0% found this document useful (0 votes)
23 views11 pages

Electronic Coil Contactor

Uploaded by

Farid.mansouri89
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
23 views11 pages

Electronic Coil Contactor

Uploaded by

Farid.mansouri89
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/3219763

An Electronic Control Unit for Reducing Contact Bounce in Electromagnetic


Contactors

Article in IEEE Transactions on Industrial Electronics · March 2008


DOI: 10.1109/TIE.2007.909073 · Source: IEEE Xplore

CITATIONS READS
39 1,952

2 authors:

Paulo Mario Dos Santos Dias de Moraes Arnaldo José Perin


Federal University of Santa Catarina Federal University of Santa Catarina
6 PUBLICATIONS 41 CITATIONS 77 PUBLICATIONS 749 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Wireless Converter with ZCS Rectifier View project

All content following this page was uploaded by Arnaldo José Perin on 24 June 2015.

The user has requested enhancement of the downloaded file.


IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 55, NO. 2, FEBRUARY 2008 861

An Electronic Control Unit for Reducing Contact


Bounce in Electromagnetic Contactors
Paulo Mário dos Santos Dias de Moraes and Arnaldo José Perin, Member, IEEE

Abstract—With the increasing concern on reliability and life to heat generation and possibly to failure—most commonly the
span of electromagnetic contactors, manufacturers are now pro- static welding of the contacts.
viding devices equipped with electronic control units, in an effort The contact bounce is a complex phenomenon. Although
to reduce problems associated with contact bounce, such as contact
wear, excessive heating, contact welding, etc. The addition of a con- mathematical models have been proposed in the past [11], [12],
trol unit, however, brings increased costs both in the development most of what is known of contact bounce is essentially empiri-
and production of this new generation of contactors. Considering cal. Several results were obtained with test rigs specifically de-
this, a cost-effective microcontroller-based electronic control mod- signed for the study of the phenomenon [8], [13], and thus, the
ule is proposed. Some fundamental equations are reviewed, and extension of those results to real-life situations should be made
the most important aspects of the control algorithms employed are
discussed. Finally, experimental results are presented. with restrictions. Other works, even though based on the oper-
ation of commercially available contactors [14]–[16], focus on
Index Terms—Contact bounce, contactors, electronic control. the contact-bounce phenomenon from a different perspective,
since several parameters related to the contact closing cannot be
I. I NTRODUCTION controlled (or can be, but in a limited fashion). It is known that
both mechanical (e.g., contact velocity before impact, contact

E LECTROMAGNETIC contactors are one of the most


widely adopted types of switchgear equipment and are
used in such diverse applications as rectifiers [1], motor drives
mass, spring forces [17], and damping coefficients) and electri-
cal quantities (i.e., load current) have influence on its behavior.
This paper proposes an electronic control module suited for a
[2], [3] and uninterruptible power-supply systems [4]. Being conventional electromagnetic contactor that not only offers the
electromechanical devices, electromagnetic contactors have a aforementioned advantages but also brings a reduction in the
limited lifetime, which is dependent on several factors [5]. contact-bounce time when the performances of the contactor,
The advantages that electronic control can bring to both the with and without control, are compared. Since the control unit
performance and reliability of electromagnetic contactors are is microcontroller-based, the CPU is responsible for all control
well known [6], [7]. Despite the added expense of the compo- and supervision tasks, keeping the remaining circuitry simple
nents used on the control module, an electronically controlled and affordable. The control unit can be fitted to conventional
contactor, which is equipped with a single coil, can be used on off-the-shelf electromagnetic contactors without any changes
a wide range of control voltages, both ac and dc, whereas a on the contactor itself [18], [19]. Unlike previous works, the
conventional apparatus would require different coils to cover detection of contactor closing is achieved by monitoring elec-
the same range, meaning increased design, production, and trical quantities rather than relying on mechanical sensors [20],
storekeeping costs. [21]. Furthermore, the bounce-reduction strategy is closely tied
One of the major concerns regarding the operation of con- to the closing detection and, thus, is more efficient than previous
tactors is the presence of contact bounces during the making solutions, which were simply “time-based” [22]–[24].
process [8]. When two solid objects collide, there is inevitably a
bounce, and in the case of electrical contacts, this bounce results
II. C ONTROL U NIT
in the formation of electric arcs, if the conditions of the load
circuit allow. These electric arcs, in turn, lead to the erosion A. Block Diagram
of the contacts. Materials like AgCdO and AgSnO2 offer some
A simplified block diagram of the control unit is shown
degree of protection against the erosion and oxidation of the
in Fig. 1. The auxiliary power supply provides two distinct,
contact surface [9], [10], but when they are worn to the base
continuous, and regulated voltages required for the operation
material, usually copper, the increased contact resistance leads
of the microcontroller and driver circuits. Since the contactor
shall operate over a wide range of control voltages, ac or dc, the
Manuscript received July 13, 2006; revised September 14, 2007. This work power supply must be of the switched-mode type. A flyback
was supported in part by the National Council of Scientific and Technological converter is most suitable and was the choice for the auxiliary
Development—CNPq—Brazil.
P. M. S. D. De Moraes is with Microsol Tecnologia S/A, 61760-000 power supply.
Fortaleza, Brazil (e-mail: [email protected]). The microcontroller is a low-cost programmable intelligent
A. J. Perin is with the Department of Electrical Engineering, Power Electron- computer device from Microchip, operating at 20-MHz clock
ics Institute, Federal University of Santa Catarina, 88040-900 Florianópolis,
Brazil (e-mail: [email protected]). frequency. Its internal digital–analog converter is used to
Digital Object Identifier 10.1109/TIE.2007.909073 measure the coil current and the control voltage Us . While
0278-0046/$25.00 © 2008 IEEE
862 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 55, NO. 2, FEBRUARY 2008

