Durham Research Online

Deposited in DRO:
19 May 2016

Version of attached le:

Accepted Version

Peer-review status of attached le:

Peer-reviewed

Citation for published item:

Widmer, J.D. and Martin, R. and Spargo, C.M. and Mecrow, B.C. and Celik, T. (2012) 'Winding
congurations for a six phase switched reluctance machine.', in 2012 XXth International Conference on
Electrical Machines (ICEM). Piscataway, NJ: IEEE, pp. 532-538.

Further information on publisher's website:

http://dx.doi.org/10.1109/ICElMach.2012.6349921

Publisher's copyright statement:

c 2012 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in

any current or future media, including reprinting/republishing this material for advertising or promotional purposes,

creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of

this work in other works.

Additional information:

Use policy
The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for
personal research or study, educational, or not-for-prot purposes provided that:
• a full bibliographic reference is made to the original source
• a link is made to the metadata record in DRO
• the full-text is not changed in any way
The full-text must not be sold in any format or medium without the formal permission of the copyright holders.
Please consult the full DRO policy for further details.

Durham University Library, Stockton Road, Durham DH1 3LY, United Kingdom
Tel : +44 (0)191 334 3042 | Fax : +44 (0)191 334 2971
https://dro.dur.ac.uk
 Winding Configurations for a Six Phase
Switched Reluctance Machine
J. D. Widmer*, R. Martin*, C. M. Spargo*, B. C. Mecrow* and T. Celik†

Abstract – Winding configurations are investigated and deriving a six phase unipolar supply from a conventional
evaluated for a six phase, 12-10 switched reluctance machine three phase inverter, investigating winding options in both
driven by a conventional three phase full bridge converter with conventional and segmental rotor SRMs.
the addition of six diodes. A new winding configuration is
proposed and shown to develop more torque with less torque
ripple than a conventional winding in this application. Finite In this paper, the six phase prototype machine and adapted
element modelling is used to investigate the electromagnetic three phase drive configuration examined in [1] are first
behaviour and compare the output of different winding summarised. Different winding patterns are then investigated
configurations. Some initial experimental tests are described in and a new configuration proposed, which utilises short flux
verification of the modelling predictions. The novel drive and paths and relies on some mutual interaction between phases
new winding configuration offer significant advantages over a
for torque production. Finite Element (FE) modelling is used
standard three phase machine and drive, giving increased mean
torque with lower torque ripple and acoustic noise, as well as to investigate the electromagnetic behaviour, illustrating the
reduced converter complexity and potentially cost. various flux paths and predicting the current, flux linkage,
and torque characteristics arising in the prototype machine
Index Terms-- AC machines, Brushless machines, Motor from different winding arrangements. Experimental tests on
drives, Rotating machines, Rotors, Stator windings, Torque, the prototype machine are described and the results used to
Variable speed drives, Windings, Reluctance motors verify the FE predictions. It is concluded that the new
winding arrangement produces more torque with less torque
I. INTRODUCTION ripple than more conventional options and that the resultant

I T has been shown in a previous paper [1] that a six phase

switched reluctance machine (SRM) can be operated from
a three phase full bridge converter through the simple
topology generates torque through a combination of self and
mutual inductance variation.

addition of six rectifier grade diodes. This arrangement was

shown to compare favourably with a three phase SRM driven II. THE PROTOTYPE MACHINE AND DRIVE
from a conventional asymmetric half bridge, offering the
following features: Fig. 1 shows how six antiparallel diodes convert the
 Standard three phase inverter drive; bipolar current output from each phase of the three phase
 Only three connections between motor and drive; inverter into two unipolar half waveforms. The machine was
connected in star for the FE modelling and experimental tests
 Only two current sensors; described in this paper, but both star and delta connections
 Low torque ripple; have been shown to work in practice [1].
 No increase in motor loss; and
 Very similar converter VA rating.
3 Ph
Inverter
The electrical winding of the prototype six phase machine Ph 1
described in [1] can be arranged in various configurations. Ph 4
Alternative SRM winding configurations and flux paths have
previously been reported. Miller [2] described short flux Ph 6
paths arising from paired groupings of stator poles, as well as Ph 5
from four-pole field configurations, giving the example of a Ph 2
three phase 12/8 machine which is effectively a 6/4 machine Ph 3

