A Single Curve That Answers the Most Expensive Question in Product Development
At a professional EMC exhibition 2026, one demo stood out for its clarity.
No complex multi-channel setup. No eight-quadrant display. Just a compact blue box, a green PCB, two coaxial cables — and a Bode plot on a screen that told the complete story of an EMI filter's behavior from 100 kHz to 300 MHz.
The instrument was the Omicron Lab Bode 500, a USB-connected vector network analyzer running the Bode Analyzer Suite software. The measurement was transmission mode — stimulus in, response out, magnitude versus frequency across nearly three decades of spectrum.
And the curve on the screen was answering the question that drives most of the late-stage redesign cost in power electronics development:
Does your EMI filter actually work — and where does it stop working?
Reading the Bode Plot: Three Frequency Regions, Three Different Stories
The transmission magnitude curve on the Bode Analyzer Suite display is a direct measurement of the filter's insertion loss — how much signal at each frequency is blocked by the filter before reaching the output. Three distinct regions tell the complete filter story.
Low frequency (100 kHz – ~1 MHz): near-flat, slight attenuation
The curve starts around +5 dB and rolls off gradually. At these frequencies, the filter components — inductors and capacitors — are not yet at their designed operating points. The attenuation is modest. For conducted emissions standards like CISPR 32 Class B, which begin at 150 kHz, this region is where many filters first encounter compliance margin issues.
Mid frequency (~1 MHz – 20 MHz): the design intent zone
The curve descends steeply into the filter's primary attenuation band. The deepest point — the notch — sits around 10 MHz at approximately −95 dB. This is where the filter is doing what it was designed to do: blocking the switching frequency harmonics that a power converter generates most aggressively. A 95 dB notch means the signal is attenuated by a factor of roughly 56,000. In practice, this is the difference between a product that passes conducted emissions testing and one that fails by 40 dB.
High frequency (>50 MHz): where reality diverges from simulation
At the cursor measurement point — 110 MHz — the attenuation has recovered to approximately −47 dB. By 190–200 MHz, the curve is approaching −20 dB and still rising. This high-frequency degradation is not a design failure. It is physics: at these frequencies, the parasitic capacitance of the filter inductors and the parasitic inductance of the filter capacitors create resonant paths that bypass the intended filter topology. The component that was supposed to block high-frequency energy instead becomes a conductor for it.
This is the region that simulation almost always gets wrong — and that only a physical transmission measurement reveals accurately.
The Hardware: Bode 500 and the Test Fixture Architecture
The Bode 500 is a two-port vector network analyzer in a compact USB-powered enclosure — output port generates the swept frequency stimulus, two input channels measure the response. Connected to a PC running the Bode Analyzer Suite, it covers 1 Hz to 40 MHz natively, with the extended version reaching into the hundreds of MHz range used in this demo.
The two small adapter boxes on top of the Bode 500 unit are Omicron Lab test fixtures — designated B-SMC and G-WIC. These are not arbitrary connectors. They are calibrated fixtures that present a controlled 50-ohm impedance environment to the device under test, ensuring the measurement reflects the filter's actual insertion loss rather than artifacts from connector mismatch or ground loop effects.
This fixture approach addresses a practical problem that engineers hit immediately when trying to characterize filters on a bench: the way you connect the filter to the measurement instrument changes the result. A long coaxial cable with a clip lead to the filter input and output will measure a different insertion loss curve than a properly impedance-matched fixture — especially above 10 MHz where lead inductance begins to matter. The B-SMC and G-WIC fixtures make the measurement repeatable and comparable across different filters or different bench setups.
The green PCB in the foreground is the device under test — a small EMI filter board with visible inductors and capacitors in an LC topology. Precisely the kind of board that a power electronics engineer would design, prototype, and need to characterize before deciding whether the component values and layout are suitable for the application.
What This Measurement Is Actually For
Frequency response measurement of EMI filters is one of the most direct connections between bench characterization and product compliance outcomes in all of electronics engineering.
The conducted emissions limits in CISPR 32, EN 55032, and automotive equivalent standards (CISPR 25, UNECE R10) define exactly how much attenuation a product's filtering must provide at each frequency. A filter characterization measurement is not abstract — it directly predicts whether a product will pass or fail a formal test lab measurement, and by how much margin.
The typical engineering workflow this instrument supports:
→ Design verification — the first physical prototype of a new filter design is measured to confirm that the component values, PCB layout, and topology produce the intended attenuation curve
→ Component tolerance analysis — measuring multiple boards built with components at different ends of their tolerance range reveals how much the filter characteristic varies in production
→ Layout optimization — changing via placement, ground plane routing, or component orientation and re-measuring shows whether the layout change improved or degraded high-frequency performance
→ Pre-compliance assessment — before committing to formal EMC testing (which costs several thousand euros per submission), a bench Bode plot measurement gives a high-confidence prediction of the compliance outcome
The last application is where the economics are most direct. A formal EMC test failure at a certified laboratory typically costs two to four weeks of redesign time plus the cost of retesting. A filter characterization bench that catches the problem in the lab costs a few hours. The Bode 500 and a calibrated test fixture represent a small fraction of one failed compliance test cycle.
The Context: An EMC-Focused Exhibition Ecosystem
The backdrop visible in the photos — "EMC Testing & more – Your All-in-One Solution" — and the surrounding exhibitor logos (Keysight, Langer EMV-Technik, Rohde & Schwarz, Tektronix, Pendulum) locate this demo precisely: a specialist EMC instrumentation event where the audience is engineers actively working on compliance problems, not general electronics development.
In that context, the Bode 500 demo is correctly targeted. The visitors to this exhibition understand what a 95 dB notch at 10 MHz means, know what their CISPR limits look like, and immediately recognize the high-frequency parasitic behavior in the curve's right-hand region as the problem they spend hours debugging.
This is the version of a test equipment demo that earns credibility without explanation: the curve on the screen is self-evidently relevant to the engineering problem the audience is in the room to solve.
The Insight That the Bode Plot Uniquely Provides
There is a persistent gap in how EMI filters are specified versus how they actually perform in hardware. Component datasheets provide insertion loss curves measured under specific source and load impedances — typically 50 ohms in, 50 ohms out, which is a convenient test condition but rarely the impedance environment the filter sees in a real power converter circuit.
The Bode 500 measurement closes that gap in two ways.
First, it measures the actual physical filter as built — solder joints, PCB parasitics, component tolerances, and all. Simulation with ideal component models doesn't capture these effects. Only measurement does.
Second, the fixture-based approach can be adapted to replicate the actual source and load impedance the filter sees in the application — allowing the engineer to measure performance under realistic operating conditions rather than idealized test bench conditions.
The curve shape that results — steep rolloff into the notch, predictable high-frequency degradation from parasitics — is the ground truth. It's what the product will look like to an EMC test receiver. Everything else is an approximation.
Instrument observed: Omicron Lab Bode 500 · USB Vector Network Analyzer / Frequency Response Analyzer · with B-SMC and G-WIC test fixtures · Bode Analyzer Suite software · measuring PCB EMI filter insertion loss, 100 kHz – 300 MHz
All photos: Thomas · @SignalByThomas