25.6 Transducer Selection

Back

A transducer is a device that senses a physical

quantity (vibration in this case, but it can also be

temperature, pressure, etc.) and converts it into an

electrical output signal, which is proportional to

the measured variable (see Chapter 15). As such,

the transducer is a vital link in the measurement

chain. Accurate analysis results depend on an

accurate electrical reproduction of the measured

parameters. If information is missed or distorted

during measurement, it cannot be recovered later.

Hence, the selection, placement, and proper use of

the correct transducer are important steps in the

implementation of a condition monitoring and

fault diagnostics program.

Considerable research and development work has gone into the design, testing, and calibration of

sensors (transducers) for a wide range of applications. The transducer must be:

* Correct for the task

* Properly mounted

* In good working order (properly calibrated)

* Fully understood in terms of operational characteristics

Transducers usually require amplification and conversion electronics to produce a useful output

signal. These circuits may be located within the sealed sensor unit or in a separate box. There are

advantages and disadvantages to both of these configurations but they will not be detailed here.

Traditional vibration sensors fall into three main classes:

* Noncontact displacement transducers (also known as proximity probes or eddy current probes)

* Velocity transducers (electro-mechanical, piezoelectric)

* Accelerometers (piezoelectric)

Force and frequency considerations dictate the type of measurements and applications that are best

suited for each transducer. Recently, laser-based noncontact velocity/displacement transducers have

become more commonplace. These are still relatively expensive because of their extreme sensitivity, and

hence are still predominantly used in the laboratory setting.

Figure 25.5 shows the relationship between the different transducer types in terms of response

amplitude and frequency. For constant velocity vibration amplitude across all frequencies, a

displacement transducer is more sensitive in the lower frequency range, while an accelerometer is

more sensitive at higher frequencies. While it may appear as if the velocity transducer is the best

compromise, transducers are selected to optimize sensitivity over the frequency range that is expected to

be recorded.

The type of motion sensed by displacement transducers is the relative motion between the point of

attachment and the observed surface. Velocity transducers and accelerometers measure the absolute

motion of the structure to which they are attached.

25.6.1 Noncontact Displacement Transducers

These types of sensors find application primarily in fluid film (journal) radial or thrust bearings. With

the rotor resting on a fluid film there is no way to easily attach a sensor. A noncontact approach is then

the best alternative. Noncontact measurements indicate shaft motion and position relative to the bearing.

Radial shaft displacements and seal clearances can be conveniently measured. Another advantage of using

0.1 1.0 10 100 1,000 10,000

0.1

1.0

10

Relative Amplitude Response

Frequency (Hz)

Velocity

Displacement

Acceleration

FIGURE 25.5 Frequency versus response amplitude

for various sensor types.

25-10 Vibration and Shock Handbook

© 2005 by Taylor & Francis Group, LLC

noncontact displacement probes is that when they are used in pairs, set 908 apart, the signals can be used

to show shaft dynamic motion (orbit) within the bearing. Figure 25.6 shows a single channel and dual

channel measurement result. When the two channels are plotted against one another, they clearly show

what are known as shaft orbits. These orbits define the dynamic motion of the shaft in the bearing, and

are valuable fault detection and diagnostic tools.

The linearity and sensitivity of the proximity probe depends on the target conductivity and porosity.

Calibration of the probe on the specific material in use is recommended. This type of sensor is capable of

both static and dynamic measurements, but temperature and pressure extremes will affect the transducer

output. The probe will detect small defects in the shaft (cracks or pits), and these may seem like

vibrations in the output signal.

Installation of these sensors requires a rigid mounting. Adaptors for quick removal and replacement

without machine disassembly can be useful. The minimum tip clearance from all adjacent surfaces

should be two times the tip diameter. Probe extensions must be checked to ensure that the resonant

frequency of the extension is not excited during data gathering. As with all sensors, care must be taken

when handling the cables and the connections must be kept clean.

25.6.2 Velocity Transducers

There are two general types of velocity transducers. They can be distinguished by considering

the mode of operation. The two types are electro-mechanical and piezoelectric crystal based.

Piezoelectric crystal-based transducers will be discussed in the next section, so the focus here will be

on electro-mechanical (see Chapter 15).

Electro-mechanical velocity transducers function with a permanent magnet (supported by springs)

moving within a coil of wire. As the sensor experiences changes in velocity, as when attached to a

vibrating surface, the movement of the magnet within the coil is proportional to force acting on the

(a)

(b)

Vibration Direction

Time

Vibration Amplitude

Output

X Direction Displacement

Vibration

Y Direction Displacement

Direction

Y Output

X Output

FIGURE 25.6 (a) Shaft displacement with one sensor; (b) shaft orbit with two sensors (x direction versus

y direction displacement — assumes shaft is circular).

Machine Condition Monitoring and Fault Diagnostics 25-11

© 2005 by Taylor & Francis Group, LLC

sensor. The current in the coil, induced by the

moving magnet, is proportional to velocity, which

in turn is proportional to the force. This type of

device is known as “self-generating” and produces

a low impedance signal; therefore, no additional

signal conditioning is generally needed.

Electro-mechanical velocity sensors require the

spring suspension system to be designed with a

relatively low natural frequency. These devices

have good sensitivity, typically above 10 Hz, but

their high-frequency response is limited (usually

around 1500 Hz) by the inertia of the system.

