Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Low Frequency Vibration Analysis & Effects of AC Coupling

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

Connection Technology Center, Inc.

590 Fishers Station Drive


Victor, New York 14564
Toll Free: (800) 999-5290
Phone: (585) 924-5900
Fax: (585) 924-4680

Low Frequency Vibration Analysis & Effects of AC Coupling


Background:
There are some vibration applications
that require the analysis of frequencies
below 2 Hz or 120 CPM. Analyzing
these low frequencies requires proper
selection of the sensor and proper set-up
of the measurement input to avoid
possible errors in the amplitude of the
vibration.

Typically these accelerometers will offer


higher sensitivity outputs, i.e. 500 mV/g,
to improve the signal to noise ratio.
Examples from CTC are the AC135
Series and AC136 Series sensors as
shown in Figure #2.

Sensor Selection - Accelerometer:


If the application calls for a casing
measurement on the machine, as
illustrated in Figure #1 with a magnet
mount accelerometer, then the user must
select an accelerometer with the best
possible low frequency response.

Figure #2 AC135 & AC136 Series


Both of these sensors offer a sensitivity
of 500 mV/g +/- 10% and a low
frequency response of 0.2 Hz or 12 CPM
as
shown
in
the
performance
specifications of Figure #3.
Sensitivity +/-10%
Dynamic Range
Frequency Response +/- 3dB
Frequency Response +/- 10%

500 mV/g
+/- 10 g peak
0.23000 Hz
0.61500 Hz

Figure #3 Performance Specifications

Figure #1 Casing Measurement

Jack D. Peters

These IEPE (Integrated Electronics


Piezo Electric) accelerometers will have
a DC bias (operating) voltage and an AC
voltage output representing the vibration

Regional Sales Manager

level. The AC vibration voltage output


rides on top of the DC bias voltage as
shown in Figure #4.

Time Record 1
12.5
V

40 mV p-p
VAC
12.23
VDC

Real

Bias

12
0 s

79.96092 ms

Figure #4 - AC Vibration on DC Bias

If you were using a 500 mV/g


accelerometer to measure the signal in
Figure #4, the result would be
40 mV/(500 mV/g) or 0.08 gs p-p.

These non-contact displacement sensors


provide measurement of the shaft gap
and shaft vibration with a low frequency
response of 0 Hz or 0 CPM.
The shaft gap is the distance from the
face of the displacement sensor to the
outside diameter of the shaft, and it is a
negative DC voltage.
Shaft vibration is the movement of the
shaft around its centerline and is a
variable DC voltage that creates a
Dynamic AC Signal. This signal is very
similar to the AC vibration voltage. It
rides on top of the DC gap voltage and
can be treated in a similar AC method
for analysis. Figure #6 illustrates these
two displacement voltages.

Time Record 1
-9.75
V

30 mV p-p
VAC (Dynamic)

Sensor Selection Displacement:


If the application calls for a shaft
measurement on the machine, using a
displacement sensor as shown in Figure
#5 would provide the best result.

-10.00
VDC

Real

Gap

-10.25
0 s

79.96092 ms

Driver

Figure #6 - AC Dynamic on DC Gap

Typical non-contact
displacement sensor
for measuring shaft
vibration on a sleeve
or journal bearing.

If you were using a 200 mV/mil


displacement sensor to measure the
signal in Figure #6, the resulting shaft
vibration would be 30 mV/(200 mV/mil)
or 0.15 mils p-p. This is a very small
shaft vibration! However, the gap
measurement (distance from the face of
the displacement sensor to the outside
diameter of the shaft) would be equal to
10.00 V/(200 mV/mil) or 50 mils. This
is a typical gap setting for most
applications.

Cable

Probe

Eddy Currents
Shaft
Journal/Sleeve
Figure #5 Displacement Sensor

Jack D. Peters

Regional Sales Manager

AC Coupling:
The electrical term AC coupling means
that the AC signal is allowed to be
measured, and that the DC signal is
blocked from measurement. Normally
AC coupling is accomplished by using a
capacitor, and this has generated the
term de-coupling capacitor in our
industry. The capacitor de-couples the
AC signal from the DC signal, and
allows the measurement of the AC
vibration signal only. About 99% of the
time, this is exactly what the vibration
analyst wants to happen. The analyst is
only interested in the vibration signal to
determine the machine faults, and by
using the AC coupling feature, the DC
bias voltage or DC gap voltage is
removed from the measurement.

frequency signals will fool the capacitor,


and as the capacitor tries to block the DC
signal it will also attenuate the low
frequency vibration signal. In fact, the
de-coupling capacitor will become a
high pass filter as shown in Figure #7.
Therefore, the measured amplitude of
the low frequency vibration signal will
be much smaller than the actual value.

