Beruflich Dokumente
Kultur Dokumente
567
1. INTRODUCTION
This paper is a summary of work done to
characterize the residential power circuit as a
communication medium. Here the term residential power circuit refers to that part of the power
circuit on the secondary side of the distribution
transformer. This circuit is often used for communication signals for controlling appliances,
security devices, and for other purposes. The
frequencies of most interest in these systems
range from about 100 to 200 kHz. The interface
to this circuit is usually a 120 V or 240 V receptacle used to deliver power. These communication systems are called residential power line carrier systems. Other types of power line carrier
are distribution line carrier and transmission line
carrier both of which are used by the power utilities for communications purposes.
The communication properties of the
residential power circuit include
(a) The impedance at the receptacle
(b) The signal transfer between power
outlets within a residence
(c) Noise
(d) Standing waves
The primary objective in most residential
power line carrier systems is to communicate
information from one power outlet in a residence
to another. The communication medium, therefore, consists of everything connected to power
outlets. This includes house wiring in the walls
568
5. Impedance modulation caused by rectifier circuits is modeled herein. It is shown how this
source of impairment alters the power spectrum
of the communications signal.
ie-
20 kHz
30 MHz
569
x-
0-
2!.4
0
z
2'00
a-
~~~~~~x
16
6~~~~~~
I
z
A. Distribution Transformer
The distribution transformer is a 3 port network consisting of a primary port and two secondary ports. The analysis of such 3 port networks
by the use of short circuit admittance parameters
(y-parameters) is given in standard texts on networks such as Van Valkenburg [7]. Fortunately,
our measurements have shown that the transfer
of a signal from one secondary side of the
transformer (Li) to the other side (L2) at frequencies above 20 kHz is affected very little by the
primary impedance Zp. This permits us to
analyze the transformer by using 2 port network
analysis as shown in Fig. 3. What follows is a
brief analysis showing how Li-12 signal transfer
and the secondary impedance of these
transformers can be calculated from the short circuit admittance parameters (y-parameters).
A
&- 20 KHZ
2)8
0)
I
x
100 KHZ
50 KHZ
I12
00
ox
a
0&
AA
A A
000
xx
0
n
lw
1
0
I,
I
II
12
16
20
240V
analysis of the components involved is quite consistent with the measurements shown in Figs. 1
and 2. Loads connected to all residences on the
distribution transformer must be considered.
For testing purposes it is often important to
simulate the impedance of the residential power
line circuit. Networks which do this are called
Line Impedance Stabilization Networks (LISN).
The LISN specified for many measurements is
discussed in [5]. This IEEE standard LISN is a
good approximation to the mean value of the
impedance measurements shown in Fig. 1 for
frequencies between about 200 kHz and 1 MHz.
LISNs which better represent the measurements
in Fig. 1 and 2 at lower frequencies where
residential power line carrier systems are likely to
operate are discussed in [6].
Li
._,-----
Vi
I2
II
L2
-3
YA
V2
570
Li I1
12 L2
12
(2)
Y21 V1 + Y22 V2
By inserting an admittance Y, on one side
of the distribution transformer as shown in Fig.
3(c) it is relatively easy to derive some important
properties of distribution transformers in terms of
the y-parameters. The details of the derivations
of the following expressions are contained in
[6,7].
