A
Discussion of the requirements of Harmonics and Flicker by David Mawdsley,
Laplace Instruments Ltd
Just
when you thought it was all under control.....
............The latest elements in
the EMC Compliance requirement.
In
order to limit the ever increasing harmonic distortion imposed on the public
mains supply the Harmonics and Flicker standards are to be introduced.
As
from Jan 1st 2001, compliance with these Harmonics and Flicker standards
becomes a mandatory part of the EMC Directive.
This
applies to all products within the scope of these standards.
Many
products which are powered from the mains electricity supply take their power
in a non-linear fashion. Current is drawn in pulses, rather than continuously.
See the typical power supply circuit, fig
1. The current is drawn in pulses as the incoming mains voltage exceeds
the voltage on the reservoir capacitors. See fig
2. The frequency content of the current waveform includes many harmonics
of the fundamental frequency and as a consequence the supply system must deliver
power at these multiples of 50Hz (harmonics) fig
3, a requirement which causes problems for the supply industry and distorts
the voltage waveform, fig 4.
Standards:
Harmonics
IEC/EN61000-3-2
(replaces IEC555-2)
Flicker IEC61000-3-3
Scope:
The
Harmonics standard applies to ALL products that are:
·
Connected
to the public mains supply
·
Rated
at less than 16A per phase
·
Supplied
at 230V nominal (eg 110V as used in
the US is not within the scope)
·
Are
single or 3 phase.
·
An
exception is any product for professional use rated above 1kW.
For
rated currents from 16A to 75A per phase, IEC/EN61000-3-4 applies. This is not
a mandatory standard. It includes a set of standard forms which can be
submitted to the local electricity supplier in order to gain acceptance that
the equipment can be connected.
The
Flicker standard applies to all products that, in the opinion of the
manufacturer, ‘could cause flicker’(!).
If you believe that any product is unlikely to cause flicker, then it
does not need to be tested and can be judged compliant without further ado.
However, the final version of the standard has possibly changed this view
somewhat.
Why
A
good question. The Americans think that we are all quite mad and have refused
to have anything to do with these standards. However, the electricity supply
companies are pushing strongly for the harmonics standard for much the same
reasons that we use power factor correction. If large numbers of products draw
significant current at, say, the third harmonic, then the electricity supply
system will become stressed at frequencies it was not designed to deliver. The
voltage waveform will become distorted and this may affect the performance of
other equipment connected to the same network.
Although
Flicker is a relatively rare phenomenon, it can be a serious health and safety
hazard when it occurs, hence the standard.
It particularly affects those who may be prone to epileptic fits, and
can cause distress in others.
Changes
Even
before the Harmonics standard is introduced, a new version has been approved.
This new version is likely to appear in the Official Journal in 18 months time,
followed by a (probable) two year transition period. This change is being introduced due to widespread dissatisfaction
with the original standard, some features of which are confusing and subject to
different interpretations.
In
practice, industry in general and many test labs are already switching to the
revised standard in anticipation of the formal introduction of the ‘mark II’
and in practice, both versions are used, but the Mk I will be dropped in due
course.
The
significant details of the revisions will be outlined later.
Implications
Having
checked many different products, we find that many already comply. Those that
are ‘at danger’ include products that take relatively high power or include
power supply circuits delivering DC outputs. Phase angle power control devices
may also be problematic.
Products
that cause flicker are actually quite rare, but include lighting systems such
as used in clubs and theaters, heavy drive systems and some heating systems. However, the latest revision of this standard
defines a max. Voltage change (dmax) and specifies a maximum value
for this change. This covers single events such as inrush current on power
up, nothing to do with ‘Flicker’! This
means that products must be tested for inrush currents, regardless of their
potential for causing flicker.
The Harmonics Standard
To begin with:
·
The
measurements are for current drawn at each harmonic frequency.
·
The
standard covers up to the 40th harmonic (2KHz)
·
Each
harmonic has a specified maximum in mA. (The limits)
·
You
are judged compliant if you are below these limits.
