Getting CE Mark: A Case History
by Analog Services, Inc.
On January 1, 1996 it became mandatory for electronic equipment sold in Europe to have the CE Mark. This is a narrative of how we went about getting it for one product -- an industrial modem. This may be of help to you if you still need to get the CE mark for one or more of your products. But, be aware that your device may have different requirements than ours. So, this may or may not be applicable.
Events Leading To Testing
We understood the underlying purpose of CE Mark: To control electromagnetic interference (EMI) and electrostatic discharge (ESD). There were various EMI and ESD requirements and standards that existed at the time in many countries. And sensible equipment vendors were running tests and trying to comply with them. But the Europeans were going all out and requiring a Mark right on the product to indicate compliance. Without it, the product could not be sold in Europe.
Starting in 1995 we knew we would have to get the CE Mark. But we hadn't a clue as to how to get it. Most of what was published in engineering/technical magazines, etc. seemed contradictory or dwelt too much on listing specifications and their dependencies on other specifications. The State of Minnesota proved to be quite helpful. We contacted the Minnesota World Trade Center [1] in St. Paul and were soon provided with a stack of papers, seminar notes, and articles on the subject. These seem to have been written more for company executives (What's the bottom line?) and were better at explaining things.
After sifting through the available information, it seemed like there were two avenues open to us. One of these was to self-certify. This just means you just go ahead and add the CE Mark, based on your own tests or beliefs that the product complies. The other was to have another company that specializes in this testing run tests on our product.
Self-certification seemed slightly reasonable in our case,
since the maximum clock frequency of the device was only 460 kHz and the total current
consumption was only a few mA. We were relatively sure it wouldn't emit much EMI.
But there was still the question of susceptibility to EMI and ESD. And we
didn't have any way of running the tests. In the information we read, there were
dire warnings to the effect that people would be arrested and hauled away to jail if it
were proven that a CE-Marked device didn't comply. We didn't have any representative
in Europe who could be hauled away. And it didn't seem likely that anyone
would ever be driven to test our device for suspected non-compliance. But, in the
end, we decided that it should be tested by another company. Not only would outside
testing "prove" that we had acted reasonably. But it would also be a good
selling point. We would become associated (in a way) with the large and well-known
testing company.
At around this time we went to another consulting company that we had used before and that we new had taken a device through CE-Mark before. We asked them to quote on getting CE Mark for the device. But their response was that it was not terribly complicated and that we might as well do it. So we decided to go for it.
The next step was to determine what specification we should meet. The required specifications were soon identified as EN55022 and EN50082. These make a distinction between "light industry" and "heavy industry." This same set of tests applies to both. But the "heavy industry" tests blast the product with stronger EMI, etc. We decided that the light industry tests probably applied to our product because they are used for office products and the like. And our device was intended to be used with personal computers. At about this time we first contacted the company that we would ask to quote on the testing. They also advised us that our device fell into the "light" category.
Then came one of the most difficult parts of the process. We needed to decide what our device should be doing during the tests and what should constitute a failure of the EMI susceptibility test. We believed that we would probably have to put this into the form of a document to give to the people who would do the testing. There really were only 3 things that our device -- a modem -- could do: sit there idle, send a message, or receive a message. We decided that we should connect two modems together through a short length of cable and have one sending out messages at regular intervals and the other receiving them. Both of the modems would be exposed to the EMI. The receiving computer would know what the transmitted message should be and would make a bit-by-bit comparison. It would keep a count of the number of received messages and the number that contained errors. It would display the message count and error count as the test progressed.
The cable could be relatively short because our purpose was to test the modem susceptibility to EMI and not the cable. The cable could also be shielded, since this is what we specified as a normal installation in the device manual.
We next got a quote from the testing company. The initial quote included safety testing. After talking with them a few times, we all concluded that safety testing didn't make a whole lot of sense for our device because it was powered from the RS232 port of a computer. If the computer port was safe, then the combination of port and our device should also be safe.
A day was set up to run the tests. There was a backlog of about 2 months, which was OK. It would take us a while to write the test software, anyway. We set about writing the test software. But we still didn't know how many errors we should consider acceptable. In the end we didn't say how many errors there should be and the testing company didn't seem to care. The test would be more or less an experiment.
