Pre-Compliance Testing — Find Your EMI Problems Before a ₹2 Lakh Test Lab Does

An accredited EMC test lab will find every problem your product has. That is not a feature — that is a very expensive diagnostic service that issues you a failure report and keeps your test fee. An accredited test session in India typically runs ₹80,000 to ₹2,00,000 depending on the standard (BIS/IS, CE, FCC) and the number of sub-tests. If you fail, you pay again for the retest. Pre-compliance testing is the discipline of finding those problems yourself, with equipment that costs 5–10% of one failed session, before the official test. Engineering teams that do pre-compliance arrive at the lab and pass. Teams that skip it arrive and spend.

What Pre-Compliance Testing Is (and Is Not)

Pre-compliance testing is informal testing using affordable equipment to approximate — not replicate exactly — the measurement conditions of accredited tests. The goal is to find problems and fix them, not to achieve a numerical result that certifies the product.

What it is:

Locating emission sources on your PCB using near-field probes
Comparing your product's spectrum to the applicable limit lines (CISPR 32, FCC Part 15B, etc.)
Applying fixes and confirming they work before spending lab money
Building engineering intuition about what your circuit radiates

What it is not:

A substitute for accredited certification
A measurement you can put in a report for a regulatory authority
A guarantee of passing the official test (test site variation, cable configuration, and orientation can shift results by 3–6 dB)

The Near-Field Probe Kit

A near-field probe kit is the most useful tool in pre-compliance work. It consists of small loop probes (for magnetic field / H-field) and electric field probes, connected to a spectrum analyser or an oscilloscope with FFT mode.

How it works: a small loop probe held 1–5 mm above a PCB surface picks up the magnetic field associated with current loops on the board. Peak readings identify noise sources. Sweep the probe across the board systematically, then zoom in on peaks.

Reading the results:

Loop probe held horizontally: detects vertical H-field from horizontal current loops (power planes, ground return loops)
Loop probe held vertically: detects horizontal H-field from vertical current loops (traces, vias)
Rotate 90° to confirm directionality — a true source will show strong directional sensitivity

Available kits:

Tekbox TBPS01: four probes (three H-field, one E-field), BNC connector, ~₹8,000–12,000. Includes calibration chart.
Rigol near-field probe set: compatible with Rigol DSA815 spectrum analyser
DIY: 20 mm diameter loop of semi-rigid coax, soldered to SMA connector. Not calibrated but functional for identifying sources.

Spectrum analyser options:

Rigol DSA815: 9 kHz – 1.5 GHz, ~₹40,000. The workhorse for pre-compliance at startups.
Rigol DSA832: 9 kHz – 3.2 GHz, ~₹1,20,000. Required if your product operates at 2.4 or 5 GHz.
RTL-SDR + SDR#: ₹1,500 USB dongle + free software. Covers 24 MHz – 1.7 GHz. Resolution and dynamic range are inferior, but it will find dominant emission peaks.
Oscilloscope FFT: adequate for finding relative changes after a fix, not for comparing to limit lines.

Radiated Pre-Compliance with a Wide-Band Antenna

For a first approximation of radiated emissions, place a wide-band antenna 1–3 m from the product in an open area (field, rooftop, outdoor space away from reflective walls) and connect to the spectrum analyser.

Antenna selection:

Biconical antenna: 30–300 MHz (covers the most problematic range for digital electronics)
Log-periodic antenna: 300 MHz – 1 GHz
Both antennas combined with an antenna switch: covers 30 MHz – 1 GHz

The official CISPR 32 test uses a 3 m or 10 m separation distance in a calibrated semi-anechoic chamber with an antenna 1 m above a ground plane and specific cable routing. Your open-area test at 1 m will not produce the same numbers. However:

Practical rule: if your product exceeds the Class B limit line by more than 10 dB in a pre-compliance 1 m test, it will almost certainly fail the accredited 3 m or 10 m test. If you are below the limit line by 10 dB in pre-compliance, you have a reasonable (not guaranteed) chance of passing. The 10 dB buffer is your test-site variation budget.

LISN for Conducted Emissions

Conducted emissions testing measures noise flowing back into the mains power line from the product. The test requires a Line Impedance Stabilisation Network (LISN) between the mains outlet and the Device Under Test (DUT).

The LISN:

Presents a stable, defined 50 Ω impedance to the DUT (so the result does not depend on the impedance of the mains grid at the test site)
Blocks high-frequency noise from the mains reaching the measurement
Provides a 50 Ω measurement port for the spectrum analyser

Without a LISN, conducted emissions measurements are meaningless — the result changes from socket to socket.

