EMC basics: I/O

EMC Basics #5: I/O as critical circuits article gives some useful tips on the EMC issues related to inputs and outputs.

Digital inputs/outputs — The key concern for digital interfaces is ESD. A secondary concern is radiated emissions. Radiated susceptibility is rare with digital I/O, although possible at very high RF levels. The solutions for both radiated problems include filtering at the interface and/or or shielding of external cables.

Analog inputs/outputs — The key concern for analog interfaces is RF. High RF levels can cause rectification in the I/O circuits causing errors and/or noise. Typical solutions include high frequency filters and/or shielding of the external cables.

Relay outputs — Since relay drivers are usually digital, the regular digital concerns apply. In addition, inductive transients from the relay coils may pose a self-compatibility problem. Snubber circuits may be needed at either the relay (best) or at the driving circuit on the boards.

Contact inputs — Since the receiving circuits are usually digital, the regular digital concerns apply.

When designing or reviewing circuit boards for EMI, ALL of the I/O circuits deserve EMI attention!

I have some additions to those suggestions:

Opto-isolators (also known as optocouplers) work to protect the receiving system at the expense of the sending system needing to drive the cables/interconnects. They are a great way to isolate digital from power circuits but have limited bandwidths. Fairchild Application Note AN-3001 Optocoupler Input Drive Circuits gives some implementation tips for optocoupler based input circuits.


Using a balanced line interface for sensitive and/or fast signal is a very good idea. Using balanced interface reduces EMI pickup and radiated EMI considerably compared to single-ended signals. Applications like telephone lines, analogue instrumentation, professional audio signals, fast serial bus standards and Ethernet all use balanced interfaces to get good noise performance.

Be careful on the grounding of cable shield when they enter the cabinet. The cable shields should be grounded at the point where they enter the metal cabinet. This will stop the RFI from entering inside the device. This advice applies especially to sensitive analogue circuits like audio interfaces. Proper grounding is essential in keeping RFI and ground loop noise away.

In many power controlling applications you can’t beat a relay for isolation or low on-resistance, as well as low cost. For relay outputs you need to carefully consider the need for snubber circuits. When talking about snubber circuits there are two kind of applications for them: Snubber cuircuit in parallel with the relay coil and snubber circuits in parallel with the relay output.

For the relay coil driven with DC voltage at known polarity an inexpensive diode in parallel with the coil works well. If the relay is switched with AC, the DC polarity is not known or you need very fast operation (parallel diode can slow down relay release time).

You need to consider snubber circuit also at the relay contact side especially if you are switching anything that is even slightly inductive. Relay contacts can arch. The end result of Contact Arc Phenomenon is shortened contact life. In addition to that arching causes lots of electromagnetic interference.

Relay Contact Life article tells that perhaps the most popular method of quenching an arc between separating contacts is with an R-C network placed directly across the contacts. Contact Protection and Arc Suppression Methods for Mechanical Relays gives information how to design a suitable R-C network for quenching an arc.


Some relay users connect a diode across the inductive load to prevent counter-voltage from reaching the contacts. In some application zener diodes are used. The MOV performs in a manner similar to back-to-back zener diodes, and can be used in both AC and DC circuits.

An added benefit of arc suppression is the minimization of EMI. An unsuppressed arc between contacts is an excellent noise generator. Arc may radiate energy across a wide spectrum of frequencies. By suppressing the arc, electromagnetic interference is held to a minimum. By quenching the arc quickly, this action is held to a minimum. The result often is a considerably lessened amount of electromagnetic and radio frequency interference. Contact arc noise can be troublesome to sensitive components in a circuit. In worst-case conditions, EMI can cause unwanted turn-on of IC logic gates, SCRs, and triacs, and can cause damage to other semiconductor devices.


  1. Tomi Engdahl says:

    How to Read Data Sheets: Relays

    When you are used to working with semiconductor devices, the data sheets for relays may pose some problems. What do terms like ‘Operate Voltage’ and ‘Release Voltage’ mean, for example?

