EMC design and pre-compliance support
Most EMC failures are designed in months before the chamber. This is the work that prevents them.
Chamber time is expensive, and failing is expensive twice, because the fix is usually a respin and another booking.
The uncomfortable part is that almost none of it is decided in the chamber. By the time a product is being tested, the outcome was determined months earlier by a handful of layout and architecture choices — most of which cost nothing to make correctly and a great deal to correct.
The two mechanisms that cause most failures
Return paths that detour. Current returns beneath its signal. When a plane split, a gap, or a change of reference layer forces that return current to travel around an obstacle, the signal and its return enclose a loop — and a loop is an antenna. This is invisible on a schematic and obvious on a stack-up drawing.
Cables. A shield terminated through a pigtail is a shield with an inductor in series, which is barely a shield at the frequencies that matter. Three hundred and sixty degree termination at the connector body, bonded to chassis, is the difference between a cable that contains its energy and one that broadcasts it. We have produced client-facing shield-termination concepts for harness and connector systems including 360-degree termination at ODU connectors.
What we do, and when
At architecture. Identify which subsystems are aggressors and which are victims. Decide the grounding and shielding strategy. Plan the enclosure’s role — chassis bonding, apertures, cable entry — with your mechanical engineer, while it is still a drawing.
At schematic. Filtering at every interface that leaves the board. Common-mode chokes where differential pairs exit an enclosure. ESD protection placed where it does not wreck the signal it is protecting. Component grading appropriate to the environment.
At layout. Continuous reference planes under high-speed routing. Stitching vias at layer transitions. Switching nodes kept physically small. Clock routing away from board edges and connectors. Guard traces where they earn their space.
Before the chamber. A review against all of the above with severity-ranked findings, so the pre-compliance session confirms what you already expect rather than discovering it.
Pre-compliance, honestly framed
Full pre-compliance measurement needs a chamber or at minimum a calibrated near-field setup. What design-stage work buys you is that the measurements come back where you expect, and that when something is marginal you already know which mechanism to look at.
Certification testing itself runs through an accredited house, against EN 55032 and EN 55035 for most commercial products, with sector-specific standards on top where the market requires them.
The eight checks worth running before you book
- Every high-speed net traced against its return path — no detours
- Plane splits mapped, and nothing critical crossing one
- Every cable’s shield termination specified, 360 degrees where possible
- Filtering on every interface that leaves the enclosure
- Switching node copper minimized and its loop area measured
- Clock nets away from board edges, connectors and apertures
- Enclosure apertures checked against the highest frequency of concern
- ESD paths defined all the way to chassis, not just to local ground
None of these needs equipment. All of them are cheaper than a second chamber booking.
Common questions
Straight answers
When should EMC work start?
At architecture, which is roughly a year before most teams think about it. The decisions that determine whether a product passes — where the plane splits are, how cable shields terminate, which node has the most copper on it, whether the connector has filtering — are all made before layout. Bringing in EMC help after a failed chamber session means paying for a respin to fix something that cost nothing to get right.
Do you perform the certification testing?
No. Testing runs through an accredited test house, and we would not want you to rely on anyone who claims otherwise. What we do is design so the session is a formality, sit in on it if useful, and turn any failures into specific design changes quickly rather than into a month of speculation.
What causes most EMC failures?
Return paths that have to detour, and cables. A signal whose return current is forced around a plane split becomes a loop antenna; a cable shield terminated through a pigtail rather than 360 degrees becomes a very effective radiator at exactly the frequencies you care about. Between them those two account for a large share of first-attempt failures, and both are free to fix in design.
Can you help with an existing product that already failed?
Yes. Send the test report and the design. A failure at a specific frequency with a specific orientation is diagnostic information, and it usually points at a small number of candidate mechanisms. We will tell you what we think it is, what to try first, and whether it is fixable without a respin.
Does this cover medical or automotive requirements?
We design to IEC 60601-1 constraints for medical electrical equipment and select AEC-Q100 and Q101 graded parts where automotive robustness is required. We are not a certification body — our role is to make sure the electronics do not make certification harder than it needs to be, and to say clearly where you need an accredited house.
Evidence
Where this has been applied
Flight-controller carrier for a tube-launched UAV
Teardown, clean-room re-implementation, and a rigid-flex carrier that fits inside a launch tube.
Edge-compute architecture for an unmanned surface vessel
Three architectures compared on real bandwidth budgets, so the platform decision was made on evidence rather than TOPS figures.