Commercial LISNs:

Tekbox TBLC08: two-line LISN, 9 kHz – 30 MHz, rated 8A, ~₹15,000–20,000. Covers CISPR 32 and FCC Part 15B conducted range.
Fischer Elektronik F-LISN V8: professional unit, used in accredited labs, ~€500.

Test setup: mains outlet → LISN → DUT. Spectrum analyser connected to LISN measurement port. Overlay the CISPR 32 Class B conducted limit line on the spectrum analyser display (stored as a reference line). Any peak above the limit line is a failure.

Common Failure Patterns and Immediate Fixes

Switching Regulator Harmonics

Symptom: comb of peaks at f_sw and multiples (2×, 3×, 4× …). Very common with DC-DC converters. At 400 kHz switching frequency, harmonics appear at 400 kHz, 800 kHz, 1.2 MHz, extending into the 30 MHz CISPR test range.

Fixes:

Add output Pi filter (C-L-C): typically reduces peaks by 20–40 dB
Add ferrite bead on regulator output
Slow down switching edges (increase gate resistance on the switching MOSFET) — reduces high-frequency content in harmonics
Add a small spread-spectrum capacitor on the RT/SYNC pin if the regulator supports it

Elevated Broadband Noise Floor

Symptom: spectrum analyser shows a broad raised floor across many frequencies rather than discrete peaks. Indicates poor power distribution network — noise is radiating from the power plane.

Fixes:

Improve decoupling capacitor placement: move 100 nF ceramics closer to IC VCC pins
Add via directly to power and ground at each capacitor — one via each, not shared
Check for a split or interrupted ground plane — any return current forced around a gap creates a loop antenna

Crystal and Clock Harmonics

Symptom: discrete peaks at f_crystal and its harmonics (3rd, 5th for fundamental crystals). A 12 MHz crystal generates peaks at 12, 24, 36, 48 MHz — the 48 MHz peak is directly inside the CISPR test range and often causes failures.

Fixes:

Add series termination resistor on the clock output trace: 22–47 Ω placed within 5 mm of the driving pin. This slows the edge, reducing harmonic content by 6–12 dB.
Ensure continuous ground plane beneath the clock trace
Add a small ferrite bead (120 Ω at 100 MHz) on the clock line if it exits the MCU to a remote peripheral
Use a chassis ground pour around the crystal on the PCB surface

Peaks Correlated with Cable Movement

Symptom: near-field probe shows elevated noise near the cable exit connector; spectrum analyser peaks change when the cable is moved.

Diagnosis: the cable is acting as an antenna. Common-mode current from the PCB is flowing onto the cable shield or the cable itself and radiating.

Fixes:

Add ferrite clamp-on core (Fair-Rite 31 material, 0431164281) at the cable exit — snap onto the cable at the chassis wall
Add common-mode choke (TDK ACM series) on the PCB before the connector
Bond cable shield to chassis at the connector point — use a 360° metal shell connector with chassis bonding

ESD Pre-Compliance Testing

Electrostatic Discharge testing is distinct from emissions testing but equally likely to cause certification failures. The IEC 61000-4-2 test fires 8 kV contact discharges and 15 kV air discharges at every metal surface and I/O connector.

Crude pre-compliance ESD test:

A piezo lighter produces ~10–15 kV with < 1 nJ stored energy. Hold it near the PCB and fire repeatedly. Not calibrated, not safe for production decisions, but it will reveal gross ESD vulnerabilities.
A commercial ESD gun (e.g., NOISEKEN ESS-2000A, ₹80,000–1,50,000 secondhand) replicates the IEC 61000-4-2 waveform at rated voltages.

What to watch for: any reset, lockup, data corruption, display glitch, or communication error during ESD application is a failure.

Fixes:

TVS diode on every I/O pin that exits the product (e.g., PRTR5V0U2X for USB, SP0503BAHT for general I/O)
Improved chassis-to-PCB ground bonding (reduces chassis potential rise during discharge)
Ferrite beads on I/O lines to slow the ESD current edge into the IC

The Fix-Measure-Fix Loop

Pre-compliance testing is not a one-time event. It is a loop.