  2. Tomi Engdahl says:

    How to Read Data Sheets: Latching and Automotive Relays

    Latching relays latch mechanically or magnetically so you don’t have to maintain current running through the coil to keep them on.

    Latching Relays

    you can buy latching relays. These relays latch mechanically or magnetically so you don’t have to maintain current running through the coil to keep them on. This sounds like a great solution for my battery cutoff switch

    Looking at this data sheet, we see that this beauty is rated for 60A at up to 250VAC, and it will break a maximum of 15kVA. The only downside is that the initial voltage drop is 1mΩ at 3A. While the 1mΩ is very good, the 3A gives me pause. This implies that it might not work so well if I have a low power radio.

    Moving on, the coil resistance is much lower than for the non-latching relay, which means that the power required is higher (which makes sense if it has to latch). The two-coil version has a resistance of 48Ω, which is 0.25A at 12V and will dissipate 3W. No wonder there’s a maximum duration.

    irst, Coil data followed by Contact data and other specifications.

    The Set and Reset voltages are presented a bit differently than in the previous data sheet in that they are given as a 75% of the rated voltage. For a 12V coil this is 9V. The coil resistance for the two-coil version is 72Ω. The contacts are rated at 50A with the minimum load specified as 100mA. So far, this part is looking better than the other latching relay for my purposes.

    Because a relay is also an electrical device, there is another component of its life relating to the load that is being switched. If you are switching a resistive load of 50A, for example, you’ll get 1 x 104 operations, but if you switch 25A, they you’ll get 1×105 operations.

    Automotive Relays

    The Panasonic Electric Works CB1A-M-12V looks similar to ones I’ve used in the past. This part is rated at 40A and I should be able to get something very similar at my local automotive store…

    …or maybe not.

    According to the data sheet, in addition to the bracket mount with tabs that I’m familiar with, these parts also come in PCB and plug-in types. They also come in heat-resistant types for use in the engine compartment.

    The Coil data section is a bit different for this relay. Instead of just giving guaranteed Pick-Up and Drop-Out voltages, we are given a range. This is extremely useful information if we want to reduce the power consumption required to keep the relay on. For example, the 12V part has a 103Ω coil. Since the dropout voltage is 4.2V or less, we can use a 150Ω resistor in series to hold the relay on with 4.8V. Of course, this assumes that the Pick-Up and Drop-Out voltages are constant over temperature.

    Moving along, for the standard 12V type, Contact Resistance is 2mΩ (measured at 1A), while the heat resistant type is 15mΩ. Both of these are good numbers. The minimum and maximum switching capacity is given for 14VDC

  3. Tomi Engdahl says:

    Relays Are Great, But There’s No Need to Let Them Waste Power

    The relay remains one of the most useful components in the designer’s kit, but its power-hungry operation can be overcome with some non-static drive and specialized ICs.

    And why not? Yes, the relay is “ancient” and may not get a lot of respect or consideration from novice engineers, yet the countless models on the market covering an enormously wide range of specifications attest to its value, and I am sure millions of units are shipped every year (and we’re not even considering at RF relays).

    Let’s face it: the relay is easy to use; easy to size; the input coil and output contacts can have wildly different ratings (current, voltage, AC, DC); the contacts are signal agnostic to a large extent; the contacts are easily “floated” (not grounded); and the contact arrangement is offered in many problem-solving variations, including normally open (NO), normally closed (NC), both NO and NC, and multiple independent poles, and various mounting styles, to cite a few. The absolute isolation between the coil drive-side circuit and the contact-closure side is another major benefit. In short: they do lots of things well, and don’t give headaches or unexpected surprises; further, if used within specifications, a quality relay has a million-plus cycle life. What’s not to like?

