I manufacture very accurate 1-axis inclinometers for marine use and now one
customer would like to add second axis to the inclinometer (not as accurate) a
gyroscope (2/3-axis) and an accelerometer (3-axis).
I plan to use the same inclinometer sensor I have used before (not from Analog
Devices) for the accurate axis (say X). Thus I'm looking for sensors that can
provide Y-axis inclinometer with rather good accuracy, very accurate 2/3 axis
(X,Y + optionally Z) gyroscope and very accurate 3-axis accelerometer.
The operating temperature will be 0-70 C. Offset will be calibrated after
installation and "long term" (a few minutes) drift of the gyros can
be compensated with the inclinometer. The inclining is limited to +- 5 degrees
in X and +-45 degrees in Y. The accelerations to all directions are +-2g, but
are typically well below 1 g (+ gravity in Z). Dynamic range is limited to
about +-45 deg/s.
The following specs are demanded:
Accuracy below is maximum error (at least 3 sigma), not RMS.
0.1 degree accuracy at +-5 degree + 0.1 degree offset drift over temperature
(0-70 C) + 0.1 degree long term (year) offset drift
1 mg (this is probably not possible, if it includes nonlinearity +sensibility +
Offset temperature (0-70C) drift 20 mg
Gyro (X,Y + optionally Z):
Absolute accuracy 1 deg/s
Offset temperature (0-70C) drift 0.5 deg/s
Bias Instability 1 deg/h (this may be impossible)
ARW 0.5 deg/sqrt(h)
Since you have a lot of products for this category, I would appreciate some
help reducing the options and also understanding how the specs are defined.
We have several products that are close to meeting your needs, the "absolute"
nature of your performance requirements makes me hesitate to offer anything
with firm conviction. With that in mind, here are some
suggestions to consider.
1) Our best performing accelerometers are from the ADXL203 family :
2) Our best inclinometer products are the ADIS16209 (www.analog.com/ADIS16209)
and ADIS16210 (www.analog.com/ADIS16210). These devices provide ~0.1 deg of
accuracy, over most conditions, but in the most extreme cases, parts can
approach 0.5 deg of end of life drift (0.1 deg would be ~1 sigma). Out of
curiosity, how do you plan to control your mechanical support system to better
than 10um (required flatness stability, on a 9mm package, for 0.1 deg). As you
know, without this, even a perfect sensor won’t meet these requirements.
3) Our best performing gyroscopes are from the ADXRS64x family. In particular,
the ADXRS646 (www.analog.com/ADXRS646) offers ~12 deg/hour of in-run bias
stability. 1 deg/hour, over all conditions, is pretty tough, even for more
expensive technologies. The ADXRS453 (www.analog.com/ADXRS453) provides ~15
deg/hour and a side-mount package, for simpler assembly.
4) We do offer high-performance IMUs. You might want to take a look at the
ADIS16445 (www.analog.com/ADIS16445) and ADIS16485 (www.analog.com/ADIS16485),