Freq Resp: Hyx 2:1


2

Amplitude attenuation
(roll off) begins at 2.5
Hz or 150 CPM

Magnitude

-3dB or -29.3% of the


amplitude occurs at
0.4 Hz or 24 CPM
0
0 Hz

Please keep in mind that the AC voltage


representing the vibration signal is much
smaller than the DC bias voltage or DC
gap voltage. By blocking these DC
voltages, the vibration data collector,
dynamic signal analyzer, oscilloscope, or
chart recorder can use their entire
dynamic amplitude range to measure and
amplify the vibration signal.
AC coupling is such a proven
measurement method in the vibration
industry that most vibration data
collectors will default to this condition
during power up.
Unfortunately, if the analysis is for a
very low frequency, less than 2 Hz or
120 CPM, AC coupling may attenuate
the amplitude of the measured signal.
The low frequency vibration will have a
very long wavelength, and the output
voltage of the sensor will simulate a DC
response as a result of the long
wavelength. Since the de-coupling
capacitor is used to separate the AC
vibration from the DC bias, these low
Jack D. Peters

5 Hz

Figure #7 Frequency Response


AC Coupling (0 5 Hz)

The frequency response illustrated in


Figure #7 is typical for AC coupling.
However, the method of measurement
you are using may be different, and
should be checked prior to making a low
frequency measurement. These results
will be translated in the FFT as shown in
Figure #8.

Auto Pwr Spec 2


0.02
V
rms

Amplitude attenuation
(roll off) in FFT

Magnitude

0
0 Hz

5 Hz

Figure #8 FFT AC Coupling

Regional Sales Manager

DC Coupling:
After reviewing Figures #7 and #8, it is
obvious that AC coupling will attenuate
the low frequency signal amplitude. If
this low frequency is the region of
interest and importance, than another
method of measurement input will be
required.
DC coupling, sometimes nicknamed
direct coupling, allows the measurement
of the DC and AC signals
simultaneously. This means that there is
no attenuation of the low frequency
amplitudes. Figure #9 illustrates the
frequency response of a DC coupled
signal.

Freq Resp: Hyx 2:1


2

No amplitude attenuation
or signal roll off. Good
response to zero Hz.

Magnitude

0
0 Hz

5 Hz

Figure #9 Frequency Response


DC Coupling (0 5 Hz)

Auto Pwr Spec 1

Zero Hz amplitude
as a result of DC
coupling the input

0.02
V
rms

Magnitude

0
0 Hz

5 Hz

Figure #10 FFT DC Coupling

The zero Hz amplitude could be reduced


if you have the capability of applying a
DC offset voltage. This offset voltage
will require special treatment through
hardware or firmware to mix a DC
signal of equal amplitude but of opposite
polarity with the sensor signal. This
offset voltage will negate the DC
signal voltage and balance the AC signal
around a zero DC amplitude value.
DC coupling with DC offset is a very
powerful combination for low frequency
measurements. The correct amplitude
will be measured, and the zero Hz bin in
the FFT will not be corrupted.

Caveat:
Similar good results can be observed in
the FFT. Figure #10 illustrates that the
response of the FFT has no amplitude
attenuation or roll-off as we observed
using AC coupling.
There is a large amplitude component in
the FFT at zero Hz. This first bin or line
of resolution in the FFT will contain
amplitude energy associated with the DC
signal value due to the DC coupling of
the input signal.
Jack D. Peters

Many data collectors offer special


programming to identify low frequency
modulation of high frequency resonance,
or stress wave analysis. These programs
identify a resonance associated with the
natural frequency of the accelerometer,
bearing, gear, or machine, and analyze
the modulating frequency disturbances
or stress waves that are produced during
impacting.
Results are not always
absolute signal amplitude, but are often

Regional Sales Manager

indicative of the fault or failure in the


machine based on the frequency. Some
of the popular names for these special
detection programs are:

Demodulation
Acceleration Envelope
gs Spike Energy TM
PeakVue TM
Shock Pulse Method TM
Envelope Demodulation
High Frequency Enveloping

Although these techniques offer an


excellent early warning of faults in
bearings and gears, they do not measure
the actual running speed or low order
harmonics of the machine.

Although AC coupling is a very


convenient method for general analysis
of vibration, it has low frequency
restrictions as a result of the de-coupling
capacitor. The de-coupling capacitor
will create a high pass filter and
attenuate very low frequency measured
amplitudes. Figure #11 illustrates the
results of AC coupling.
DC coupling will allow very low
frequency
measurements
without
degradation of the measured amplitude.
DC coupling does require sufficient
dynamic range, or offset capabilities, so
that the small AC variations can be
analyzed in the presence of large DC
signals. Figure #12 illustrates the results
of DC coupling.

Summary:
If the measurement and analysis has
frequencies of interest that are less than
2 Hz or 120 CPM, then careful
consideration needs to be given to sensor
selection and input coupling.

Freq Resp: Hyx 2:1

Freq Resp: Hyx 2:1


2

Magnitude

0 Hz

5 Hz

Auto Pwr Spec 1


0.02
V
rms
Magnitude

Magnitude
0
0 Hz

5 Hz

Auto Pwr Spec 2

0.02
V
rms

0 Hz

5 Hz

Figure #12 DC Coupled

Magnitude

0
0 Hz

5 Hz

Figure #11 AC Coupled

Jack D. Peters

Regional Sales Manager

You might also like