Fig. 1. Block diagram of the control unit.

of the main switch current are taken at the beginning of each


switching period, shortly after the switch is turned on. The time
taken to complete the analog-to-digital (A/D) conversion of the
shunt-resistance voltage and to process the results then defines
the minimum on-time of the chopper, and thus, a high-speed
A/D converter is preferable to allow the switching frequency to
be increased.

C. Current Measurement
Contactors are peculiar in that the current level required for
closing (closing current) and the current level that allows the
contactor to remain closed afterward (maintenance current) are
very different—frequently being an order of magnitude apart.
Considering this fact, instead of one, two shunt resistances
are actually used for the current measurement, as shown in
Fig. 3: Rshunt1 and Rshunt2. Rshunt2, being a higher value
Fig. 2. Power circuit.
than Rshunt1, is rated to allow a precise measurement of the
the coil current is the controlled variable, the measurement maintenance current. An auxiliary switch allows Rshunt2 to be
of the control voltage is used for two purposes: to guarantee bypassed, which is required during contactor closing because of
a safe-operating region for the contactor (for instance, the higher currents involved.
shutting down the contactor in an undervoltage condition, thus
preventing contact “chattering”) and for the detection of the
D. Demagnetizing Circuit
control-voltage type—ac or dc.
The driver circuits are used to adapt the low-voltage low- The way the coil is demagnetized after removing the control
power signals from the microcontroller to the requirements for voltage is of great importance to preserve the integrity of the
driving the various MOSFET transistors of the power circuit. electronic components of the control circuit and to guarantee
a contact’s long life. McBride and Sharkh [25] investigated
the relationship between opening velocity and erosion of the
B. Power Circuit
contacts. They concluded that to reduce the arc energy under
The power circuit is shown in Fig. 2. It consists of a full- ac loads, the contact opening velocity should be diminished,
bridge input rectifier and a chopper circuit designed to control whereas the opening velocity must be as large as possible under
the contactor coil current. The chopper is formed by the main dc conditions so that both the energy and the duration of the arc
switch (with a series shunt resistance for current measurement), are reduced. Either way, they observed that material transfer
a free-wheeling diode, the demagnetizing diode, and the de- between the electrodes (contacts) was lower when the opening
magnetizing switch. To drive the main switch, a fixed-frequency speed was increased, despite the greater arc energy.
PWM-modulated signal is generated by the microcontroller. A To assure a high-speed opening of the contactor, the de-
current control is afforded by the comparison—made within magnetizing circuit was developed. When the main switch is
the microcontroller—of the main switch current and an inter- turned off, the coil current circulates through the free-wheeling
nal reference. The main switch is turned off either when the diode and—when the demagnetizing switch is also blocked—
measured current is higher than the reference value or at the through a demagnetizing power Zener diode. In this situation,
end of the switching period, whichever occurs first. Samples the demagnetizing diode forces a voltage on the coil that, in
DE MORAES AND PERIN: CONTROL UNIT FOR REDUCING CONTACT BOUNCE IN ELECTROMAGNETIC CONTACTORS 863

Fig. 3. Schematic diagram of the electronic control unit (auxiliary power supply not shown).

takes into account the two poles of the electromagnet, and the
minus sign indicates that the force is attractive. The flux density
is related to the current I circulating through the coil by its
inductance L
λ N ·Ψ
L= = (2)
I I
where λ is the magnetic flux concatenated by the N turns of
the coil. The relationship between flux ψ and flux density B is
given by

