Preamble . Our former EMC articles reviewed four of the 5 principal conduction and radiation coupling mechanisms, as they affect equipment/ system susceptibility and emissions. The last ones (EMC Articles #6 and 7) were addressing the shielding of equipment boxes, from the smaller hand-held devices up to large cabinets or even entire rooms. The present article is covering the 5th, very important coupling mechanism: How the various disturbances that exist on the power mains (public AC distribution, ship or aircraft AC distribution, vehicle dc distribution… etc) can find their way from their source down to the electronic components of a system?
Reciprocally, some active elements in our equipments (dc-dc and ac-dc switch mode regulators, fast IGBT and thyristors, variable speed drives, fast logic circuits, etc…) can in turn become EMI sources, polluting the power distribution to the prejudice of other users. And like for the other coupling paths, many solutions that solve power line suceptibility problems will also reduce conducted emissions.
1. Equivalent Circuit for Typical Distributions
Whatever your installation, you must know at least approximately what is the wiring and grounding scheme of your power distribution, from the power source down to the users. This knowledge is important for:
- understanding and estimating the coupling factors
- understanding the CM (Common Mode) to DM (Differential Mode) conversion
- selecting the most appropriate, and economical filter
- selecting the proper characteristics for surge protection devices.
Fig.1 shows the most common configurations for power distribution schemes. Depending on the applications, a power distribution can be fairly simple: in motor vehicles the battery (+) is running in the entire harnessing, the (-) return conductor being the metallic car body. It is generally more complex with the (+) and (-), or Phase and Neutral, running together in the entire Power Distribution Network. In houses and professional buildings, the Ph and N wires are run along with a third wire, acting as a safety lifeline to the earthing electrode(s), where the Neutral of the power source (Entry Transformer or Power Plant) is also grounded. Therefore as shown on the schematic, we have in general:
a) a DM impedance between the hot wire and its return.
b) a CM impedance between either one (or both) power wires and the local structure ground or chassis.
c) If a there is safety earth (green/yellow) wire, two values for CM impedance can be considered:
- The Ph+N wires vs the earthing wire, which is only valid as long as the earthing wire is sufficently short and the frequency low enough for considering this wire as the earth potential.
- The Ph+N+ Earth wire altogether vs the real local ground, which can be a structural ground, concrete slab, ship hull etc ….In this later case, the CM currents will flow in the Power + Earthing wires in the same direction, returning to the Power source via the local ground reference.
2. Power Mains Impedance
This is the impedance that will be seen by an equipment when looking toward its power source. Figure 2 shows the general impedance profile of urban power distribution, made after compilation of wide statistical surveys in major cities, worldwide. The thin dotted plots are the upper and lower 10% of the measured impedances (that is less than 10% of the measured data points were above or below these values). Shown for comparison is the profile of a Line Impedance Stabilization Network that is deemed to represent the average impedance value of urban power mains (Re. our Article #2, Military & Civilian EMC Norms). Both CM and DM Power Mains impedances must be considered.
Values depends on the application: public network, vehicle, aircraft, ship etc…. which dictate a typical length of wiring from the source to the users and the Neutral, or (-) conductor grounding practices. In any case, the overall shape of Power Mains impedances is the same, when seen from the end user Power Entry terminals, or simply the wall outlet:
- from dc to a few kHz, it is basically the ohmic resistance of the wires plus the source resistance, this latter being obtainable from alternator or substation transformer short-circuit current.
- From a few kHz to ≈ MHz, the impedance increases linearly with frequency, since it is dictated by the self inductance of the mains wires, the wide spread of values being due to the overhead or buried type of wiring, the proximity of a ground plane, the number of users in parallel, etc …
- Above the MHz region, the impedance stabilizes around a mean value of 50Ω for CM (line-to-ground) and 100Ω for DM (line-to-line), corresponding to the line characteristic impedance.
3. Equipment Input Impedance
Knowing the equipment input impedance is important too, because it affects the behavior of the equipment to EMI coming in, as well as filters and surge suppressors selection. Without EMI filter, the impedance at input terminals is that of the front-end transformer (often the1st regulator), with its secondary load transferred to the primary. If there is no front-end transformer, the power input being dc or rectified ac, the DM input impedance is low, being that of the direct-polarized rectifiers and associated buffer capacitor.
