CONDITION MONITORING OF A CYCLOID GEARBOX

Eric Bechhoefer
GPMS Inc.
40 Ridge Rd
Cornwall, VT 05753
[email protected]

Abstract: A cycloid drive for gearboxes allows for high reduction ratio and zero or very
low backlash. The cycloid gear design is based on compression, whereas most gear
interactions are based on shear. Further, the contract of a cycloid gear is typically subject
to rolling forces vs. sliding, which are seen in traditional gearboxes. These features of a
cycloid gearbox allow for high shock load capacity, high torsional stiffness, and quiet
operation. This paper details the modeling required for correct configuration to perform
analysis on the cycloid gearbox and then is demonstrated on a 51:1 ratio, run to failure
test. This paper documents the sensitivity of standard condition indicators for
gear/bearing during the run to failure test.

Keywords: Cycloid; Smart Sensor; vibration diagnostics; Resonance; Spectral

Estimation; Model Bases Dynamics;

Introduction: There are many non-standard, condition monitoring application which use
relatively unusual gearbox design. While most reduction (example: epicyclical)
gearboxes have well-understood dynamics, other, such as a Variator (continuously
variable transmission) or cycloidal gearbox, have not been reported on. This paper covers
the dynamics, configuration and some test observation of work done on a cycloidal
gearbox.
The interest in cycloidal gearbox is derived in that they are used in many applications
were a low-cost drive motors is needed. For example, many conveyer belt systems
(sorting, moving bulk media, slew drives) use cycloid drives. When the drive itself is a
relatively low-cost asset, the process, which they support, can have significant economic
impacts if they fail. For example, one of the larger courier delivery services has a
distribution center with 3000 cycloid drives to move packages. The loss of a drive unit
halts sorting, which could impact revenues of up to $200,000 per day.
The cycloid drive for gearboxes allow for high reduction ratio and zero or very low
backlash. The cycloid gear design is based on compression vs. shear forces, where the
contact is typically rolling vs. sliding. These features allow for high shock load capacity,
high torsional stiffness, and quiet operation. Single stage ratios more than 200:1 are
possible.
The gearbox chosen for the test was an integrated induction motor and gearbox. This
gearbox is rated for 0.75 kW, approximately 1 Hp drive. For 60 Hz operations, using a
four-pole motor, the drive unit has a 100% input shaft rate of approximately 1795 rpm.
The gearbox a 51:1 reduction ratio. Figure 1 shows the typical arrangement of a cycloid
gearbox.

1
Figure 1. Typical Cycloid Gears (copied from
http://mechanicaldesign.asmedigitalcollection.asme.org/data/journals/jmdedb/27951/08
5001_1_1.jpeg)

The main components of the gearbox are: 1) inputs shaft, 2) support bearing, 3) eccentric
bearings, 4) the cycloid gears, 5) the pin teeth-case, 6) the pins, 7) output rollers, 8)
output shaft, 9) support bearing.

The ratio for the gearbox is given as:

( nteeth−1 ) × n pins
ratio= (eq 1)
nteeth−n pins
The test gearbox has dual disc (4) with 26 teeth and 51 pins.
Equations of Motion and Configuration: Configuration is driven by the equations of
motions for the monitoring components. This consists of describing synchronous motion
analysis of the shafts and gears, and the asynchronous motion of the bearings.
Bearing Motion: The simple input/output gearbox design uses three bearings on the
input shaft: bearing D (the eccentric bearing) and input shaft bearing C. The output shaft
was supported by two bearings: bearing B and bearing A.
The bearing rate fault frequencies are determined by the shaft rate, and:
 the number of rolling elements (b),
 the roller element diameter (d),
 the bearing pitch diameters (e), and
 the bearing contact angle ().
The fault features are related to damage accumulated on the bearing itself.
There are typically six fault features calculated for the bearing associated with bearing
elements: cage, ball, inner race, outer race. For mechanical looseness, the bearing may
also generate signatures associated with whip/whorl (in the base spectrum) or a
1/revolution impact (tick) in the heterodyne analysis. The bearing feature rates are
calculated as::