Ψ = B · dS = B · S (3)
Fig. 4. Electromagnet.
where B and S are the vector quantities whose magnitudes are,
turn, causes the current to rapidly decrease. On the other hand, respectively, B and S. Applying (2) and (3) in (1) results in
when this behavior is not desired—for example, during the
establishment of the contactor’s closing current—the micro- L2 · I 2
F =− . (4)
controller simply keeps the demagnetizing switch turned-on, µ0 · N 2 · S
suppressing the demagnetizing diode.
As can be seen in (4), the attractive force is related to the
current I, which justifies its choice as the controlled variable.
E. Schematic Diagram of the Control Unit
However, the force also depends on the coil inductance L. This
A simplified schematic diagram of the control circuit is inductance is not a fixed quantity but rather dependent, among
shown in Fig. 3. other factors, on the current I itself and mainly on the air-
gap length l between the fixed and movable parts of the core.
III. F UNDAMENTAL E QUATIONS Making the simplifying hypotheses that the magnetic flux in
the air gap is uniform and that the magnetic material of the
The contactor is basically an electromagnet, as shown in core operates in the linear portion of its B–H curve and has
Fig. 4. The attractive force F of the electromagnet is given by considerably greater permeability than the air, the inductance L
 2  can be expressed in terms of the air-gap length l as
B ·S
F = −2 · (1)
2 · µ0 N 2 · µ0 · S
L= . (5)
where B is the flux density in the air gap, S is the cross- 2·l
sectional area of each of the pole faces, and µ0 is the air mag- Equation (5) states that the inductance of the contactor coil is
netic permeability. The factor of two multiplying the fraction an inverse function of the air-gap length or the distance between
864 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 55, NO. 2, FEBRUARY 2008

experimental coil-current waveforms during the closing process


were divided into four stages.
In the first stage, from ta to tb , the coil current is in a
transient state, reaching a steady state about tb . Its waveform
remains practically unchanged throughout the second stage, and
during this interval, the initial attractive force is established,
but the contactor stays at rest. At time tc , the coil current (or
its envelope, in the ac case) starts to drop, characterizing the
movement of the contactor armature. The first occurrence of
contact making happens, in both figures, at t = 100 ms. While
there is a series of bounces, the coil current starts to increase,
again reaching a steady state at time td .

B. Coil-Current Control
Fig. 5. Coil current and contact voltage, with dc control voltage.
In both Figs. 5 and 6, there was no control of the coil current,
i.e., the main switch of the power circuit was kept turned-on
during the entire closing process. However, some sort of control
must act upon the coil current. Traditionally, the technique
mostly used is current limiting [6]. It consists of closing the
main switch until the coil current, which is measured by means
of the shunt resistance, reaches a set point. Then, the switch is
turned off—for a fixed time (variable frequency) or until the
start of the next switching cycle (fixed frequency)—when it is
turned on again.
Current limiting, at least the way it is commonly deployed,
works very well for dc control voltages, where it can actually
control the average current value, since the current ripple is very
small. However, under ac conditions and with the absence of a
large filter capacitance at the output of the rectifier bridge (as
it is the case of the control unit proposed, see Fig. 2), there
Fig. 6. Coil current and contact voltage, with ac control voltage. is a pronounced ripple, at twice the control-voltage frequency,
the fixed and movable parts of the core. As the contactor leaves on the current waveform. Furthermore, the peak-to-peak value
its position of rest and starts to close, the length l is dimin- of the ripple depends, among other factors, on the magnitude of
ished, followed by an abrupt rise of the inductance. In Fig. 4, the rectified control voltage. What this all means is that, using a
considering that the air gap is divided between the two poles of fixed set point, the traditional current-limiting scheme cannot
the electromagnet, a factor of two is included on the denomina- maintain either the average or the rms value of the closing
tor of (5). current at a fixed desired level over a wide range of control
If the voltage is imposed on the coil, the rise of the coil voltages. As the control voltage increases, the ripple becomes
inductance causes the coil current to instantaneously decrease; smaller, but the maximum current value is kept the same. The
therefore, the equilibrium of the equation is maintained. Every effective average and rms values of the coil current tend to
contactor exhibits this transient behavior during the closing increase as well, and the resulting dynamics of the contactor
process regardless of the nature of the voltage applied to its (e.g., the closing time) is completely different.
coil. This can be seen in Figs. 5 and 6 and is the basis of the To attenuate the aforementioned problem, the algorithm
closing-detection algorithm that will be presented shortly. shown in Fig. 7 is presented. Ilim is the coil-current-limit value
Instantaneous voltage vL (t) and instantaneous current iL (t) and can vary between Ilim_ min (the initial value) and Ilim_ max .
in the coil are related by the following equation, where the Iavg is the measured average current value calculated over one
winding resistance r is also considered: half cycle of the ac control voltage. The goal is to adjust Ilim so
that Iavg gets sufficiently close to Iref , which is the reference
diL (t) value for the average closing current. On each iteration, which
vL (t) = L · + r · iL (t). (6)
dt begins at the start of a new control-voltage half cycle, the
average current value is calculated and compared against Iref ,
IV. C ONTROL S TRATEGY and then, the Ilim value is adjusted according to the bisection
method, setting a new limit value for the coil current during
A. Characterization of the Coil Current
the following half cycle. Even if the magnitude of the error
During Contactor Closing
ε = Iref − Iavg cannot be constrained at a desired margin, the
The process of contactor closing involves a number of stages algorithm “gives up” after niter iterations, assuring that the
and modifications in the coil current. In Figs. 5 and 6, the contactor will not end up closing during this regulation stage.
DE MORAES AND PERIN: CONTROL UNIT FOR REDUCING CONTACT BOUNCE IN ELECTROMAGNETIC CONTACTORS 865

