By Roger Nichols May 1, 1996
Spike, noise, surge, swell, transients, harmonics and sag are not the names
of members of a new alternative rock group; they are characteristic problems
encountered in power management.
AC power is often the most overlooked area in recording studio design. If
you were a farmer and your horse was your livelihood, you would probably
pay attention to how well he is doing. AC power is the main source of your
income, and also the primary cause of all the hums and buzzes you must deal
with on a day-to-day basis. They say that if you build a better mouse trap
that they will beat a path to your door. Well, just wait until you have
the quietest studio in town and see how fast everyone wants to work there.
Power quality can be measured by the Duration vs. Magnitude of a disturbance.
Short fault durations, like transients, can damage sensitive electronic
devices such as diodes, transistors and ICs. Lower level transients slowly
eat away at internal semiconductor junctions within electronic equipment,
eventually causing failures.
High-frequency noise can cause digital data errors in both digital audio
and computer equipment, and can interfere with clock signals, causing timing
errors and excessive jitter.
Voltage fluctuations effect motor operation and electronic equipment that
require a steady power source.
Electric utilities must generate the service capacity to meet the peak demand,
kVA (kilovolt amps), whether or not the customer is using that current efficiently.
Utilities can only charge for the active power, or kW (kilowatts).
The ratio of kW (active Power) to kVA (apparent power) is called the power
factor.
Utilities now charge a penalty to companies when the power factor is low.
This penalty can be avoided with the use of power factor correction.
Induction loads, such as induction motors used to power fans in air handling
systems, may operate at less than their full rated load because of poor
power factor.
Under these conditions the motor inductance causes the current to lag, or occur later in time than the applied voltage (Fig. 1). Some portion of the current is doing the actual work demanded of the motor, kW, while some is supporting the reactive, inductive load. This is known as kilovolt Amps Reactive, or kVAR(Fig. 2).
The ratio of the kW to kVA at the power line frequency of 50 or 60 Hz is known as Power Factor, or Displacement Power Factor (Fig. 3). The current is displaced in time from the voltage. This refers specifically to the phase shift, described as the cosine of the phase angle (phi). In these cases apparent power, kVA, can be larger than active power, kW. Due to this phase shift of the fundamental current. The current must be larger to produce the same active power. In this way kVA becomes larger because of this larger current.
Starting from an ideal sine wave with current in phase with the voltage,
as the phase angle increases, the current waveform occurs later and increases
to a higher current. The RMS current increase produces a higher apparent
power. With linear loads, both the Displacement Power Factor and Total Power
Factor change at the same rate as the phase angle changes. Apparent power
can also be larger than active power when non-linear loads are present.
These loads produce harmonic currents which circulate back through the distribution
system, and the secondary of the distribution transformer. Harmonic current
adds to the RMS value of the fundamental current supplied to the load, but
does not produce any significant power. Using the definition for total power
factor, the kW is essentially that of the fundamental only, while the kVA
is made larger because of the higher RMS current.
Total Power Factor also includes the effects of any phase difference between
the fundamental voltage and current. In many cases, when the distribution
system is serving only single phase receptacle loads, the phase difference
at the fundamental is minimal. DPF is near 1.0 and PF represents the contribution
of harmonics to the current. As the total harmonic distortion increases,
the current waveform changes to a pulse with higher peak current. The RMS
current increase produces a higher apparent power. The active power, or
Watts, and displacement power factor do not change since they are based
only on fundamental voltage and current. Total Power Factor decreases. More
current must be carried by the system to deliver the same amount of active
power. The different responses of DPF and PF can lead the way to the proper
power factor correction methods.
Induction motors can suffer from harmonic current heating if the supply
voltage is distorted. And the presence of negative sequence harmonic currents
reduces motor torque. The combination of these effects causes motors to
burn out. To test the proper operation of the motor and its power factor,
use a true RMS tool (such as the Fluke 8060, Fluke 73, Protek 506, Wavetek
2030, Fluke 41B, Tektronix Wavemeter, and many others) to measure the three
phases for proper voltage balance (Fig. 4). Look for obvious distortion
in the waveform. Most motor manufacturers recommend less than 5% distortion
for a fully loaded motor. Measure the three phases for proper current balance,
then measure power and power factor at full or normal load. If total power
factor and Displacement power factor are the same, you may need to add Power
Factor correction capacitors. To minimize the effects of harmonics on induction
motors, you can reduce the voltage distortion on the terminals by connecting
the motor to a distribution center supplying only linear load. Or, you can
consult with a power management expert and connect a harmonic filter at
the source of harmonics.
