure strip.  The two flexures should not touch.

This should all be clarified by a drawing of the details. And a picture is
worth 10^3 words next time I play with the camera and the scanner. Actually,
I am inclined to redesign the whole seismometer to about 1/2 the size using
the soldered box flexures and the bought parts coil/magnet assembly I have 
previously described.

regards,
Sean-Thomas

____________________       ____________________
					\     /
					 -----
					 -----
____________________/     \____________________


__________________________________________________________

Public Seismic Network Mailing List (PSN-L)