Scan: take a complete spectrum from 30 MHz to 1 GHz (radiated) or 150 kHz to 30 MHz (conducted). Screenshot and note the 3–5 tallest peaks and their frequencies.
Identify: correlate each peak to a circuit element. Is it at the crystal frequency? At the SMPS frequency? Does it disappear when you disconnect the cable?
Fix: apply one fix at a time. Do not change multiple things simultaneously — you will not know what worked.
Re-scan: compare to the screenshot from step 1. Did the peak go down? By how much?
Margin check: is the peak now at least 6 dB below the limit line? 6 dB is the minimum margin to allow for test-site variation and measurement uncertainty.
Repeat: move to the next peak.

flowchart TD A[Build Board / Assemble DUT] --> B[Near-Field Scan\nover PCB surface] B --> C[Identify Top 3\nEmission Peaks] C --> D[Apply Fix to\nHighest Peak] D --> E[Re-Scan] E --> F{Peak ≥ 6dB\nbelow limit?} F -- No --> D F -- Yes --> G{More peaks\nabove limit?} G -- Yes --> C G -- No --> H[Book Accredited\nTest Session] H --> I{Pass?} I -- Yes --> J[Certificate Issued\nProduct Ships] I -- No --> K[Lab Failure Report\nIdentify New Issues] K --> D classDef std fill:#dbeafe,stroke:#1d4ed8,color:#1e3a8a classDef warn fill:#fee2e2,stroke:#dc2626,color:#7f1d1d classDef ok fill:#dcfce7,stroke:#16a34a,color:#14532d classDef caution fill:#fef9c3,stroke:#ca8a04,color:#713f12 classDef hw fill:#ffedd5,stroke:#ea580c,color:#9a3412 classDef proto fill:#f3e8ff,stroke:#9333ea,color:#581c87 class A,B,C hw class D,E std class F,G caution class H,I proto class J ok class K warn

Common EMC Failures — Cheat Sheet

Failure Type	Near-Field Symptom	Likely Cause	Fix	Typical Reduction
SMPS harmonics	Peaks at f_sw × N, near inductor	Switching edges, inadequate output filter	Pi filter + ferrite bead on output	20–40 dB
Crystal harmonics	Peaks at f_xtal × N, near crystal	Fast clock edges	33 Ω series termination + ground plane	6–15 dB
Cable radiation	Peaks change with cable position	CM current on cable	Ferrite clamp at cable exit	10–20 dB
Broadband floor	Elevated across 30–500 MHz	Poor decoupling, PCB loop areas	Better cap placement + via density	6–15 dB
GPIO harmonics	Peaks correlate with I/O activity	Fast edges on long traces	Series resistor 22–100Ω on GPIO	6–12 dB
ESD failure	Reset/lockup on discharge	No TVS, poor chassis bond	TVS diodes + chassis bonding strap	N/A

What Pre-Compliance Cannot Replace

Accredited tests are conducted in a semi-anechoic chamber (SAC): a metal-lined room with RF-absorbing foam on the walls and ceiling, a rotating turntable for the DUT, and a calibrated antenna on a height-adjustable mast from 1 to 4 m. The test engineer uses a calibrated receiver with a specific detector type (peak and quasi-peak for CISPR, peak for FCC) and scans through every frequency at every DUT orientation and antenna height to find the worst case.

None of this is replicable with a spectrum analyser on a bench. The chamber eliminates reflections; your bench does not. The calibrated antenna has a known gain curve; yours may not. The test protocol specifies exact cable routing; yours is arbitrary.

Pre-compliance testing cannot produce a certificate. It can only prevent you from failing the official test. The certificate comes from the accredited lab, and no shortcut exists.

This Is the End of the Certifications Section

You started this section at the ISI helmet — the mark on the box that a consumer trusts without understanding, because the testing system has already done the work. You have now walked through every layer of that system from the inside.

You have seen how a ferrite bead dissipates the noise your switching regulator produces. How a Pi filter cleans the power line before it enters the board. How a Faraday cage terminates the fields that escape the PCB entirely. And now how pre-compliance testing lets you find and fix these problems yourself, systematically, before the test lab charges you to discover them.

The ISI mark, the CE mark, the FCC ID — they are not bureaucratic obstacles. They are the evidence that someone went through this entire stack and proved the product does not interfere with everything around it. The helmet passed the drop test. The charger passed the conducted emissions test. The radio passed the spurious emissions limit. None of it happened by accident.

Now you know why it happens — and how to make it happen for what you build.

Key Takeaway

Pre-compliance testing is the highest-leverage activity in the certification process — it converts expensive lab failures into cheap bench discoveries. A ₹12,000 near-field probe kit and a ₹40,000 spectrum analyser can find 80% of the problems that a ₹2,00,000 lab session would bill you to identify. Run the fix-measure-fix loop until every peak is 6 dB below the limit. Book the lab only when your pre-compliance results are clean. The certificate at the end is not a reward for effort — it is the consequence of a disciplined process that you now know how to run.