85 V, the coil-current control becomes saturated, i.e., the input


voltage conditions do not allow for the average coil current to
reach the desired value Iref even though the current-limit value
is set all the way up. This can be further attested to when the
results obtained from the fixed current-limit scheme and those
from the proposed variable current-limit scheme are compared,
showing that both the rms and the average coil current from
both schemes are practically the same.

C. Detection of Contactor Closing


After the closing current is adequately established on the coil,
either using a fixed limit value (in case of a dc control voltage)
or at the end of a regulation stage (when an ac control voltage
is applied), the next step of the proposed control strategy is
to detect the start of the armature movement. As shall be
discussed, this detection must take place fast enough so that
it anticipates the instant of first contact touch. Late detections
do not keep the contactor from closing securely but render the
bounce-reduction algorithm useless. Premature detections, on
the other hand, should be avoided, as they may cause reopening
under certain circumstances.
As it was mathematically demonstrated, as well as easily
identified on the preceding figures, the movement of the arma-
ture provokes a drop in the coil current despite the control effort
to sustain its value. In the case of dc control voltages, the current
ripple is very small; therefore, the reduction of its value below
a certain threshold value clearly indicates that the contactor is
about to close. The algorithm used for the detection of contactor
closing, in this case, is thus very simple.
However, when the control unit is supplied with an ac control
voltage, merely comparing the current value against a fixed
minimum threshold, although functional, is inefficient. Contac-
tors usually take several control-voltage cycles to completely
close, and the start of the armature movement can happen at any
point in the coil-current waveform. If it started, for instance, at
an instant when the current was at its peak value, more time
would be needed for the current to fall below the detection
threshold than it would be for the case if the coil current was at
its valley. This additional time delays the detection, which may
be enough to compromise the bounce-reduction procedure.
A more sophisticated technique for the closing detection
regarding the ac control voltages is shown in Fig. 8. It can be di-
vided into two parts. In the first part, the microcontroller records
both the measured current value at the end of each switching
cycle and the instant the measurement was made—referenced
to the start of the ac control-voltage half cycle. This time stamp
tref is cleared at the beginning of every new half cycle of the
ac control voltage and serves as the pointer for the table that
Fig. 7. Algorithm for regulating the current-limit value (ac control voltage). holds the current measurements created in the microcontroller
internal memory.
Table I summarizes the results obtained with the proposed al- After an interval t2 has elapsed, the second part of the
gorithm versus the traditional fixed-level current-limit scheme. detection algorithm kicks in. The table where the measured
As the data show, with the fixed value scheme, both the average values were formerly recorded now provides a reference current
and rms current values increased while the control voltage was waveform against which every new current value is compared.
raised. The proposed algorithm, on the other hand, managed The tref pointer is still refreshed the same way as before to
to keep these variations below tighter limits. While there was a assure that the reference and present waveforms stay in sync.
substantial variation from 85 to 110 V, it is due to the fact that, at The microcontroller then interprets any considerable deviation
866 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 55, NO. 2, FEBRUARY 2008

TABLE I
COMPARISON BETWEEN FIXED SET-POINT AND VARIABLE SET-POINT CURRENT-LIMIT CONTROL SCHEMES

Fig. 8. Closing-detection algorithm (ac control voltage).

