Start Here:
CISPR 32: “Electromagnetic compatibility of multimedia equipment - Emission requirements”
CISPR 32 is the international standard most comparable to the FCC standard on consumer electronics.
CISPR 32 governs emission, both conducted and radiated, from “multimedia equipment”, which encompasses most consumer electronics, such as computers, televisions, radios, etc. It is the international standard most similar to the testing of unintentional emitters required by the FCC and found in ANSI C63.4. CISPR 35 is the corresponding immunity requirement document. In the EU, CISPR 32 is incorporated as a regulatory document with the number EN 55032.
The most current version is from 2015 with amendments in 2019 and can be purchased here. It superseded CISPR 13 and 22, which makes sense--CISPR 13 had covered broadcast media equipment and CISPR 22 had covered IT/computer equipment, and they were managed by different committees. That was fine until digital TV, digital broadcasting, and then streaming became ubiquitous--now broadcast equipment was IT equipment and it all got very confusing (no one likes testing their equipment twice). CISPR 32 is the result of a harmonization effort. It uses CISPR 16 to govern the standards of its measurement equipment. Officially the standard covers the frequency range 9 kHz - 400 GHz, although most equipment does not need to be tested to the extremes of that range. Generally equipment will be tested up to 6 GHz, but check the tables carefully--the upper testing limit depends on the frequency usage of the equipment under test. The conducted emissions testing will generally start at 150 kHz.
Like the FCC, CISPR 32 classifies units as “Class B” if they are intended for use in residential environments; those have stricter emissions requirements. Everything else is “Class A”. There are limit lines for both quasi-peak (QP) and average detector methods; see this explainer for information on quasi-peak and how to minimize QP test time. Also like FCC, limits are provided for measurements at both 3 m and 10m distances.
Interference Technology has good articles on the 2012 version of CISPR 32 by Dan Hoolihan and the 2015 version by Ghery Pettit.
TIP:
The actual limits that are required to be met are found in Annex A of the document.
FCC 47 CFR Part 15
This is the part of FCC regulations that covers normal electronics, as opposed to systems that have intentional RF transmitters and receivers.
Title 47 of the Code of Federal Regulations, Chapter I, Subchapter A, Part 15, “Radio Frequency Devices” is one of the main ways the Federal Communications Commission controls electronic devices from interfering with other systems in the same area. Part 15 is publicly available. It deals with items that are capable of generating RF interference and aren’t operated by licensed users (an example of licensed operators would be Ham radio enthusiasts).
Part 15 defines Class A and Class B digital devices. Class A devices are specifically for use in industrial environments, such as a controller unit for an industrial furnace. Class B devices are meant for residential use (even if they’re often used in business environments, such as a laptop). The limits on emissions are more strict for Class B devices, since they are used in open environments, whereas the Class A units are assumed to be used in environments where stricter controls on unit-to-unit interference are possible. Subpart B has the quasi-peak and average limits for unintentional radiators, units that do not have an intentional RF transmitter capability. (For more on quasi-peak measurement, here’s an explainer.)
For consumer electronics, the main route of compliance is the Supplier’s Declaration of Conformity. The idea is that the equipment manufacturer has a product sample tested to ANSI C63.4 for both conducted and radiated emissions, and keeps a copy of the test report that shows passing results. (Products that have intentional radiators, such as a Bluetooth or WiFi element, need to test to ANSI C63.10.) If there is ever a question about the product causing unwanted interference problems, the FCC will contact the supplier for the evidence of compliance. Every year several companies are fined for operating non-compliant equipment that ends up jamming neighboring systems, and the fines can add up quite quickly.
There is no immunity requirement from the FCC--their priority is to make sure equipment available on the market does not cause RF interference to other systems. They don’t particularly care whether your product works or not, which is what immunity standards fundamentally address.
TIP:
FCC testing is most similar to CISPR 32. Testing to CISPR 25 or MIL-STD-461 (RE102 or CE102) is not directly comparable to FCC limits, since the test setups are too different.
Quasi-Peak Explainer
Peak and Average measurements are pretty self-explanatory, but what on Earth is Quasi Peak (QP)?
If you think about old-school over-the-air radio (AM/FM), you’ve almost certainly heard static on a weak channel. If you think about the noise that could be coming through your car stereo speakers, which is more annoying, background/hissing static, or a continuous shrill tone?
The idea that a continuous screech is more annoying than background pops and clicks is the rationale behind QP measurements. In this method, you dwell at each frequency step long enough to apply weighting to the signals you find there. If you have a continuous wave (CW) signal, that will come through as a single tone. However if you have transient noise that is bouncing around and only present during part of the dwell time, that would be the popping and hissing type of noise. QP measurements weight CW signals more heavily. Thus for FCC or CISPR 25 testing, you have to show measurements compared to the QP limit line, in order to be less annoying to the public.
In a world where all our noise sources are stable, which is what we’re ideally trying to achieve in testing, then there can’t be any signal “worse” than a CW signal when measuring for QP. However, a peak detector will easily capture the peak of a CW signal, no problem. And a peak sweep is much faster than QP, since you don’t have to dwell at each frequency so long. So what is typically recommended is to do a peak sweep that covers the full frequency range first. If you pass that, it’s accepted that you will obviously pass the QP measurements as well. If there is a peak value over the limit, then you can return to that specific frequency range and do a longer-dwell QP sweep only in that area, saving test time.
TIP:
I’d like to note that the rule that QP is always less-than-or-equal-to Peak only holds when testing conditions are static. If you are testing a more complex system that may have functions turning on and off during testing, it is possible to see QP values higher than Peak at the same frequency. That’s because the measurements are taken at different times, and the vehicle or other DUT may have turned on something in a later test that wasn’t there originally (e.g. a continuously running electric vehicle (EV) turns on its cooling system for thermal management partway through a test). That’s not to say that you should always run a full scan on QP detection; that would take forever. But if you see QP greater-than Peak, that can be an explanation.