[RFI] Switching Power Supplies and EMC Requirements

[Dale]

All,

Switching power supplies ("switchers") and the RF hash they create are a frequent topic on this reflector.  In an effort to help users of this group better understand the requirements levied upon switchers by regulatory bodies (FCC in the US and many European countries who use the CISPR 22 standard as their local requirement), I am supplying the URL to a very interesting document, "Electromagnetic Compatibility Considerations for Switching Power Supplies".  It was authored by personnel at CUI, Inc., a well-known manufacturer of switchers, especially those embedded into other end products where CUI small switchers are used to provide local regulation for certain circuits.  

The URL is:  http://www.cui.com/catalog/resource/emi-considerations-for-switching-power-supplies.pdf

Let me say two things: 1) I have no connection whatsoever with CUI, Inc.  2) This document is NOT a cover for a big product advertisement.  Almost surprisingly, it is a reasonably decent compilation of information, INCLUDING A TABLE OF FCC AND CISPR 22 EMC REQUIRMENTS, which allows the reader to gain a better understanding of what it takes to meet these requirements.  You can also get a feel for why, in so many instances, the requirements are not adequate to address the RFI needs of radio operators.

PLEASE NOTE:
In the table of requirements, pay careful attention to the fact that FCC Part 15, Class B radiated emissions are measured at 3 meters, and the Class A emissions are measured at 10 meters.  Remember, Class B is for home and office situations.  So, the Class B numbers look larger than the Class A numbers, but that's due to the difference in measurement distance.  Keep in mind the inverse square law as it applies to field strength distribution, as well.

Similarly, the CISPR numbers are also taken at different distances, with 3m again for Class B, but 30 meters (!) for the Class A standards.  To further complicate things, the units of measurement ("UoM") are not the same between FCC and CISPR, and there are also variances within each for conducted emissions (simple uV for FCC) and radiated emissions (uV/m for FCC).  Nevertheless, since so many users here (including myself) have voiced some uncertainty about various factors of the FCC (or other) emission requirements), the tables in this document at least provide a one-stop compilation of the most important factors affecting RFI from the various devices on the market.

CUI provides factual basic information on what needs to be done in the design of end user equipment to meet the various requirements.  The fact that so many of us end up having to battle the result of poor (or even non-existent) RFI design in a lot of equipment is ample proof that many manufacturers either circumvent the requirements to save costs (rouge third party suppliers in many cases) or don't have staff well enough versed in good EMC design practices.  Further, and worst of all, we are stuck with regulatory requirements that are not adequate to prevent RFI to radio systems that are co-located with the various products.  

In the end, the best cure for a device that has a noisy wall wart power supple is most likely to be the replacement of that switching wall wart with a linear equivalent - either homebrewed or made up by recycling some old wall warts from 25+ years ago when linears were the supplies of choice.  Yes, switching wall warts can be made better than they are for RFI, but if you read this document carefully, you can understand that doing so often involves cutting open that wart and placing by-pass caps (and maybe ferrite inductors) inside the case.  In certain situations, following the advice and practices given by Jim, K9YC, in his tutorials may do the trick; if not, it's either cut it open and perform internal surgery or replace with a linear supply.

73, Dale
WA9ENA
Sr EMC Engineer, E-N-A Systems, LLC