from the recorded data as an indication of armature movement, tween the present measured value and the one recorded in the
and thus, the closing of the contactor is detected. memory begins. At time tdet , the closing is detected, which
An experimental example of a controlled coil-current wave- happened approximately 1.6 ms before the first contact touch1
form during the regulation and transition stages can be seen was registered by the oscilloscope at tclose .
in Fig. 9, where n is the iteration counter. In this example,
four iterations took place before the algorithm considered the 1 In double-break contactors, the movable contact at one end of the contact

average coil current satisfactory at instant treg . Then, during bridge usually touches the corresponding fixed contact before the opposite side
does [15]. However, the oscilloscope could only register the instants when both
period t2 , the microcontroller records the current measurements sides of the bridge made contact, making the test current, which is a different
in its memory. After this period elapses, the comparison be- situation.
DE MORAES AND PERIN: CONTROL UNIT FOR REDUCING CONTACT BOUNCE IN ELECTROMAGNETIC CONTACTORS 867

Fig. 11. Coil current and contactor voltage without transition stage.
Fig. 9. Coil-current waveform during regulation and detection stages.

Fig. 12. Transition of the coil current from the closing level (Ilim ) to the
Fig. 10. Coil current and contactor voltage during transition stage. holding level (Ihold ).

D. Transition Stage t3 should also not be set excessively large as the antireopening
protection would not be afforded in this case.
Once the armature movement is detected, the electronically
The effects of the bounce-reduction technique that takes
controlled contactor enters a transition stage where measures
place during the transition stage can be seen in Fig. 10. The
are taken to help reduce the contact bounce. Since the impact
voltage across one of the three pairs of contacts, as well as
velocity of the movable contacts has major influence both on the
the coil-current waveform, were experimentally obtained and
number of bounces and total bounce time [6], if the velocity of
depicted, together with indications of the time instants related
the armature could be decreased just before the contact touch,
to the transition and holding stages. In contrast, Fig. 11 shows
bouncing would be potentially reduced.
the experimental results from the same contactor and control
The way the proposed control strategy accomplishes this is
circuit but without any changes on the current-limit value after
by rapidly reducing the coil current, anticipating its natural
the closing detection. As one can see, the total contact-bounce
behavior, once the armature movement is detected. At the
time was reduced when the appropriate measures were taken
beginning of the transition stage, the demagnetizing switch is
after the armature movement was detected (Fig. 10).
shut off, and the current limit is changed from the closing
value (Ilim ) to a transition value (Itran ) (see Fig. 10). Once the
E. Holding Stage
measured coil current reaches down Itran , the demagnetizing
switch is again turned on, and the coil current is maintained at At last, after another time delay t4 , the reference current
the transition level. Itran sits about halfway between the holding value is again modified, this time down to Ihold —the holding
current level Ihold and the closing current limit Ilim . current level. Its value, although insufficient to move the con-
After some time period t3 , the current-limit value is once tactor armature, keeps the apparatus closed, reducing the power
more raised to the closing value Ilim . Time period t3 is chosen required from the control-voltage supply. In most cases, the
to be sufficiently long so that the bounce-reduction effect is coil-current reduction is absolutely necessary given the fact that
obtained. However, the idea behind raising the coil current is to the coil is designed to withstand (indefinitely) only the holding
avoid contactor reopening in case the armature movement was current level. Fig. 12 shows the transition of the coil current
erroneously detected. Considering this, the duration of interval from the closing level (Ilim ) to the holding level (Ihold ).
868 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 55, NO. 2, FEBRUARY 2008

TABLE II Based on the test results, some conclusions can be made.