Capacitors installed to correct low power factor caused by induction
motors can fail if harmonics are present. KVAR correction capacitors can
form resonant circuits at frequencies above the fundamental. When combined
with the inductive reactance of the distribution network this can cause
premature motor failure due to excessive heat and random breaker tripping.
This is normally not a problem if harmonics are not present. Harmonic currents
produced by non-linear loads may find a resonance involving the kVAR capacitor.
Resulting high current may cause the capacitors to fail.
To verify proper circuit operation, measure the three phases for proper
voltage and balance. Then measure the three phase power and power factor.
Notice the difference between the DF and DPF readings. If the Total PF reading
is lower than the DPF reading, a portion of the load is probably non-linear.
Examine the drive current for harmonics, typically 5th and 7th. Adjustable
speed drives are a common source of 5th harmonic. The need for correction
capacitors may be reduced when adjustable speed drives are installed on
existing motors. Line reactors can be applied at drive inputs to reduce
harmonic currents. Or to avoid harmonic frequencies from resonating with
correction capacitors, filter networks can be designed to de-tune the resonant
system.
At the load center, harmonic currents can cause circuit breakers to trip.
Thermal magnetic breakers may trip prematurely from excess heat in the panel
caused by harmonic currents. Breakers may also trip erratically when non-linear
currents with high peak values are present. A peak sensing circuit breaker
responds to the peak of the current waveform. Since the peak may be higher
due to harmonics, this type of breaker may also trip prematurely at a lower
RMS current.
To detect harmonics at a load center, check the phase voltage for flat-
topping (a condition where instead of a sine wave, the waveform becomes
flat on top resembling digital clipping).Then measure the current in the
feeder conductors using a true RMS instrument. Remember that these instruments
indicate the actual heating value of the current, including harmonics. Verify
that you are operating within the load rating of the panel. Measure the
feeder neutral current. If it reads high, triplen harmonics (see below)
may be present. Then compare the current with the ratings for the conductors,
lugs, and buss bars. Compare the individual branch circuit currents to the
breaker ratings. Check the branch neutrals for overloads due to triplen
harmonics. The same process can be repeated at other load centers fed from
the same source.
Once you are sure that a receptacle panel is effected by harmonics, there
are a few options to correct the situation:
If the neutral conductor is overloaded, a larger conductor will be required
by code, or the existing receptacle panel can be replaced by a panel and
main breaker that is rated for non-linear loads.
Excess heat caused by harmonics in lighting circuit conduit can cause conductor
insulation to fail. In energy-saving electronic ballasts with solid-state
power supplies, the phase and neutral currents can contain harmonics. Existing
standards for the number of conductors in a conduit don't always account
for the heat caused by these harmonics. To find harmonic overloads in lighting
circuits, you can make the same measurements you made at the receptacle
load center. Measure the current in the feeder neutral. If the levels are
high, compare the measured currents to the ratings of the conductor, lugs
and buss bars. Feel the conduit for excess heat. To determine the overall
level of harmonics, measure the total harmonic distortion in the phase currents.
The THD generally refers to the RMS value of all the harmonic currents,
divided by the fundamental (Fig. 5). The total harmonic distortion may be
a problem if it exceeds 20%.
To prevent harmonics from effecting a lighting load center, specify fewer conductors per conduit. Or you can install new high performance ballasts which produce lower harmonic currents and also improve Power Factor.
Harmonics on the AC line are usually caused by non-linear electrical loads.
Some of these non-linear loads are: personal computers, certain types of
lighting ballasts, electronic studio and office equipment, and adjustable
speed motor drives. These devices draw non sinusoidal current in abrupt
pulses when connected to a sinusoidal voltage source. These pulses form
a distorted current waveshape which contains harmonics.
The harmonic current drawn by non-linear loads acts in an Ohm's law relationship
with the source impedance of the supplying transformer to produce voltage
harmonics. The source impedance includes the supplying transformer and branch
circuit components. For example, a 10 amp harmonic current being drawn from
a source impedance of 0.1 Ohm will generate a harmonic voltage of 1.0 volt.
Any load sharing this transformer or branch circuit can be affected by the
voltage harmonics generated.