When the power input is isolated from the equipment frame or signal ground, as often the case with 230V/50Hz power, the CM input impedance is high at low frequency, decreasing progressively when F increases, because of the parasitic capacitance of the floated circuits
4. Major Power Mains Disturbances
Power line disturbances are a frequent cause of equipment malfunction or even damage. They can be quasi continous like HF noise superimposed to the normal mains voltage, waveform distorsion and notches, or incidental, like transient overvoltages, spikes, short partial or total voltage drop (Fig.3). A majority of them is not due to the quality of the power source, but to the operation of the various users, whose current demands are often irregular or pulsed, causing voltage dips or surges on the distribution. Lightning strokes on, or nearby the power lines are also a frequent cause of severe overvoltages.
- Slow or steady fluctuations of line voltage can reach -/+ 10%, and are regarded as normal. All modern equipment are equipped with regulators that can easily handle such variations
- Flickering: periodic, very low frequency undervoltage, causing lamps blinking
- Ripple or HF noise riding over the supply voltage can be caused by poor filtering of switch-mode regulators, or strong ambient RF field coupling onto the power lines. These rarely exceed a few volts and are easily reduced by EMI filtering
- Harmonic distorsion and ”notches” in the AC voltage sinewave are caused by currents peak drawn by non-linear loads of the various users
- Unipolar or ringing transient overvoltages and spikes can often exceed hundred Volts, superimposed to the AC (or dc) voltage. Extreme values of 4-6kV can be reached with severe lightning indirect effects or Fast Switching transients. These can be damaging and require transient suppressors.
5. Improving Equipment Immunity to Power line disturbances
Several variables need to be considered when hardening an equipment, or an entire system:
- the nature of the disturbance: under or overvoltages, exceeding the normal +/- 10% fluctuations, energetic- high voltage transients, short spikes, total outage?
- the duration of the disturbance: sec? msec? more?
- the risk assessment vs the probability of occuring: how many times a day, a week, a year?
- the DM or CM nature of the incoming disturbance, as seen by the victim equipment
5.1 Filters for Reduction of Powerline Interference
While practically no modern equipment, with its fast digital circuits, switch mode power suply regulators and eventually RF devices, could meet EMC requirements without an efficient filtering, EMI filters are too often choosen on an empirical basis or a vague belief in manufacturers catalog performances.
A good EMI filter must:
a) attenuate the incoming power line noise to make the equipment immune to the most severe DM and CM agressions expected in its normal environment.
b) attenuate the HF noise caused by the equipment to its own power input, both DM and CM, such as it does not violate conducted EMI emissions limits.
Both a) and b) performances must be achieved in a system whose source and load DM and CM impedances are poorly defined. On top of these challenges, the filter must comply with size, weight and cost limitations, plus constraints on maximum permitted values for line-to-ground capacitance and high voltage withstanding tests. Needless saying, a filter is generally the result of a tight trade-off …
5.2 Selecting a filter
Given that an EMI filter is a low-pass element, the following parameters dictate our choice :
a) the required attenuation (DM/CM) for a given frequency, rigourously termed Insertion Loss (IL)
b) the cut-off frequency
c) the number of poles, which itself depends on a) and b)
d) the impedances on the source and load sides of the filter, that in turn will guide the filter scheme
e) capacitive-only, inductor-only, L-C, Tee or Pi type of filter
f) the normal service voltage/ current, to be considered for the filter capacitors and inductors
– Filter Criteria regarding EMI Immunity:
The attenuation must be such as the most suceptible circuits inside the equipment, must not receive on their dc input a noise voltage greater than their threshold of sensitivity.
Attenuation: A(dB) = 20log Vemi (w/o filter) / Vload (with filter)
This after having accounted eventually for the built-in attenuation of the power-supply regulator.
An equipment must not exhibit malfunction when submitted to bursts of 2kV(CM) 300kHz ringing transients on its 230V AC input. The noise immunity of internal digital circuits is 0.5V on their dc input. The existing filtering on the regularor output provide already 20dB attenuation.
