Regarding comments on hinge flexures: my 2cents:
My experience is that things aren't that difficult:

1: Cutting the material: I bought a roll of 0.005" phosphor-bronze
from McMaster, as well as stainless and spring-steel sheets. I
have had no problem cutting the thinner material, up to 0.010",
with an office type paper cutter. It makes very clean, burr-free,
and straight cuts. I do sharpen the cutter (flat file the edges) once
every few years. To make a single 1/4" hole in the thin material
(for clearance of the single clamping screw) I use a common paper 
punch. Most will make a clean hole.

2: Gluing the flexures: industrial-strength, like DEVCON brand, 
epoxies are stronger than the hinge material IF the surfaces are
properly cleaned and prepared. I "un-polish" the surfaces with
400-grit emery paper. The shear strength of these epoxies is
remarkable; so under normal conditions they work well, since the 
forces on crossed-hinge flexures are entirely in shear. 
Glued flexures are the ultimate in a noise free (no micropositioning)
design. The ones where I prepared the surfaces properly and haven' t 
abused have lasted many years and show no signs of coming apart. But the
abused ones, where the flexure actually gets bent from such clever
acts as dropping the seismometer, are a problem. The bent flexure starts
to come unglued at the hinge side. So I have installed a single screw
clamp thru a hole punched in the hinge strip to a tapped hole in
the aluminum angle to control any accidental damage. I use a 1/4",
6-32 pan head screw, a #10 washer, and a 1/4" washer with the side
toward the hinge filed straight. The straight edge is aligned exactly
over the filed back corner edge of the aluminum angle. (figure needed!)
(Excerpt from previous explanation:
When I prepare the aluminum angle brackets that the hinges are epoxied 
to, I file off the outside corners of the angles in the area where the
hinges go from one bracket to the other. I file the corner at a 45deg
angle, removing about 1/16 or more from each orthogonal side, so when 
the hinges are assembled, the free air distance is at least 0.125 inches. 
The actual pivot point is within the hinge thickness (0.005") of the 
corner of the angles.  This method of providing the hinge clearance has 
been common practice in many of the hinge designs I have seen in large 
mechanical seismometers over the years. )
When the flexure is visibly bent it has to be replaced.  It can usually
be pealed away from the aluminum angle and the epoxy sanded off.
This usually means re-doing all the flexures of a hinge assembly;
a clamping arrangement without gluing could avoid this.

3: I have had no problem in using extruded aluminum angle stock from
hardware stores or McMaster supply. THe tolerances are very good.

4: I have had no indication of the "oil can" distortion. Even after
cutting, the flexures are flat enough to mirror your face in. However,
once a flat piece has been bent or twisted, it is history, since the
original stress-free surface is lost. A handy source of thin brass 
sheet is the hobby materials section of your ACE hardware store.

5: Flat or crossed flexure hinges are generally not used in compression. 
In the case of a flat hinge, especially used as the lower hinge of a "garden
gate" horizontal, this can result in unstable and non-linear response,
since as the center of mass moves away from the centerline of the boom,
the hinged end moves in the opposite direction (at the scale of the 
ground motions we are interested in .... microns or less). Large
horizontal boom seismometers use a short taught wire for both upper and
lower hinges, where its dimension is minimal compared with boom length
and transducer clearance. More compact horizontals with close tolerance
transducers use crossed hinges for both, with the flexure elements under 
tension (even when using the Lucas-Bendix assemblies)..
In general, the spring constant of the hinges should be much less than
the gravitational restoring force that determines the instrument
characteristics. In fact, forces caused by the hinges are usually
ignored, and the period of the horizontal is calculated by:
(assuming that the restoring force by the hinges and/or pivot are minimal):

The natural period: Tn = 2*pi*sqrt(l/(g*sine i))
 
where l is the boom length in cm, g=980cm/sec^2, i is the angle
that the boom makes wrt the horizontal, measured in radians,
(where sine i = i). For example, a 40 cm boom hanging vertically (like the
SG design) as a simple pendulum ( an angle of 90 degrees) has a period of 
1.3 seconds. (a one second clock pendulum is 24.8 cm). When tilted
horizontally to about 4 degrees, the period is 5 seconds. At
about a 1 degree angle, it is 10 seconds, and at about 0.23 deg.
it is 20 seconds. So the period is changing with inverse of the
square root of the angle, which is why long periods are so unstable.
But less stable also means sensitive to smaller ground motions.
THen the tilt sensitivity of a horizontal  is a function of the
square of the operating period Tn.
		(where z is the displacement, and phi is the tilt)
		z = (g * Tn^2 / 4 * pi^2) * phi (horizontal)

6: Inverted pendulums have been experimented with since the Weichert
of 1903. Even with a mass of 80 kg 1 meter above the hinge, the operating
period was only 6 seconds. There was a "Reef" design that was about 1 
meter high but was only usefully stable to 1 second. Horizontal-boom
designs (as above) were preferred because their response does not
depend on the elastic behavior of the hinge.

Regards,
Sean-Thomas


_____________________________________________________________________

Public Seismic Network Mailing List (PSN-L)