How to Read an EMC Test Report: Margin, Detectors and the Limit Line
Learn to read the numbers in an emission test report: reading vs. limit, correction factors, QP and average detectors, margin, and judging with uncertainty.
A formal emission test report is full of tables and plots, but the pass/fail decision comes down to a few columns. This guide explains what each number means, how to read the limit line and how to judge margin.
1. Where to look first
- Standard and class: check which standard, edition, equipment class (A or B) and test distance the limit was taken from. A Class A result cannot support a Class B claim.
- Test configuration: check the operating mode and the cables and ports that were connected. A report only speaks for the configuration that was tested.
- Equipment and calibration list: antenna, LISN, receiver and cables with their calibration dates.
- Ambient scan: the evidence for telling peaks from broadcast and communication signals apart from peaks from the EUT.
2. Reading a result table column by column
| Column | Meaning |
|---|---|
| Frequency | Frequency of the emission found in the scan |
| Reading | Raw value read by the receiver (for example in dBµV) |
| Correction factor | Radiated: antenna factor + cable loss + attenuator − preamp gain. Conducted: LISN factor + cable loss |
| Corrected level | Reading + correction factor. This is the value compared with the limit |
| Limit | Limit at that frequency for the detector, class and distance |
| Margin | Limit − level. Positive means below the limit |
| Detector | QP (quasi-peak) or AV (average). Each detector has its own limit |
| Polarisation / height / angle | Radiated tests only: the antenna and turntable position that gave the maximum |
3. How to read the limit line
- The limit is a function of frequency, not one number. In radiated tests the line steps up at a band boundary. For example, Class B at 10 m is 30 dB(µV/m) for 30–230 MHz and 37 dB(µV/m) for 230–1000 MHz.
- Mains-port conducted limits slope down with the logarithm of frequency at the low end. Class B QP goes from 66 to 56 dBµV across 0.15–0.5 MHz and Class B average from 56 to 46 dBµV. Read the limit at the exact frequency, not from the nearest flat part.
- QP and average have separate limit lines and the average line is lower. A peak can pass QP and still fail average.
- Check the distance. The 3 m and 10 m limits differ (Class B, 230–1000 MHz: 47 dB(µV/m) at 3 m, 37 dB(µV/m) at 10 m). Converting between distances is for estimates only; judge against the limit for the actual test distance.
4. Worked example
The numbers below are made up to show the arithmetic and are not real measurements. A Class B product, mains port, final measurement at 0.5 MHz.
| Detector | Level (dBµV) | Limit (dBµV) | Margin (dB) | Result |
|---|---|---|---|---|
| QP | 52.0 | 56 | 4.0 | Pass |
| Average | 47.1 | 46 | −1.1 | Fail |
The QP margin looks comfortable, but the average margin is negative, so the product fails at this frequency. Always check every detector.
5. How much margin is enough
- Margin = limit − level. Positive means below the limit and negative means over it.
- A positive margin smaller than the measurement uncertainty does not prove compliance. If the margin including expanded uncertainty is negative, the true value may exceed the limit even though the reading is below it. (Calculators: limit margin, combined uncertainty)
- Whether the reading is compared directly or with uncertainty added follows the uncertainty criteria of CISPR 16-4-2 and your certification body’s policy. Ask the lab which rule the report uses.
- For design, aim for at least 3–6 dB of margin to absorb lab-to-lab spread, and flag items under 3 dB as countermeasure candidates.
6. Common reading mistakes
- Mixing units: compare dBµV with dBµV and dB(µV/m) with dB(µV/m). Do not compare an uncorrected receiver reading with a field-strength limit. (Calculators: dBµV ↔ V, dBµV/m ↔ V/m)
- Looking only at the highest peak in the plot: the plot is often a peak pre-scan, and the judgement is made on the final QP and average table.
- Ignoring the test mode: a pass in one mode does not cover a worse mode.
- Reusing a report after a design change: a change in layout, cables or firmware may require a new evaluation.
This guide is a general reference. Report formats differ between labs. For judgements used in certification, follow the official standard and the policy of your certification body.
Related calculators
Results and summaries are for reference. For certification and test reports use the latest official standard text and calibrated instrument data.