with a ‘multiplicity’ of 2. Michaelides and Pollock described

five and seven phase machines with short flux paths arising
from adjacent stator poles with opposing magnetic polarities, Fig. 1: Six phase SRM driven by a three phase full bridge converter
[3]. Fully pitched windings giving rise to torque production
entirely from changing mutual inductance between phases The design of the SRM for this application was based on
were introduced by Mecrow [4]. Later, Mecrow et al [5] standard best practice for a conventional, doubly-salient
described a novel, segmental rotor SRM where stator slots SRM having twelve stator teeth (two per each of six phases),
contain only the winding of a single phase, and segments on and ten rotor teeth. In order to maximise the torque
capability, the tooth width to rotor pole pitch ratio was set to
the rotor modulate the permeance of short flux loops around
0.4. The core backs of the machine are relatively deep
individual slots. Liu et al [6] compared unipolar and bipolar
compared to the tooth width since three phases will be
excitations in SRMs and Celik [7] described the principal of
conducting at any given time in this drive configuration.
Increased core back depth also increases the stiffness of the
* J. D. Widmer, R. Martin, C. M. Spargo, and B. C. Mecrow are with the
machine and thereby helps to reduce acoustic noise due to
School of Electrical and Electronic Engineering, Newcastle University, UK torque ripple. Machine dimensions are summarised in Table
(e-mail: james.widmer@ncl.ac.uk). 1.
† T. Celik is with Dyson, UK (e-mail: tuncay.celik@dyson.com).
 TABLE 1
SIX PHASE SRM DESIGN PARAMETERS

Number of Stator Teeth 12

Number of Rotor Teeth 10
Axial Length (Lamination Stack) 150.0mm
Stator Outer Diameter 150.0mm
Stator Inner Diameter 91.4mm
Stator Core Back Depth 11.0mm
Stator Tooth Width 11.4mm
Airgap Length 0.3mm
Rotor Outside Diameter 90.8mm
Rotor Insider Diameter 36.0mm
Fig. 4: General distinction between ‘long’ flux paths (left) and ‘short’ flux
Rotor Coreback Depth 18.0mm
paths (right) in a six phase 12-10 SRM
Rotor Tooth Width 11.4mm
Turns per Phase 100
Winding each phase for conventional, ‘long’ flux paths
Photographs of the prototype machine are shown, fig. 2 would yield an arrangement where ten of the twelve slots
and fig. 3. would contain the ‘go’ conductors of one phase and the
‘return’ conductors of the adjacent phase. Hence this is
referred to as the ‘dot-cross’ configuration. With an even
number of phases and where phase MMFs reinforce it is not
possible to wind the machine so as to give symmetry, so two
of the twelve slots contain ‘dot-dot’ and ‘cross-cross’
orientations. However, unipolar six phase operation would
give rise to a field consisting of predominantly ‘long’ flux
paths. This is referred to as the asymmetric ‘dot-cross’
winding and is illustrated in fig. 5 where the dotted line
indicates the discontinuity in the pattern.

‘dot-cross’ ‘dot-dot’
Asymmetric

Fig. 2: Constructed six phase prototype SRM

Symmetric

Fig. 3: Rotor from the six phase prototype SRM

III. WINDING CONFIGURATIONS

The stator of the prototype machine has twelve teeth and
Fig. 5: Illustration of possible six phase winding configurations in the 12-10
so six phases comprise two tooth-wound coils each. This prototype SRM where the dashed line indicates the discontinuity in the
gives a number of possibilities for winding the machine and winding pattern arising in the cases where phase MMFs reinforce
the prototype was built with this in mind, having
interchangeable coil connections as well as oversized stator ‘Short’ flux paths can be realised by simply reversing the
and rotor core backs. orientation of every second coil in the conventional
configuration so as to yield a predominantly ‘dot-dot’
Conventionally, the two coils of each phase would be configuration. Again, where phase MMFs reinforce, the six
connected in series such that the resulting MMFs reinforced phase configuration is asymmetrical and in this case,
each other. Thus, single phase energisation would give rise unipolar six phase operation would give rise to a field
to ‘long’ flux paths crossing the rotor and utilising the full consisting of predominantly ‘short’ flux paths. This is
rotor core back so as to give a two-pole field. Alternatively referred to as the asymmetric ‘dot-dot’ winding and is
‘short’ flux paths, resulting from other winding similarly illustrated in fig. 5.
configurations, are defined as paths where the flux does not
fully cross the rotor but rather makes shorter loops utilising In both configurations described, the asymmetry in the
more proximate teeth for the return path. The general conductor arrangement gives rise to discontinuities in the
distinction is illustrated in fig. 4, and four specific winding winding pattern which could cause excess torque ripple as
configurations are illustrated in fig. 5 and described below. well as localised saturation in the stator and rotor core backs.
Michaelides and Pollock [3] stated that: “Motors with an
even number of phase windings cannot be wound to ‘dot-cross’ ‘dot-dot’
successfully implement the short flux loops. There are
always discontinuities in the flux pattern, forcing some flux
across the rotor”, and went on to suggest that five or seven
phase machines are better suited to short flux paths.

Asymmetric
However, in the case of the six phase machine and drive
under consideration here, the authors believe that this
asymmetry may be avoided by connecting the two coils of
each phase such that the resultant MMFs oppose. In this
case, single phase energisation would give rise to ‘short’ flux
paths in both ‘dot-dot’ and ‘dot-cross’ configurations. Hence
this gives the two ‘symmetric’ windings shown in fig. 5.

Symmetric
IV. FE MODELLING
The prototype machine was modelled using a
commercially available FE package. Initially, magnetostatic
models were developed in 2D for the visualisation of the
various winding configurations. Assuming sinusoidal current
waveforms for simplicity, the six phase machine driven by a Fig. 7: Comparison of FE derived flux lines for different winding
three phase bridge as described in a previous paper [1] can configurations.
be modelled magnetostatically in different states of
energisation by recognising that each phase ideally conducts A transient FE solution with rotation at 1000rpm was used
a half-sinewave and therefore only three adjacent phases to show how these field patterns rotate with sinusoidal
conduct at any one time, as shown in fig. 6. current sources of 28A peak, which roughly equates to a
current density of 10A/mm2. This demonstrated the effect of
asymmetry which gave similar effects in both ‘dot-dot’ and
‘dot-cross’ configurations, consisting of the predominant
field rotating for the majority of the mechanical cycle, with a
temporary switch to the counterpart field at the point of
discontinuity. Transient FE results also yielded the torque
output for the different configurations, shown in fig. 8.

Fig. 6: Illustration of three phase bipolar currents (dotted lines) and the six
phase unipolar currents arising from this drive configuration (dashed lines)
with one of the six phases shown with a heavy solid line for clarity

Considering the instant where a given phase is in the

position of maximum torque production (e.g. as shown by
the vertical line at 90 degrees electrical in fig. 6): that phase
would conduct peak current and the two adjacent phases
would conduct half the peak current. Fig. 7 compares flux
plots arising from such energisation with different winding Fig. 8: FE simulated torque output for different winding configurations
configurations, where the phase receiving peak current is on driven from a three phase bridge under sinusoidal current control
the horizontal with the rotor in the position for maximum
torque from that phase. Some initial observations can be Fig. 8 confirms that the symmetric ‘dot-cross’ arrangement
made. Firstly, the symmetric ‘dot-cross’ may be discounted produces no useful torque. On the basis of the FE flux
as flux lines with regard to the rotor position in fig. 7 density illustrations (not shown) it was suggested that the
indicate that the torques developed at different teeth act in stator core back was a limiting factor in the ‘dot-cross’
almost perfect opposition. Secondly, both of the ‘dot-dot’ configuration. Fig. 8 shows the associated reduction in
configurations appear to be quite similar, giving rise to average torque and increase in torque ripple. Stator core
predominantly short flux paths, with flux density shading back saturation may also explain the slight reduction in
(not shown) indicating saturation in the teeth as a limiting torque in the asymmetric ‘dot-cross’ configuration by
factor on the flux. Finally, the asymmetric ‘dot-cross’ gives comparison with the symmetric version; namely that the
rise to predominantly long flux paths with flux density temporary switch to long flux paths at the discontinuity in the
shading (not shown) indicating lower flux density levels in winding configuration gives a temporary reduction in
the teeth, with the stator core back appearing to be the average torque. This was verified by repeating the simulation
limiting factor.
for the asymmetric winding options, with the stator core back
thicknesses increased by 50%. The torque waveforms from
the three configurations were very similar, with the new
symmetric ‘dot-dot’ arrangement still producing slightly
more torque. Hence it is concluded that, on the basis of this
FE modelling with sinusoidal current control, the new
arrangement is preferable in the case of this six phase SRM
driven from a three phase bridge, since the torque output can
be achieved with a reduced stator core back thickness and
therefore a smaller overall machine. In order to understand
this new winding configuration, further FE simulation work
was carried out.

Fig. 9 shows the flux linkage versus current (psi-i) Fig. 10: FE simulated psi-i characteristics for a single phase with the
characteristics measured in a single phase in this new asymmetric ‘dot-cross’ winding configuration, showing static
winding arrangement. The dynamic loop under single phase characteristics in the unaligned and aligned positions (dotted lines), a single
operation is what might conventionally be expected with phase dynamic loop (dashed line), and a dynamic loop arising from full six
phase operation (solid line)
reference to the static characteristics. However, the dynamic
loop arising from full six phase operation is rather different,
indicating significant mutual effects whereby adjacent phases
increase the flux linkage in a given phase thus increasing the
torque output. This mutual activity can clearly be seen in the
flux plots of fig. 7 where those arrangements with short flux
paths exhibit considerable interaction between phases. The
dynamic loops in fig. 9 also indicate that the peak current
under six phase operation exceeds that under single phase
operation; this is considered further, below.

a - static characteristics

Fig. 9: FE simulated psi-i characteristics for a single phase with the

symmetric ‘dot-dot’ winding configuration, showing static characteristics
in the unaligned and aligned positions (dotted lines), a single phase
dynamic loop (dashed line), and a dynamic loop arising from full six phase
operation (solid line)
b - dynamic loops under single phase operation
It was previously explained that the thickness of the stator
core back was a limiting factor in the torque output of the
asymmetric ‘dot-cross’ configuration and that increasing the
core back thickness in the FE model caused the torque output
to equal that of the new winding configuration. Since that
configuration was shown to give predominantly ‘long’ flux
paths with little apparent mutual interaction, some key
differences should be apparent from the psi-i characteristics.
Fig. 10 shows the psi-i characteristics for the conventional
‘dot-cross’ configuration and fig. 11 makes comparisons with
the new configuration. In order to avoid the limitation of
core back saturation and thus facilitate a useful comparison,
the ‘dot-cross’ version was again modelled with the stator
core back thickness increased by 50%.
c - dynamic loops under full six phase operation

