32

CHAPTER 3

MODELING OF HYBRID STEPPER MOTOR AND

INVESTIGATION OF RESONANCE

3.1

A SURVEY

INTRODUCTION

A mathematical model is a description of a system using

mathematical concepts and language. Mathematical models are used not only
in the research and engineering disciplines, but also in the curriculum of social
sciences. A model may help to explain a system in a better way to investigate
the effects of different components and to make predictions about behaviour.
Mathematical models are usually composed of variables, which are abstractions
of quantities of interest in the described systems, and operators that act on these
variables can be algebraic operators, functions, differential operators etc. If all
the operators in a model exhibit linearity, the resulting mathematical model is
defined as linear.

A model is considered to be nonlinear otherwise. A static model

does not account for the element of time, while a dynamic model does.
Dynamic models typically are represented with difference equations or
differential equations. A discrete model does not take into account the function
of time and usually uses time-advance methods while a continuous model does.
Continuous models typically are represented with f(t) and the changes are
reflected over continuous time intervals.

33

3.2

MATHEMATICAL MODEL OF HSM

The two phase HSM model equations are explained by Kenjo (1984)
and expressed as

Electrical equations

(3.1)
Mechanical equations

Where ia, ib, Va and Vb are the currents (Amp) and voltages (Volt) in
phases A and B respectively. Km is the motor torque constant (Nm/A). J is the
rotor inertia (kgm2), Nr is the number of rotor teeth, B is the viscous friction
(Nms2/ rad),

rad). R
L

is the load

torque (Nm).

The mathematical model is derived based on the following

assumptions as explained by Kenjo (1984)

1. Mutual inductance is neglected as being dominated by the self

Inductance.

34

2. Variation of self inductance with the mechanical angle is assumed

to be sinusoidal around the mean value.
3. Fluxes due to permanent magnet are sinusoidal with the same
periodicity of the reluctance flux.
4. Detent torque is assumed to be negligible compared to magnitude
of holding torque.

3.3

THEORY OF VIBRATION AND RESONANCE

Stepper motor makes a move from one step to the next and the rotor

forth (overshoots) until it comes to rest. This is known as ringing and it occurs
every time due to the commanded move of the motor to the next step before it
comes to rest. When motor under no load condition exhibits a fair amount of
ringing, it gets converted into motor vibration. The motor will often stall due to
the high vibration and cause the motor to lose its synchronism if loaded heavily
as discussed by NMB Technologies Corporation (online) data.
Vibration is inherent in all motor systems resulting in the imbalances of
moving mechanical parts. Vibration is due to small step movement, detent
torque of PM and pole to pole variation. Effect of vibration includes the
Error in actual step movements
Decrease in torque
Loss of step or even a change of direction of rotation
Vibration is similar to synchronous motors or as a spring oscillating
system. Load applied on the motor also dampens the vibrations as discussed by
NMB Technologies Corporation (online) data.

35

3.4

RESONANCE INVESTIGATION
A stepper motor will exhibit much stronger vibrations when the

input pulse frequency matches the natural frequency of the motor. This is called
resonance. During resonance conditions, oscillations will be more and there is
chance of skipped steps. The resonance range may change slightly due to the
All stepper motors suffer from resonance.
It is the oscillatory phenomena which disturb the normal operation of the motor.
In some cases the magnitude of oscillation increases with time and eventually
the motor loses synchronism. Resonance and instabilities may be classified into
three categories as explained by Acarnely (2002) and Athani (2005) and they
are

1. Low-frequency (occur below 200 Hz)

2. Medium-range instability (range of 500 Hz to 1500 Hz)
3. Higher-range oscillation (range of 2500 HZ to 4000 Hz).

During resonance the motor may even stop in the lower frequency
range up to approx. 250 Hz. Medium and higher frequency resonances are
mainly caused by electrical parameters such as inductance of the motor
winding. High frequency resonance has no significant effect on the torque and
can be controlled relatively easily through appropriate timing of switching
sequence. The resonances in the lower frequency range are mainly caused by
mechanical parameters of the motor. Irregular movement of the motor also
leads to significant loss of torque and also may even obstruct it through loss of
step.