·
THD
(total harmonic distortion) is not used.
Now lets complicate it:
There
are 4 classes, each with differing limits and other criteria:
A -
covers all types of product NOT covered by the other three classes.
B -
Portable equipment, eg. hand tools, lawnmowers etc...
C -
Luminaires
D -
Any equipment having a certain current waveshape. (Except phase angle
controlled motor driven equipment)
The
original intention of class D was to apply more stringent limits to equipment
such as computers and other electronic devices that are used in large numbers
and which draw current from the supply in a short pulse in each half cycle of
the mains voltage.
See
fig 5 for the selection flow
diagram and fig 6 for the
definition of the ‘special waveshape’.
The
standard tabulates the limits for all these Classes.
Transitory
harmonics occurring for less than 10s when equipment is switched on or initialised
can be ignored.
And add more confusion:
·
If
the harmonic level is fluctuating, you are allowed to exceed the limits by up
to 150% for up to 10% of the time in any 2.5 minute period. This applies only
to odd harmonics up to the 19th and even harmonics up to the 10th.
·
If
the waveshape is Class D, but the power consumption exceeds 600W, then Class A
applies.
·
Class
A and B have absolute limits, quoted in amps (or ma). Class C limits are calculated
as a percentage of the fundamental current and Class D limits are calculated
from the Wattage consumed by the product.
·
If
Class D and the power consumption is below 75W, no limits apply, but ‘4 years
after the implementation date of this standard’ (whatever that means!) this
threshold is reduced to 50W.
Fig
7 shows how the selection of the class affects the limits that apply.
This
causes various differing interpretations:
·
What
class applies if the waveshape is Class D, but the power consumption varies
above and below 600W.
·
When
calculating the Class D limits, the limits are defined for ‘the rated load
condition’. Is the rated wattage used, or the actual as calculated in real time
under ‘rated load condition’, or simply measured once at the beginning of the
test?
·
What
is the implementation date of the standard?
Some of these ambiguities are clarified or
eliminated in the proposed changes.......
Class
D is redefined as applying to:
·
Personal
computers and monitors
·
TV
receivers and VCRs
·
Printers,
fax machines and multimedia devices with a specified power less than 600W
The
waveshape test has been eliminated.
Other
products (not covered by Classes B and C now default to Class A.
The
transition from 75W to 50W is postponed indefinitely.
Regarding
the 600W cut-off point for Class D, the manufacturer is allowed to specify
what power level is used in the calculation of the limits and selection of
the Class. This power level must be within 10% of the actual. This avoids situations where fluctuating power
consumption may cause step changes in the harmonic limits during a test.
Note
that then AC2000 harmonics analyser
is fully compliant with the standard and copes with all the above factors,
including automatic waveshape detection and both original and revised standards.
Test Technique
The
conditions to be set for testing many types of equipment are defined in the
standard. These include TV receivers, audio amps, lamps, ballasts, vacuum
cleaners, washing machines, ovens and air conditioners. Other products should be set so as to
measure the maximum harmonic current for each successive harmonic in turn. In practice, this is such a time consuming
process that many test engineers simply set the product to the condition in
which they produce the worst THD.
Any
measurement of harmonic currents will be affected by the harmonic content of
the incoming mains voltage. ‘Clean’ mains supplies are very rarely available
from your wall outlet!
In
order to accurately measure these harmonic currents taken by your product, a
clean mains should be supplied. The standard specifies the maximum harmonic
content of the supply voltage as follows (expressed as a ratio of the nominal
test voltage):
|
Harmonic number |
Maximum (%) |
|
3 |
0.9 |
|
5 |
0.4 |
|
7 |
0.3 |
|
9 |
0.2 |
|
Even, 2 - 10 |
0.2 |
|
All, 11 - 40 |
0.1 |
Values
for crest factor and phase angle are also specified.
Generally,
these requirements will only be met if a ‘clean’ mains generator or source
is used. See the AC1000
power source. Fig 8 shows
the arrangement generally used.
