Case Study: Building a Bridge Between Product Safety and EMC
Richard Georgerian and Ron Duffy
Instead of tackling product safety and EMC issues independently
of each other, engineers can analyze and meet both sets of requirements
together.
Line
filters help products meet electromagnetic compatibility (EMC)
requirements. Unfortunately, these same filters make it more
difficult to meet product safety requirements. Engineers must
optimize the line filter while meeting both sets of requirements,
but finding a balance can be challenging. Using an ac filter
as an illustration, this article examines product safety and
EMC issues together. The experimental results are based on safety
standards for information technology equipment (ITE) and measuring
instruments. The same methods, however, can be employed with
any equipment that uses a line filter, although the values would
need to be adjusted to meet appropriate standards requirements.1
The product safety analysis of an ac filter includes testing
for capacitor discharge, current leakage, and dielectric strength,
as well as inspection of the filter's chokes. The EMC analysis
of an ac filter involves determining whether the filter provides
adequate attenuation. The filter lies between the line impedance
stabilization network and the intended load. By combining the
product safety and conducted emissions tests, compliance engineers
can determine optimum filter component values and subsequently
the minimum and maximum component values for different loads.
To follow the methods used for this article, the influence
of the shunt resistor, X and Y capacitors (specifically regarding
capacitor discharge), current leakage, and hipot tests must
first be understood. The values of the X capacitors and shunt
resistor determine capacitor discharge rate. (The load, if turned
on, also influences discharge rate, but for these experiments
no loads were used.) The values of the Y capacitors determine
the amount of current leakage from the mains power to ground.
UL 1950 limits current leakage for ITE to 3.5 mA.2
The Y capacitors also determine the hipot test current. There
is no limit to the hipot current, but the current must be determined
to properly set the hipot equipment. Setting the equipment's
upper-current detection too low will result in defective operation.
The capacitor discharge test ensures that if an ac plug is
abruptly removed from its receptacle, the voltage across the
line and neutral terminals will not exceed a safe level. Per
UL 1950, voltage across a capacitance greater than 0.1 µF
must decay to 37% of the ac-input peak voltage in 1 second for
type A equipment and 10 seconds for type B equipment. IEC 61010-1
requires that the pins not be hazardous (live) at 5 seconds
after disconnection from the supply.3
Time Constant Calculation
Before each measurement, the X capacitance of the filter was
measured using a network impedance analyzer. The X-capacitor
value is the capacitance of the filter construction plus the
parallel total of the actual values of the physical X capacitors
(C3||C4). The resistance of the shunt resistor, R1, was measured
with a multimeter to obtain R1meas. C = capacitance
of the X-capacitor system, as listed in Table I.
(1) C = (capacitance of the ac filter construction) + C3||C4
(2) Time constant tc = C(R1meas)
Table I uses the UL 1950 capacitor discharge criteria. The
peak voltage, Vpk, was calculated using equation
3, with a 20% safety margin for the ac mains input voltage.
(see
Table I)
Vpk = ((check)2)[240 V rms + 0.2(240 V rms)]
Vpk = 407 V rms
The capacitor voltage, Vcap, is to be 37% of Vpk
after one time constant. Equation 4 derives the calculated capacitor
voltage at any given time.
The measured Vcap is within a few volts of the calculated
value. Even though R1 was not physically placed in the ac filter,
10 MW was measured. The resistance
was the resistance of the scope probe. Product safety engineers
still must take into account the resistance value of the probe
when calculating t, be-cause the probe is parallel with R1.
Use a probe with a resistance at least 10 times greater than
that of the shunt resistor. Figure 2 shows the discharge voltage
of filter no. 1 after one time constant.
 |
| Figure 1. Schematic of an ac filter. |
 |
| Figure 2. Discharge voltage of filter no.1 after one
time constant. |
Current Leakage
To minimize the calculated error for Ileakage,
each Y capacitor was measured with a network impedance analyzer.
To maintain a safety margin of 10% for the ac mains input frequency,
the analyzer was set to 66 Hz. Ileakage included
the contribution of current from the power cord and the ac filter
construction, 18 µA and 3 µA, respectively. For Ileakage
(equation 5) the variables are as follows: Eleakage,
the input voltage to the circuit when measuring for leakage
current, equals 288 V rms; fleakage, the frequency
of the ac voltage input when using a variable ac power source,
equals 66 Hz; and C represents the capacitance of C1 or C2 in
nF. Ileakage, the leakage current through
one of the Y capacitors, was calculated and measured for each
Y capacitor, C1 and C2. This was done by reversing the mains
polarity.
(5) Ileakage = Eleakage/(1/2pfleakageC)
Typical Y-capacitor values of 3.3;4.7 nF result in leakage
less than the 3.5 mA limit (see Table II). Because filter no.
4, however, has larger Y capacitors, its leakage current exceeds
most safety standard requirements. Product safety engineers
must account for more than the stated value of capacitance.
The positive capacitance tolerance value is equally important.
The Y capacitors set the current leakage, but the added tolerance
could cause excess leakage.
|
Ac filter
|
Ileakage
|
Y-capacitor
system measured at 66 Hz
|
|
No.
|
Calculated
|
Measured
|
Y capacitor,
C1
|
Y capacitor,
C2
|
| 1 |
570 µA |
568 µA |
4.6 nF |
4.6 nF |
| 2 |
558 µA |
556 µA |
4.5 nF |
|
| 606 µA |
603 µA |
|
4.9 nF |
| 3 |
439 µA |
435 µA |
3.5 nF |
|
| 427 µA |
432 µA |
|
3.4 nF |
| 4 |
4.07 mA |
4.05 mA |
33.9 nF |
33.9 nF |
|
Table II. Current leakage.
|
Dielectric Strength (Hipot)
From the previous current leakage experiment, we obtained the
Ileakage for each of the ac filter configurations:
Ehipot, the voltage to which the hipot equipment
is set, equals 1500 V rms; Eleakage = 288
V rms; fhipot, the frequency of the hipot
equipment, equals 60 Hz; and fleakage= 66
Hz. If the Y capacitors were not equal, the Ileakage
was calculated to be the average measured current flowing through
each Y capacitor.
