You should contact the Fredericks Company, or
Spectron Glass and Electronics. Both of these
make electolytic vial-based tilt sensors, with
resolutions that can be below a tenth of a
percent of slope.

These devices consist of a curved glass tube,
with an electrode at each end, and a center
electrode. A bubble in the electrolyte moves
back and forth based on slope (the traditional
bubble level, electrified) and the center
electrode is used to measure the bubble position.

There are some constraints; these devices will
fail if exposed to DC currents (that means even
an ohmmeter!) so they must be completely capacitor-
coupled. The usual method is to drive the ends
from a differential output amplifier, then use
a precision bridge to rectify the center electrode
signal, amplify the resulting DC and meaure
the amplitude to derive slope. Note that a
conventional diode bridge won't work; too much
deadband in the middle.

Non-linearities can occur because of the drive
electronics, or because the glass vial isn't *exactly*
a uniform curve. In the application I used these on,
we bought vials to a specified degree of precision.
We could get others for less money, but they weren't
as precise. Nowadays, I'd buy the cheap ones and
linearize them with a table in ROM....but back then,
that was still a bit exotic.

Precision slope measurement is a surprisingly tricky
problem, and the choice of sensing technology is
really dependent on what precision you need and
what the intended environment is.


Mark G. Forbes, R & D Engineer  |  Acres Gaming, Inc.    (541) 766-2515
KC7LZD                          |  815 NW 9th Street     (541) 753-7524 fax
forbesm@peak.org                |  Corvallis, OR 97330
http://www.peak.org/~forbesm
mforbes@acresgaming.com

"There has been an alarming increase in the number of things I know nothing
about."
---Anomalous