Conducted emissions test setup (LISN): 7 wiring mistakes that cause “false fails”

If your conducted-emissions plot changes when you move a cable, touch a ground lead, or swap outlets, you’re not doing “EMC engineering” yet—you’re fighting the measurement. A LISN (line impedance stabilization network) is meant to isolate the mains from your DUT and give you a repeatable measurement port, but only if the wiring, bonding, and test chain are disciplined. A major clue: do a DUT-OFF baseline first. If the baseline is already close to your limit or full of mystery peaks, your setup is lying to you.

What a LISN does (so you don’t “measure the wall outlet”)

A LISN is not just a “box in the loop.” Its job is to separate what the DUT generates from what the power source (and building wiring) already contains. In Keysight’s conducted-emissions workflow, the LISN (with a limiter) sits between the power source and DUT, and the receiver measures the disturbance at a defined port. :contentReference[oaicite:1]{index=1}

Keysight also describes three practical LISN functions: (1) isolate the mains from the DUT, (2) keep DUT noise from back-feeding onto the mains, and (3) couple DUT noise to the receiver through a defined path (commonly presenting a 50-ohm load in the measurement band). :contentReference[oaicite:2]{index=2}

That only works if your wiring doesn’t create extra antennas, extra coupling paths, or “mystery grounds.” CISPR 16-1-2 exists because these coupling devices matter enough to standardize. :contentReference[oaicite:3]{index=3}

Baseline first: the 60-second check that catches most false fails

Before you blame the power supply, filter, or PCB: run the exact setup with the DUT powered off and look at the trace. Keysight explicitly recommends measuring the signals on the power line with the DUT off; if you are close to limits with the DUT off, your environment/setup is contributing heavily. :contentReference[oaicite:4]{index=4}

Use this fast baseline checklist:

1. DUT OFF scan: record the baseline and note any fixed peaks.
2. Shorten DUT-to-LISN power cord: if peaks drop, you were measuring antenna behavior, not DUT noise.
3. Confirm bonding: LISN, DUT chassis, and reference plane should be consistent and repeatable.
4. Enable worst-case modes only after baseline: switching loads, comms traffic, and max CPU activity change the signature.

The 7 most common LISN setup and wiring mistakes

These are ranked by how often they create “fake failures” on a lab rack: results that look like a problem, but disappear when the setup is corrected.

Mistake 1: Long DUT-to-LISN power cord (it becomes an antenna)

This is the classic: you plug the DUT into the LISN with a convenient cord length and unknowingly build an antenna. Keysight explicitly warns to keep the power cord between the DUT and LISN as short as possible because the cord can act as an antenna if longer than necessary. :contentReference[oaicite:5]{index=5}

Fix: shorten the cord, keep it close to the reference plane, and avoid loose loops. Rerun DUT OFF first; then DUT ON.

Mistake 2: No stable reference ground plane (or weak bonding)

Conducted measurements are extremely sensitive to reference impedance. Keysight notes that more attention to detail—like using a good ground plane—improves measurement accuracy. :contentReference[oaicite:6]{index=6} In practice, inconsistent bonding turns your return paths into variables, and the trace drifts with tiny physical changes.

Fix: bond the LISN case and DUT chassis consistently; keep bond straps short and wide where possible. If you’re chasing cabinet issues, review typical grounding/bonding failure modes in control panels: control panel grounding and bonding failure modes.

Mistake 3: Using ferrites during measurement (you “improve” the data by changing the DUT)

Ferrites are valid mitigation tools, but adding them during measurement changes the conducted signature and can hide common-mode behavior. Keysight specifically advises not to use ferrites on the power cord in the measurement step because common-mode signals from the DUT may be suppressed, producing a lower measured value. :contentReference[oaicite:7]{index=7}

Fix: measure the “as-is” configuration first; document it. Then test ferrites as a controlled design change with a clear hypothesis (which cable, which mode, what symptom).

Mistake 4: Mixing noisy and quiet wiring (routing errors inside the rack)

On a lab rack, it’s easy to bundle everything neatly and accidentally couple noise into the very lines you’re measuring. Typical offenders: DC/DC outputs bundled with comms lines, long I/O harness loops, relay/contactor wiring routed alongside measurement-sensitive power leads.

Fix: separate “dirty” switching paths from measurement and sensing paths; reduce loop area; keep harnesses tight to the reference plane. If your DUT includes DIN-rail PSUs and distribution, treat the distribution harness as part of the DUT system, not “just wiring.”

