Interview with Ronny Kirsch, Managing Director at Schlöder GmbH, on successful EMC testing
‘Those who test internally have the upper hand in the laboratory.’
More and more development departments are conducting their EMC tests internally. However, not every measurement meets the requirements, and not every test device delivers results that are relevant for approval. So where does meaningful pre-compliance testing begin? What is required for official certification? And what are the most common reasons for in-house EMC tests failing in practice?
We spoke to Ronny Kirsch, Managing Director at Schlöder GmbH, about typical pitfalls, the right testing strategy and the importance of suitable test equipment. His company develops and sells specialised EMC test systems, so he is familiar with both the technical side and the typical challenges faced by users. A conversation for anyone who wants to leave nothing to chance in product development and navigate the approval process with confidence.
Mr. Kirsch, who actually needs to concern themselves with EMC testing?
Anyone who wants to place an electronic device on the market in the EU. Without CE marking, such products may not be sold or put into service, and a key component of this CE marking is proof of electromagnetic compatibility.
In technically demanding industries in particular, detailed EMC reports are now a matter of course in tenders or approval processes. Anyone who cannot provide clear documentation in this area will be at a disadvantage in case of doubt – regardless of how well the product actually works.
So what exactly is tested in an EMC test?
Essentially, there are two things that need to be ensured: Firstly, that the device does not cause any impermissibly high electromagnetic interference, i.e. that it does not interfere with other devices. And secondly, that it is sufficiently robust to withstand external interference.
There are defined EMC tests for this purpose: Emission tests measure the interference emitted. Immunity tests check how the device reacts to defined disturbance variables. The aim of both is to verify that no malfunction occurs under standard conditions.
Many developers start testing early on. How exactly do pre-compliance and compliance tests differ?
During pre-compliance or development testing – usually carried out in a dedicated internal laboratory – the product is tested during development to see if there are any initial indications of EMC problems.
This is essentially a pragmatic preliminary check designed to avoid any major surprises in the subsequent approval test.
The actual compliance test is then about proving compliance with the standards – which is why these EMC tests almost always take place in an accredited laboratory.
Could I carry out these EMC tests myself if I had the right equipment?
In principle, yes. Many tests that accompany development, such as ESD, burst, surge or simple emission measurements, can be carried out effectively in your own laboratory with the appropriate equipment.
The test equipment is often identical to that used in an accredited laboratory, i.e. it complies with standards in terms of measurement accuracy, calibratability and signal generation. We recommend that your own test setup should be as close as possible to the final setup in the test laboratory. Does the laboratory test in accordance with IEC 61000-4-6, e.g. with CDNs or with an EM coupling line?
Due to the technically high tolerance ranges in the EMC standards, we always recommend over-testing your own products. At least 10%, preferably 30%.
Most customers today are well aware of the importance of testing during development. This realisation comes at the latest when a planned product release has to be postponed because the product has failed the final EMC tests.
Which standards are relevant for developers?
Each target industry refers to different industry-specific EMC product standards. In most cases, these in turn contain references to generic basic standards such as the IEC 61000-x series or dedicated ISO standards, which describe the test procedures, test setups and devices in detail. Selecting the right test equipment is therefore no trivial task.
It is not enough to simply buy any interference generator or measuring receiver – the device must comply exactly with the required standard, cover the right test types and pulses, provide the necessary test levels and be reliably calibrated. Making the wrong decision here not only risks incorrect measurement results, but also failure in the approval process.
So what is the best way to proceed?
First, you need to ask yourself: Which EMC tests are relevant for my product, and which ones am I willing and able to carry out in my own laboratory? In many cases, a pre-compliance setup is sufficient to identify the most common weak points during development. For the final test – i.e. the standard compliance test – EMC testing in an accredited laboratory with appropriately certified equipment is usually required.
Once it is clear which EMC tests you want to cover internally, the next step is to identify the relevant standards. These specify the types of disturbances that must be tested, such as ESD, burst/EFT, surge or even conducted disturbances, voltage interruptions or magnetic fields. Next, you need to determine which test levels are required: What voltages, currents or pulse shapes must the test device generate?
Only now does the question of device selection arise: Does it make sense to use a combination device that can cover several types of interference, e.g. burst, surge and voltage interruption in one device? In any case, it is important that the device can actually perform the required EMC tests in accordance with the standards.
Those who analyse these steps carefully will save a lot of effort afterwards and arrive at reliable results more quickly.
What technical specifications must the devices meet in order to be used in such EMC tests?
The requirements are clearly defined in the respective standards. For example, what pulse shape, repetition rate or impedance the device must deliver, how the test voltage is regulated or what accuracy is required. It is equally important to prove that the device has been regularly tested and calibrated, e.g. by a factory calibration or an accredited body.
It is also important to consider the service life of the devices. In the EMC field, devices often last 10, 15 or even more years. Here, it is important to ensure that the devices can still be converted or adapted as far as possible even if the standard changes (Schloeder GmbH | EMV-Systeme & Komponenten - 28 year old SFT 400 successfully repaired!), and, if necessary, achieve higher test levels than those specified today.
The EMC test result depends not only on the device, but also on the test setup. What needs to be considered here?
Exactly. The test setup is crucial for reliable measurement results. The complexity of the test setup usually depends on the test item and the standard. There are differences here, for example, between whether a household appliance can be tested relatively easily via thesafety socket of a burst or surge generator, or whether complex control electronics are tested in the area of conducted interference via various CDNs and EM coupling lines.
It is not enough to make the setup for the EMC test ‘roughly correct’. Even the smallest deviations, such as a missing terminating resistor, can significantly influence the measurement result. Absolute precision is required, especially for conducted immunity tests or field strength measurements.
