Windings and Axes

1.0 Introduction
In these notes, we will describe the different windings on
a synchronous machine. We will confine our analysis to
two-pole machines of the salient pole rotor construction.
Results will be generalizable because
• A machine with p>2 poles will have the same
phenomena, except p times/cycle.
• Round rotor machines can be well approximated using a
salient pole model and proper designation of the
machine parameters.
We will also define an important coordinate frame that we
will use heavily in the future.
2.0 Defined axes
The magnetic circuit and all rotor winding circuits (which
we will describe shortly) are symmetrical with respect to
the polar and inter-polar (between-poles) axes. This
proves convenient, so we give these axes special names:

1
 • Polar axis: Direct, or d-axis
• Interpolar axis: Quadrature, or q-axis.
The q-axis is 90° from the d-axis, but which way?
• Ahead?
• Or behind?
Correct modeling can be achieved either way, and some
books do it one way, and some another. We will remain
consistent with your text and choose the q-axis to lag the
d-axis by 90°, which is “consistent with the
recommendation and rationale of [15]1” (p. 92, VMAF).

1
IEEE Committee Report, Harrington, D. (chair), “Recommended phasor diagram for synchronous machines, IEEE
Trans. Power App. Syst. PAS-88: pp. 1593-1610, 1969.

2
Fig. 1 is from your text, and shows the q-axis lagging the d-
axis, consistent with our assumption.
Fig. 2 is from Kundur, and shows the q-axis leading the d-
axis, which we will NOT do.

Fig. 1 Fig. 2
Example 5.1 in Chapter 5 is worked with the q-axis lagging
the d-axis; Example 5.2 reworks the same problem with
the q-axis leading the d-axis. Example 5.2 states that ‘”The
objective of this example is to illustrate, though the
calculations differ slightly, that the results are identical;
thus, the choice of which coordinate reference frame is
assumed is arbitrary, as long as calculations are consistent
with the approach.”
3
3.0 Physical windings
There are typically 5 physical windings on a synchronous
machine:
• 3 stator windings (a-phase, b-phase, and c-phase)
• 1 main field winding
• Amortissuer windings on the pole-faces
The stator windings and the field winding are familiar to
you based on the previous notes. The amortissuer winding
might not be, so we briefly describe it.
Amortissuer is a French word for “shock absorber.” These
windings provide positive sequence damping under
transient conditions and so are good for stability2. Under
steady state, they are dead, i.e., they have no influence.
Amortissuer windings are not usually used on smooth-
rotor machines, because the solid steel rotor cores of such
machines provide paths for eddy currents and thus
produce the same effects as amortissuer windings3.
Amortissuer windings are often used in salient-pole
machines, but even when not, eddy currents in pole

2
E. Kimbark, “Power system stability, Vol. III: Synchronous Machines,” IEEE Press, 1995 (orig pub 1956), p. 219.
3
Ibid, p. 215.
4
faces contribute the same effect, although greatly
diminished.
Amortissuers have a number of other good effects, as
articulated by Kimbark in his Volume III book on
synchronous machines4 (he lists 10 such effects). For our
purposes, they affect machine subtransient data5.
Amortissuer windings are embedded in the pole-face (or
“shoe” of the pole) and consist of copper or brass rods
connected to end rings. They are similar in construction to
the squirrel cage of an induction motor.
Figures 3 (from Sarma) and 4 (from Kundur) illustrate
amortissuer windings. Note that they may be continuous
(Fig. 3a and Fig. 4) or noncontinuous (Fig. 3b).

