CISPR 25 conducted failures are rarely “just a noisy DC-DC.” Most early failures are a system problem: harness geometry, return paths, and how your input filter interacts with the wiring. This guide shows a repeatable pre-check workflow you can run on an automotive test bench: pick a measurement method, lock the harness controls, baseline the setup, measure in a worst-case operating mode, then use a fix-mapping table to choose the right first correction.
CISPR 25 defines limits and measurement procedures for radio disturbances intended to protect on-board receivers, covering 150 kHz to 5 925 MHz. (Your customer/OEM spec selects bands and classes.)
CISPR 25 is written to control radio-frequency disturbances in vehicles and components so on-board receivers (AM/FM, etc.) are protected. The standard includes both limits and measurement procedures across a wide frequency range (150 kHz to 5 925 MHz).
Your OEM spec typically chooses a “class” and a set of frequency bands that match the vehicle’s market and receiver set. Your pre-check’s job is not to “perfectly replicate” the lab. Your pre-check’s job is to catch the dominant coupling path early (differential-mode vs common-mode, harness resonance, filter interaction) so your first fix is the right fix.
For CISPR 25 conducted work, many teams use either a voltage measurement method (with an automotive LISN/AN/AMN) or a current-probe method to characterize noise on the harness. Texas Instruments’ overview of CISPR 25 conducted approaches discusses both voltage and current measurement methods for conducted emissions work and focuses on the conducted portion of the standard.
Automotive conducted EMI is “wiring-sensitive” by design: the harness is part of the system. If you don’t control harness geometry, you’ll get impressive-looking plots that are not predictive.
A practical example: TI’s CISPR 25 conducted setup guidance for an automotive reference design notes a battery cable length of 200 mm to 400 mm (about 8–16 inches), including the LISN connector, and demonstrates a setup using about a 9-inch cable. If your bench uses “whatever cable is convenient,” your resonance and coupling can shift dramatically.
Many CISPR 25 test setups are performed in controlled environments and include ground plane and harness routing/positioning requirements (so results are repeatable). A COM-POWER application note summarizing CISPR 25/ISO 11452 automotive EMC testing calls out the controlled enclosure concept and test setup requirements such as ground plane and harness routing/positioning.
Use this section to avoid the most common bench mistake: treating every peak as “need more filtering” without first deciding whether the dominant problem is differential-mode (DM) or common-mode (CM).
If you see “peaks move when the harness moves,” treat that as a diagnostic: your coupling path is geometry-sensitive, which usually means CM is winning.
| What you see on the bench | Likely dominant mechanism | Likely root cause in automotive context | First fix to try (bench-friendly) | Fixes that often waste time |
|---|---|---|---|---|
| Peaks shift a lot when you move harness or change lead length | Common-mode | Parasitic capacitance to chassis/plane + harness acting like part of the return path | Lock harness geometry; add CM choke (proved on affected lead); improve chassis/plane return path and bonding | Randomly increasing input capacitance without addressing CM return |
| Peak frequency tied to switching frequency/harmonics; grows with load | Differential-mode | Input ripple current + filter resonance with harness impedance | Tune/damp input filter (add damping R/C as needed); minimize loop area; verify stability across operating modes | Stacking ferrites everywhere without confirming current path |
| Wideband “grass” that changes with bench equipment | Setup/environment | Bench supply noise, grounding inconsistencies, uncontrolled cable routing | Run DUT-OFF baseline; clean the supply chain; document and freeze placement and routing | Redesigning hardware before the baseline is stable |
If your automotive bench includes DIN-rail power conversion in supporting racks or fixtures, treat the distribution harness and mounting as part of the EMI system. Hardware context: DIN-rail power supplies. For examples of compliance-focused projects, see safety & compliance cases.
CISPR 25:2021 contains limits and measurement procedures for radio disturbances from 150 kHz to 5 925 MHz.
The voltage method measures disturbance voltage through a defined network (LISN/AN/AMN) and is useful for limit-line comparisons. Current probe methods help identify noise current on harness segments and are often used for rapid triage and common-mode dominance checks.
Harness geometry changes impedance and coupling, shifting resonances and common-mode return paths. Even “small” setup details (like battery cable length between LISN and DUT) can be controlled explicitly in CISPR 25 conducted setups (for example, 200–400 mm in a reference setup).
First decide whether the problem is differential-mode (switching harmonic fingerprints, load-linked ripple) or common-mode (peaks move with harness/placement). Then apply the matching fix pattern and remeasure with identical harness geometry.
References:
