Inside Our EMC Chamber: CISPR 25 Emissions Debugging for an Industrial Vehicle Safety System

Our customer manufactures fully illuminated stop arms for school buses: an LED-lit stop sign on a motorized arm, with a sensor that confirms whether the arm is open or closed. Their team designed the electronics, motor gearing and mechanical assembly in-house.
When their first round of regulatory testing flagged CISPR 25 conducted and radiated emissions, the standard for electronics installed in vehicles, they brought the system to us.
The product worked. It needed to clear emissions compliance before it could ship. They needed to know where the noise was coming from and what to change.
That's where our EMC chamber comes in. Having one in-house means we can set up a test, measure, change one thing on the board, and measure again, as many times as it takes. Here's how we worked through it: set up, test, debug, solve.
STEP 1: SET UP AND REPLICATE THE FAILURE

You can't fix noise you can't see. The first job was to reproduce what the customer's test lab had measured, so we knew we were chasing the same problem.
We set the system up in the chamber and took baseline radiated and conducted emissions measurements across the frequency range. Then we compared the "shape" of our results against their lab report.
We reproduced the same emission peaks the lab found, which gave us a reliable baseline to debug against.
STEP 2: DEBUG BY SNIFFING THE BOARD

The probe work pointed to three main noise sources: two driving the radiated emissions and one driving the conducted emissions. We tackled them in order of impact.
RADIATED EMISSIONS
The boost converter
The boost converter takes the bus's supply voltage and roughly triples it to drive the LED lighting. Almost all of the system's current flows through this circuit, so it was the most likely source of noise.
We slowed its switching down by increasing the gate resistance. That one change brought the overall noise down significantly.
The inductive sensor coil
The next RF pattern we spotted matched the frequency and harmonics of the customer's inductive coil. The system uses it to sense whether the stop sign is open or closed: a voltage wave runs through the coil, and when metal gets close, the wave changes shape.
The customer's design constraints called for this sensing approach, so our focus was on keeping the noise it can generate out of the rest of the system. We tried several options. The one that worked was adding a grounded shield around the small cable running from the PCB to the coil.
The result

Together, the boost converter and coil cable fixes dramatically reduced the radiated emissions. In some parts of the spectrum, the measured signal dropped by around 40 dBμV. The two sweeps were taken six days apart.
CONDUCTED EMISSIONS
Noise escaping on the power lines
That left the conducted emissions: noise travelling out of the system on the power cables. Fully fixing that would take a board layout change, so the next step was to choke it off before it left.
We added aggressive pi filters to each power line running from the main PCB to the bus power system. We deliberately used oversized, high-value parts. They weren't meant for production. The goal was to prove that filtering at that point would work, and it did, with a noticeable improvement. It also showed that filtering combined with layout changes is what will get the design across the finish line.
STEP 4: BUILD IT IN AT THE LAYOUT LEVEL
The bench fixes showed the customer where the noise came from and which changes made a difference. The lasting fix belongs in the PCB itself.
We gave the customer layout recommendations for their next board revision:
Move noisy circuits away from everything else on the board
Use better routing practices to keep noise contained
We reviewed their revised design and gave a second round of recommendations. From there, the customer took the redesign forward into their next round of regulatory testing.
WHY IT MATTERS
An EMI failure late in development is stressful. The product is built, the schedule is tight, and the test report only tells you that you failed, not why.
With our own chamber, we can replicate the failure, trace it to the circuit causing it, and prove out fixes before you go back to the regulatory lab. That's fewer expensive formal test runs and a clearer path to a compliant design.
Failing EMI testing or want to check your design before you book a formal test? Get in touch with our team.