Fig. 3

4
Ibid, p. 216.
5
Ibid, p. 217-219.
5
 Fig. 4
4.0 Modeled windings and currents
Although there are typically 5 physical windings on a
machine, we will model a total of 7, with associated
currents as designated below.
• 3 stator windings: ia, ib, ic
• Field windings: There are 2: one physical; one fictitious
o Main field winding: carrying current iF and
producing flux along the d-axis.
o G-winding: carrying current iG and producing flux
along the q-axis. This is the fictitious one, but it
serves to improve the model accuracy of the round-
rotor machine (by modeling the q-axis flux
produced by the eddy-current effects in the rotor
6
 during the transient period), and it can simply be
omitted when modeling the salient pole machine (in
salient pole machines, there is little q-axis flux
produced by the eddy current effect in the rotor).
The G-winding is like the F-winding of the main field,
except it has no source voltage in its circuit. Kimbark
suggests modeling it in his Vol. III, pg. 73.
• Amortissuer winding: This one represents a physical
winding for salient-pole machines with dampers, and a
fictitious winding otherwise. Because these produce flux
along both the d-axis and the q-axis, we model two
windings:
o d-axis: amortissuer winding carrying current iD
o q-axis: amortissuer winding carrying current iQ
It is important to understand the difference between the F
and G windings and the D and Q windings, respectively,
driven by the fact that D and Q windings have higher
resistance than F and G windings. Therefore:
• Both the F and D produce flux along the D-axis, but D is
“faster” (lower time constant or L/R ratio) than F.
• Both the G and Q produce flux along the Q-axis, but Q is
“faster” (lower time constant or L/R ratio) than G.

7
5.0 Flux linkages and currents
So we have seven windings (circuits) in our synchronous
machine. The flux linkage seen by any winding i will be a
function of
• Currents in all of the windings and
• Magnetic coupling between winding i and winding j, as
characterized by Lij, where j=1,…,7.
That is
7
i =  Lij ij (1)
j =1

For example, the flux linking the main field winding is:
F = LFa ia + LFb ib + LFc ic + LFF iF + LFD iD + LFQ iQ + LFG iG (2)

Repeating for all windings results in Equation (4.11) in

VMAF. (Note in older editions the G-winding is not
represented).

8
   a   Laa Lab Lac LaF LaG LaD LaQ   i 
 a 
    L Lbb Lbc LbF LbG LbD LbQ  ib
stator   b   ba  
  c   Lca 
Lbc Lcc LcF LcG LcD LcQ   ic 

    
 F  =  LFa LFb LFc LFF LFG LFD LFQ  iF 
 G   L  
    Ga
LGb LGc LGF LGG LGD LGQ  iG (3)
rotor 
D   LDa  i 
 LDb LDc LDF LDG LDD LDQ  D
     i 
 Q   LQa LQb LQc LQF LQG LQD LQQ   Q 

Note the blocks of the above matrix correspond to

• Lower right-hand 4×4 are rotor-rotor terms.
• Upper-left-hand 3×3 are stator-stator terms;
• Upper right-hand 3×4 are stator-rotor terms;
• Lower left-hand 4×3 are rotor-stator terms;
VMAF summarize the expressions for each of these groups
of terms on pp. 94-96. I expand on this summary in the
next section.
6.0 Inductance blocks
6.1a Rotor-rotor terms: self inductances
Recall (see eq (15) in notes called “Preliminary
Fundamentals) that the general expression for self-
inductance is

9
 iN i2
Lii = = (4a)
ii Ri
where Ri is the reluctance of the path seen by λi, given by
l
Ri = (4b)
A
where l is the mean length of the path, μ is the
permeability of the path’s material, and A is the cross-
sectional area of the path.

At any given moment, the stator and the rotor present a

constant reluctance path to flux developed by a winding
on the rotor, i.e., the reluctance path seen by any rotor
winding is independent of the “position” angle θ. This is
illustrated in Fig. 5 for the main field (F) winding.
Rotation Rotation
φF

φF
N S

Fig. 5a: θ=0° Fig. 5a: θ=90°

Fig. 5
10
Thus, since Lii=(Ni)2/R, rotor winding self-inductances are
constants, and we define the following nomenclature,
consistent with eq. (4.13) in VMAF.
• d-axis field winding LFF = LF (5)
• q-axis field winding LGG = LG (6)
• d-axis amortissuer winding: LDD = LD (7)
• q-axis amortissuer winding: LQQ = LQ (8)
Note your text’s convention of using only a single subscript
for constant terms.
6.1b Rotor-rotor terms: mutual inductances
Recall (eq. (15) in “Preliminary Fundamentals”) that:

i Ni N j
Lij = =
ij Rij (9)

where Rij is the reluctance of the path seen by λi in linking

with coil j or the path seen by λj in linking with coil i (either
way – it is the same path!). Again, by similar reasoning as
in section 6.1a, these mutual terms are constants (i.e.,
independent of θ).

