What happens when a measurement stands out?
During development, an exceeded limit is not a setback but a finding. What matters is what happens next – and that in the end it is documented which measure actually worked.
Short answer
We stay with the device. The measured value turns into a narrowed-down interference source, that turns into a reasoned measure, and that turns into a re-measurement under comparable conditions. What you end up with is not the assumption that something helped, but a documented before-and-after comparison.
The process in four steps
The steps repeat as often as development requires.
Step 1
Put the value in context
Which frequency range, which operating state, how pronounced? Only once that picture is clear is it worth searching for the cause.
Step 2
Locate the source
With near-field probes we work over the board, the cables, the connectors and the openings in the enclosure, and narrow down the place.
Step 3
Implement a measure
One reasoned change, implemented on its own – to the setup, the cable routing, the filtering or the shield bonding.
Step 4
Re-measure
Verification under comparable conditions. Only the re-measurement shows whether the measure works, and how strongly.
Why only one measure at a time
The most common mistake in troubleshooting is to implement several changes at once: a ferrite added, a cable rerouted, a shield bond tightened. If the level improves afterwards, nobody knows which of the three changes did it.
For series production that is expensive: when in doubt, all three measures end up in the product, even though one would have been enough. So we change one thing at a time and measure in between.
What “comparable conditions” means
A re-measurement only says something if the measure is the only thing that changed between the two measurements. In particular, these have to stay the same:
- Position and orientation of the device under test
- Routing and length of all cables
- Connected peripherals and loads
- Operating state and operating mode
- Measurement configuration and instruments
That is why we document the setup with photos and boundary conditions already during the first measurement – without that documentation, a later re-measurement is worthless.
When the measure does not work
That happens, and it is not a wasted appointment: a refuted hypothesis rules out a path and shrinks the search space. We record what was tested and what it was not, and move on to the next hypothesis.
Honestly, that is part of troubleshooting. Anyone who promises to eliminate every anomaly on the first attempt is promising something that cannot be kept – the effect of a measure is shown by the measurement, not before it.
How a root cause analysis runs in detail is described under root cause analysis and optimisation. Typical patterns are collected under common EMC problems.
Narrow down the anomaly together
Send us the finding and the boundary conditions of the measurement – we will propose the next measurement step.