2
 cage=0.5 ( 1−d /e∗cos ( α ) ⁡) (eq 2)
ball=e / d ( 1−( d /e )2∗cos ( α )2 ⁡) (eq 3)
inner race=b/2 ( 1+ ( d /e )∗cos ( α ) ⁡) (eq 4)
outer race=b/2 ( 1−( d /e )∗cos ( α ) ⁡) (eq 5)
Because the outer rate of the eccentric bearing is in contact with the cycloid gear and the
input shaft, the total rate seen by the bearing is the input shaft + output shaft. The
eccentric bearing analysis was assigned to the input shaft. To capture the effect of the
relative motion, the bearing rates were multiplied by 1 + 1/51 = 1.0196. This speed
correction is needed in order to determine the correct bearing rate fault features.
Shaft/Gear Motion: Shaft and gear analyses are based on the time synchronous average,
which requires an accurate rate from the tachometer. The tachometer is used to resample
the vibration data and correct for any changes in shaft rate. Gear analysis, and more
importantly gear mesh tones, is a function of the shaft rate and the number of teeth on the
gear. A conventional input shaft with a 29.23 Hz rate with 26 teeth would have a gear
mesh frequency of 29.23 x 26 = 759.96 Hz. However, the cycloid gear has a relative
motion to the shaft, driven by the eccentric gear and the output shaft. The relative motion
of the cycloid to the ring gear means that, for each revolution, there is one extra gear
mesh. Hence, the actual gear mesh frequency would be 789.19 Hz. Figure 2 shows the
corresponding TSA order of 27 (26 teeth + 1) and the accompanying raw spectrum (in
Hz).
For this reason, gear analysis is based on 27 teeth, not 26 teeth.

Figure 2, Comparison of the Raw to TSA Spectrum for the input shaft.
The ring gear analysis would usually be associated with the number of ring gears teeth,
but because there are two cycloid gears (of 26 teeth), the measured mesh is 51 x 2 or 102

3
mesh. The TSA spectrum and raw spectrum then show frequencies at 29.23/51 x 102 =
58.46 Hz. Due to the modulation of two disc, there are sidebands at 102 +/-51 = 51 and
153, or 29.22 and 87.69 Hz, as seen in Figure 3.

Figure 3, Comparison of the Raw to TSA Spectrum for the output shaft.
Window Selection for the Envelope Analysis: The envelope analysis is based on
demodulation of high-frequency resonance, which is based on impact. Poor selection of a
window results in poor envelope/bearing analysis. Initially, with no faulted bearings to
test, a "rap" test is conducted to excite a modal response. This is sampled, and the
spectrum calculated, (see Figure 4 below).

Figure 4. Impact and resulting modal response.

4
Window selection is based on a sample rate of the sensor. The sample rate also affects the
length of the TSA:
ceil log ( SampleRates /Shaft Rate ) )
TSA length=2 ( 2
(eq 6)
Given the low output shaft rate of approximately 0.57 Hz, the measured acceleration will
be low. For this reason, the acquisition length must be adequately long to capture perhaps
20 revolutions. Hence, a high sample rate taken over an extended period results and a
very large data set, which takes more time to process and to download raw data (if
needed).
For this reason, the sample rate of the output shaft was taken at 2930 sps, for 60 seconds.
Note: at 0.57 Hz, this is only 34 revolutions. The TSA length is then 8192. For the input
shaft, which is close to 30 Hz, only eight seconds of data was taken at 23438. This allows
a Nyquist frequency of 1465 Hz for the output shaft, and 11719 for input shaft. From the
model response in Figure 4, the window for output shaft analysis was taken at 300 to
1300 Hz, which covers the small resonant mode at 1000 Hz. For the input shaft, the
window was taken from 9000 to 11000 Hz, covering the modal response at 10000 Hz
(Figure 5).

Figure 5. Envelope Response from Impact

Gear Feature Analysis: The following CIs were used on the cycloidal gear:
Residual RMS, Residual Kurtosis, Residual Crest Factor, Energy Ratio, Energy Operator
Kurtosis, Energy Operator Crest Factor, Figure of Merit 0, Side Band Lifting Factor,
Side Band Analysis, Narrow Band Kurtosis, Narrow Band Crest Factor, Amplitude
Modulation RMS, Amplitude Modulation Kurtosis, Frequency Modulation RMS,
Frequency Modulation Kurtosis, Gear Mesh Energy [2].

