Time to Read 5 min
EMC (ElectroMagnetic Compatibility) is a red rag for many hardware developers and project managers, as are other hardware certifications. They often take longer and, above all, they often require changes late in the project, where they are very expensive and delay the market entry.
Certification can also be done efficiently. The following approach has been established at Solcept because it has proven to be fast and low-risk:
Clarify the Starting Point
Clarify the standards situation early on and create a system specification and architecture (frontloading), then start hardware and software development. Otherwise it can happen that a single paragraph of the standard ruins the whole concept.
Clarity through Modeling/Simulation
Define the critical EMC measurements and set up modeling/simulation for these, including the measuring instruments or equivalent networks in accordance with the EMC or safety test standard. The primary aim here is not to determine quantitatively where the measured (EMC) values will lie exactly. The main issue is to develop an understanding of the components and parameters that influence the measured values and how they do so.
Creating Degrees of Freedom in the Layout
Provide for filter, backup and protective components in the schematic/layout, especially at interfaces but also on critical networks (e.g. reset), like ferrites or protective diodes. These are bridged (e.g. a zero-resistance bridge instead of a ferrite) or not fitted (protective diodes) in the first prototype, because these elements only serve to create degrees of freedom for debugging.
If the measurements on the first prototype show that, for example, an interface radiates too much, you can easily fit the filter elements at this interface and use measurements in the internal EMC laboratory to set the necessary suppression or protection level. If you are planning to use protective or filter elements anyway, make sure that the space in the layout is dimensioned so that larger components can also be fitted.
First Prototype
Now you can have the first prototypes manufactured. By the first prototype, we do not mean a functional model or a demonstrator, as you can put together from evaluation boards, but a PCB that is already ready for series production.
For this prototype, it is important that the essential hardware concepts (security, EMC, clock, reset, ground, power) are defined and implemented (hardware module architecture) and that the environment of the electronics (e.g. housing, overall device) is defined and, above all, built to such an extent that meaningful measurements can be carried out in the laboratory.
EMC Measurements in an Accredited Laboratory
With the first prototypes, you should carry out as many safety, EMC and performance measurements as possible in an accredited laboratory, but at least the critical ones. Do not make any changes to the hardware in the laboratory; the only thing to do is to record the current state.
Accredited laboratories are expensive and usually there is no time for long cause analysis or optimizations, so the next step takes place internally:
Optimization and Elimination of Weak Points in the Iinternal EMC Laboratory
The internal laboratory is usually not accurate enough to determine absolute levels. Often it consists only of the necessary measuring instruments and setups, and there is no shielded measuring chamber. However, if you know the measured values from the accredited laboratory, you can work with relative measurements and thus eliminate the weak points.
The internal laboratory is close by, so optimization measures can be quickly verified. The simulations set up beforehand show which parameters need to be adjusted to achieve the desired effect.
In practice, for example, this looks like this:
- In the case of lightning protection (surge), the permissible voltage limits for the semiconductors are exceeded. You consider setting the varistor at the input to a lower value, but are aware that this can have an influence on standby consumption. You solder in the varistor, measure the voltage at the semiconductor (surge generator, oscilloscope) and then measure the standby consumption (wattmeter).
- The conducted interference is too high in a certain frequency range. You consider increasing the size of the Y capacitors, but know that you have to keep an eye on the leakage currents. Both can be verified very quickly internally (LISN (Line Impedance Stabilization Network), measuring receiver or spectrum analyzer, voltmeter)
- The radiated interference is well below the permissible limits. You are considering eliminating certain filter elements for cost reasons. Using an antenna and a measuring receiver, you first measure a reference (which corresponds to the measured values from the accredited laboratory), then eliminate the filter elements and test whether there is still sufficient margin by taking another measurement.
Pilot Series / Second Prototype
After debugging and optimization, remove the unneeded EMC and protection components (degrees of freedom, see above) from the schematic and layout and implement all the corrections from the optimization phase.
If massive or conceptual layout adjustments do not have to be made due to functional changes in the product requirements, then this release already has the character of a pilot series.
Final EMC Measurements in an Accredited Laboratory
You then take this pilot series/prototype to an accredited laboratory. Thanks to the optimization, everything should now be in order. If a large number of changes have been made, especially conceptual ones, a second round of optimization is required. Here, too, it is very efficient to use an internal laboratory.
Series Production and Measurement of the Series Product
Many standards require safety and EMC measurements on (randomly selected) series products. If you have worked carefully in the previous steps, there should be no surprises here.
We can support you in all steps, be it modeling, creating an (EMC) concept, debugging or setting up an internal laboratory.
Samuel Leemann
Do you have additional questions? Do you have a different opinion? If so, email me or comment your thoughts below!
No Comments