What is the required filter attenuation? Answer:
A(dB) = 20 Log (2000/ 0.5) – 20dB
= 72 – 20 = 52dB @ 300kHz
– Filter criteria Regarding EMI Emissions
The attenuation must be such as the HF noise generated by our equipment on its power input (CM and DM) does not violate conducted EMI emissions limits.
Attenuation: A(dB) = Vemi (dBμV(w/o filter) – VdBμV(spec. limit)
Early testing on a prototype of our equipment has shown spurious harmonics of the switchmode power supply with amplitudes of 104dBμV from 350kHz up to 3MHz. The EN 55022 class B limit is:
- 350kHz: 50dBμV
- 1MHz: 46dBμV
What is the required filter attenuation? Answer:
- 350kHz: AdB = 104 dBμV – 50 dBμV = 54dB
- 1MHz: AdB = 104 dBμV – 46 dBμV = 58dB
* Remarks: – When substracting dBμV from dBμV, result is in dB, since arithmetically we are making a ratio of two voltages. – The filter requirement can be based on 54dB @ 350kHz, which will also satisfy the 58dB goal at 1 MHz: filter attenuation normally increases with frequency.
5.3 Number of elements in a filter
The number of poles in the filter (that is the number of L,C elements that are cascaded) is determined by the desired attenuation at frequency Femi, and by Fco, the cut-off frequency (or -3dB point), below which the filter has no attenuation. At a first glance, it would seem, for instance that a powerline filter,
– must exhibit no attenuation at all for 50/60Hz
– start attenuating any undesired frequency above say100 or 150Hz.
This would be a perfect, but enormous and expensive filter, because of the physical size of the capacitors and inductors. These are only found near large loads (greater than tens or hundreds of kWatts) for correcting the power factor of heavy inductive loads, or reducing harmonic distorsion on the utility side, upstream. So generally, the cut-off frequency of line filters for individual equipments have Fco values in the 10-30kHz range. Fig.6 give the attenuation of any filter, given its cut-off Fco and numer of poles (n).
With the filter of example 2, we were looking for 54dB at 350kHz.
Assuming a cut-off at 10kHz, how many poles our filter should have? Answer:
From Fig.6, we see that, given Femi /Fco = 350kHz / 10kHz = 35, a 1st order filter is not enough, our filter must be at least a 2nd order (n=2).
5.4 Influence of actual source and load impedances
One must beware of filter manufacturer’s data. They are generally measured in a 50Ω/50Ω set-up (per Mil Std 222), and actual in-situ attenuation may differ significantly. Actual insertion loss is strongly dependent on the value of the impedances seen on both sides, to such extent that the performance shown on the data sheet is probably the only one you will never get! Table Fig.7 shows the preferred choice for the 4 possible impedances configurations (ref.1). One simple, flawless rule that suffers no exception is: ”With filters, capacitors should look toward high impedances both sides, inductances should look toward low impedances, both sides”.
5.5 Coarse approximation for 50/60Hz filter components
By default of a more accurate calculations, the following rule of thumb can be used to define conservative, maximum values for the filter elements. The rationale is that, for an ac power mains, we do not want the filter capacitor to derive uselessly too much 50Hz or 60Hz current from the power mains. All the same, we would not like our filter inductance to cause unacceptable 50Hz or 60Hz voltage drop. If we define 1% as a tolerable impact on ac current consumption and input voltage, we get:
with Xc, XL being the impedances of filter capacitor and inductance at the power mains frequency, and ZL the equivalent load imedance. For a 50Hz ac input, using more practical units for filter components:
If there are several capacitors like in a ”Pi” filter, the formula applies to all the capacitors in parallel.
For the CM mode filter capacitors mounted line-to-chassis, that is to the earth connection, the constraint is a safety issue. In European countries the maximum permitted 50Hz leakage current is 0.5mA for the most severe applications (domestic appliances), medical devices excepted. So, these capacitors, termed ”Y” type are generally limited to 5nF, with a 2.7kVdc overvoltage survival test.
