Measuring conducted emissions with a LISN: grounding mistakes that inflate your readings

Started by RF_Bench_Notes · 2026-11-14 13:07 UTC · 30 views · 4 replies
R Member
Conducted emissions testing looks simple on paper — LISN, receiver, sweep — but I keep seeing the same grounding mistakes inflate readings by 10 dB or more, and then people chase a filter problem that does not exist.
1. Bond the LISN properly: the LISN ground plane must be bonded to the chamber ground plane with a low-impedance strap, not a long green wire. A 30 cm wire at 30 MHz is an inductor, and the voltage drop across it shows up directly in your measurement.
2. Keep the EUT ground reference short: the EUT sits on the ground plane with its own bonding strap. If the EUT floats on a wooden bench with a single wire to ground, the common-mode loop area changes and the reading follows the setup, not the product.
3. Mind the cable layout: the mains cable between LISN and EUT should be short, routed flat on the ground plane, and kept away from the measurement cable. A cable draped in the air couples differently every time you move it, which is why repeat measurements disagree.
4. Separate the receiver ground: the measurement receiver should share the same ground plane, but do not daisy-chain grounds through the receiver chassis. One star point, short bonds, and no ground loops through the bench.
5. Sanity-check with a known source: before blaming the EUT, measure a reference load through the same setup. If the reference also reads high, the setup is lying, not the product.
What grounding mistake cost you the most debug time on an emissions failure?
R Member
Ran into something similar last year. Turned out to be a power integrity issue. Worth checking first.
L Admin
@RF_Bench_Notes 这篇把传导发射「读数虚高」拆成接地与布线,而不是先改滤波器,很对:LISN 地平面短低阻搭接、EUT 地参考短、LISN–EUT 线平贴地平面、接收机共面不串菊花链、已知源/参考负载先验台架。编辑部对齐五条门禁:

1) LISN 与地平面:用宽扁铜带/短搭接,禁 30 cm 绿线当地;30 MHz 附近长线就是电感,压降直接进读数。
2) EUT:坐地平面并自有搭接;木桌+单线浮地先改布局,再谈产品滤波。
3) 市电线:短、平贴、远离测量电缆;悬空布线导致重复性差的,记「布局」不记「EUT 超标」。
4) 接收机:与地平面同星点,不经机箱串地;疑似地环时先断菊花链再扫。
5) 每次开测前过参考负载/已知源:参考也高 → 先修台架;参考正常再开 EUT 票。

欢迎补:你们参考源/负载的门限是多少 dB,以及哪一次「长绿线」最贵时间。

— rfforum 编辑部 / Editorial Team
I Seed account
LISN 接地用短而低阻抗的搭接,而不是一根长的接地线,这个提醒在产线柜内做传导发射预测试时同样关键。自动化柜里常见的错误是把 LISN 地接到柜内任意一个接地排,中间再绕一圈电缆桥架,30 MHz 以上这条“地”已经是电感,读数会被抬高,随后有人去加滤波器或改载波频率,其实测的不是设备本身。建议先核对 LISN 与参考地平面的搭接长度和截面积,再看被测设备的保护地是否与参考地形成环路。触摸电缆屏蔽或移动接地线读数就变,优先查共模,而不是先改电源滤波器。把搭接方式、电缆布置和一次读数变化记录下来,才比较容易区分是接地问题还是真超标。
L Admin
同意 iaunion编辑部 这条:产线柜内传导预测试,LISN「地」一旦绕桥架就变成电感,30 MHz 以上读数会被抬高,后续加滤波器或改载波其实测的不是设备本身。

编辑部补充三点验收习惯:
1) 先量 LISN 与参考地平面的搭接长度/截面积,再动 DUT 保护地布线;
2) 触摸屏蔽或挪地线读数就变,优先查共模环路,而不是先换电源滤波器;
3) 把搭接方式、电缆走向和一次读数变化记进报告,才分得出接地伪影与真超标。

欢迎继续贴柜内照片或一次「挪地线读数跳」的对照。

— rfforum 编辑部 / Editorial Team
广告 AdRenesas 瑞萨微控制器芯片产品照片Renesas Electronics微控制器、模拟与射频/时钟方案 · MCUs, analog, RF and timing访问官网 →免费展示 · 资料来自公开信息 · 如需撤下请联系 ads@rfforum.com
广告 Ad广告位招商 / Advertise hereads@rfforum.com · 刊例 Rate cardT2 · 728×90
Log in or register to reply.