    Actually, there is one thing, a shortcoming that is inherent in the relay’s physics (and its close cousin, the solenoid): they can use a lot of power. As a device based on magnetic-field energy, it takes current to energize the coil, and using current means dissipating power (obviously, a waste) and self-heating (which has important consequences).

    Fortunately, there are ways around this dilemma. First, it’s important to know that the relay “hold” current is typically about half the pull-in current.

    IC vendors know this, and also know that relays are a big opportunity. That’s why there are application-specific ICs such as the Texas Instruments DRV110 PWM current controller which offloads the microcontroller, (Case 3). It allows the user to set initial activation current, time at that peak current, hold current, and other factors,

    It’s important to have independent control over the activation and hold currents due to that unavoidable irritation of circuit design: high temperature. As the relay coil heats up due to the applied current, the resistance of its copper windings increases substantially

    This, in turn, reduces the current to the coil, and thus the ampere-turns which develop the magnetic field.

    So, don’t be afraid of the relay, it can be a viable solution – and often, the best or even the only solution –to a situation where the controlling signal and the load are very different or need to be isolated.

  4. Tomi Engdahl says:

    Blinds (Or Any High Power Motor) Control

    How to control several blinds with inexpensive relay boards (not ruining them) with physical buttons and remotely simultaneously.

    wanted to automate everything, starting with the blinds. What seemed to be an easy project turned out to be a bit of a nightmare: the relative high power motors (150W) were destroying my relays and triacs, what wireless communication and controller to use and how to make it work was not obvious, noise in the lines were causing random activation of the blinds (pretty scary in the middle of the night)…

    When controlling motors, an important issue is their inductance, which causes that when trying to open the circuit, the current insists on keeping flowing through your breaking device, causing a very high voltage. If you try to break the circuit with no precautions with a small relay, their contacts will stick together, and if you use a triac (solid state relay) the over-voltage (in my case I measured peaks of more than 1600V) will destroy the semiconductor.

    I realized by googling, that other people had issues with this, but they took the easy, expensive and voluminous way, they just get bigger relays, still needing the cheap relays just to activate the bigger ones, while the contacts will still suffer and may fail eventually. As an engineer I could not allow myself not to get the most efficient solution. :) In the schematic below you have the solution to spare this big relays just by adding one resistor, one capacitor and one varistor.

    The varistor protects the triac from an over-voltage. The resistor plus capacitor forms a RC Snubber circuit that absorbs the energy during the breaking commutation.

  5. Tomi Engdahl says:

    Ask the Engineer: Why Should a Relay Meet the New NEMA 410 Standard?

    NEMA is a standard that defines the worst case inrush current expected from Lighting Control Switches/Relays to be used with electronic drivers and discharge ballasts. This includes both self-ballasted compact fluorescent lamps and integrated LED lamps. NEMA specifies that any 20A devices are tested with a 16A load representing 80% of the branch circuit rating.

    Panasonic Solutions for Lighting Controls

  6. Tomi Engdahl says:

    Beware of Zero-Crossover Switching of Transformers

    A zero-crossover solid-state relay may be the worst possible method of
    switching on a transformer or a highly inductive load. Evidence1
    has come to light that zero-crossover turn-on of such loads can cause a surge
    current of perhaps 10 to 40 times the steady state current, whereas
    turn-on at peak voltage results in little or no surge.

    Surge currents of such magnitude can seriously shorten the life of the
    zero-crossover SSR, unless the SSR has a current rating well in excess
    of the load.

    Additionally, these surge currents create thermal and
    mechanical stress on the windings of the inductance and on the
    transformer core laminations. These stresses can lead to early failure of
    the device.
    The cause of inrush currents of such magnitude is core saturation.

    A 150 VA transformer has a 120 volt primary DC resistance of
    approximately 1.5 ohm, and a 500 VA transformer, a 120 volt primary
    resistance of approximately 0.3 ohm. One might think a 5 amp zerocrossover
    SSR would be more than sufficient to switch the current of
    the 150 VA transformer. However, during core saturation, primary-winding
    inrush is 80 amps

    Under such conditions, the SSR is severely overloaded, and the
    transformer overheats. (Power expended in the primary during this 400
    amp surge would be approximately 40 KVA.)