What are the maximum value of filtering capacitor (Line-to-Line) and inductor we can accept for a 230V equipment with 10A of normal line current?
The 10A / 50Hz current can be translated into an equivalent filter load of 230V / 10A = 23Ω. Thus, maximum values are: C(μF) ≤ 32/ 23 = 1.4 μF , L(mH) ≤ 0.032 x 23 = 0.73mH
5.6 Recommendations for Filters mounting
As much as its performances, the way a filter is mounted is crucial to its effectiveness. It should be installed as close as possible to the equipment wall opening where the power cord passes, the best being to have it mounted through the wall itself (See Fig 8). Filter case must be the metallic type, making a tight metal-to-metal contact with the equipment box, or at least with a metal barrier. Wiring on the line side of the filter should never be bundled with those on the load side.
5.7 Why Transient Suppressors must be preferred to filters for high energy transients
Although often thought of as ”clean-it-all” components, filters, being low-pass elements are unefficient against energetic pulses, with duration exceeding a few μsec. This is because long pulses have a large energy content in the low frequency range, where the filter has no attenuation. This is pictured on Fig.9, showing the response of a filter for two pulse waveforms having a same 1kV amplitude.
The short 100ns pulse falls down much before the filter has been charged up to the peak value, so the pulse is reduced to 10v, with its duration stretched over > 10μs. By comparison, the same 1kV pulse with 50μs duration (a lightning induced transient for inst.) is lasting long enough for the filter to get charged up to the peak value. What is needed is a component that is not frequency- selective but amplitude selective, as seen next.
6. Transient Voltage Suppressors
Suppressing high voltage, energetic surges is the role of Transient Voltage Suppressors (TVS) like Varistors, Transzorbs or Gas tubes. These non-linear devices are an open circuit up to a breakdown value, above which they become abruptly shunting elements with large current capacity. We will review the 3 principal types of TVS: MOV, TransZorb and Gas Tube
6.1 Metal Oxyde Varistors (MOV)
They are avalanche devices, with an ”Off” resistance ≥ 10^6 Ω. When the applied voltage reaches the avalanche threshold (Fig.10), their ”ON” resistance drops to a few Ohms. Their main features are:
- fast response, in the nsec range
- low cost
- fair energy handling (20J for a 14mm device)
- open circuit destructive mode when Imax is exceeded: can be a drawback, or an advantage, depending on the application
- slanted ΔI/Δ slope, causing the actual clamping voltage for Imax to be 2 or 3 times the Vbreak
- large parasitic capacitance (a few nF) that must be taken into account
6.2 Zener Diode-based TVS (Transzorb®, Transil®)
These are special Zener diodes with enhanced power dissipation capability. Their Vbr threshold is more accurate, with a much sharper ΔI/ΔV characteristic.
- fast response < nsec
- almost vertical ΔI/ΔV slope, thus a tight clamping voltage at Imax, typ ≤ 1.5 xVbreak
- fair energy handling
- short-circuit destructive mode when Imax is exceeded: can be a drawback, or an advantage, depending on the application
- large parasitic capacitance (a few nF) that must be taken into account
6.3 Gas Tubes
Gas tubes are a modern version of the air gap arrestor that was based on the arcing of a calibrated gap when the air breakdown voltage is exceeded. Modern gas tubes consist in a sealed envelope filled with low pressure gas and a precise interval between the electrodes
- high current handling capability since the terminals are pactically shorted during the arcing
- a low parasitic capacitance (a few nF) that must be taken into account
- slow to react, with delay > μsec
- puts the line in short-circuit when arcing, until the voltage waveform crosses zero
- not to be used on dc power lines, because the arc will be sustained by the dc voltage and will not extinguish, until a fuse blows, requiring a manual action for re-powering the equipment.
- does not exist for voltages below ≈ 90V. ”Solid State gas-tubes”: Solid State versions of the spark gap exist in form of thyristor-based crowbar devices that turn on when a certain
voltage threshold is reached, putting the line in short-circuit, protecting the equipment components and eventually triggering a breaker or fuse.
6.4 Selecting the proper surge suppressor
The selection steps are summarized below:
- select a triggering voltage (Vbr) ≈ 10-20% above the highest peak voltage of the power line.
