Before WWII the EMC business was pretty much on telephone poles and in radio broadcasting studios. Although it was nearly impossible to cause a relay to open or close with crosstalk, you sure could, and folks often did, interfere with each other’s conversations on the phone. There is an urban legend that the official start of EMC was when the Air Force found that the tail gunner in a B29 was having trouble hearing the pilot’s commands over the intercom because of noise in the system.

The U.S. military decided they had to get electrical noise under control so they wrote specifications setting acceptable levels of it in their systems. How do you characterize electrical noise? By the field amplitude, of course. So the specifications required that dipoles be used in an open field in order to insure a legitimate measurement of the far field amplitude. But the military still wanted their equipment to perform properly and they couldn’t care less what the far field amplitude measured. Did the equipment work was the question.

The problem was that the functioning of the equipment seemed to be independent of the field measurements. To the extent that that was true the specifications were useless. They found out by experiment in the field that the interference was caused by near field effects. Even RFI from distant transmitters was often dependent on local conductors. Like the directors and reflectors of a Yagi affect the antenna gain, the presence of building steel was often more important to interference problems than the far field amplitude of the signal.

When dealing with crosstalk, radiatively coupled power supply interference and local transmitters (in the same cockpit or relay rack, for instance), the far field amplitude was nearly irrelevant. Besides, it rained and snowed out in the open field and, perhaps even worse, in Philadelphia ambient interference hid the signals of interest. These signals were not present on Guam or for that matter in a screen room. (Yes, Virginia, before we had shielded enclosures we had screen rooms.)

So the military developed a specification for measuring noise that did correlate with the operation of the equipment in the presence of noise. When the government types found that they were getting wet when it rained and uncomfortably cold when it snowed, they decided it would be a good idea to figure out a way to do this indoors. If they used an indoor screen room, they didn’t have to sweat separating the signals from the ambient.

We’re talking about a lot of trial and error here and comparing the performance of a lot of different sorts of equipment, but they finally wound up with an RFI specification that worked very well in the sense that equipment that passed the RFI specs acted pretty much as expected when the weapons systems were integrated. That specification was called MIL-E-6181D and later was rewritten and called MIL-STD-461. It didn’t even try to measure far fields. For example, the HF interference was measured with a loop or a whip at a distance of one meter. I know the whip seems ridiculous till you realize that HF receiving equipment involved a short whip antenna mounted on the back deck of a Jeep or the fantail of a ship. This configuration mimics the military systems almost exactly.

After that things went along just fine in the military arena until suddenly along came the TV and its arch enemy the computer. Suddenly bureaucrats got into the EMC business. What few engineers were involved in speck writing were either unaware of what the military had learned or were unwilling to learn from the (gasp, choke) military. So they decided that to quantify the response of electronic equipment to electrical interference they needed to measure the far field amplitude of the interfering signals. Obviously this needed to be done with dipoles in an open field. Worse, because the civilian bureaucrats are always underfunded, even if they had figured out how to write specifications that would get interference problems under control, they did not have the money for the experimentation and trial and error needed to quantify interference in the near field.

So now we’re stuck with rain, snow and intolerable ambients at our measurement facilities or with buying anechoic chambers that cost as much as the Golden Gate Bridge. And we’re still depending on far field amplitudes to properly characterize near field behavior, which for all practical purposes is totally independent of the measured values. But this isn’t the worst of it.

The bureaucrats, realizing that EMI measurements seem to have little to do with the interference, decided that the lawyers were right: “The guy with the best argument is always right.”Why bother with measurements? Let’s skip right to the arguments. Interference can be controlled without all this technical hocus-pocus. The problem will go away if we require the manufacturers to write a declaration that they always do good work. And that’s the truth, Ppzzzttt.