I’ve been following the excellent series on EMC by the gurus at Kimmel-Gerke Associates. Every engineer should have their materials close to hand when designing any kind of electronic equipment.
And the engineer should obey their advice and – here’s the hard part – convince management to obey their advice.
A couple of years ago, I began to experience significant interference to my ham equipment, mostly on the 80- and 160-meter bands. Signals on these bands from long distances are generally pretty weak, and it doesn’t take much interference to wipe them out. And this interference was huge…obviously originating close by, and with the typical signature of a noisy switching power supply: wideband crud with peaks every 60 kHz or so. After some trial-and-error, I tracked it down to the new PC my wife was using, and specifically to the USB-powered speakers. Unplug speakers, no interference; plug back in, lots of noise. I figured that unplugging the speakers every time I wanted to operate on the low bands was inconvenient, so I decided to take the system approach and ask my wife to turn off her PC when I wanted to operate.
Lo and behold, shutting the PC off with the speakers plugged in did NOT make the interference go away! It seems that this PC (and maybe others do the same) keeps the USB supply alive when everything else is shut down…with the PC “off”, I could make the interference go away by unplugging the speakers OR unplugging the PC power cord.
When my wife wasn’t looking, I took the machine apart, and examined the power supply. Right behind the IEC connector was a big gaping hole, about the same size as a packaged Corcom or similar line filter or a couple of chokes and caps. I installed a filter I bought for a buck at a local surplus place in a small external box with a short power cord and the interference problem went away completely.
I figure that the manufacturer of the power supply submitted one unit to a testing lab with a filter installed, then after getting certified as meeting the specs for conducted and radiated noise, modified the BOM to remove the filter. Yeah, it saved a buck or two of cost, and most customers would never know the difference. Does your company do this?
This power supply manufacturer broke the rules, cheated, and caused the kind of interference to a licensed service (in this case, the Amateur Service) that Class B computing devices are not supposed to cause.
I can still hear lots of noise sources in my receiver, but some of them are likely in neighbors’ houses and I figure it’s easier to live with the noises (as long as they are small) than approach neighbors who have no idea how illegal their PC/TV/modem/phone-charger/electric blanket really is.
FCC regulations would not prevent this supplier from removing the filter when shipping its products...but it was probably low cost item...sometimes you get what you paid for, unfortunately I see no solution for this problem...Kris
Quite possible. Even likely.
I was told by a local _US electrician_ that US companies routinely buy all the necessary regulatory stickers they need including those that are nominally "serialized" by standards bodies and agencies and slap them on products without any testing. He apparently knew directly of a dozen or so in Silicon Valley.
Of course when you do actually though the real certification, you have to buy the same raw stickers too.
Printing stickers is not rocket science and you can't do the "Soviet method" of controlling them like printing presses and copy machines without destroying what little innovation still goes on in the US.
Trouble is brewing in Britain where BT (our Ma Bell) has been selling Comtrend Power-Line Transmission (PLT) home systems that are 30dB over the permissible limits. OFCOM (=your FCC) have been refusing to acknowledge the problem and when called upon to intervene, our public representatives (MPs) just quote OFCOM's press releases claiming there is no problem.
Effects like blanking of car remotes, garage openers and other 2.4GHz products are increasingly widespread, not all due to PLT but emissions 1000x over the limit is like fouling the well in your village, nothing to be proud of, surely!
Check out www.theemcjournal.com if interested...
I had a similar problem here but mine was a electric lawn mowers charger, so much for green technology huh, It makes me want to go back to gas mowers.
This charger knocked out HF and VHF receivers from 3 Mhz to 138 MHz and even created a nice buzz in the TV audio, it was detectable and still killing receivers at over a 1/8 mile from the source.
Within 1 block the wideband noise floor increased over 40 dB.
Does the FCC really care about issues like this anymore what with the selling of our GPS frequencies to lightsquared all in the name of creating jobs which we all know will only create them overseas ?
Nice EMI troubleshooting story! If your speculation is proven to be correct, then the manufacturer must be penalized. My doubt is whether the tests required for certification was performed at all. Now a days any electronic equipment we buy from anywhere, however small it might be, we see those fancy markings indicating compliance to various standards. Are all of them sincerely getting the tests performed before those markings are stickered on the product?
At one company the management types had a wonderful idea - buy product from an established manufacturer and under license re-label and re-sell under their own name. (Yeah, I know, but these were MBAs).
The product was an Ethernet hub for UTP (unshielded twisted pair) cable. I asked for the EMI test reports.
Was very interesting to see that the radiated emissions testing had been done using SHIELDED twisted pair cable...
I remember a software department director who at first would not believe me when I told him the product could not be sold until it had been tested and proven to comply with radiated and conducted regulatory limits. Then he learned just enough to be dangerous - he asked what was the clock frequency, I told him 20 MHz, he said no problem, FCC doesn't care unless the clock is above 30 MHz. (Sigh....)
Another company called on me to solve an emissions problem with a product they had modified, sold, then finally got around to the emissions testing and discovered extreme levels. After I cleaned up the problems with highly visible hardware fixes, the head honcho of marketing asked how they were going to explain the recall and rework to their customers. My answer was very straightforward and succinct - "That's not MY problem."
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.