- estimate the maximum peak current of the transient pulse. It can be found from the transient immunity specification applicable to the equipment. A better figure can be derived from the I,V load curve of the test generator, after substracting the clamp voltage from the generator open voltage.
- Check this peak current against the maximum, no damage peak current for the selected TVS.
- Check that the peak power: (Vclamp x Ipeak) is tolerable for the applied pulse duration
7. Isolation Transformers
An isolation transformer has distinct primary and secondary windings, breaking the CM ground loop between the external power lines and the equipment or system. The primary-to-secondary magnetic coupling allows the normal transfer of the DM (line-to-line) voltage, while the galvanic isolation prevents the CM voltages from passing thru the transformer. An ordinary isolation transformer is an adequate barrier for opening low frequency ground loops between an equipment (or an entire installation) and the power mains. But as frequency increases, the high value of the barrier isolation (typically > 10^6 Ω) becomes shunted by the inter-winding capacitance.
This capacitance is ranging from 50-100pF for small equipment transformer up to several nF for transformer in the kVA range, causing the primary/secondary protection against loop currents to almost disappear around a few MHz. Special winding sequence and more physical separation between the winding stacks can improve the CM rejection by about 10dB.
7.1 Faraday shielded Isolation Transformers
For more HF isolation the best technique is to use a Faraday-shielded transformer. (Ref.2) A non-closed foil is wound between the primary and secondary windings and grounded locally, such as a CM voltage appearing on either side is stopped by the shield, and the capacitive current is sunk to the chassis or structural ground. (Fig.11). An other advantage of an isolation transformer is to allow recreating locally an earthed Neutral on the secondary side, in a certain zone of a large facility.
Back-Up and Uninterruptible Power Supplies
When the problems are caused by serious disturbances, or total outages of the normal ac mains, the ultimate resource is to locally re-create a more dependable power delivery, using line conditioners or Uninterruptible Power Supplies (UPS).
A line conditioner is a special transformer with an automatic regulation circuit on its secondary. For power line over or under voltages like +/- 20%, these devices can provide a secondary voltage within 5%of the nominal value.
With modern electronic equipments using switch-mode power supplies, this feature has less interest because switch-mode can keep a constant dc output even with primary voltage varying within 100- to-250V. Yet, there are still many electrical equipments (motors, machine-tools, lighting appliances) whose longevity rely on an ac supply within 5% tolerances. But, except for ferro-resonant transformers that can keep-up with one missing cycle (10ms for 50Hz ac), a line conditioner has no energy storage and cannot make-up for a long 50% voltage drop or a total outage.
A UPS is an ac/dc + dc/ac converter coupled with a battery bank. It can provide a well regulated ac output, even during a momentary, full interruption of the regular ac mains. Two basic principles are used: ”Off-line” (or standby) and ”On-line”. For both devices, the autonomy is depending on the size, in Amp-hours of the battery storage. With the Off-line type, the loads are normally fed by the regular ac supply, while the ac/dc converter is keeping the battery bank fully charged. In case of ac mains shortage, the ac distribution to the secured loads is automatically switched on the dc/ac inverter, until the regular ac mains return to normal.
Notice that this system provides no protection to HF disturbances and transients of the normal ac mains. It is the simplest system, typically used for small installations, like small offices, shops etc… where most of the sensitive equipments like appliances, PCs and the like are normally equipped with EMC protection. With the On-line type, the secure loads are always supplied from the dcac inverter, hence decoupled from the regular ac mains, which serve only to charge the battery bank. In case of inverter failure, a no-interruption static switch is transferring the privelegied load line to the regular ac.
This static switch is, in turn, a possible cause of EMI problems because the solid-state switches are a mediocre isolation barrier against high frequency. This can be avoided by careful EMC filtering and transient protection of the back-up line.
9. Emission Aspects with Power line coupling
One earliest cause of user-created disturbances to the power mains have been the ac-dc rectification and gate-controlled rectifiers used in many equipments. The increasing use of high frequency converters like Switch-mode power supplies and inverters, variable speed drives, light dimmers, fluorescent lights, has shifted the noise emissions towards higher frequencies, causing both conducted and radiated EMI issues. In this respect, EMC filters should be regarded not only as a cure against conducted interference up to ≈ 30 MHz, but also against radiated emissions from the power cord at frequencies > 30 MHz.
