DEPARTMENTS

ADRE* Tips

ADRE 408* DSPi Signal Processing

AC vs. DC Mode
One of the most basic concepts
Rusty Lane in vibration monitoring is the
Bently Nevada Product Engineer importance of understanding the
rusty.lane@ge.com
signal processing settings that
are being applied to the measured
signal during sample collection.

This tip is based on an inquiry that we Background

When configuring data collection settings for a Dynamic
received from a customer. The ADRE user
(waveform) Channel in the ADRE 408, it is important
was wondering why the signal waveforms to take into account the signal conditioning features,
they were seeing were different for the same and understand the impact that they will have on the
signal processing that will be performed. The coupling
very low-frequency vibration signal that
and sampling modes are shown in Figure 1 as they
was being sampled in both the AC and DC are displayed in the ADRE SXp software interface.
Modes. The cause ended up being that these
Coupling
two modes have some subtle differences Observe that there are two available signal coupling
in signal processing, which the user hadn’t options in the ADRE 408 unit – ac and dc. These
correspond to classic analog oscilloscope display settings,
realized. This is a great example of a case
where ac-coupling samples only the dynamic component
where readings differ for some reason that of the signal, relative to a line that represents the average
isn’t initially obvious (References 1 & 2). value of the waveform. Similarly, dc-coupling samples the
signal with the dynamic component as well as its actual
dc offset value, as displaced from the 0 volt reference.

Figure 1: Dynamic Channel Configuration dialog in ADRE SXp software. The red rectangle indicates where coupling and sampling mode
settings are selected.

 O R B I T Vo l.34 • N o.1 • J an. 2014

 DEPARTMENTS
ADRE* tips

Sampling
Coupling Mode Range
Associated with the Coupling setting is the Sampling
AC High 1.6 Hz to 50 kHz
Mode, which can be selected for either “High” or “Low.” The
high sampling mode is optimized for collecting samples AC Low N/A
with high frequency content, while the low sampling mode DC High 1 Hz to 50 kHz
is more appropriate for collecting low-frequency signals.
DC Low 0.167 Hz to 20 kHz
The selected sampling mode determines whether a high
or low frequency bandpass filter is applied to the signal. Figure 2: Two sampling modes (High and Low) are supported at
the channel level. Observe that the low frequency bandpass filter is
Note: A low bandpass filter has a longer settling time
only available in the DC mode. These filter frequency specifications
than a high bandpass filter. The table in Figure 2 shows are ±1 % of Full Scale Range and ± 0.011V below 1 V pp.
the frequency corners for the different modes.

Figure 3: Signal processing paths for the DC mode are shown in the upper half of this drawing. The AC mode paths are shown in the
lower half of the drawing.

J a n. 2 014 • N o. 1 • Vo l . 3 4 ORB I T 
DEPARTMENTS
ADRE* Tips

Figure 4: Tabular List Plot shows values after signal processing was performed on the identical input signal by three Dynamic Sampler
channels with slightly different configuration settings.