AVERAGE BOUNCE TIMES
1) The bounce-reduction strategy lowered the bounce times
in all cases.
2) Bounce times were longer for dc control voltages than for
ac voltages.
3) There was a tendency of longer bounce as the control
voltage was increased.
4) The bounce-reduction algorithm allowed a reduc-
tion of the average mechanical bounce of approxi-
mately 30%.
The performance of the algorithm was worse for dc control
voltages. Although the detection of contactor closing is rela-
tively simple in the case of dc control voltages, the coil-current
waveform presents the unfavorable characteristic that the fall of
the current is almost immediately followed by the first contact
bounce. On the other hand, for ac control voltages, the fall of
the current envelope can be detected earlier, which is relative to
the first bounce. This fact by itself already explains the inferior
As can be seen in Fig. 12, the transition between Ilim and performance for dc voltages, although other factors may have
Ihold includes a “preholding” intermediate stage. This exists contributed to it as well.
to ensure that even after the turn off of the auxiliary switch
(and, consequently, the rise of the equivalent shunt resistance), VI. L IMITATIONS OF THE P ROPOSED A LGORITHMS
the voltage on the current-measurement node is within safe
limits for the microcontroller. Therefore, the auxiliary switch The bounce-reduction strategy proposed in this article funda-
is actually turned off only after the coil current decreases and mentally depends on the fast detection of the instant in which
remains for a few switching cycles at the Iph level. The instant the contactor armature starts moving. Some factors, however,
the auxiliary switch turns off is referred to as thold . t4 is may limit the efficiency of the closing-detection algorithm.
the time interval during which the coil current remains at the One of these is related to the validity of the current samples
closing level so as to avoid contactor reopening, as mentioned taken by the microcontroller. The absence of a large filter
earlier. capacitance in the power circuit bus results in a discontinuous
conduction of the bridge rectifier diodes during the instants of
zero crossing of the control voltage. In other words, when the
V. P ERFORMANCE OF THE instantaneous control voltage is null, so is the bus voltage, and
B OUNCE -R EDUCTION T ECHNIQUE there is no energy transfer to the coil. At this moment, part of
To verify the effectiveness of the adopted bounce-reduction the coil current goes through the free-wheeling diode.
technique, a series of tests was performed as follows. Although the main switch remains closed, it only conducts
part of the total coil current. Therefore, the measurement of the
1) For each of the five dc control-voltage values, ten closing coil current by means of the shunt resistor is not accurate. Thus,
sequences were made with the bounce-reduction algo- the current measurement at the instants close to the zero cross-
rithm in place, and in other ten make operations, the ings should be ignored, or else a false closing detection might
algorithm was deliberately inhibited (i.e., the transition occur. Fig. 13 shows an example of the current measurement
stage was suppressed). near the zero crossing of the control voltage.
2) The same set of tests was repeated for the ac control Since the algorithm does not consider all current measure-
voltages in a total of 200 making operations. In the case of ments to be valid, the aforementioned measurement errors may
ac control voltages, there was no syncing with the voltage delay the closing detection. During the tests, it was observed
waveform whatsoever. that this delay—in some making operations—was enough to
3) One of the three contact pairs was connected in series allow the first contact collision before the closing detection
with a circuit containing a small resistive load and a dc took place. In these situations, the total bounce time was longer
power supply, allowing the observation of the mechanical than average. Either way, even under unfavorable conditions,
bounces during the contact closing. the contact bounce was never worse than if no form of control
4) The total bounce time was measured during each make was used for the coil current.
operation. The operations during which the second Another issue regarding the closing-detection-algorithm ef-
bounce phenomena took place were ignored, and the ficiency is related to the coil-current-waveform analysis. In
operation was repeated in that case. the case of ac control voltages, for the contactor closing to
be adequately detected, the current measurements must exhibit
For each test condition, the arithmetic averages of the bounce a consistent fall compared to the values that were previously
times were calculated and are shown in Table II. stored in the microcontroller memory (reference waveform).
DE MORAES AND PERIN: CONTROL UNIT FOR REDUCING CONTACT BOUNCE IN ELECTROMAGNETIC CONTACTORS 869

ble for the contactor closing. Furthermore, in comparison to the


control systems that act directly on the coil voltage, the current
control is not sensitive to the typical electrical resistance vari-
ations that the winding exhibits over temperature. In addition,
a unit equipped with such a current control system can use a
single coil on a wide range of control voltages, both ac and dc,
whereas a conventional contactor would require different coils
to cover the same range, meaning increased design, production,
and storekeeping costs.
Among the algorithms that were specifically developed in
this paper are the variable-limit current control strategy and
the closing-detection algorithm. It is important to mention that,
although the efficiency of the bounce-reduction technique has
been experimentally verified under several different conditions,
Fig. 13. Control voltage and current-measurement voltage waveforms. it is subject to limitations on certain circumstances.

