EMC Pre-Compliance Testing: Save Money Before the Lab

Why pre-compliance testing exists

Roughly half of consumer electronics products fail EMC testing on their first attempt at a formal accredited lab. When that happens, the math gets ugly fast: $2,000 -- $10,000 in retest fees, plus $1,000 -- $15,000 for the engineering fix (which might require a board re-spin), plus 4 -- 12 weeks of added timeline while you redesign, rebook the lab, and retest.

Pre-compliance testing exists to catch those failures before you spend the money. You run informal emissions measurements on your prototype using either your own equipment or a rented pre-compliance bench. When something fails -- and on a first product, something usually does -- you fix it on your bench with fast iteration cycles, not in a $400/hour accredited chamber.

The concept is simple: spend $500 -- $2,000 to avoid spending $5,000 -- $30,000.

What pre-compliance testing is

Pre-compliance testing is informal EMC measurement performed before formal certification testing. It is not required by the FCC, not performed to accredited standards, and the results have no regulatory standing. You cannot use pre-compliance data in an SDoC declaration or a TCB application.

What it does give you:

• A realistic picture of your product's emissions profile against FCC limits
• Early identification of specific failure modes (which frequency, which source, which path)
• The ability to iterate fixes quickly and cheaply
• Confidence that your product will pass formal testing

Pre-compliance testing typically covers the same measurements as formal testing -- radiated emissions and conducted emissions -- but with less expensive equipment, in less controlled environments, and with wider measurement uncertainty. The goal is not perfect accuracy. The goal is to find problems that would cause a definitive failure at the lab.

A typical rule of thumb: if your product passes pre-compliance with 6 -- 10 dB of margin against FCC limits, it will almost certainly pass formal testing. If it is within 3 dB of the limit or over it, you have a problem to fix.

The two measurements that matter most

Radiated emissions (30 MHz -- 40 GHz)

Radiated emissions testing measures the electromagnetic energy your product unintentionally radiates into the surrounding space. This is the measurement that catches:

• Clock harmonics: Your microprocessor, FPGA, or SoC runs at a clock frequency (say, 48 MHz). Harmonics of that clock radiate from PCB traces, cables, and enclosure gaps at multiples of the fundamental: 96 MHz, 144 MHz, 192 MHz, and so on. These harmonics are the most common cause of radiated emissions failures.
• Switching power supply noise: DC-DC converters and switch-mode power supplies generate broadband noise from their switching edges. This noise couples to nearby traces and cables and radiates.
• Cable radiation: Cables act as antennas. USB cables, HDMI cables, ribbon cables, and power cords are often the dominant radiating structures, not the PCB itself. A product that passes radiated emissions with short cables may fail with longer ones.
• Display emissions: LCD and OLED displays with high-speed data interfaces (LVDS, eDP, MIPI DSI) generate significant broadband emissions from the display cable and the panel edges.

FCC Part 15 Subpart B Class B limits for radiated emissions at 3 meters measurement distance:

For pre-compliance, you want to be at least 6 dB below these limits to account for measurement uncertainty.

Conducted emissions (150 kHz -- 30 MHz)

Conducted emissions testing measures the noise your product puts back onto the AC power line. This is primarily a power supply test. If your product runs on battery or USB power only (no AC mains connection), conducted emissions testing may not apply.

The measurement is performed using a Line Impedance Stabilization Network (LISN), which presents a defined impedance to the device under test and routes the conducted noise to the measurement receiver.

Common sources of conducted emissions failures:

• Switching power supply fundamental and harmonics: A flyback converter switching at 65 kHz produces harmonics at 130 kHz, 195 kHz, 260 kHz, and so on -- all within the conducted emissions measurement range.
• Insufficient input filtering: Missing or undersized common-mode chokes, X-capacitors, and Y-capacitors on the AC input.
• Ground loops: Multiple ground connections creating loops that couple noise to the power line.

