Start Here:
CISPR 25: “Vehicles, boats and internal combustion engines - Radio disturbance characteristics - Limits and methods of measurement for the protection of on-board receivers”
CISPR 25 is an automotive standard that covers the main emissions testing methods needed to ensure self-compatibility between a vehicle and its own on-board receivers..
CISPR 25 governs emissions related to automotive components, both for ground vehicles and boats. It includes both component level and vehicle level tests. The current version is the 5th edition published in 2021, and you can purchase it here. Some of the component level test methods in CISPR 25 are very similar to comparable methods in MIL-STD-461, but they are not at all similar to emissions testing done for the FCC via ANSI C63.4. (You can check out one of my presentations that dives pretty deep into that comparison.) And obviously MIL-STD-461 does not have an equivalent of the full vehicle testing found in CISPR 25.
More than anything, CISPR 25 is aimed at self compatibility. Historically this meant limiting emissions that would interfere with onboard radios (AM/FM/DAB) so as to avoid customer complaints. Today with vastly more complex vehicle electronics and the HV components of EVs, there are a lot more “victims” that can be affected by emissions than just AM/FM radio. [Aside: I once worked on troubleshooting an EV where a CISPR 25-noncompliant inverter interfered with the vehicle’s twisted/shielded CAN lines to the extent that whenever the driver stepped on the “gas” pedal, the battery control module shut down, requiring a full restart. In that case the solution was improving the shield terminations on the CAN lines.]
In general, vehicle level tests are done in four possible modes: key on, engine off; internal combustion engine (ICE) in driving mode (spinning wheels on a dyno); EV in charging mode, and EV in driving mode. (A plugin hybrid electric vehicle would need to test in all four modes.) For testing on the full vehicle, you need to disconnect each antenna from its receiving radio module, and feed the antenna coax into the measurement receiver, often through an impedance matching unit.
At the component level, CISPR 25 includes both conducted and radiated emissions tests. Conducted emissions can be measured via voltage measured from an artificial network (AN, equivalent to a LISN in the aerospace/defense world), or via a current probe measured at a minimum of two locations along the cabling. For radiated emissions the main test method is conducted in a semi-anechoic chamber, with an informative Annex F describing a stripline method. It currently does not provide for a reverb chamber method.
TIP:
Pay close attention to the flowchart in Figure 1 of CISPR 25. Because of the different limit lines for different detectors(peak, quasi peak, and average) and how they apply to narrowband vs. broadband noise sources, there are a lot more steps before saying something definitively “passes” or “fails” than in a method like MIL-STD-461 RE102.
TIP:
CISPR 25 is a critical standard for the HV components of electric vehicles as well as the traditional 12 Vdc systems. Take this as one consultant’s experience, but every EV on which I’ve done troubleshooting had problems with an HV component that failed CISPR 25 testing (particularly inverters and DC/DC converters). I have not yet had to troubleshoot on an EV where every component passed CISPR 25. It is a major challenge to design an HV inverter that passes CISPR 25 limits, but it can be done–and it’s likely worth investing the extra time to design a compliant system, or the extra money to buy one. You can also see my presentation on Noise Sources in EVs for more.
TIP:
Getting to the audio head unit (radio) to disconnect the AM/FM radio for vehicle level testing can be painful, and the sweeps can take a long time, especially if there are a lot of frequencies where quasi peak measurements are needed. CISPR 25 is not a regulatory test, so judgment of compliance is up to the manufacturer. It may make more sense to run the vehicle and tune through stations looking for audible problem areas, then only run the CISPR 25 sweeps on frequency ranges where there are issues. Making sure the audible test is legitimate has its own set of challenges, but once set up, it can run considerably faster than full CISPR 25 sweeps.
CISPR 12: “Vehicles, boats and internal combustion engines – Radio disturbance characteristics – Limits and methods of measurement for the protection of off-board receivers”
CISPR 12 is a key automotive standard that details the test method for ensuring cars and other vehicles don’t emit too much RF energy at a distance, usually 10 m away. It’s often referred to as an “off-board radiated emissions test”.
CISPR 12 is one of the critical standards that apply to full vehicles, whether using traditional internal combustion engines, electric vehicles, or hybrids. The current edition is from 2007, with amendments in 2009, and you can find it for purchase here. The intent of the requirement is to protect radio receivers that are external to the vehicle, such as an FM radio in a nearby house, or the two-way communication system of a passing ambulance. It applies to boats and ground vehicles, with both internal combustion or high voltage EV engines. The test is generally at 10 m distance, although if only a 3 m separation between the vehicle under test and the test antenna can be achieved, the limits can be increased by 10 dB to account for the difference (this linear extrapolation is doable since both 3 m and 10 m measurements are reliably in the far field in the 30 MHz - 1 GHz range--don’t try this with 1 m measurements such as in CISPR 25).
TIP:
CISPR 12 has limits that are expressed in Average and Quasi Peak. Quasi Peak (QP) testing takes forever, so the standard recommends speeding things up by doing a Peak sweep first. If all the Peak values are below the QP limit lines, then it’s accepted that any QP measurements will also pass the requirement. (This is true probably 95% of the time, but not always.) In that case you’re done and you’ve saved yourself a lot of time compared to taking QP measurements. If there are specific frequencies where Peak values are over the QP limits, you can go back and take QP measurements at only those frequencies. Quasi Peak detection is a little strange and can be hard to track down information on, so see this Quasi Peak article for a quick explainer.
TIP:
Another way to speed up testing is to increase the bandwidth of your initial scan, keeping dwell time the same. If you pass the specified limit with the wider bandwidth, you can be assured that you would have passed with the smaller bandwidth. Then if you run into exceedances at a specific frequency range, you can go back and test that range with the smaller spec bandwidths.