Update to CISPR-25 in 2002: Impact of Changes

CISPR-25, Second Edition, 2002-08 - Radio disturbance characteristics for the protection of receivers used on board vehicles, boats, and on devices Limits and methods of measurement

The 1995 edition of CISPR-25 is a widely regarded standard, adopted by many automotive OEMs and used as the basis for other standards, both OEM and ISO standards.  The LISN defined in CISPR-25 is used in virtually all automotive component test standards and often referred to as the CISPR-25 LISN, hence CISPR-25 is potentially the most important single standard for automotive component EMC testing. 

The 1995 standard has some idiosyncrasies that plague both test services and equipment designers, some are due to the frequency bands which are non-contiguous, some are due to the undefined classifications in the standard, some are due to modern technology exceeding the frequency limit.  There is also a lack of correlation of the test set-up with the EU legislative directive 95/54/EC due to minor set-up differences for radiated emissions testing.  Despite these idiosyncrasies CIPSR-25 is the de-facto global standard for conducted emissions and the prime global standard for radiated emissions.

The standards was up-issued and published in 2002, this article examines the changes from the 1995 edition and discusses which, if any, of the previous editions idiosyncrasies have been addressed.


The first things to note are the lack of changes to most of the test conditions; scan times, resolution bandwidth (RBW), use of peak, average and quasi-peak (QP) detectors remains the same as with the 1995 edition.  The definitions of frequency classes (LW, MW, SW, FM, VHF and mobile services) is consistent between the older and newer revision, the test methods both on-vehicle and of components remains the same as do the test set-ups, this includes the use of TEM cell testing that some commentators expected to be removed from the revision.

Power supply regulation for component testing is the same as the previous edition, however the 2002 revision include regulation specification for the supply during on-vehicle testing (12V +2/-0V and 24V +4/-0V).  Many will be disappointed with the lack of contiguity in the standard, the standard still has non-contiguous frequency bands (non joined-up frequency ranges), the previous standards bands remain intact apart from a few minor changes given below.

Conducted Emissions

The lower limit of the high VHF band has reduced by 2MHz from 70MHz to 68MHz; the upper span is now 68MHz to 108MHz.  The note that used to be at the bottom of table 6 (broadband peak and QP detector limits) specifying the 6dB addition to the limits for short duration disturbances is removed, this clarifies the situation significantly and does not cause the previous confusion on short duration versus narrowband signals.  The note on table 7 (narrowband, peak detector limits) specifying a similar 6dB addition for the 76MHz to 108MHz band has been removed and the table explicitly lists 68MHz to 87MHz as a mobile services band with the previous limits intact and 76MHz to 108MHz as a broadcast band with the 6dB addition tabulated.

The artificial network (AN or LISN) is maintained, the component tolerances are still not specified, but a lower limit of attenuation is provided for signal line AN's as well as the typical curve used in the 1995 edition.  The lower limit curve starts at 10dB at 1MHz, rising to 40dB at 30MHz, then continues at 40dB up to cut-off frequency of the measurement (400MHz in the provided example where attenuation drops to 20dB up to 1GHz).  The specifying of the minimum attenuation across the range 1MHz to 1GHz is a welcome improvement and should help with repeatability of results from different AN makes for signal line results (not normally a problem for conducted emissions as these were within the previously specified frequency band (100kHz-100MHz).  As the same AN is used for radiated emissions this limited attenuation could have an effect on these results as the harness termination will have a defined impedance over the complete measurement frequency band.  Not applying this lower limit of attenuation to the power line AN is a serious omission in this revised standard.

Another addition to the specification that makes measurements easier for fixed harness length installations is the ability to use a harness length other than 0.2m, although this is still specified as the standard length for conducted emissions.  If using a harness other than the standard 0.2m the upper measurement frequency (fc) is limited by the equation;

fc = 30 / lp

Where lp is the length of the harness in metres.

Although most testing will still be performed with 0.2m harnesses, if a component has integral fixed power leads these could be used and result obtained at the terminal end.  This may be more useful for testing after market products that come with a fixed cable and accessory socket plug.

Radiated Emissions

The same 2MHz frequency change has been implemented in the radiated emissions testing, for both on-vehicle and component test methods, increasing the mobile services VHF band from 70MHz-108MHz to 68MHz-108MHz.    Other upper frequency bands for mobile services have had their spans increased.  VHF band 4 spans 142MHz-175MHz (-2MHz and +3MHz from first edition), UHF band 1 spans 380MHz-512MHz (-40MHz and +0MHz from first edition) and UHF band 2 is unchanged.

On-vehicle tests have unchanged limit levels for radiated emissions testing.  The frequency bands below 54MHz have unchanged limit levels for both peak and quasi peak detector (table 10) and for narrowband signals (table 11) for component tests.  However, above 70MHz in the first edition of CISPR-25 all bands had the same limits set for each CISPR-25 class for components, the second editions separates out these upper bands and applies different levels above 175MHz.  Above 175MHz the limit levels for radiated emissions are raised by 7dB in the band 380MHz to 512MHz and by 13dB for the band 820MHz to 960MHz, for all detectors and broad and narrowband disturbances.


One of the more forward looking changes implemented, that is new to the second edition, is a flow chart to determine the applicability of CISPR-25 (Annex A).  What is interesting and forward looking is not just the inclusion of such an applicability flowchart, but the implication that both compression engine vehicles (diesel) and electric vehicles are covered by the standard (and therefore hybrid and fuel cell vehicles).

Missed Opportunity

There are several very disappointing aspects of this revision of CISPR-25, the main being the minute level of changes that may clarify some of the minor idiosyncrasies, but do nothing to significantly improve this test standard.  There is virtually no need to buy this new edition since the changes are so minor as to question the need for a revision at all.  The one caveat for this is if you have a product that is showing failure at frequencies above 175MHz where the increased limits may help a non-compliant to CISPR-25: 1995 product pass the revised standard CISPR-25: 2002.

The main missed opportunity has been the lack of extension of the upper test frequency for radiated emissions to encompass mobile telephony services at 1800MHz and 1900MHz and bluetooth technology at 2.4GHz.  Many OEM's have already increased their radiated emissions upper frequency limit to between 2GHz and 3GHz using the CISPR-25 set-up, to have this encompassed in the generic standard alone would have made the revision worthwhile.  Another missed opportunity was to specify limits in the out-of-band regions of the frequency scales for both conducted and radiated emissions.  Even the simple step of joining the end points of the existing bands to produce a ramped limit set would have been a step forward, this revision adds no significant change to the frequency bands or the limit lines.

Other more minor changes that would have made some improvement would have been the increase of the test table height to 1m.  Increasing the test table height by 10cm would not only align this standard with 95/54/EC, but increasing the antenna height from the chamber floor does improve signal reception for large log-periodic and biconical antennae, making the repeatability better for 1m antenna calibrations in the vertical polarisation.

Conclusion: a wasted opportunity to significantly improve the CISPR-25 standard, the revision is nothing more than a minor clarification of the existing document, that after 7 years was well understood anyway, and a reduction of the radiated emission limits above 175MHz.


www.AutoEMC.net 2003                                   TOP OF PAGE                                        HOME