    A “zero-crossover” SSR does not always turn on at precisely zero voltage.
    It takes perhaps a millisecond or more for the circuitry to react. Therefore,
    the load switch may not be fully on until load voltage is perhaps 15 to 20
    volts. In this event, surge current isn’t as great, but it is still potentially
    destructive. Also, a random turn-on SSR may, at times, turn on at or near
    zero cross-over. The best method of turning on transformers and other
    saturable, highly-inductive loads is by use of a peak voltage turn-on device.
    Turn-on at peak voltage results in minimal surge, if indeed any surge is
    present at all.
    Zero-crossover SSRs are excellent switches for resistive capacitive, and
    slightly inductive loads.

    Even so, inrush current must be taken into
    consideration. That is, an incandescent lamp can pull a “cold-filament”
    inrush current of 10 to 20 times the steady-state “hot filament” current.
    A motor can pull a “locked rotor” current of perhaps 6 times its running
    current. And the inrush of a capacitor, or the inrush of a circuit in which
    significant stray capacitance is present, is limited solely by the DC
    resistance of the circuit.

  7. Tomi Engdahl says:

    Understanding Signal and Power Isolation Techniques

    Electronic isolation is a means of preventing the transfer of direct current (dc) and unwanted alternating current (ac) between two parts of a system while still enabling signal and power transfer between those two parts. This kind of isolation is required in a number of instances, such as:

    Protecting industrial operators from high voltage.
    Protecting expensive processors and related circuits from high voltage.
    Preventing ground loops in communications networks.
    Improving noise immunity.
    Communicating with high-side devices in a motor drive or power-converter systems.
    Industrial equipment that requires isolation includes programmable logic controllers (PLCs), motor drives, medical equipment, solar inverters, electrical vehicles (EVs), and some special power supplies.

  8. Tomi Engdahl says:

    Have Your Cake and Eat It, Too: Overcoming Conflicting Isolation and EMC Standards

    Sponsored by: Texas Instruments. Applying techniques to improve EMC performance plus integrating compact data and power isolators can bring together the often-at-odds benefits of small size and excellent emissions performance.

    Designers of power systems often must satisfy sets of seemingly incompatible goals. Small size. High power. High efficiency. Low emissions. And, of course, low cost. Add a requirement that your small, high-power, efficient, quiet, inexpensive design also include isolation, too, and the harried designer might start to feel like there’s too much cake—sorry, functionality—to fit into too little a mouth—er, space.

    An isolated design prevents the flow of both direct current (dc) and unwanted alternating current (AC) between different sections of a system while still allowing signal and power transfer between them. Applications employ isolation for many purposes: to protect human operators and low-voltage circuitry from high voltages; to improve noise immunity; and to reduce the effects of ground differences between communicating subsystems.

    Digital isolators provide input-to-output isolation of CMOS or TTL logic-level signals, but complete isolation between two system blocks requires isolated power, too.

    Low-Emissions Design Challenges
    A highly integrated design may satisfy the isolation and space requirements, but the total solution must also satisfy the applicable EMC regulations. Thus, the designer must pay attention to several system-level issues.

    Radiated emissions stemming from the integrated transformer are a particular cause for concern.

  9. Tomi Engdahl says:

    Reinforced Isolation Passes the Test

    Protecting users from potentially lethal voltages and currents is a key requirement in any electrical design. The latest generation of reinforced-isolated devices is allowing designers to replace two separate levels of protection with a single component that performs the equivalent function. This article will discuss the semiconductor technology that enables this advance and the tests that ensure user safety isn’t compromised.