5
Health Indicator Paradigm: In order to reduce maintainer workload, and improve fault
detection, the concept of the health indicator was introduced in [2]. In general, the Health
Indicator (HI) is a function of distributions of n CIs, where the HI is scaled such that:
 HI ranges from 0 to a positive value, where the probability of exceeding an HI of
0.35 is the PFA (probability of false alarm),
 A warning (yellow) alert is generated when the HI is greater than or equal to 0.75.
Maintenance should be planned by estimating the RUL until the HI is 1.0.
 An alarm (red) alert is generated when the HI is greater than or equal to 1.0.
Continued operations could cause collateral damage.
Note that this nomenclature does not define a probability of failure for the component, or
that the component fails when the HI is 1.0. Instead, it suggests a change in operator
behavior to a proactive maintenance policy: perform maintenance before the generations
of collateral or cascading faults. For example, by performing maintenance on a bearing
prior the bearing shedding extensive material, costly gearbox replacement can be
avoided.
Hence, from a maintainer perspective, this is a stoplight-based threshold setting/alerting
system: when a component is yellow, plan maintenance, and when the component turns
red, do maintenance.
Controlling for the Correlation Between CIs: All CIs have a probability distribution
(PDF). Any operation on the CI to form a health index (HI) is then a function of
distributions. The HI function is defined as The norm of n CIs (energy):
0.35
HI = √ YTY (eq 7)
crit
where Y is the whitened, normalized array of CIs, and crit, is the critical value.
The function is valid if and only if the distribution (e.g., CIs) are independent and
identical (e.g., IID). As an example, for Gaussian distribution, subtracting the mean and
dividing by the standard deviation will give identical Z distributions. The issue of
ensuring independence is much more difficult. In general, the correlation between CIs is
non-zero. For instance, Table 1 shows the correlation coefficients for 6 CIs used for gear
fault analysis: most are statically significant.
Table 1: Correlation Coefficients for the Six CIs Used in the Study
ij CI 1 CI 2 CI 3 CI 4 CI 5 CI 6
CI 1 1 0.84 0.79 0.66 -0.47 0.74
CI 2 1 0.46 0.27 -0.59 0.36
CI 3 1 0.96 -0.03 0.97
CI 4 1 0.11 0.98
CI 5 1 0.05
CI 6 1

This correlation between CIs implies that for a given function of distributions to have a
threshold that operationally meets the design PFA, the CIs must be whitened (e.g., de-
correlated). A whitening solution can be found using Cholesky decomposition. The
Cholesky decomposition of Hermitian, positive definite matrix results in A = LL*, where

6
L is a lower triangular, and L* is its conjugate transpose. By definition, the inverse
covariance is positive definite Hermitian. It then follows that if:
LL* = -1, then Y = L x CIT (eq 8)
The vector CI is the correlated CIs used for the HI calculation, and Y is 1 to n
independent CI with unit variance (one CI representing the trivial case). The Cholesky
decomposition, in effect, creates the square root of the inverse covariance. This, in turn,
is analogous to dividing the CI by its standard deviation (the trivial case of one CI). In
turn, it can be shown that Y = L x CIT creates the necessary independent and identical
distributions required to calculate the critical values for a function of distributions.
Finding the Critical Value: The critical value is taken from the inverse cumulative
distribution function for the HI. The CIs used have Rayleigh like PDFs (e.g., heavily
tailed). For magnitude based CIs, it can be shown that for the nominal case, the CI PDF is
Rayleigh. For Gear CIs, and Bearing CIs (where magnitudes which are biased by RMS),
a transform is used by make the CI more Rayleigh like. The transform “left shifts” the CI.
For example, a shift such that the .05 CDF (cumulative distribution function) is assigned
to 0.0.
Consequently, the HI function is based using the Rayleigh distribution. The PDF for the
Rayleigh distribution uses a single parameter, , defining the mean  = *(/2)0.5, and
variance 2 = (2 - /2) * 2. The PDF of the Rayleigh is: x/2exp(x/22). Note that when
applying these equations to the whitening process, the value for  for each CI will then
be: 2 = 1, such that:  = 2 / (2 - /2)0.5 = 1.5264.
For the HI equation in (7), the normalized energy of the CIs, it can be shown that the
function defines a Nakagami PDF [2]. The statistics for the Nakagami are  = n, and  =
1/(2-/2)*2*n, where n is the number IID CIs used in the HI calculation.
Test Stand Results: Using a nominal gearbox, run at approximately 50% load for 45
hours. Acquisitions were taken every 5 minutes, Figure 6.

Figure 6. Nominal Gearbox

7
Initiation of Run to Failure: After the initial test run to set threshold, the gearbox was
run at 150% for one hour, then at 100% until failure. Note the resulting substantial step
change. The step-change is seen on most components. It is likely that the eccentric
bearing was damaged during the overload (Figure 7, 8).

Figure 7, Input Shaft driven by 1/Rev. Step change occurs during overload.

While the HI for both bearing C and D increase drastically (input shaft), it seems this is
more a function of the generalized wear in the gearbox. This wear caused a low-
frequency resonance within the gearbox, confounding the gear analysis for bearings A
and B (Figure 9).