Fig. 11: FE simulated psi-i comparisons for a single phase between the
asymmetric ‘dot-cross’ (dotted line) and the symmetric ‘dot-dot’ (solid line)
winding configurations
 The following observations can be made from fig. 10 and case of this six phase SRM driven from a three phase bridge.
fig. 11: The next section describes some experimental tests of the
 In the asymmetric ‘dot-cross’ case (fig. 10) the single prototype machine with different winding configurations.
phase dynamic loop remains within the static aligned and
unaligned characteristics;
 Similarly, the full six phase dynamic loop (fig. 10) V. EXPERIMENTAL TESTING OF THE PROTOTYPE MACHINE
remains within the static characteristics and encloses a The prototype six phase SRM was tested in the laboratory
similar area to the single phase loop whilst having a with the different winding configurations described above,
slightly different loci; all in star connection and with a floating star point.
 The symmetric ‘dot-dot’ configuration has slightly
inferior static characteristics (fig. 11a) and a smaller A Control Techniques SP3410 three phase drive was used
single phase dynamic loop (fig. 11b) owing to the under speed control operation. The drive is rated at 18kW
increased reluctance of the short flux paths where the air and has a dc link voltage of 560V and a switching frequency
gap is crossed twice per coil; and of up to 16 kHz. In the absence of a directly relevant setting,
 The symmetric ‘dot-dot’ configuration exhibits a larger the drive was configured to feed a twenty-pole permanent
six phase dynamic loop (fig. 11c) which arises from self magnet machine, with ten magnet pole-pairs being emulated
plus additional mutual effects and accounts for the by the ten rotor teeth. It was then possible to ‘autotune’ the
increased torque capability overall. drive parameters as a basis for testing each winding
configuration, although manual adjustment of the encoder
It is concluded that, whilst the two machines are broadly phase offset angle was required so as to minimise the motor
similar (allowing for extra back iron in the ‘dot-cross’ case), phase currents for a given torque/speed operating point.
there are subtle differences in the mode of operation, giving
rise to slightly improved torque in the symmetric ‘dot-dot’ Initially, the symmetric ‘dot-cross’ configuration was
configuration through mutual interaction between phases. tested. This was predicted to produce zero useful torque on
Hence it is suggested that the new winding configuration is the basis of FE modelling and consideration of the flux paths
superior (giving enhanced torque capabilities for a smaller present with respect to the rotor teeth. In practice, the drive
machine), and that the conventional approach of ignoring did manage to develop a small amount of torque although the
mutual interaction between phases is questionable in this machine was acoustically very noisy and the phase currents
particular case. quickly exceeded the rating.

Lastly, it is clear from fig. 10 and fig. 11 that the dynamic For each of the remaining three winding configurations,
loop from six phase operation reaches higher current levels the machine was tested from standstill up to 4000rpm. The
than both the single phase loop and the peak value of 28A load at each operating speed was steadily increased until
from the current source. This must be a result of circulating torque could not be sustained. The results are shown in fig.
currents between the two, 180 electrical degree separated 13. It should be noted that the limit of the load machine is
phases sharing a single leg of the three phase bridge. Fig. 12 40Nm and this restricts the measurable performance of the
illustrates this. The resultant phase currents in fig. 12 do sum ‘dot-dot’ configurations at low speed. In the case of the
to the current source but are not perfect half sine waves, thus conventional ‘dot-cross’ winding, the performance at low
indicating the presence of currents circulating between the speed was limited by the maximum current of the inverter
phases. drive.

Fig. 12: FE simulated current plots for the six phase symmetric ‘dot-dot’
configuration driven from a three phase bridge, showing a single sinusoidal Fig. 13: Measured peak torque capabilities of the six phase prototype driven
current source (dotted line), and the resultant 180 degree separated phase from a three phase full bridge with different winding configurations (note
currents (solid line) 40Nm load torque limit)