Voltage transfer function:
V1
Y12
V2
Z2
Y21 12/V1 V2
short circuited)
=
(apply voltage
to
Li
with L2
V2
/2
(yl+Yll)Y22-Yl22
V1
Y12
'2
Y12 Y22(Y1Y11)
/2
(4)
Transfer impedance:
(3)
Secondary impedance:
Yii
Y1 +Y11
_1
(Y1y112
Y22(y1
+Yll)
y2
Y~2
Y12
(5)
(6)
B. Transformer Measurements
Since no measurements of the yparameters could be found which would give the
signal transfer from Li to L2 at frequencies of
interest, some measurements were made by two
students from N.C. State University on six distribution transformers in a transformer yard at the
Carolina Power and Light Co. The transformers
were not energized with 60 Hz power when the
y-parameters were measured. Previous work
has shown that energizing the transformers
doesn't change these measurements at frequencies above 2 kHz. The values Of Y11,Y12 and Y22
were measured at 7 frequencies between 20 and
300 kHz. The transformer sizes were 5, 10, 25
and 50 kVA with 2 transformers measured for
each of the 5 and 10 kVA cases. The detailed
571
dB
20
50
100
150
200
250
300
-6.09
- 14.20
-19.18
-22.74
-25.26
-27.40
-29.14
Freq
(kHz)
dB
20
50
100
150
200
250
300
+ 1.47
-8.24
-12.63
-17.03
-18.94
-21.24
-22.64
Phase
-63.74
-80.28
-79.04
-82.53
-84.82
-84.94
-85.84
d
dB
-6.36
-12.34
-19.02
-21.90
-24.65
-26.30
-28.00
Phase
dB
-87.7
-84.12
-89.44
-88.75
-88.74
-89.60
-89.82
+ 1.74
-8.15
-12.34
-16.82
-18.66
-21.01
-22.40
Phase
-62.66
-79.70
-78.07
-81.80
-84.34
-84.33
-85.36
(kHz)
dB
20
-3.52
-11.26
-16.89
-20.36
-22.98
-25.07
50
100
150
200
250
300
-26.82
Yii
Phase
-74.50
-80.39
-83.91
-84.78
-86.14
-86.67
-87.31 1
dB
- 6.02
-13.04
-18.98
-22.28
-24.67
-26.58
-28.08
Y2
Phase
dB
-75.46
-3.47
-11.47
-17.07
-20.57
-23.22
-25.34
-27.13
-81.87
-84.36
-85.94
-86.54
-87.30
-87.81
Y2
Phase
dB
-72.18
-79.87
-76.04
-81.84
-84.60
-85.58
-86.09
-9.38
-15.70
-21.62
-24.88
-27.34
-29.09
-30.48
Y12
Phase
dB
-74.78
-79.86
-87.31
-85.90
-85.78
-86.43
-87.10
-5.55
-13.79
-18.82
-22.39
-24.90
-27.02
-28.72
-2 -
-77.96
-82.07
-84.96
-85.57
486.82
-87.28
Phase
-70.78
-79.54
-76.48
-81.90
-84.46
-85.44
-86.00
cm
cm1
-6
coJc-
-8
-10
\y CD
lro
i;
-12
m
:r
-14
1
I
LaJ
-16
CM
-2
CO1.
LO
<0
-
LU
az
:.r
-6
I-
-8
co
wCJ
-10
m
-12
>
...
-14i
>yO
cm
-4
ui
-6
Phase
Y2
Y1 = 0
-4-
10
Table 1
Secondary y-parameter measurements for three
distribution transformers
(kHz)
-84
-1tO
-12
100
150
F-ECUENCY IN KHz
-87.85
vs. frequency
on L1.
572
Fig. 4 is a plot of calculated values of voltage transfer ratio V1/V2 when Li is loaded with
values of Y, = 0, 0.125 and 0.250 mhos. When
there is no load (YV =0), there is very little loss of
signal. As the load increases so does the voltage attenuation from Li to L2. The attenuation
also increases with frequency implying that lower
frequencies will have better L1-L2 signal transfer
than higher frequencies. The value Y1 = 0.125
represents a full load on a circuit in a residence
with a 15 amp. circuit breaker since
Zl=120/15=8. Some 120 V electric heaters can
have an impedance only slightly greater than 8
ohms. The value Y1 = 0.250 mhos, representing a resistance of 4 ohms, was used because
the impedance appearing across a 120 V secondary circuit of the transformer is the impedances
of all residences across this circuit. The Y1 =
0.250 case could occur when two residences
have 8 ohm heaters operating on the same circuit. It is not uncommon for a distribution
transformer to serve 4-8 residences each of
which may have loads across the secondary.
The greatest attenuation from Li to L2 shown is
about 16 dB for the 10 kVA transformer. The
values of V1/V2 shown in this figure were calculated using equation (3). The time allocated to
this study did not permit calculations of the voltage transfer ratio for a wider variety of loads.
Calculations of the secondary impedance of
the transformers are easily made from equation
(4). Care must be taken to avoid having confidence in such calculations when the difference
Y1l Y22 - Y22 is much smaller than the quantities involved. The accuracy of the measurements can dramatically affect the calculation of
secondary impedances when this is the case. A
similar situation holds for calculations of transimpedance V1/12.
The y-parameters can be used to determine
signal transfer for a wide variety of circumstances not covered here. These yparameters determine the characteristics of the
transformer exactly when its operation is linear.
Previous work on many occasions has shown
that energizing the transformer at 60 Hz does not
change the y-parameters for frequencies above
20 KHz.
573
050
sechine
%1 sanider
backg r'und
-100
1.
20
40
i-
HZ 60
80
+-
1 00
574
A. Synchronous Noise
The solid state light dimmer is used with up
to 600 watts of incandescent lighting to provide
continuously variable lamp brightness. The dimmer is wired in series with the incandescent
lights and controls lamp brightness by switching
on and off rapidly through the use of SCR's or
triacs. The National Electrical Manufacturers
Association has established limits on allowable
noise as a function of frequency, but the lowest
frequency specified is 500 kHz. The noise generated by this device is a function of the triac and
the elements of the switching network.