36

Figure 3.1 Typical pull-out Torque Vs Speed characteristic curve

At high speeds the motor/drive system can become unstable and so it is

impossible to operate the system continuously in a certain stepping rate band.
Furthermore the effects are reduced in heavily loaded open-loop systems with
substantial viscous damping. The instability is often represented by dips in the
steady-state pull-out torque Vs speed characteristics at higher stepping rates.
The instability occurs close to those stepping rates where the angular frequency
of the phase winding excitation is equal to the phase resistance (R) divided by
the phase winding inductance (L). Thus, for the hybrid stepping motor which
executes four steps per excitation cycle, the critical stepping rate is

Instability rate =

The following points are observed from the typical pull-out Torque Vs
Speed characteristics as shown in Figure 3.1.
1. Limited torque at higher speed
2. Torque will decrease with increasing speed

37

3. Torque drops at the resonant speed

4. Resonant speed will vary on the type of driving circuits and load
5. Maximum torque is achieved at low speed.
Two major negative side effects of resonance Viz;

1.

Rough motion
- When step motors resonate, they generally vibrate more.
- This can be bad for sensitive equipment.

2.

Loss of torque, which can lead to

- Loss of position and often leads to stalling

3.5

TROUBLESHOOTING VIBRATION AND RESONANCE

A two-phase HSM can only skip the steps in multiples of four full
steps (equivalent to one tooth pitch or pole pitch) as shown in Figure 2.9. If the
skipped step is a multiple of four, vibration or overloading may be causing a
loss of synchronism. When the skipped step is not a multiple of four, it may
indicate circuit problem as discussed by NMB Technologies Corporation
(online) data.

3.6

PERFORMANCE FACTORS

Maximum torque is proportional to the winding current and the

number of turns of wire as discussed by NMB Technologies Corporation
(online) data.

38

To increase torque by 20%, the current be increased by about 20%.

Due to magnetic saturation, there is a limitation for increasing the current to
more than 2 times the rated current. Inductance reduces a high speed torque
performance. At low speeds, high inductance is not a problem. Current can
easily flow into the motor windings fast enough that the stepper motor has rated
torque. At high speeds, however, sufficient current cannot get into the winding
fast enough before the current is switched to the next phase, thereby reducing
motor torque. Increasing the driver voltage can overcome this loss of torque at
higher speeds by forcing current into the windings of the motor at an increased
rate.

The electrical time constant is the amount of time it takes a motor

coil to charge up to 63% of its rated value. When a stepper motor is rated at 1
amp, after one time constant, the coil will be at 0.63 amps, giving the motor
about 63% of rated torque. After two time constants, the current will increase to
0.86 amps and the rated torque of motor also increases to 86%.

t = L/R

(3.2)

time constant t (ms). The resonance in stepper motors is due to their design, it
can only be reduced or diminished in part by the application of following
techniques.

3.7

EXISTING TECHNIQUES TO TACKLE RESONANCE

Different techniques are available to reduce resonance and vibration

as discussed by Kenjo (1984) and Acarnely (2002). Each technique has its
advantages and disadvantages. The key to eliminate its effects lies in either
controlling where the resonant point falls or in reducing its severity.

39

3.7.1

Micro Step Mode of Excitation

Due to the smaller step angle, the oscillation is reduced and the

system has less resonance points. The major disadvantage of the mini-step drive
is the cost of implementation due to the need for partial excitation of the motor
windings at many current levels, using a chopper drive circuit in which the
reference current level for each phase is changed every mini-step. If no current
compensation is provided in the micro step driver, a torque reduction of the
motor may occur which is disadvantages in some applications. However, even
if the motor is driven with sinusoidal pulse, vibration and resonance still exist
due to the motor inherent characteristics.
3.7.2

Change in Phase Current

Resonance excitation is strongest during no-load operation and

therefore, brings problems during testing. The phase current reduction also
minimizes the stiffness and must be taken into account while positioning
accuracy.
3.7.3

Shifting the Step Frequency

Resonance occurs during no-load operation using full step mode at

approximate 70-100 Hz and appears at multiples or harmonics of the resonance.

It is easier to avoid the resonant frequency by choosing a frequency that is
somewhat higher or lower. Small deviations from the critical step frequencies
reduce the resonance.
3.7.4

Increasing the Friction

Friction generally has a damping effect on the system and oscillation

becomes smaller. However, the torque is reduced by this and the motor
efficiency also reduces.

40

3.7.5

Dampers
Damper reduces the oscillation and absorbs the vibration energy.

The resonant frequencies are very much reduced because of the speed
difference between the oscillating rotor and the external mass. Although this
results in the significant reduction of overshoot, the use of dampers is
unsatisfactory during high frequencies as it is limited by the friction torque.
Above techniques are either for reducing the resonance or for
avoiding the resonant frequency. The solution to operate the motor during
resonance conditions is left undiscussed.

3.8

PERFORMANCE LIMITATIONS
At low step rates, increasing the supply voltage produces

proportionally higher torque until the motor magnetically saturates. Nearing

saturation, the motor becomes less efficient and increasing the power is
unjustifiable. The maximum speed of a stepper motor is limited by inductance
and eddy current losses. At a certain step rate, the heating effect of these losses
limits any further attempt to get more speed or torque out of a motor by driving
it harder as discussed by NMB Technologies Corporation (online) data.

3.9

TECHNIQUES TO IDENTIFY THE RESONANCE

The following identification techniques are compared to show their

delineating capability

and NR condition.

1. Eigenvalue based identification

2. Acceleration sensor based vibration identification
3. Direct position measurement based identification

41

3.9.1

Eigenvalue Based Identification

The linearized model about a particular operating condition, with all

electrical perturbations measured in the synchronous frame, is time invariant

and takes the form explained by Verghese et al (1986) is
(t) = Ax(t)
Where

(3.3)
is a constant matrix. Complete information about local

stability of the operating condition is embodied in the eigenvalues of A in (3.3).

These eigenvalues are the natural frequencies of the system and are the roots of
the characteristic polynomial det (sI - A).

any eigenvalues in the open

right-half plane, the machine is locally unstable at the operating condition under
consideration.

For most Park-transformable electrical machines, A has a pair of

eigenvalues close to the imaginary axis that accounts for electromechanical
ringing about nominal operation. Local instability is predicted by the linearized
model when this eigenvalue pair lies in the right-half plane. Midrange
instability corresponds to this pair crossing into the right-half plane for some
intermediate range of operating speeds.

The first pair of eigenvalues given for each

in the table

corresponds to the electrical time constants of the phases and is very closely
approximated by - R /L jN . The presence of the imaginary part jN reflects
the fact that monotonically decaying currents in the stator phases appear
oscillatory. The second pair of eigenvalues

may be associated with the

mechanical hunting oscillations that give rise to the instability of interest here.

42

This technique is applied for various sets of motor data and results are
tabulated in Table 3.1. From the table, it is observed that the resonant frequency
based on Eigenvalue method and simulated value is not matching except for the
first set of motor data. This method does not reveal the resonance condition for
all type of stepper motor parameter.

0.012
0.108

5. Balakrishnan,
et al (2011)

6. Balakrishnan, et
al (2013)

0.55

3. Kim, et al (2009) 0.0012

0.015

1.2

2. Burg, et al(1994) 0.0007

4.Yang, (2001)

7.35

0.0107

1.Verghese, et al
(1986)

52

1.55

( )

L
(H)

S.No

0.001

0.00008

0.00001

0.0008

0.002

0.0016

B
(Nms/rad)

0.5

0.19

0.01

0.00331

0.00436

0.00938

Km
(v/(rad/s)

0.00001

0.000045

0.000025

0.00045

0.000187

0.00541

J
(kg-m2)

2.6

10.67

9.476

38.44

15.523

1025

230

595.99

71.9

138.45

54

Resonant
frequency range
based on
Eigenvalue

44

60

18.2

25

40.7

15.6

85

80

596.1

71.9

120

54.1

Resonant
frequency
range found by
simulation

Table 3.1 Eigenvalue based resonant frequency range for different sets of motor data

43

44

3.9.2

Accelerometer Sensor Based Vibration Identification

Tinghsu (2002) has discussed a method to reduce the mechanical

vibration in the open loop controlled three phase HSM motor by the repetitive
control system using accelerometer or microphone as the vibration sensor. The
detected signal containing various kinds of noises other than the signal due to
the vibration torque has to be extracted. The detected vibration signal is input to
the Fourier Transform (FT) through a low-pass filter to cut-off higher
frequency sensor noise. The FT is a kind of filter to cut off all higher and lower
order harmonics except a particular vibration frequency signal. The
compensation signal to suppress the vibration of the motor frame is obtained by
the repetitive controller installing an online Fourier transformer and utilizing an
acceleration sensor attached to the motor frame or an acoustic sensor such as a
microphone placed close to the frame. The sensor is used only for the
acquisition of the feedforward compensation data. The feedforward
compensation signal at an arbitrary operating point is derived from the
amplitude and phase data of the frequency components and the operating point
data.
Compensation data obtained by the repetitive controller is applied to
the operating point changed by reference frequency and load condition in
steady state. The compensation signal for the new operating point will be
generated from compensation data utilizing polynomial equation approximation
and linear interpolation method. Demerits are requirement of accelerometer or
microphone sensor and also complex control circuit.

3.9.3

Direct Position Measurement Based Identification

Tsui et al (2009) explained the micro step control where the

simplified model predicts only one resonant speed, which is caused by the
detent torque. However, this does not match with that of a commercial stepper

45

in real life. For example, the studied motor has three resonant speeds. The
system is configured to open-loop control, and holding current is set at 1.9 A.
The natural frequency of the system calculated is 142 Hz. On driving the motor
with smooth sinusoidal current, three resonant speeds are observed at about 43,
86, and 173 r/min. These correspond to the driving current at 36, 72, and 144
Hz. They are shown to be caused by the detent torque. At the first resonance
(43 r/min/36 Hz), the frequency of the detent-torque ripple is 36 Hz 4 = 144
Hz, which matched with the calculated system natural frequency (142 Hz). The
first resonance is shown to be excited by the fourth-harmonic detent-torque
ripple, the second and third resonances, occurring at doubled speed of the
previous resonance. Similarly, the first harmonic and second harmonic detent
torque components are shown to excite the resonances. Model-based
compensation of the low-order torque ripple is shown to be effective to remove
vibration and resonance. This method requires changes in mathematical model
and measuring sensor adding to cost and complexity.

Comparisons of the three identification methods are tabulated in

Table 3.2. Demerits of the above methods include requirement of sensor to
measure the position and speed and changes in the mathematical model. The
identification scheme proposed and discussed in the chapter 4 will overcome
this problem.

46

Table 3.2 Comparison of various identification methods

Eigenvalue

Accelerometer

Direct position

Verghese et al
(1986)

Tinghsu (2002)

Tsui et al (2009)

Change in the
motor model

Required

Required

Required

Resonance
identification

Resonant
frequency
range based
on Eigenvalue

Sound
amplification

Specific harmonic
frequency
extraction from
phase current.

Measured
variable

Speed

Position

Position, speed &

current

3.10

CONCLUSION

This chapter explained the mathematical model of HSM. The theory

of vibration and resonance has been discussed. Investigation of resonance with
typical pull out torque speed curve has been explained. Techniques to tackle the
resonance and its pros and cons have also been discussed. Some of the
identification techniques of resonance have been investigated and comparison
results tabulated.