ACKNOWLEDGMENT
What is considered consistent is the succession of a minimum
amount of consecutive samples that are relatively smaller than The authors would like to thank WEG S.A. and its engineers
the ones stored in the memory. These measures avoid occa- for having supplied the contactor (model CWM112) used in
sional current waveform variations—not actually associated the experiments and also for providing invaluable literature and
to the armature movement—from being considered by the information for the research.
detection algorithm. Moreover, it was seen that the control-
voltage zero crossing is used for synchronizing the reference R EFERENCES
waveform. However, in practice, this zero-crossing detection [1] I. D. Kim, “Module-type switching rectifier for cathodic protection of un-
is also subject to small variations, and it has a limited preci- derground and maritime metallic structures,” IEEE Trans. Ind. Electron.,
sion. Thus, it is possible that the relative time instants of the vol. 52, no. 1, pp. 181–189, Feb. 2005.
[2] J. Rodríguez, P. W. Hammond, J. Pontt, R. Musalem, P. Lezana, and
reference waveform and the current-measurement sample being M. J. Escobar, “Operation of a medium-voltage drive under faulty con-
compared against it do not correspond. This is another reason ditions,” IEEE Trans. Ind. Electron., vol. 52, no. 4, pp. 1080–1085,
that makes it essential to establish a certain amount of toler- Aug. 2005.
[3] X. Zhou, L. Zou, and E. Hetzmannseder, “Asynchronous modular contac-
ance in the comparison. Obviously, these two constraints—the tor for intelligent motor control applications,” in Proc. 51st Meeting IEEE
tolerance in the comparison between the current measurements Holm Conf. Elect. Contacts, 2005, pp. 55–62.
and the reference waveform, and the demand for a mini- [4] J. H. Choi, J. M. Kwon, J. H. Jung, and B. H. Kwon, “High-performance
online UPS using three-leg-type converter,” IEEE Trans. Ind. Electron.,
mum amount of consecutive measurements that are inferior vol. 52, no. 3, pp. 889–897, Jun. 2005.
to the reference—result in a minimum time interval for the [5] Y. Kawakami, M. Takashima, M. Hasegawa, Y. Watanabe, and K. Sawa,
closing detection to take place. If the comparison tolerance “An evaluation method of the contact erosion based on the DC mode
test of electromagnetic contactor,” in Proc. 50th IEEE Holm Conf. Elect.
and/or the minimum amount of samples are reduced, the time Contacts, 22nd Int. Conf. Elect. Contacts, 2004, pp. 83–89.
spent between the closing detection and the first contact touch [6] T. Roschke, “Electronic control of electromagnetic contactors,” in Proc.
will be smaller but at the cost of higher probability of early 19th ICEC, 1998, pp. 295–299.
[7] G. Griepentrog, F. Kalvelage, N. Elsner, and N. Mitlmeier, “Increase
misdetections. of lifetime of electromagnetically actuated contactors by avoiding self-
synchronization,” in Proc. 50th IEEE Holm Conf. Elect. Contacts, 22nd
Int. Conf. Elect. Contacts, 2004, pp. 408–415.
VII. C ONCLUSION [8] J. W. McBride, “Electrical contact bounce in medium duty con-
tacts,” in Proc. 34th Meeting IEEE Holm Conf. Elect. Contacts, 1988,
This paper presented a low-cost microcontroller-based pp. 141–149.
[9] H. Manhart and W. Rieder, “Erosion behavior and ‘erodibility’ of Ag/CdO
control-unit fit for conventional electromagnetic contactors, and Ag/SnO2 contacts under AC 3 and AC 4 test conditions,” IEEE
together with relevant parts of the control algorithm. As can Trans. Compon., Hybrids, Manuf. Technol., vol. 13, no. 1, pp. 56–64,
be seen in Fig. 3, the circuitry is simple and inexpensive and Mar. 1990.
[10] W. Rieder and V. Weichsler, “Make erosion on Ag/SnO2 and Ag/CdO
can be used with contactors of various sizes requiring minimum contacts in commercial contactors,” in Proc. 36th IEEE Holm Conf. Elect.
changes in the power components. The microcontroller brings Contacts, 15th Int. Conf. Elect. Contacts, 1990, pp. 110–116.
flexibility to the design, allowing additional features (e.g., con- [11] H. Nouri, N. Larsen, and T. S. Davies, “Contact bounce simulation
using Matlab,” in Proc. 43rd IEEE Holm Conf. Elect. Contacts, 1997,
tactor control using low-power signals from industry-standard pp. 284–288.
buses) to be easily incorporated. [12] X. A. Morera and A. G. Espinosa, “Modeling of contact bounce of
The decision to use the coil current as the control variable AC contactor,” in Proc. 5th Int. Conf. Elect. Mach. Syst., 2001, vol. 1,
pp. 174–177.
brought the following advantages: It can be measured by means [13] J. W. McBride and S. M. A. Sharkh, “Electrical contact phenomena
of low-cost shunt resistors, and the current-limit feature gives during impact,” in Proc. 37th IEEE Holm Conf. Elect. Contacts, 1991,
the circuit an intrinsic protection against faults such as a coil pp. 132–140.
[14] B. Z. Sandler and A. A. Slonim, “Experimental investigation of relay
short-circuit condition. Moreover, the coil current is directly contact dynamics,” IEEE Trans. Compon., Hybrids, Manuf. Technol.,
associated with the magnetic attraction force, which is responsi- vol. CHMT-3, no. 1, pp. 150–158, Mar. 1980.
870 IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS, VOL. 55, NO. 2, FEBRUARY 2008