The
harmonic currents are measured using a shunt, but the value of the shunt is not
specified, it being such that the voltage drop due to the input current of the
load is less than 0.15V
Flicker
The
test for Flicker needs only to be applied to a product if it is likely to
cause voltage fluctuations or flicker. (Fig
9).
Flicker
is a measure of the amount of mains voltage fluctuation caused by a product
which loads the mains cyclically. The aim of the standard is to reduce the
likelihood that such fluctuations would cause perceptible flicker in lighting
circuits which in turn can cause the onset of epileptic fits and other undesirable
effects.
The
standard is based on a considerable body of research, summarised in EN60868-0,
which produces a technique for the statistical evaluation of voltage
fluctuation. This evaluation leads to two parameters which defines the severity
of the flicker.
Essentially,
flicker severity is dependent on the amount of voltage change DV, and the frequency of the change. The perception of flicker is very dependent
on frequency, and is maximum at 8.8Hz. Thus a flickermeter includes measurement
of DV and the application of a frequency weighting
function. After statistical processing both a short term (Pst)
and a long term flicker (Plt) indicators are derived.
Generally,
Pst is evaluated over 10 minutes and Plt is evaluated
over 2 hours.
The
limit for Pst is 1.0 and for Plt is 0.65.
In
addition the relative steady state voltage change shall not exceed 3%
The
maximum relative voltage change shall not exceed 4%. (This applies to single events such as inrush current)
Test technique
The
product is fed from a zero impedance supply in series with a specified
reference impedance. This impedance simulates the effective output impedance of
the mains supply.
The
voltage between Phase/neutral is measured on the product side of the reference
impedance. See fig
10 (3 phase) and fig 11
(single phase). For single phase applications,
the two parts of the reference impedance can be combined as shown in fig
12.
The
product must be cycled or configured such that it produces the most unfavorable
sequence of voltage changes.
Voltage
fluctuations are recorded and:
-
converted to an estimate of light variation from an incandescent bulb.
-
an estimate of how this light variation is perceived by the eye/brain
- this
variation is ‘weighted’ according to it’s frequency.
-
This is converted to an instantaneous flicker reading. (Pinst))
-
The instantaneous flicker reading is assessed over a 2.5 minute period and a Pst
is obtained.
-
From measurements over a longer period, a measure of Plt is derived.
The
details of all the data processing and statistical evaluation involved in all
this is quite beyond the limited mathematical understanding of the author, but
if you really what to dig deeper, full explanations are given in 50868-0 and
-1. Good luck!
One
problem with the above is that most testing will use the mains supply from
the wall socket. See fig 13.
This probably has an output impedance
similar to the reference impedance, thus the voltage change is likely to be
doubled. The standard anticipates this by advising that the voltage drop
on the supply side of the reference impedance is also measured, and
only if this is less than 20% of the volt drop on the output side of the reference
impedance are the results valid. Quite what you are supposed to do if the
change exceeds the 20% is not specified!
The
AC2000 overcomes this problem
by measuring the current rather than voltage, and converting this by calculation
to volts drop as would be obtained across a perfect reference impedance as
shown in fig 14.
The sting in the tail
As
was stated earlier, Flicker testing is only conducted on products that may
cause flicker. This effectively means that 99% of all products are outside the
scope of this standard (hooray!).
However,
in the final version which has now been ratified, the term dmax has
been introduced. This is the maximum
voltage change
The
implication is that any product that may have a relatively high inrush current
could be construed as having to be tested under this standard. This suddenly
expands the types of products which have to be tested very significantly. This
change, mandatory from next month, appears to have caught most unawares. Even
test labs have often omitted the ‘inrush current test’ on products that may now
require it.
If you have any comments, suggestions regarding our site or require further product information please use the feedback form
3B, Middlebrook Way,
Holt Road, Cromer,
Norfolk, NR27 9JR, UKRevised: 6 March, 2002
Copyright © Laplace Instruments 1998, all rights reserved