Figure 8, Bearing A (Output Shaft Bearing) health – likely not due to damage but
measuring a noise floor increase due to damage in the high-speed section of the gearbox.
General Comments on the Run to Failure Test: The large number of failed
components, as indicated by most of the components being in Alarm, suggests a high
level of generalized wear. The high-speed input shaft shows an increase in SO1

8
(imbalance). It is suspected that there was damage to bearing C or D cage (output shaft
bearings), allowing the shaft to no longer be centered. Again, this may have occurred
during the overload (Figure 6).
The evidence to suggest that neither bearing A, nor B (output shaft) were damaged is that
all four bearing components (Cage, Ball, Inner Race, Outer Race) are correlated – no
individual component indicating damage. This supports the hypothesis that the noise
floor for the envelope analysis was increased due to damage in some other part of the
gearbox.
The Eccentric bearing showed bearing wear due to mechanical looseness (as evident by
the 1/Rev "Tick" at 29.75 Hz and a Ball failure, approximately 24 hours before failure
(Figure 9)

Figure 9, Eccentric bearing Envelope analysis, one day before failure. Note 1/Rev (29
Hz) tone indicating mechanical looseness.
The unfiltered bearing health eventual spiked at a level of 100, one hour before failure. It
is not unusual to see a dip in bearing fault feature energy with severe faults. This is due to
the ball elements no longer rolling, but sliding – hence the slip is increased, and the
frequency of the bearing fault feature drops, exceeding the search window of the bearing
energy algorithm (Figure 10).

Figure 10. Eccentric bearing health spikes at 100.

The cycloid gear shows in alarm level gear mesh, from 50 to 20 hours before failure.
Gear mesh energy is not usually a consistent indicator (Figure 11).

9
Figure 11. Cycloid Gear Health.

Note that from 20 hours before failure, Residual RMS, Energy Ratio, and FM0 is
sensitive to the impending fault. From this, it was learned that the best five indicators for
the cycloid gear health: Residual RMS, Energy Ratio, FM0, AM Kurtosis, and Gear
Mesh. This suggests that during the last 10 to 20 hours of the run, the cycloid gear
experienced a second failure mode that was detected by the more traditional gear faults
(Figure 12).

10
Figure 12, Cycloid gear condition indicators
The ring gear showed damage as well. Because the gear health is based on the filtered HI,
it does not reflect the step change HI of 2 in the raw HI data just prior to failure (Figure
13)

Figure 13: Ring Gear HI.

The gear CIs that were sensitive to this fault was found to be: Residual RMS, Energy
Ratio, FM0, AM Kurtosis, and Gear Mesh. These condition indicators will be used in the
future.
Conclusion: The cycloid gearbox has unique dynamics, which require using the correct
ratios for the TSA and bearing rate calculation. The bearing analysis for the Cycloid
gearbox is relatively standard. The eccentric bearing rate was multiplied by a correction
factor to account for the rotating frame of the outer race. The eccentric bearing is the
weak point of the gearbox. During the run to failure test, the eccentric bearing roller
elements were found to be faulted. Towards the end of the run to failure test, the high
level of damage (resonance energy) associated with the eccentric gear raised the noise
floor of the envelope analysis, which contributed to high HI level of all the bearings in
the gearbox.
The cycloid gear itself posed a challenge in that the apparent gear mesh frequency is
based on the gear eccentric behavior and not on the gear mesh frequency alone. The
observed gear mesh frequency is the gear tooth + 1 vs. gear tooth, multiplied by the shaft
rate.

11
Because the gears are under compression vs. shear, care must be used in the CI selection
used for the health indicator, as many of the traditional analyses were not sensitive. It
was found that residual RMS, amplitude modulation RMS and gear mesh energy for gear
tooth + 1 the shaft order worked well.
References:
[1] Zhu, J., Nostrand, T., Spiegel, C., Morton, B., "Condition Indicators for Condition
Monitoring Systems," Annual Conference of the Prognostics and Health Management
Society, 2014
[2] Bechhoefer, E., He, D., Dempsey, P., "Gear Health Threshold Setting Based On a
Probability of False Alarm," Annual Conference of the Prognostics and Health
Management Society, 2011.

Bibliography:

[1] Eric Bechhoefer is the president of GPMS, Inc., a company focused on the
development of low-cost condition monitoring systems. Dr. Bechhoefer is the author of
over 100+ juried papers on condition monitoring and prognostics health management and
holds 23 patents and five patent applications in the field of CBM.

12