Measurement Computing WavePort rev.3.0 - Manual (Page 132)

A-2 04-06-00 Using Accelerometers
Transverse Sensitivity
An accelerometer is designed to have one major axis of sensitivity, usually perpendicular to the base and
co-linear with its major cylindrical axis. The output caused by the motion perpendicular to the sensing axis
is called transverse sensitivity. This value varies with angle and frequency and typically is less than 5% of
the basic sensitivity.
Base-Strain Sensitivity
An accelerometer’s base-strain sensitivity is the output caused by a deformation of the base, due to bending
in the mounting structure. In measurements on large structures with low natural frequencies, significant
bending may occur. Units with low base-strain sensitivity should be selected. Inserting a washer (smaller
in diameter than the accelerometer base) under the base reduces contact surface area; and can substantially
reduce the effects of base-strain. Note that this technique lowers the usable upper frequency range.
Acoustic Sensitivity
High-level acoustic noise can induce outputs unrelated to vibration input. In general, the effect diminishes
as the accelerometer mass increases. Use of a light, foam-rubber boot may reduce this effect.
Frequency Response
An accelerometer’s frequency response is the ratio of the sensitivity measured at frequency (f) to the basic
sensitivity measured at 100 Hz. This response is usually obtained at a constant acceleration level, typically
1 g or 10 g. Convention defines the usable range of an accelerometer as the frequency band in which the
sensitivity remains within 5% of the basic sensitivity. Measurements can be made outside these limits if
corrections are applied. Care should be taken at higher frequencies because mounting conditions greatly
affect the frequency range (see Mounting Effects, in upcoming text).
Dynamic Range
The dynamic measurement range is the ratio of the maximum signal (for a given distortion level) to the
minimum detectable signal (for a given signal-to-noise ratio). The dynamic range is determined by several
factors such as sensitivity, bias voltage level, power supply voltage, and noise floor.
Bias Level
Under normal operation, a bias voltage appears from the output signal lead to ground. There are two basic
MOSFET configurations commonly used. One exhibits a 7-8 V bias and the second a 9-12 V bias.
Operation of the two circuits is identical except for the available signal swing. The low-voltage version
typically exhibits 5-10 µVrms versus 10-20 µVrms for the high voltage.
Thermal Shock - Temperature Transients
Piezoelectric accelerometers exhibit a transient output that is a function of a temperature’s “rate-of-change.”
This “thermal shock” is usually expressed in g/°C and is related to:
• Non-uniform mechanical stresses set up in the accelerometer structure.
• A pyroelectric effect in piezoelectric materials—an electrical charge is produced by the
temperature gradient across the crystal.
This quasi-static effect produces a low-frequency voltage input to the MOSFET amplifier. This voltage is
usually well below the low-frequency corner, but the effect can reduce the peak clipping level and cause
loss of data. This effect does not affect the accelerometer’s basic sensitivity or the data unless the thermal
shift in the operation bias level results in clipping. Where drastic thermal shifts are expected, use 12 V bias
models. The effect’s severity is related to the mass of the accelerometer. In 100 mV/g industrial units, the
effect is usually negligible. Using rubber thermal boots can reduce the effect significantly.
Overload Recovery
Recovery time from clipping due to over-ranging is typically less than 1 ms. Recoveries from quasi-static
overloads that generate high DC bias shifts are controlled by the accelerometer input RC time constant that
is fixed during manufacture.