Voltage and current waveforms which occur
in the operation of a particular dimmer with a 100
watt (144 ohm) lamp are shown in [10]. The
switch (triac) closes early in the first half-cycle for
maximum brightness and late in the first halfcycle for minimum brightness, and then opens
during the zero crossing. This repeats for the
second half-cycle. The high-frequency voltage
produced is a damped oscillatory waveform with
natural resonant frequency of approximately 125
kHz [11]. When the dimmer is adjusted for maximum brightness, voltage is not applied to the
lamp during the complete cycle so that sharp
leading edges are still present in the current
waveform. Since the switching of the triac
(dv/dt= 1 OOv/usec) is much faster than the
change in 120 volt 60 Hz excitation voltage,
extremely high frequency components are produced.
Vines' analysis of the light dimmer shows
that both even and odd harmonics of 60 Hz are
produced and their relative strength depends on
the setting of the dimmer. Fig. 5 shows a spectrum of the noise voltage produced by a 100 watt
light dimmer. Notice that the spectrum actually
increases with frequency. This is true because
the impedance of the power circuit which feeds
the light dimmer increases with frequency. Fig. 6
is an expanded plot of the spectrum about 12480
Hz. This figure shows the sharp spectral lines
produced at both even and odd harmonics of 60
Hz.
-20
-40
&A
19
0
bd
-60 _
a,11
a10
208tb Harmonic
of 60 Hz
odd envelope
eten envelope
P.
-80
11 IL-I-I
-4t1-Ti TitlE~~~~~~~~~
i~1T[ias
W I
l 1968
IsZZ"
s~~~~~~~~~~~~lf I.
124ao
Frequency ln kHz
1Z736
LZ99z
575
vm(t)
V. (t)
A2S(f+60)
f0-60
/TA2So(f-60)
- t A0
/2
fo
fo+ 60
-120)
f+120
576
The power spectrum is composed of the spectrum of the unmodulated signal SO(f) plus
attenuated replicas of SO(f) that occur at frequencies f oi60, where f0 is the center frequency of the PLC signal. For a full wave rectifier the values of Ai for i odd will be very small so
that the spectral lobes will exist primarily at f0
plus multiples of 120 Hz. Experimental evidence
on the severity of this impedance modulation
could not be found.
A more thorough analysis of this impairment
results from specifying the current source
impedance and being more specific about the
value of Z(t). This would not change the general
shape of the resulting spectrum shown in Fig.
7(c).
5. THE DESIGN OF RESIDENTIAL POWER
LINE CARRIER SYSTEMS
Given the characteristics of the residential
power circuit as a communications medium, what
design features would permit residential power
line carrier systems to best utilize this medium?
Following are four suggestions.
1. The most severe noise seems to occur near
the peaks of the 60 Hz power signal. Thus
transmitting a signal at or near the zero crossing
of this signal would seem to be an advantage.
The X-10 signal design utilizes this idea. This
signal design would also avoid the impedance
modulation impairment which switches capacitors
across the line near the peaks of the 60 Hz signal. Thus, time domain techniques at the receiver
can be used to "listen" for the signal only near
the 60 Hz zero crossings where the noise is low.
spread spectrum [12] or the use of multiple carriers would help combat this problem.
4. Since the most severe noise is impulsive,
clipping the signal at the receiver may be advan-
tageous.
6. REFERENCES
(4)
(5)
(6)
(7)
84-86.
J.A. Malack and J.R. Engstrom, "RF
Impedance of United States and European
Power Lines," IEEE Transactions on Electromagnetic Compatibility, Vol. EMC-18,
No. 1, February 1976, pp. 36-38.
R.M. Vines, J. Trussell, K. Shuey, and J.B.
O'Neal, "Impedance of the Residential
Power Distribution Circuit," IEEE Transactions of Electromagnetic Compatibility, pp.
6-13, February 1985.
"Radio Interference: Methods of Measurement of Conducted Interference Output to
the Power Line from FM and Television
Broadcast Receivers in the Range of 300
kHz to 25 MHz," ANSI/IEEE Standard 2131961, IEEE Institute of Electrical and Electronics Engineers, Inc. 345 East 47th
Street, New York, NY 10017.
J.B. O'Neal Jr., "The Transmission Medium,
Impairments and Testing of Residential
Power Line Carrier Systems," EIA report
dated Jan. 1, 1986.
M.E. Van Valkenburg, Network Analysis,
577
He is currently