[15] A. A. Slonim, “Bouncing of contacts under current load (the influence Paulo Mário dos Santos Dias de Moraes was
of mechanical system parameters and load current on the closing process born in Rio de Janeiro, Brazil, in 1977. He re-
of electrical contacts),” IEEE Trans. Compon., Hybrids, Manuf. Technol., ceived the B.E. and M.Sc. degrees in electrical en-
vol. CHMT-10, no. 1, pp. 122–126, Mar. 1987. gineering from the Universidade Federal de Santa
[16] W. Li, J. Lu, H. Guo, W. Li, and X. Su, “AC contactor making speed Catarina, Florianópolis, Brazil, in 2002, and 2004,
measuring and theoretical analysis,” in Proc. 50th IEEE Holm Conf. Elect. respectively.
Contacts, 22nd Int. Conf. Elect. Contacts, 2004, pp. 403–407. He is currently working with Microsol Tecno-
[17] W. Shujuan, R. Wanbin, and F. Weiwei, “Simulation of spring logia S/A, one of the leading manufacturers of un-
system’s operation process for electromagnetic relay,” in Proc. 50th interruptable power supply systems and automatic
IEEE Holm Conf. Elect. Contacts, 22nd Int. Conf. Elect. Contacts, 2004, voltage regulators in Brazil. His research interests
pp. 453–458. include power supplies for computer and telecom-
[18] T. Roschke, “Indirect position measurement on electromagnetic actua- munications equipment and electromagnetic compatibility of power supplies
tors,” in Proc. 5th Int. Conf. New Actuators, 1996, pp. 441–444. in general.
[19] F. Shuhua and L. Heyun, “Magnetic field analysis and control circuit
design of permanent magnet actuator for AC contactor,” in Proc. 5th Int.
Conf. Elect. Mach. Syst., 2005, vol. 1, pp. 280–283.
[20] D. Chen, Y. Liu, and W. Tong, “Dynamic characteristic analysis and
optimization for the energy-saving and bounce-reducing double-coil con-
tactor,” IEICE Trans. Electron., vol. E89-C, no. 8, pp. 1194–1200,
Aug. 2006. Arnaldo José Perin (M’86) was born in Nova Prata,
[21] C. T. Chi, “A study of closing adaptive control in electronically con- Brazil, in 1953. He received the B.E. degree in elec-
trolled intelligent contactor,” in Proc. IEEE Region 10 Conf. TENCON, tronic engineering from the Pontifícia Universidade
Nov. 2006, pp. 1–4. Católica do Rio Grande do Sul, Porto Alegre, Brazil,
[22] T. S. Davies, H. Nouri, and F. W. Britton, “Towards the control of contact in 1977, the M.Sc. degree in electrical engineering
bounce,” IEEE Trans. Compon., Packag., Manuf. Technol. A, vol. 19, from the Universidade Federal de Santa Catarina,
no. 3, pp. 353–359, Sep. 1996. Florianópolis, Brazil, in 1980, and the Dr.Ing. de-
[23] J. H. Kiely, H. Nouri, F. Kalvelage et al., “Development of an applica- gree from the Institut National Polytechnique de
tion specific integrated circuit for reduction of contact bounce in three Toulouse, Toulouse, France, in 1984.
phase contactors,” in Proc. 46th IEEE Holm Conf. Elect. Contacts, 2000, Since 1980, he has been with the Department of
pp. 120–129. Electrical Engineering, Power Electronics Institute,
[24] X. Zhihong and Z. Peiming, “Intelligent control technology of AC con- Federal University of Santa Catarina, Florianópolis, where he is currently
tactor,” in Proc. IEEE/PES Transmiss. Distrib. Conf. Exhib.: Asia Pac., engaged in education and research on power electronics analysis and design.
2005, pp. 1–5. His research interests include power electronics, modulation, ac converters, and
[25] J. W. McBride and S. M. A. Sharkh, “The effect of contact opening power-factor correction. Since 1993, his research interest has been concentrated
velocity and the moment of contact opening on the AC erosion of more specifically on electronic ballast to use with fluorescent lamps and with
Ag/CdO contacts,” in Proc. 39th IEEE Holm Conf. Elect. Contacts, 1993, high-intensity discharge lamps. In this area, many researches were published in
pp. 87–95. Brazilian and International conferences.

View publication stats

You might also like