  10. Tomi Engdahl says:

    Choosing the Right Switch: Know Your AC from Your DC

    When picking a switch to use in electrical design, many people believe they can use any switch, provided its current rating is more than the maximum load in the circuit. This, of course, isn’t true.

    Alternating current (ac) or direct current (dc) circuits are capable of carrying very different currents

    This is why it’s so important for designers and engineers to understand how to pick the right switch for their product.

    Imagine you have two circuits, each carrying the same current—one is an ac circuit and the other is a dc circuit. When you switch off power to an ac circuit, a voltage spark (or arc) is created inside the switch that’s quickly extinguished—a desirable condition. This is because an ac sine wave is naturally at zero amps twice per cycle. So, there’s a 50% chance that the power to the circuit will not be at peak levels when the power is switched off.

    However, this isn’t the case in a dc circuit, where both the current and voltage are constant. When power is switched off to a dc circuit, it can take much longer for the voltage arc inside the switch to extinguish itself. This is why the switching speed (how fast the switch contacts open and close) of a switch in a dc circuit is so important. The objective is for the switch contacts to separate as fast as possible when turning off power to the circuit. This helps minimize the arc’s time to develop and extinguish itself.

    The longer a switch takes to open or break the power to the circuit, the more prolonged the arc. This can result in the switch contacts becoming pitted, potentially lead to overheating, premature switch failure, or even a fire.

    Know your Load Type

    As well as the speed at which the switch works, the nature of the electrical load is important in determining whether a switch will be suitable. Will it be switching an inductive or resistive load? This will affect both the voltages and currents your switch has to deal with.

    an inductive (L) load, such as a transformer or electric motor, will initially draw a large amount of “inrush current” when first switched on before settling back down after a few seconds to the load’s full running current. Also, when an inductive load is switched off, a huge voltage develops across the switches’ contacts in the form of an arc. This arc voltage can be much higher than what the switch is rated for, which can lead to contact pitting and shorten the switch’s life.

    For the same current rating, the dc voltage rating on a switch will generally be much lower than its ac voltage rating. For example, a switch rated for 15 A at 250 Vac will only be rated for 15 A at 12 Vdc. The only difference is whether you are dealing with ac or dc.

    it’s always best to choose one that’s been certified to a recognized standard. This will ensure it performs—both electrically and mechanically—the way it’s intended. Two important standards bodies in this regard are the Underwriters Laboratories (UL), and Verband der Elektrotechnik, Elektronik und Informationstechnik (VDE).

  11. Tomi Engdahl says:

    Shocking protection with reinforced digital isolators

    If you’re a person that’s working in a dangerous, high voltage area, you might take a pause and think about whether you have the right electrical protection. If the designers of the systems around you have selected devices that meet the international safety standards for two-barrier or reinforced isolation, you stand an excellent chance of being safe. Both types of barriers protect system electronics and, most importantly, the people using the equipment.

  12. Tomi Engdahl says:

    #127 Relay Comparison – Electromechanical, SSR, Latching

    All relays provide some form of switch. Relays come in various shapes, sizes and purposes. They are operated electrically, electronically or even manually. In our Arduinite world, they are used to enable higher voltages and/or currents to be switched by something that could not handle that requirement. So your Arduino, PIC or Raspberry Pi can now switch high current household voltages without a problem.

    So which one to choose? And why choose one type over another?

    In this video, we look first at the traditional, electromechanical relay.

    We next look at a Solid State Relay (SSR).

    Finally, we take a look at a mechanical latching relay.

    Now be aware that some Far Eastern manufacturers offer “latching” relays that are not the same as the one used here. Their “latching” mechanism still requires power to be applied at all times, and the flip-flop between contacts is controlled by a further on-board component.

  13. Tomi Engdahl says:

    Inductive spiking, and how to fix it!

    A description of inductive spiking, why it happens, and how a diode can save your circuits.