11
Therefore, we have the following:

• F (field)–D (amort): LFD = LDF = M R (10a)

• G (field)-Q (amort): LGQ = LQG = M Y (10b)

But we have four other pairs to address:

• F (field)-G (field): LFG = LGF = 0 (11a)

• F (field)-Q (amort): LFQ = LQF = 0 (11b)

• G (field)-D (amort): LGD = LDG = 0 (11c)

• D (amort)-Q (amort): LDQ = LQD = 0 (11d)

The last four pairs of windings are each in quadrature, so
the flux from one winding does not link the coils of the
other winding, as illustrated in Fig. 6. Therefore the above
four terms are zero, as indicated in eqs (11a-11d).

Fig. 6
12
6.2a Stator-stator terms: self inductances
We can derive these rigorously (see Kundur pp. 61-65) but
the insight gained in this effort may not be great. Rather,
we may be better served by gaining a conceptual
understanding of four ideas, as follows:
1. Sinusoidal dependence of permeance on θ: Due to
saliency of the poles (and to field winding slots in a smooth
rotor machine), the path reluctance seen by the stator
windings depends on θ, as illustrated in Fig. 7.
Rotation Rotation
φa φa

a' a'
a a
N S

Fig. 7a: θ=0° Fig. 7b: θ=90°

Fig. 7
From Fig. 7a, we observe that when θ=0°, the path of
phase-a flux contains more iron than at any other angle
0→180°, and therefore the reluctance seen by the phase-

13
a flux in this path is at a minimum, and permeance is at a
maximum.
From Fig. 7b, we observe that when θ=90°, the path of
phase-a flux contains more air that at any other angle
0→180°, and therefore the reluctance seen by the phase-
a flux in this path is at a maximum, and permeance is at a
minimum. This suggests a sinusoidal variation of
permeance with θ.
2. Constant permeance component: There will be a
constant permeance component due to the amount of
permeance seen by the phase-a flux at any angle. This will
include the iron in the middle part of the rotor (indicated
by a box in Figs. 7a and 7b), the stator iron, and the air gap.
Denote the corresponding component as Ps.
3. Double angle dependence: Because the effects
described in 1 and 2 above depend on permeance (or
reluctance), and not on rotor polarity, the maximum
permeance occurs twice each cycle, and not once.
Taking (1), (2), and (3) together, we may write that
P = Ps + Pm cos 2 (12)
14
4. Inductance: Because L=N2/R=N2P, the self inductance
of the a-phase winding can be written as
Laa = Ls + Lm cos 2 (13)
Likewise, we will obtain:
Lbb = Ls + Lm cos 2( − 120) (14)
Lcc = Ls + Lm cos 2( − 240) (15)
Equations (13), (14), (15) are denoted (4.12) in your text.
Note that because θ is a function of t, then (13), (14), (15)
imposes that stator self-inductances are functions of t!
(Recall our discussion in “Preliminary Fundamentals”.) This
means in Faraday’s law, e=d(Li)/dt, L is not constant.
6.2b Stator-stator terms: mutual inductances
We will identify 3 important concepts for understanding
mutual terms of stator-stator inductances.
1. Sign:
First, we need to remind ourselves of a preliminary fact:
➔ For any circuits i and j, Lij is positive if positive currents
in the two circuits produce fluxes in the same direction.
15
With this fact, we can state important concept 1:
➔As a result of defined stator current directions, the
stator-stator mutual inductance is always negative.
To see this, we can observe that the flux produced by
positive currents of a and b phases are in opposite
directions, as indicated in Fig. 8.