FCC Part 15.207 Class B conducted emission limits (quasi-peak):

Equipment for in-house pre-compliance

You do not need a $2 million anechoic chamber for pre-compliance. Here is what a practical in-house setup looks like, from minimum viable to fully capable.

Minimum viable setup ($3,000 -- $8,000)

This setup lets you identify the worst offenders and iterate fixes. You will not get accurate absolute measurements (no calibrated antenna factors, no controlled environment), but you will find the 10 -- 20 dB failures that would sink your formal test.

Near-field probes are particularly valuable. By holding a probe close to different parts of the PCB, you can pinpoint the specific IC, trace, or connector that is the dominant emission source -- information that makes fixing the problem dramatically faster.

Full pre-compliance bench ($10,000 -- $25,000)

This setup provides measurements accurate to within 3 -- 6 dB of a formal lab, which is good enough to make confident pass/fail predictions.

Rent, do not buy

If you ship fewer than 3 -- 4 products per year, buying equipment may not be justified. Many accredited labs rent their pre-compliance chambers for $500 -- $2,000 per day, giving you calibrated equipment in a controlled environment with a lab technician available for guidance. Some universities with EMC research programs and hardware incubators (HAX, Highway1/SOSV) also offer access at reduced rates.

Even a single day of rented pre-compliance time is high-ROI if it prevents one formal lab retest ($2,000 -- $10,000+).

Common failure modes to catch early

These are the problems that most frequently cause formal lab failures. Check for all of them during pre-compliance.

1. Clock harmonic radiation

What it is: Harmonics of your system clock (or any periodic signal) radiate from PCB traces, connectors, and cables. A 48 MHz clock produces harmonics at 96, 144, 192, 240 MHz, and so on. One of these harmonics will land near an FCC limit.

How to find it: Use a near-field probe over the clock oscillator, PLL, and traces carrying the clock signal. Look for narrowband spikes at exact multiples of the clock frequency on the spectrum analyzer.

Common fixes:

• Add series resistance (22 -- 33 ohm) on clock lines to slow edges
• Use spread-spectrum clock generators (SSCGs) to spread the harmonic energy across a wider bandwidth, reducing the peak
• Shorten clock traces and avoid routing them near board edges or connectors
• Add local decoupling (100 nF + 10 nF) close to the clock source

2. Switching power supply noise

What it is: DC-DC converters generate broadband noise from fast switching edges. This noise conducts back onto the AC input (conducted emissions) and radiates from the inductor, switching node, and input/output cables (radiated emissions).

How to find it: Conducted emissions measurement with a LISN. For radiated, near-field probe over the power supply section -- look for broadband noise humps centered around the switching frequency and its harmonics.

Common fixes:

• Add or improve the input EMI filter (common-mode choke + X-cap + Y-cap)
• Slow the switching edges with gate resistors or snubber circuits (RC snubber across the switch)
• Use a shielded inductor instead of an unshielded one
• Minimize the hot loop area (switch, inductor, output capacitor)
• Add a copper pour under the power supply section connected to the ground plane

3. USB and HDMI cable radiation

What it is: High-speed digital interfaces (USB 2.0/3.x, HDMI, DisplayPort) put high-frequency signals on cables that act as efficient antennas. Common-mode current on the cable shield radiates.

How to find it: Measure radiated emissions with and without the cable attached. If emissions jump 10+ dB when you plug in the cable, the cable is the dominant radiator.

Common fixes:

• Add a common-mode choke or ferrite bead on the cable at the connector
• Improve the ground connection between the cable shield and the device chassis/ground plane
• Use a shorter cable (less efficient antenna)
• Add filtering on the data lines at the connector (common-mode filter ICs for USB 3.x, HDMI)

4. Enclosure leakage

What it is: Gaps and seams in a metal or shielded enclosure leak RF energy. A slot acts as a slot antenna -- a 15 cm slot is an efficient radiator at 1 GHz.