(6) Ihipot = 2(Ehipot/Eleakage)(fhipot/fleakage)Ileakage
Table III indicates that the typical Y-capacitor values of
3.3;4.7nF result in hipot currents of 4.3;5.9 mA.
The high-value Y capacitor of filter no. 4 results in a high-level
hipot current. Product safety engineers should make allowances
for a variety of ac filters that could be tested on the production
line. When setting hipot equipment for the current, a minimum
current should also be included to alert an operator that the
on-off switch is in the on position.
|
Ac filter
|
Ihipot
|
Total capacitance
of Y capacitors
|
Y capacitor,
C1
|
Y capacitor,
C2
|
|
No.
|
Calculated
|
Measured
|
| 1 |
5.3 mA |
5.9 mA |
9.2 nF |
4.6 nF |
4.6 nF |
| 2 |
5.5 mA |
5.6 mA |
9.4 nF |
4.5 nF |
4.9 nF |
| 3 |
4.1 mA |
4.3 mA |
6.9 nF |
3.5 nF |
3.4 nF |
| 4 |
38.0 mA |
39.1 mA |
67.8 nF |
33.9 nF |
33.9 nF |
| Table III. Current hipot. |
|
Ac filter
|
R1
|
X capacitors
Y capacitors
|
Choke
|
|
No.
|
C3
|
C4
|
C1
|
C2
|
|
1
|
330 kW
|
0.33 µF
|
0.33 µF |
4.7 nF
|
4.7 nF
|
1 mH
|
|
2
|
417 kW
|
0.10 µF
|
0.10 µF |
4.9 nF
|
4.9 nF
|
2.5 mH
|
|
3
|
1 MW
|
0.47 µF
|
0.47 µF |
3.0 nF
|
3.0 nF
|
2 ¥ 50 µH
|
|
4
|
1 MW
|
0.47 µF
|
0.47 µF |
33 nF
|
33 nF
|
2 ¥ 50 µH
|
| Table IV. Ac filter components. |
Chokes
The chokes of an ac filter are to be constructed to meet the
intended safety standard. Per UL 1950, operational insulation
is needed from line to line. UL 1950 requires 1.5 mm for clearance
and, depending on the material, 1.6;3.2 mm for creepage.
There is no requirement for distance through the insulation.
The dielectric strength from line to ground is 1500 Vpk.
Per IEC 61010-1, basic insulation is needed from line to line.
IEC 61010-1 requires 1.5 mm for clearance and, depending on
the material, 1.6;3.0 mm for creepage. The distance through
the insulation is not addressed in IEC 61010-1. The dielectric
strength from line to ground is 1900 Vpk.
If a product safety engineer intends to evaluate open-framed
ac filters, then accessibility to the components must be considered
to ensure operator and service personnel safety. Off-the-shelf
filter components are most likely potted and metal encased,
so the accessibility of the filter's ac input and output terminals
is the primary concern.
Conducted Emissions Testing
The EMC analysis of an ac filter consists mainly of performing
conducted emissions testing. Performing the emissions tests
after the components have been selected based on safety testing
minimizes testing time. Using this order, if a filter does not
meet relevant conducted emissions requirements, its components
can be easily changed and retested or EMC compliance. Once new
values have been selected, a filter can be retested to ensure
that it still meets safety requirements. Because ac filter components
can be easily swapped and tested, a variety of components should
be tried prior to conducted emissions testing. Because EMC equipment
tends to be more expensive than safety test equipment, access
to EMC equipment may not be as readily available as safety equipment.
The components used in a line filter can create problems for
safety engineers, but with capacitance reduction and proper
choke selection, most product safety problems can be avoided.
Selecting parts based on safety requirements provides the parameters
for identifying filter options for conducted emissions testing.
By fabricating an ac filter with components that can be easily
substituted, engineers can experiment with a variety of configurations
and quickly determine whether specific component values will
meet both safety and EMC requirements.
1. Richard Georgerian and Ron Duffy, "Product Safety and Electromagnetic
Compatibility Requirements of an A.C. Filter," in Proceedings
of the IEEE EMC Symposium (Washington, DC: IEEE EMC Society,
2000), pp 845;850.
2. UL 1950, 3rd ed., "Safety of Information Technology Equipment,
Including Electrical Business Equipment," Underwriters Laboratories,
Englewood, CO, 1995.
3. EC 61010-1, "Safety Requirements for Electrical Equipment
for Measurement, Control, and Laboratory Use," International
Electrotechnical Commission (IEC), Brussels, 1990;1999.
4. CISPR 22:1997, "Information Technology Equipment‹Radio Disturbance
Characteristics‹Limits and Methods of Measurement," IEC, International
Special Committee on Radio Interference (CISPR), Geneva, 1997.
5. "Unintentional Radiators," Code of Federal Regulations,
47 CFR Part 15, Subpart B (October 1, 1999): pp 667;692.
Richard Georgerian is a product compliance
engineer at Exabyte Corp. (Boulder, CO) and can be reached at
richardg@exabyte.com. Ron Duffy is a product safety engineer
at Agilent Technologies Inc. (Colorado Springs, CO) and can
be reached at ron_duffy@agilent.
com.
Back to January/February
Table of Contents
|