I therefore always recommend consulting an EMC expert. They can help to implement the test setup in accordance with standards and tailor it to the specific product application.
Fortunately, the days when customers wanted to save money on an ESD generator and use an electronic lighter for ESD tests are over.
At the latest for formal CE approval, EMC testing in an accredited laboratory is required. When is this step due?
Proof of electromagnetic compatibility is required by law in the EU and is part of the CE marking process. In order to carry this out in a formally correct and legally compliant manner, an accredited EMC laboratory is usually required.
‘Usually’ because it is theoretically also possible to carry out the verification in your own laboratory – provided that the test environment meets the same strict requirements as an accredited laboratory and the manufacturer can fully demonstrate compliance with all standards. In practice, however, this is rarely the case because it involves considerable effort and expense.
During development, on the other hand, a lot can be covered in-house: initial burst, surge or ESD tests, emission measurements or magnetic field tests. In-house EMC tests are so popular because they identify weak points before the product enters the final approval process. After all, no one is happy about expensive redesigns in the final development phase.
Which EMC tests are particularly suitable for the development phase?
A classic starting point is the ‘holy trinity’ of EMC tests.
ESD tests – i.e. testing for electrostatic discharge – are useful in this phase. They show how sensitive a device is to contact, e.g. on USB ports or control buttons. These tests can be carried out on the open assembly or later on the finished housing.
Burst tests, which simulate fast transient disturbances on supply lines and signal paths, are also useful. These can be carried out directly on the circuit board, even before the device has a housing. This allows, for example, the resistance of power supplies or communication interfaces to be tested – a common weak point.
In the later stages of development, surge tests are also worthwhile to simulate overvoltage scenarios, such as those that can occur due to mains switching or lightning strikes. This test also makes sense in-house, especially for devices with a mains connection or long cables.
And how much time should be allocated for such pre-compliance tests?
Those who already have experience with their own product category and have suitable testing equipment available can complete initial tests in just a few hours. For more complex systems with many interfaces or demanding standards, it can sometimes take several days.
You already mentioned that testing is sometimes carried out at the circuit board level and sometimes at the device level. What does this depend on?
It essentially depends on the development phase and the objective of the EMC test. In the early concept and prototype phase, testing is often carried out directly at the circuit board or component level. There is a simple reason for this: you are still close to the layout and can make quick and targeted changes if problems arise, before a housing even exists. Burst tests with coupling probes or ESD tests on the bare circuit board are particularly efficient and helpful in this phase.
However, this is no longer sufficient for the final evaluation of a product with regard to CE approval. The bad news is that just because the individual components have passed the EMC tests, this does not necessarily apply to the overall system. This is why the complete system is required, i.e. the fully assembled device unit including housing, control elements, cables and power connection. This is the only way to realistically assess how the product behaves under actual operating conditions and whether it is sufficiently interference-free and low-emission in practice.
However, testing at the circuit board level early on creates the best conditions for a successful final test at the device level.
How can the test engineer determine whether the device is functioning correctly under EMC influence?
That depends entirely on the device and its function. In some cases, it is very simple: a status LED lights up green, a display shows measured values, or a beeper emits an acoustic signal. Then, during the interference simulation, the test engineer can directly see or hear if something goes wrong during the EMC test.
For more complex devices, this is not enough. Automated test sequences are required here: for example, special software can run through specific functions during the EMC test and log the results. Or camera systems can be used when a visual inspection is necessary, such as for displays or user interfaces.
It is important that, before the EMC test, developers and the laboratory jointly define what is considered ‘correct function’. This can vary greatly depending on the device. Only when it is clear what the device is allowed to do in the event of a malfunction – and what it is not allowed to do – can its immunity be objectively assessed. Often, the problem is not the failure of a machine, but an undefined or uncontrolled state. Imagine a robot arm on an assembly line that develops auncontrolled life of its own after a brief power failure.
Cables in particular often cause uncertainty in EMC testing. What needs to be considered here?
EMC testing must be carried out under realistic conditions. This means that the exact cable types and lengths specified in the manual or technical documentation for the intended use must be used. Exceptions, such as those in IEC 61000-4-6, are clearly defined.
During development, the device is often tested with shorter or better shielded cables, which leads to supposedly better measurement results. In the subsequent approval test with the actual cable configuration, interference or limit value exceedances suddenly occur. Such surprises can be avoided by working with the ‘real’ cables from the outset, provided this is permitted by the standard.
Conducted immunity tests in accordance with IEC 61000-4-6 are particularly complex. Here, each interface must be tested individually.
That's right – every power cable, every data connection, every control cable. And each one must be tested with a suitable test device (CDN) that is precisely described in the standard. Unused CDNs must then be terminated correctly (open or terminated with 50 ohms). The more connections a device has, the more extensive, complex and time-consuming the EMC test becomes.
The approach of simply placing all cables in an EM coupling section is also difficult, as this is only permitted for supply lines in exceptional cases and, on the other hand, decoupling is more complicated than with a setup using CDNs.
So let's sum up: is it worth having your own EMC laboratory in a company?
That depends heavily on the company's development focus – but in many cases, yes, having your own EMC laboratory can be a major strategic advantage.
If new products are developed regularly and these are also technically similar, internal test setups can be optimised and standardised. This saves an enormous amount of time, as you can test independently and immediately when necessary. Having your own laboratory is also financially worthwhile. Although the initial investment is not insignificant, those who test regularly often recoup this investment after just a few projects. Not to mention how much time and money a failed approval test can cost.
I would also argue that an in-house laboratory strengthens the company's development expertise. Developers can see directly how layout changes or new components affect EMC behaviour, respond in a targeted manner and build up know-how and experience that is valuable to the company.
And for customers who do not want to purchase the equipment for an internal laboratory right away, we have rental equipment for various standards.