How to find it: Near-field probe over seams, ventilation openings, and cable entry points. Compare emissions with enclosure open vs closed.

Common fixes:

• Add conductive gaskets at enclosure seams
• Break up long ventilation slots with crossbars
• Use conductive paint on plastic enclosure interiors
• Ensure cable shields make 360-degree contact at connectors

5. Ground bounce and decoupling

What it is: Insufficient decoupling near ICs causes voltage bounce that radiates from PCB edges and cables.

How to find it: Near-field probe over high-speed ICs. Look for broadband elevation in the 100 MHz -- 1 GHz range.

Common fixes:

• Decoupling capacitors (100 nF ceramic) at every IC power pin
• Bulk decoupling (10 uF -- 100 uF) near high-current ICs
• Solid ground plane with no splits under high-speed signals

Cost comparison: pre-compliance vs lab failure

The math is clear. Even the most expensive pre-compliance investment ($25,000 for a full in-house bench) pays for itself if it prevents a single board re-spin and retest cycle on your first product. For a company that ships 2+ products per year, the payback period is under 12 months.

Practical tips for small teams

Start with near-field probes

If you can only buy one piece of pre-compliance equipment beyond a spectrum analyzer, buy a near-field probe set. Near-field probes tell you where the problem is, not just that a problem exists. Knowing the source is 80% of the fix.

Measure with cables attached

Your product will be tested at the formal lab with cables plugged in -- USB, HDMI, power, Ethernet, whatever ships in the box. Test the same way during pre-compliance. Cables are often the dominant radiating structure, and a board that looks clean on its own may fail badly with cables attached.

Test in the worst-case operating mode

The FCC requires testing in the operating mode that produces the highest emissions. For a product with multiple modes (idle, active, high-performance, charging), find the noisiest mode during pre-compliance and use that for your measurements. Typically, the worst case is maximum CPU load + maximum display brightness + active USB data transfer + active wireless communication.

Do not ignore below 30 MHz

Radiated emissions limits start at 30 MHz, but conducted emissions limits start at 150 kHz. Many teams focus on radiated emissions and forget conducted. Switching power supply noise in the 150 kHz -- 30 MHz range is one of the most common failure modes, and it is also one of the easiest to fix with proper input filtering.

Keep a pre-compliance log

Document every measurement: date, firmware version, hardware revision, cable configuration, spectrum analyzer settings, and results. When you iterate on fixes, this log tells you exactly what changed and whether it helped. It also helps your formal lab understand your product when you get to testing.

Use ferrites as diagnostic tools

Before committing to a board re-spin, clip snap-on ferrites onto cables and power lines to see if they reduce emissions. If a ferrite on the USB cable drops your worst emission by 10 dB, you know the USB cable is the dominant path and you can add a permanent common-mode choke to the design. Ferrites cost $1 -- $5 each. Board re-spins cost $1,000 -- $15,000.

Design for EMC from the start

The cheapest pre-compliance fix is the one you never need. Basic EMC design practices that prevent most failures:

• Solid ground plane on layer 2 of your PCB (no splits under signal traces)
• Decoupling capacitors at every IC power pin
• Series termination resistors on clock and high-speed data lines
• Input EMI filter on any AC-powered supply
• Keep high-speed traces away from board edges and connectors
• Use shielded connectors with proper ground contact

These practices add minimal BOM cost ($0.50 -- $2.00) and zero schedule impact. Skipping them creates $5,000 -- $30,000 problems downstream.

When to skip pre-compliance

You can go directly to formal testing when your product is a minor revision of a previously certified design, you are using an unmodified SoC vendor reference design, or your product is very simple (single-board, battery-powered, no high-speed interfaces).

For a first-time product from a team without prior EMC experience, pre-compliance is not optional in practice. The probability of first-pass failure is too high and the cost of failure is too steep to skip it.