Some devices may obtain a portion, or all, of

their damping electrically. This type must be

loaded with resistance of a specific value to meet

design constraints. These are usually designed for use with a specific data collection instrument and

must be checked and modified if they are to be used with other instruments. Figure 25.7 shows a plot

of the sensitivity vs. frequency for an electro-mechanical velocity transducer.

While electro-mechanical velocity transducers can be designed to have good dynamic range within

a specific frequency range, there are several functional limitations. Because a damping fluid is

typically used to provide most of the damping, this type of transducer is limited to a relatively

narrow temperature band, below the boiling point of the damping liquid. The mechanical reliability

of these sensors is also limited by the moving parts within the transducer, which may become

worn or fail over time. This has resulted in this type of transducer being replaced by piezoelectric

sensors in machine condition monitoring applications. The orientation of the sensor is also limited

to only the vertical or horizontal direction, depending on the type of mounting used. Finally, as a

damped system, such as an electro-mechanical velocity transducer, approaches its natural frequency,

a shift in phase relationships may occur (below 50 Hz). This phase shift at low frequencies will affect

analysis work.

25.6.3 Acceleration Transducers

By far the most commonly used transducers

for measuring vibration are accelerometers (see

Chapter 15). These devices contain one or more

piezoelectric crystal elements (natural quartz or

man-made ceramics), which produce voltage when

stressed in tension, compression or shear. This is

the piezoelectric effect. The voltage generated

across the crystal pole faces is proportional to the

applied force.

Accelerometers have a linear response over a

wide frequency range (0.5 Hz to 20 kHz), with

specialty sensors linear up to 50 kHz. This wide

linear frequency range and the broad dynamic

amplitude range make accelerometers extremely

versatile sensors. Figure 25.8 shows the sensitivity vs. frequency relationship for a typical accelerometer.

In addition, the signal can be electronically integrated to give velocity and displacement measurements.

This type of transducer is relatively resistant to temperature changes, reliable (having no moving parts),

produces a self-generating output signal meaning no external power supply is needed unless there are

onboard electronics, is available in a variety of sizes, is usually relatively insensitive to nonaxial vibration

Amplitude Response

Natural Frequency

Frequency

≈10 Hz

FIGURE 25.7 Output sensitivity vs. frequency for an

electro-mechanical velocity transducer.

Amplitude Response

Forcing Frequency

Natural Frequency

0.1 0.2 0.4 0.6 0.8 1.0

Natural Frequency

3

FIGURE 25.8 Typical accelerometer sensitivity vs.

frequency.

25-12 Vibration and Shock Handbook

© 2005 by Taylor & Francis Group, LLC

(,3% of main axis sensitivity), and can function

well in any orientation. Signals from this type of

transducer contain significantly more vibration

components than other types. This means that

there is a large amount of information available in

the raw vibration signal.

Installation of accelerometers requires as rigid

a mount as possible. Permanent installation with

studs or bolts is usually best for high speed

machinery where high-frequency measurements

are required. The close coupling between the

machine and the sensor allows for direct

transmission of the vibration to the sensor.

Stud mounting requires a flat surface to give

the best amplitude linearity and frequency

response. This type of mounting is expensive

and may not be practical if large numbers of

measurements are being recorded with a portable

instrument. Magnetic mounts have the advantage

of being easily movable and provide good

repeatability in the lower frequency range, but

have limited high-frequency sensitivity (4 to

5 kHz). Hand-held measurements are useful

when conducting general vibration surveys, but

usually result in significant variation between

measurements. The hand-held mount is least

expensive but only offers frequency response

below 1 kHz. Figure 25.9 shows a plot of the

response curves for the same accelerometer with

different mounts. For machine condition monitoring and fault diagnostics applications, there will

typically be a combination of all three mounting methods used, depending on the equipment being

monitored and the monitoring strategy employed.

As with the other types of vibration sensor, accelerometers have certain limitations. Because of their

sensitivity and wide dynamic range, accelerometers are also sensitive to environmental input not related

to the vibration signal of interest. Temperature (ambient and fluctuations) may cause distortion in the

recorded signal. General purpose accelerometers are relatively insensitive to temperatures up to 2508C. At

higher temperatures, the piezoelectric material may depolarize and the sensitivity may be permanently

altered. Temperature transients also affect accelerometer output. Shear-type accelerometers have the

lowest temperature transient sensitivity. A heat sink or mica washer between the accelerometer and a hot

surface may help reduce the effects of temperature.

Piezoelectric crystals are sensitive to changes in humidity. Most accelerometers are epoxy bonded or

welded together to provide a humidity barrier. Moisture migration through cables and into connections

must be guarded against. Large electro-magnetic fields can also induce noise into cables that are not

double shielded.

If an accelerometer is mounted on a surface that is being strained (bent), the output will be altered.

This is known as base strain, and thick accelerometer bases will minimize this effect. Shear-type

accelerometers are less sensitive because the piezoelectric crystal is mounted to a center post not the base.

Accelerometers are designed to remain constant for long periods of time; however, they may need

calibrating if damaged by dropping or high temperatures. A known amplitude and frequency source

(or another accelerometer that has a known calibration) should be used to check the calibration

of accelerometers from time to time.

Frequency

Frequency ≈ 7kHz

Accelerometer Output

≈2kHz Frequency

Accelerometer Output Accelerometer Output

(a)

(b)

(c)

≈28kHz

FIGURE 25.9 Accelerometer response vs. frequency

for various types of mounts (a — stud; b — magnet; c —

hand held).

Machine Condition Monitoring and Fault Diagnostics 25-13

© 2005 by Taylor & Francis Group, LLC

Навигация