The Tests
The day of the test we brought a disk containing the software, a cable pre-wired with spade lugs, and 4 production modems to the test house. The testing company wanted to provide the two computers for the test. The reason for this is that they want to use computers that they know, from previous testing, to be compliant. We were permitted to observe all of the testing and to help set up parts of each test.
The first test was a radiated EMI susceptibility test. The computers, two of the modems, etc. were set up in a small room with ferrite walls for EMI absorption. The room was also equipped with antennae for irradiating the equipment under test. The antennae were connected to an RF power amplifier located outside the room. A video camera set up to observe the computer displays was connected to a monitor outside of the room. Prior to the start of the test, the software was exercised in the presence of the examining engineer to show that it would indeed count a message error when one occurred.
The EMI susceptibility test was relatively automated, with the frequency being slowly swept from 27 MHz to 500 MHz in 2.5 kHz increments at 40 millisecond per increment. The field strength used is 3 volt/meter. The test standard is IEC-801-3. There was a sweep with a vertical antenna and another with a horizontal antenna. It turned out that there were no errors throughout the test.
The second test was a conducted EMI susceptibility test. This test took place in the same little room. The test was almost the same as before, except that now the cable connecting the two modems was placed in a clamp device so that it was between two pieces of metal, each about 1 meter long. Pulses of 500 volt amplitude were applied to the plates. The test standard is IEC-801-4. This time there were some errors. But the examining engineer was only concerned that the devices resumed normal operation after pulsing. This was a relief because they did resume normal operation.
The next test was an emissions test. This was conducted in a large open area with the two modems on a turn-table about 30 ft by 30 ft. A receiving antenna is placed off to one side of the turn-table. The antenna is connected to a spectrum analyzer. With the modems and computers running the same test software as before, the spectrum analyzer output is examined over a range of 30 MHz to 1 GHz. At one point the antenna had to be changed. The examining engineer pointed to various spectral lines that he knew to be radio stations or emissions from the computers. But he did not identify anything that originated in the devices under test. So far, our products were batting 100%.
Next came the ESD test. The units under test are placed on a table with a copper surface and having a short copper "wall" nearby. The wall is perpendicular to the table top. An ESD "gun" is used to zap various spots. The ESD strengths used are 2 kV, 4 kV, 6 kV, and 8 kV. The test standard is IEC-801-2. The first spot zapped is the table top. Next is the wall. Next is the device case. The top, bottom, and sides of the device case are all zapped.. Again, the criterion is that the device should resume normal operation after the zapping. The first unit tested didn't. So the examining engineer tested the remaining three devices that we had brought. These kept on working. His conclusion was that the design was acceptable and that there was a "manufacturing defect" in the one that failed. We didn't argue.
The Envelope Please
We received a final write-up of the test results several days later. This was in the form of a booklet that explained the testing in quite a bit of detail. It included color photographs of the tests being performed. The conclusion by the test engineer was that the device had passed all tests. A certificate to this effect was included.
The Label
The next step was to make an adhesive label for the device. The label contained the little "CE" made according to dimensions that we'd gotten in information from the State of MN.
The Declaration
A declaration document needed to be included with device shipments. So we decided to put it into the device's instruction manual. Then at least one copy would ship with every unit. The declaration was based on a template and read as follows:
DECLARATION OF CONFORMITY
Application of Council Directives ..................... 89/336/EEC
Standard(s) to which Conformity is Declared ........... EN55022:1987
EN50081-2:1992
Test Laboratory ........................... AAAA Testing Services
Test Report No. ........................... 560239
Manufacturer's Name ....................... BBBB, Inc.
Manufacturer's Address .................... CCCC
Type of Equipment ......................... DDDD
Model No. ................................. MMMM
Serial No(s) _________________________________________________________
I, the undersigned, hereby declare that the equipment specified
above conforms to the above Directive(s) and Standard(s).
_________________________ _________________________________
(signature)
DATE: ___________________ _________________________________
_________________________________
(position)
References
1. Minnesota Trade Office, 1000 Minnesota World Trade Center, 30 E. 7th St., St. Paul, MN 55101-4902 USA.
For more information on how Analog Services, Inc. can help solve your circuit/system problems, call or e-mail us today.
Contacting Analog Services, Inc.:
VOICE 612-927-7112 FAX 612-929-7503
E-Mail: stevea@analogservices.com