Computers used in console automation or hard-disk recording can crash or
reset from excessive harmonic voltages in the supply power. Remember, the
harmonics can come from devices anywhere on the same transformer or branch
circuit.
Each harmonic has a name, frequency and sequence. The sequence refers to
the phasor rotation with respect to the fundamental. In an induction motor,
for instance, a positive sequence harmonic would generate a magnetic field
that rotated in the same direction as the fundamental. A negative sequence
harmonic would rotate in the reverse direction.
Harmonic |
Frequency |
Sequence |
First |
60 |
+ |
2nd |
120 |
- |
3rd |
180 |
0 |
4th |
240 |
+ |
5th |
300 |
- |
6th |
360 |
0 |
7th |
420 |
+ |
8th |
480 |
- |
9th |
540 |
0 |
Zero sequence harmonics are called "Triplens." These are odd
multiples of the 3rd, such as 3rd, 9th, 15th, 21st, etc.
A simple way to determine the extent of harmonic distortion caused by a
single-phase non-linear load would be to make two separate current measurements.
Make the first measurement using an average responding current clamp or
meter with clamp on probe. Make a second measurement of the same circuit
using a true RMS current clamp meter. Divide the results of the first measurement
by the second measurement. This will give you the A/R ratio. A ratio of
1.0 would indicate little or no harmonic distortion. A ratio of 0.5 would
indicate substantial harmonic distortion. This test method works because
an averaging meter will read a true sine wave correctly, as will the true
RMS meter. If the waveform is distorted, the true RMS meter will read correctly,
while the averaging meter will read up to 50% low, depending on the amount
of distortion.
The above measurement method is not a substitute for a harmonic analyzer,
but it is a simple way to determine if there is a need for more sophisticated
equipment.
As you can see by the material that has been covered, AC power contains
more than just that pure mythical 60 Hz sine wave that you read about in
text books. In most cases, filters added to the power line add noise of
their own. The capacitors in the filter circuit leak current into the ground
system. This noise is usually in the form of a reactive, non-linear leading
current. The same type of noise on the ground is caused by switching power
supplies found in most computers and digital audio gear. This ground noise
usually shows up as hum in audio gear. Class A tube amps and balanced mic
pre amps are particularly susceptible to this ground noise.
All of the power-consuming devices in a studio are connected to unbalanced
power (Fig. 7). There are two wires supplying the 120 V power, with the
ground for safety (and noise). If you measure between the two feed wires
the results will be 120 V. If you measure between ground and one of them
you will see 120 V. If you measure between ground and the other lead, you
will see zero V. Well, you are supposed to see zero, but because of ground
noise and currents, you will measure a couple of volts. Just remember, with
unbalanced power, all of the power generated garbage ends up in the ground.
Balanced power consists of three wires (Fig. 8). The same three wires that
are connected to most studio equipment. If you measure the voltage between
the two feed wires, you get 120 V. If you measure between either one of
the feed wires and ground you will see 60 V.
If we take any of the noise generating equipment and connect it to the balanced power source, the noise generated in each leg of the power will be out of phase with each other at the ground. The ground will be quiet as a clam. Balanced power provides the same common mode rejection we are all familiar with in balanced audio.
Quiet grounding schemes in studios sometimes border on the occult. I asked
one studio why they had a water cooler in the control room with no water
in it. The said that for some reason, when the water cooler was plugged
into the same branch circuit as the guitar amps, that there was less hum
in the amps. I unplugged it once. They were right.
Grounding circuits were never meant to carry current except during a short
circuit. Objectionable ground currents are those that will provide you with
a shock. Anything less than that is OK as far as Underwriters Laboratories
is concerned.
We have all experienced ground loops in the studio. The really bad ones,
with hum levels above the signal level, we try to cure. The ever present
little hums, that make the DAT meters stick one segment up from the bottom,
we try to ignore. We try breaking grounds in balanced cables at one end
so that we do not have multiple ground paths for ground loops. We lift chassis
grounds with special plugs and make sure that metal chassis do not touch
each other. If we removed the currents from the ground, then we would have
no current to loop.
With balanced power, you can use any type of grounding configuration you
wish. Star, schmar. You can leave the grounds connected at both ends of
your audio cables. You can throw away all of the ground lift adapters. You
can finally plug everything in the way it was meant to be plugged in.