Note: AC Coupling removes the DC component within between filter settings that were processing the
the Dynamic Signal path. DC coupling is required signal in different ways as summarized here:
to obtain accurate data at very low shaft speeds • The AC High Mode attenuated everything
(for example, 1 rpm), while AC-Coupling allows below 1.6 Hz (96 rpm). With an input at
maximum data resolution due to channel gain. 0.05 Hz (3 rpm) every reading is attenuated
From an instrumentation standpoint, Figure 3 shows (synchronous, asynchronous, 1X and Direct).
the different electrical paths that produce the • The DC High Mode only attenuated the
Synchronous, 1X and 2X filtered values, Asynchronous Direct values below 1 Hz (60 rpm).
waveforms and “Direct” (broadband) values:
• The DC Low Mode is similar to the DC High
Laboratory Verification Mode but the attenuation for the Direct
We reproduced the customer’s settings in the lab using a value starts at 0.167 Hz (10 rpm).
function generator programmed to produce a sine wave Figure 5 shows the same electrical paths as Figure 3, with
signal at 2 Vpp @ 0.05 Hz (the synchronous frequency the effects of the signal processing on the input signal.
corresponding to a slow-speed machine turning at 3 rpm).
Figures 6 through 8 are the actual waveform plots, which
This signal was fed into three different Dynamic show the effects of different amounts of amplitude
Sampler channels configured as shown in Figure attenuation and phase distortion from the low frequency
1. Observe that in the Tabular List plot (Figure 4), all corners of the bandpass filters in the signal path.
three of the channels are displaying different readings
Observe that both of the samples that were originally
for Direct, and 1X Amplitude and Phase values.
collected in DC Coupled mode are being displayed here in
Question: What caused the observed AC Coupled mode for more consistent plotting comparison
discrepancies in these measured values? with the sample that was collected in AC Coupled mode.
Answer: These apparent inconsistencies were The DC Coupled samples are still intact with their dc
caused by the differences in the low frequency offset values in the database, but the SXp software
cutoff frequencies for each of the channels. simply provides the ability to display these samples in
an ac-coupled format for convenience when needed.
All three of the channels were working exactly
as designed, and it was simply the differences

 O R B I T Vo l.34 • N o.1 • J an. 2014

 DEPARTMENTS
ADRE* tips

When configuring data collection settings for a

Dynamic (waveform) Channel in the ADRE 408, it is
important to take into account the signal conditioning
features, and understand the impact that they will
have on the signal processing that will be performed.

Figure 5: This drawing shows the same signal paths as

Figure 2 with an added test signal passing through the
processing steps.

AC High Mode Results

Figure 6: AC High Mode plots showing synchronous waveform (left) and asynchronous waveform (right).

J a n. 2 014 • N o. 1 • Vo l . 3 4 ORBI T 
DEPARTMENTS
ADRE* Tips

DC High Mode Results

Figure 7: DC High Mode plots showing synchronous waveform (left) and asynchronous waveform (right).

DC Low Mode Results

Figure 8: DC Low Mode plots showing synchronous waveform (left) and asynchronous waveform (right).

Note also that the difference in displayed time (20 seconds go “back-to-basics” and take a close look at what the data
for the synchronous waveform and 2 seconds for the collection settings actually are, and what they mean.
asynchronous waveform) results from the difference
in sample rates and frequency span associated with References
the waveform data collection settings. One complete 1. Orbit Vol.25 No.2, 2005. Sabin, Steve, “Understanding
sinusoidal vibration cycle is visible in the synchronous Discrepancies in Vibration Amplitude Readings
waveforms, since 20 seconds corresponds to one Between Different Instruments, Part 1 of 2.”
complete shaft rotation at 3 rpm. However, only one tenth 2. Orbit Vol.26 No.1, 2006. Chitwood, Randy; Hamad,
of the full cycle is visible in the asynchronous waveforms, Akram; Qureshi, Munir; Sabin, Steve, “Understanding
since the total time of the waveform record is shorter. Discrepancies in Vibration Amplitude Readings
Between Different Instruments, Part 2 of 2.”
Conclusion
One of the most basic concepts in vibration monitoring is
the importance of understanding the signal processing *Denotes a trademark of Bently Nevada, Inc., a wholly
settings that are being applied to the measured signal owned subsidiary of General Electric Company.

during sample collection. This case was a great example Copyright © 2013 General Electric Company. All rights reserved.

of how readings may appear to be anomalous, until you

 O R B I T Vo l.34 • N o.1 • J an. 2014

 
Copyright 2014 Baker Hughes, a GE company, LLC ("BHGE") All rights reserved.
Bently Nevada, Orbit Logo, ADRE, Keyphasor, Promimitor, Velomitor and System 1 are registered
trademarks of BHGE in the United States and other countries. All product and company names are
trademarks of their respective holders. Use of the trademarks does not imply any affiliation with or
endorsement by the respective holders.
The information contained in this document is subject to change without prior notice.
1631 Bently Parkway South, Minden, Nevada USA 89423
Phone: 1.775.782.3611 Bently.com

ORBIT 2014 Q1