  14. Tomi Engdahl says:

    Automotive EMC Measurements with Oscilloscopes

    Today’s automobiles are an amalgamation of technology from three different centuries: combustion engines from the 19th, electrical systems from the 20th, and electronics from the 21st. The rapid rise in electronic complexity within vehicles means many more active components and assemblies than ever, all of which emit some amount of RF noise that can affect other active components and assemblies.

    As a result, the automotive electromagnetic environment is rife with both radiated and conducted interference in frequencies that range from low (LF) to super high (SHF). Sometimes the imposition of one device’s RF noise on another device can result in unanticipated, and unwelcome, changes in system operation.

    With so much electronic content within vehicles, electromagnetic-compatibility (EMC) testing has become both more essential and more challenging. In this article, we’ll discuss some techniques for using oscilloscopes to assist with the rigorous testing requirements.

    This article provides some insight into making effective automotive electromagnetic-compatibility measurements using an oscilloscope.

  15. Tomi Engdahl says:

    Design for EMI testing step-by-step guide

    Do you want to pre-test your design for EMI performance? Or do you want to correlate time and frequency domain in your signals to quickly find the root cause of your EMI problems?

    The challenges in testing EMI during early product cycle are multifold.

  16. Tomi Engdahl says:

    5 techniques for fast, accurate power integrity measurements

    This guide describes five tips for making accurate power integrity measurements with oscilloscopes from Rohde & Schwarz.

  17. Tomi Engdahl says:

    Scott Swaaley On High Voltage
    Hackaday Supercon – Scott Swaaley : Lessons Learned in Designing High Power Line Voltage Circuits

  18. Tomi Engdahl says:

    Solid State Relay (SSR) – Types of SSR Relays – Construction & Operation

  19. Tomi Engdahl says:

    The connector is often the EMI problem

    When it comes to reducing EMI problems, most coax cables behave the same. They are all just as good antenna for common currents to radiate and fail an FCC certification type test. It’s not the cable that is the chief source of EMC test failures, it’s usually the connector.

    If there were any net current, or common current on a cable, to return through the stray fringe fields between the entire cable and the floor, back to the chassis, it would radiate. In the most sensitive FCC part 15 test condition—for 88 MHz and below—in a class B test, the largest acceptable far field at 3 m from the product is 100 µV/m. This is an important number to remember.

    This means that if there is a larger field at 3m from the product than 100 µV/m, at 88 MHz, within the 120 kHz bandwidth of the FCC test, the product will fail EMC certification and not be allowed for sale in the US. Other countries have similar certification requirements.

    In a typical 50 Ω coax cable, with a 1V signal, having a 1ns rise time, the signal and return current is about 1 V/50 Ω = 20 mA. Even if the asymmetry is so slight as to generate only 0.1 nH of total inductance around the return path of the connector, the ground bounce voltage generated would be

    If the impedance the common current sees returning through all those fringe field lines is about 200 Ω, this 2 mV of ground bounce voltage will drive I = 2 mV/200 Ω = 10 µA. Is this a lot or a little?

    Remember, it only takes 3 µA of common current to fail an EMC certification test. This ground bounce driven current in the cable shield will cause an EMC failure.

  20. Tomi Engdahl says:

    Switching Inductive Loads with Safe Demagnetization

    Abstract: The purpose of this application note is to provide the system engineer with details of the unique features of Maxim’s MAX14912/MAX14913 products, and in particular, to explain how these products can safely handle 24V DC loads of “Unlimited Inductance” using Maxim’s patented SafeDemag feature.

  21. Tomi Engdahl says:

    Digital Output Drivers: Understanding Key Features and Challenges

  22. Robert says:

    I visited a few computer stores, looking for DB25M-DB25F cables, and really had a hard time finding them. Interesting how technology marches on – something that was once ubiquitous and easy to find is now difficult. So, I ended up placing an order with Famaga ( https://famaga.ch ), who we deal with at work. They’re a good supplier – and the cables arrived in a day.


Leave a Comment

Your email address will not be published. Required fields are marked *