9.1 Poor Power Factor and sinewave distorsion caused by non-linear loads
With a single wave ac rectification like on Fig.12, the current Ic charging the tank capacitor C is not a sine wave: It is a short current peak whose the product Ic x duration (that is Coulombs) is in turn, delivered to the load for a minimum ΔVr ripple on the dc output. Since power mains is not a zero impedance source, this brief, but high amplitude current is causing a small voltage collapse on the tip of the V0 ac sinewave.
Similar phenomena occurs with full-wave rectification or 3-Phase rectification (6 diodes or 6 gate-driven thyristors). All these schemes are causing ”notches” in the ac mains voltage waveforms. Distorsions appear as harmonics, spoiling the power factor (the ratio of the fundamental current to the rms addition of all the harmonic terms I1, I2, I3 etc …➔ In).
In civilian applications, maximum distorsion is ruled by IEC 555-2 norm, which require a control of the power factor by the end users of the power utility. The solutions consist in adding inductances in front of the capacitor, smoothing the risetime of the current demand, or using a Power Factor Controller (PFC), which is an active regulator acting as a dynamic current source.
9.2 HF emissions caused by Switch Mode Power Supplies and Inverters
Switchers are the main cause of conducted emissions, violating the Military and Civilian RFI limits by as much as 40 or 50dB for unfiltered items. Several mechanisms, internal to the switcher, are contributing to these emissions (Fig 13):
- The CM conducted emission path is generally the major limits violator, for equipments supplied in 115v or 230V ac. The leakage current is caused by the stray capacitance of the switching transistor, IGBT or MOSFET to their heatsink or nearby chassis. Leakage current (Icm) closes by the chassis ground, returning via the power mains impedance then back to the equipment Ph and Neutral input wiring. For testing, the noise level is measured at the LISN socket.
- Next to the CM current, the DM current that circulates in the power input wires is due to the ESR, ESL of the switcher input capacitor bank. Since these capacitors have parasitic series resistance (ESR) and self inductances (ESL) the flow of the main switching current creates a differential voltage across the primary input.
a) For the first harmonics of the switcher, the CM current driven by the dV/dt of the switches into the stray cap C1, behaves as a high impedance source (current source). Therefore, according to the table Fig.7, the filter should have at least line-to-chassis (CM) capacitors looking toward the high impedance.
b) To the countrary, the DM voltage, appearing across a low impedance – typically < 1Ω below a few MHz, behaves almost as a perfect voltage source. Therefore, using the table Fig.7, the filter should have at least, an inductance looking toward the low value ESR of the primary capacitor.
As a result of a) and b), the most appropriate filter structure is shown in Fig. 14) It is optimized against CM and DM, as a 2nd order filter (40dB/decade). The DM choke is sometimes provided by the leakage inductance of the double-wound CM choke, representing substantial space and cost savings. Some manufacturers of ”Off The Shelf” (OTS) SwitchMode PowerSupplies incorporate an EMI filter in their modules, some others do not.
So, equipment designers using OTS regulators should check if the EMI conducted emissions are documented by the vendor. Whatever they design their own filter or they plan for a commercial item, they should look at the most appropriate filter schematic, as described above.
Solutions for reducing conducted emissions of Switch Mode Power supply:
- CM current to Ground—> Use electrostatic screens, Use CM filter on input.
- CM noise on output dc loads —> Use transformer with screen, Apply special precautions in winding the transformer, Use CM filter on dc output.
- DM Noise —> Use low ESR / ESL capacitors —> DM filters.
- Recovery Spikes —> Use fast recovery devices —> Install RC snubbers.
Michel Mardiguian EMC Consultant, France firstname.lastname@example.org
- Mardiguian, M. Controlling Radiated Emissions”, Springer NY, 2014
- Mardiguian, M. EMI Troubleshooting Techniques, Mc Graw Hill, 1999