In summary, the FE modelling has given some insight into The results of these initial load tests broadly confirm what
the different winding configurations both in predicting the the FE modelling suggested, namely that the torque
performance and assisting with the understanding of the capabilities of both ‘dot-dot’ configurations are similar and
differences. The new symmetric ‘dot-dot’ winding that the conventional ‘dot-cross’ winding is inferior in this
arrangement described here is the preferred option in the context owing to stator core back saturation. It was also
observed that the conventional ‘dot-cross’ winding was VII. REFERENCES
acoustically more noisy by comparison with the ‘dot-dot’ [1] J. D. Widmer, B. C. Mecrow, C. M. Spargo, R. Martin, and T. Celik,
configurations. With reference to the FE modelling, it is “Use of a 3 Phase Full Bridge Converter to drive a 6 Phase Switched
expected that this excess noise arises from the increased Reluctance Machine,” IET PEMD Conference, Bristol, 2012, to be
published
torque ripple, again as a result of stator core back saturation. [2] T.J.E. Miller, “Switched Reluctance Motors and their Control,”
Clarendon Press, Oxford, 1993
Hall-effect current probes were also used to capture the [3] A.M. Michaelides and C. Pollock, “A new magnetic flux pattern to
line current and the two associated phase currents arising improve the efficiency of the switched reluctance motor,” IEEE IAS
from the anti-parallel diode set up. Sample waveforms are Annual Meeting, Houston, Texas, pp. 226-233, 1992.
[4] B.C. Mecrow, “New Winding Configurations for Doubly Salient
shown in fig. 14 for illustration of the concept and Reluctance Machines,” IEEE Transactions on Industry Applications,
comparison with the FE modelled waveforms shown in fig. Vol. 32, No. 6, pp. 1348-1356, Nov/Dec. 1996.
12. [5] B.C. Mecrow, J.W. Finch, E.A. El-Kharashi, and A.G. Jack,
“Switched reluctance motors with segmental rotors,” IEE Proc.-
Electr. Power Appl., Vol. 149, No. 4, pp. 245-254, Jul. 2002
[6] X. Liu, Z.Q. Zhu, M. Hasegawa, A. Pride, R. Deohar, T. Maruyama,
Z. Chen, “Performance comparison between unipolar and bipolar
excitations in switched reluctance machine with sinusoidal and
rectangular waveforms”, IEEE Energy Conversion Congress and
Exposition, pp. 1590-1595, 2011
[7] T. Celik, “Segmental Rotor Switched Reluctance Drives”, PhD
Thesis, Newcastle University, August 2011

VIII. BIOGRAPHIES

James Widmer joined Newcastle University in December 2009 from a

senior position in the aerospace industry. James coordinates the Newcastle
University Centre for Advanced Electrical Drives, which provides industry
with expert design and research services in electrical machines and their
associated electronics. He also undertakes research into novel electrical
Fig. 14: Measured current waveforms for the six phase symmetric ‘dot-dot’
machines. James has a MEng in Electrical and Electronic Engineering from
configuration driven from a three phase bridge and developing 10Nm
the University of Bristol and is in the process of completing a PhD in
torque at 500rpm. The line current and the two associated motor phase
electrical machines.
currents are shown.

Again, it was verified that the resultant phase currents do Richard Martin obtained MEng and PhD degrees from Durham University
and is now a Research Associate in the Power Electronics, Drives and
sum to the line current from the drive, but the non-sinusoidal Machines Research Group at Newcastle University.
nature of the motor phase currents indicates the presence of
currents circulating between the phases. Christopher Spargo was awarded a BEng (Hons) degree in Electrical and
Electronic Engineering from Newcastle University in 2011; he is currently
a PhD student in the Power Electronics, Machines and Drives research
group at the same university in the School of Electrical, Electronic and
VI. CONCLUSIONS Computer Engineering. His research interests include switched and
A prototype 12-10 six phase SRM driven from a three synchronous reluctance machines.
phase full bridge inverter has been described. Potential
winding configurations for this machine have been illustrated Barrie Mecrow is Professor of Electrical Power at Newcastle University
and is also the Head of Power Electronics, Drives and Machines Research
and investigated using FE modelling. Modelling results were Group. His research interests include permanent magnet and reluctance
verified through some initial experimental tests on the machines and drives for aerospace, automotive and consumer products.
prototype.
Tuncay Celik has been with Dyson Ltd. since 2006. He did his MSc degree
In particular, a new winding arrangement has been at the University of Newcastle in 2001 for which he received British
proposed and shown to develop more torque with less torque Council scholarship. He was then sponsored by the UK Overseas Research
Scholarship (ORS) for his PhD work at the University of Newcastle. His
ripple than a conventional arrangement. Investigation of the PhD research was on the six-phase segmental rotor switched reluctance
electromagnetic behaviour of this machine has shown that drives. His research interests are design and control of novel PM and SR
torque is developed through a combination of self and machines.
mutual inductance variation.