Mistake 5: Wrong measurement chain (no limiter discipline, overload, missing corrections)

If the receiver input overloads or the correction factors are wrong, you’ll see peaks that aren’t real (or miss peaks that are). Keysight’s process calls out adding correction factors for the LISN and limiter, and also checking for overload by adjusting attenuation and verifying the display doesn’t change. :contentReference[oaicite:8]{index=8}

Fix: verify limiter is installed, load the correct limit lines for your target standard, apply LISN/limiter corrections, and validate no overload condition. Then re-check DUT OFF baseline.

Mistake 6: LISN placement and mains noise back-feeding (you’re measuring the building)

A LISN helps isolate the DUT from the mains and vice versa, but if the setup is sloppy, you can still “see” the environment. Keysight’s LISN purpose section emphasizes the need for clean supplied power and notes that line noise can be interpreted as DUT noise if it couples into the measurement chain. :contentReference[oaicite:9]{index=9}

Fix: keep the LISN close to the DUT, maintain consistent bonding, and run a baseline scan with the DUT off. If the baseline is high, improve shielding/placement before touching the DUT design.

Mistake 7: Wrong operating mode (you never turned on the “noisy” state)

Conducted emissions are often mode-dependent: switching loads, high CPU activity, high comms throughput, and worst-case I/O patterns change the spectrum. If you only test an idle mode, you may “pass” on the bench and fail later in a more realistic configuration.

Fix: define and document at least two modes: typical and worst-case. For control racks, include relay/solenoid switching, max comms traffic, and maximum load on PSU rails.

A repeatable LISN test plan for lab racks

Use this plan when you’re testing a rack subsystem (power + control + harnessing) and need the data to hold up when you compare revisions.

1. Define the “DUT system boundary”: include power supply, distribution harness, and any loads that represent the real application (not just a dummy load).
2. Lock the physical setup: mark LISN placement, cable routes, and bond points so reruns are physically identical.
3. Run DUT OFF baseline: record trace and ambient peaks. (If it’s high, fix setup before DUT ON.) :contentReference[oaicite:10]{index=10}
4. Run DUT ON in two modes: typical + worst-case switching/comms. Record both.
5. Change one variable at a time: shortest path to real insight (and avoids “fixes” that only work accidentally).

If your rack uses DIN-rail power conversion, treat the PSU selection and integration as part of the emissions story (layout, grounding, wiring). See our DIN-rail power supply collection for platform context.

When to escalate: turning messy bench data into lab-ready evidence

Escalate when: (1) baseline won’t stabilize, (2) results change with small physical movement, (3) you need to align to a customer test plan, or (4) you must prove a fix across configurations. Tektronix’s pre-compliance guidance frames the goal well: improve measurement accuracy with the right accessories and a disciplined setup, then use troubleshooting tools to accelerate debugging. :contentReference[oaicite:11]{index=11}

FAQ

What does a LISN actually do in conducted emissions testing?

It isolates the DUT from mains noise, prevents DUT noise from back-feeding onto the mains, and couples the DUT’s disturbance to the receiver through a defined path (commonly with a 50-ohm measurement port in the conducted band). :contentReference[oaicite:12]{index=12}

Why do my results change when I move the DUT power cord?

Because the cord can act like an antenna or change coupling geometry. Keeping the DUT-to-LISN power cord as short as possible is a standard practical control because longer cords can radiate/pick up and distort the measurement. :contentReference[oaicite:13]{index=13}

Should I add ferrites before measuring conducted emissions?

Measure the “as-is” setup first. Ferrites can suppress common-mode signals and make the reading look better without proving the underlying cause. Keysight specifically warns against using ferrites on the power cord during measurement because it can suppress DUT common-mode signals and lower the measured value. :contentReference[oaicite:14]{index=14}

How do I tell ambient noise from DUT noise in a LISN test?

Run a DUT OFF baseline in the exact setup. If the trace is already near your limit or full of peaks, fix setup/environment before blaming the DUT. :contentReference[oaicite:15]{index=15}

What frequency range should I scan for conducted emissions?

It depends on the regulation and product category; Keysight notes commercial conducted emissions commonly span up to 30 MHz depending on regulation. For vehicle-related electronics, CISPR 25 includes procedures across a wider frequency range (starting at 150 kHz). :contentReference[oaicite:16]{index=16}

Authoritative references (external):

• IEC CISPR 16-1-2: coupling devices for conducted disturbance measurements (includes AMNs/LISNs)
• IEC CISPR 25: limits and procedures for measuring radio disturbances (includes conducted methods)
• Keysight application note: making conducted and radiated emissions measurements (LISN setup details)
• Tektronix application note: conducted and radiated emissions testing (pre-compliance setup + troubleshooting)