“X” shows current into the plane; “●” shows

b' current out of the plane. RHR gives flux
φa
direction.

Observe that physical location of the b-phase

φab a' will cause its voltage to lag the a-phase
a
●
voltage by 120°, as, for counter-clockwise
φb φba (CCW) rotation, the “leading edge” of the
CCW-rotating mag field is seen first by the “a”
●b pole of the a-phase winding and then, 120°
later, by the “b” pole of the b-phase winding.
Fig. 8

Observe the following in Fig. 8:

• The component of flux from winding-a that links with
winding-b, φab, is 180° from φb.
• The component of flux from winding-b that links with
winding-a, φba, is 180° from φa.
The implication of the above 2 observations are that
16
➔Mutual inductance is negative.
➔This implies that mutually induced voltages are negative
relative to self induced voltages.
2. Function of position:
2a. Maximum Permeance for Mutual Flux:
Recall conditions where, for self-flux, the amount of iron in
the path yields a maximum permeance (minimum
reluctance) condition (remember θ specifies rotor location).
This condition for phase-a self-flux is θ=0°.
This condition for phase-b self-flux is θ=-60°.
Therefore the condition for maximum permeance for the
mutual flux between phases a and b (which maximizes flux
produced from one winding that links with the other
winding) is halfway between these two at θ=-30°6.
2b. Periodicity of Permeance for Mutual Flux:
Starting at the maximum permeance condition, a rotation
by 90° to θ=60° gives minimum permeance. See Fig. 9.

6 The thinking here is that if θ=0° results in max flux (min reluctance path) from a-current seen by a-winding, and if θ=-60° results
in max flux (min reluctance path) from b-current seen by b-winding, then halfway between the two will result in max flux (min
reluctance path) from a-(b-) current seen by b-(a-) winding.
17
 Rotation Rotation
φa φa
b' b'

N
a' a'
a● a
●

S
φb S φb
●b ●b

Fig. 9a: θ=-30° Fig. 9b: θ=60°

Fig. 9
Starting at the maximum permeance condition, a rotation
by 180° to θ=150° gives maximum permeance again.

The implication of these observations are that permeance,

and therefore inductance, is a sinusoidal function of
2(θ+30°).
3. Constant term:
There is an amount of permeance that is constant,
independent of rotor position. Like before, this is
composed of the stator iron, the air gap, and the inner part
of the rotor. We will denote the corresponding inductance
as MS.
From above 1, 2, and 3, we express mutual inductance
between the a- and b-phases as
18
 Lab = −M s + Lab
 cos 2( + 30) (16)
One last comment: The amplitude of the permeance
variation for the mutual flux is the same as the amplitude
of the permeance variation for the self-flux, therefore
L’ab=Lm.

And so the three mutual expressions we need are

Lab = −M s + Lm cos 2( + 30) (17)
Lbc = −M s + Lm cos 2( − 90) (18)
Lca = −M s + Lm cos 2( + 150) (19)
We see that stator-stator mutuals, like stator self
inductances, are functions of time!

6.3 Stator-rotor terms

These are all mutual inductances. There are four windings
on the rotor (F, G, D, and Q) and three windings on the
stator (a, b, c phases). Therefore there are 12 mutual terms
in all.
Central idea: Recall that for stator-stator mutuals,
• windings are locationally fixed, and…
• the path of mutual flux is fixed, but…

19
• the rotor moves within the path of mutual flux and
causes the iron in the path to vary, and for this reason,
the path permeance varies.
Now, in this case, for stator-rotor terms (all mutuals),
• the rotor winding locations vary, the stator winding
locations are fixed, and so…
• the path of mutual flux varies, and so…
• the iron in the path of mutual flux varies, and for this
reason, the path permeance varies.
To illustrate, consider the permeance between the a-phase
winding and the main field winding (F).
• When the main field winding and the stator winding are
aligned, as in Fig. 10a, the permeance is maximum, and
therefore inductance is maximum.

20
 φF
N
φa φa

a' φF a'
a a S
N

Fig. 10a Fig. 10b

• When the main field winding and the a-phase stator

winding are 90° apart, as in Fig. 10b, there is no linkage
at all, and inductance is zero.
• When the rotor winding and the a-phase stator winding
are 180° apart, as in Fig. 11, the permeance is again
maximum, but now polarity is reversed.

S
φa

a'
a

N
φF

Fig. 11

21
This discussion results in a conclusion that the mutual
inductance between a-phase winding and the main field
winding should have the form:
LaF = M F cos (20a)
The d-axis damper (amortissuer) winding is positioned
concentric with the main field winding, both producing flux
along the d-axis. Therefore, the reasoning about the
mutual inductance between the a-phase winding and the
d-axis damper winding will be similar to the reasoning
about the mutual inductance between the a-phase
winding and the main field (F) winding, leading to
LaD = M D cos (21a)
Now consider the mutuals between the a-phase winding
and the windings on the q-axis, i.e., the G-winding and the
Q damper (amortissuer) winding.
The only difference in reasoning about these mutuals and
the mutuals between the a-phase winding and the
windings on the d-axis (the F-winding and the D damper
winding) is that the windings on the q-axis are 90° behind
the windings on the d-axis. Therefore, whereas the a-
22
phase/d-axis mutuals were cosine functions, these
mutuals will be sine functions, i.e.,
LaQ = M Q sin  (22a)

LaG = M G sin  (23a)

Summarizing stator-rotor terms for all three phases, we
obtain the equations on the next page.

23
LaF = M F cos (20a)
LbF = M F cos( − 120) (20b)
LcF = M F cos( − 240) (20c)
LaD = M D cos (21a)
LbD = M D cos( − 120) (21b)
LcD = M D cos( − 240) (21c)
LaQ = M Q sin  (22a)
LbQ = M Q sin( − 120) (22b)
LbQ = M Q sin( − 240) (22c)

LaG = M G sin  (23a)

LbG = M G sin( − 120) (23b)
LbG = M G sin( − 240) (23c)

24
7.0 Summary
Summarizing all of our needed expressions:
Rotor-rotor self terms: 5, 6, 7, 8
Rotor-rotor mutuals: 10a, 10b, 11a, 11b, 11c, 11d
Stator-stator self terms: 13, 14, 15
Stator-stator mutuals: 17, 18, 19
Rotor-stator mutuals: 20a, 20b, 20c, 21a, 21b, 21c, 22a,
22b, 22c, 23a, 23b, 23c
Counting the above expressions, we see that we have 28.
But let’s look back at our original flux linkage relation (3):

  a   Laa Lab Lac LaF LaG LaD LaQ   i 

 a 
    L Lbb Lbc LbF LbG LbD LbQ  ib
stator   b   ba  
  c   Lca 
Lbc Lcc LcF LcG LcD LcQ   ic 

    
 F  =  LFa LFb LFc LFF LFG LFD LFQ  iF 
 G   L  
 LGb LGc LGF LGG LGD LGQ  iG
rotor     Ga  i 
 D   LDa LDb LDc LDF LDG LDD LDQ  D
     i 
 Q   LQa LQb LQc LQF LQG LQD LQQ   Q 
(3)
We have 49 terms! Where are the other 21 expressions?
25
Note because Lij=Lji, the inductance matrix will be
symmetric. Of the 49 terms, 7 are diagonal. The other 42
terms are off-diagonal and are repeated twice. So we are
“missing” the 21 expressions corresponding to the off-
diagonal elements for which we did not provide
expressions. But we do not need to, since those “missing”
equations for the off-diagonal elements Lij are exactly the
same as the expressions for the off-diagonal elements Lji.
We will look closely at this matrix in the next set of notes.

26