Like a ghostly apparition in the dead of night, EMI isn’t normal. But though EMI-related issues are on the increase, there are ways to avoid them in your designs.
As if design engineering weren’t already challenging enough, EMI-related design problems continue to rise, driven by the relentless increase in clock/edge rates and shrinking component geometries.
A sort of ghost-buster who has set up shop in the electromagnetic spectrum, EMI consultant Daryl Gerke says that the problem is compounded by the fact that for many design engineers, EMI lies well outside of their comfort zone. “We in the EMI business are obsessed with the frequency domain rather than the time domain, and we absolutely love decibels.
‘If highway speeds had increased from 60 mph to 600 mph in the past ten years, you can just imagine the carnage,” says Gerke, co-founder of Kimmel Gerke Associates, Ltd., an electrical engineering consulting firm that specializes in electromagnetic compatibility. “Next, let's have ten times as many cars going ten times as fast. Finally, let's make each of the cars one tenth the size. Well, all that has certainly happened with electronics. The EMI design challenge is to stay ahead of the rapid changes, and the problems they can cause.”
A major reason that EMI is so vexing for engineers, Gerke says, is that it is not normal. Radiated emissions problems, he explains, can be caused by microvolts and microamps well below normal circuit levels and stuff you might not even be able to see on an oscilloscope.
ESD problems, on the other hand, he says, may start out as thousands of volts at the source, well above normal circuit levels and also very difficult to measure.
Also problematic is the fact that EMI tends to be about exceptions to the rules--it’s about things not working when you think they should. And expect them to. For example, when does a cable, or even a wire trace, become an antenna?
“That certainly is not noted on any schematics. Unfortunately, in school we are taught how things work, not the many ways in which they don't work. Fortunately, experience helps--one of the few benefits of getting older, I guess,” muses Gerke, whose EMI experience dates back to 1960 when he was a teenage ham radio operator.
His first “EMI event” involved wiping out the family television set. No great surprise there, as his ham antenna was less than four feet from the TV antenna. He suspects he ended up in the EMI business as some sort of penance for his early escapades.
So what are the key threats for EMI engineers should be on the lookout for and how can they guard against them? In the training classes his company puts on for engineers, Gerke and his partner William Kimmel discuss the top five:
Emissions (both conducted and radiated). The primary victims are licensed users of the RF spectrum (radio, television, navigation systems, etc.) The primary sources are repetitive signals, such as clocks. Secondary sources include switch mode power supplies. Because digital signals are not sine waves, they create harmonics, which can directly interfere with intended RF communications.
- RFI (Radio Frequency Interference). The primary victims are analog circuits and power circuits. The primary sources are portable RF transmitters (hand held radios, cell phones, wireless devices). The primary failure mechanism is upset due to rectification. Failures are predominately upsets; damage is rare.
- ESD (Electrostatic Discharge). The primary victims are digital reset, control, and I/O circuits. The primary sources are human generated ESD, followed by machine generated ESD (belts, printing presses, etc.). Failures include both damage (I/O) and upset (digital circuits.)
- Power Disturbances. The primary victims are power supplies, followed by digital reset and control circuits. The primary EMI sources are lightning and power transients (low frequency/high energy), followed by arcing and sparking (high frequency/low energy) such as the Electrical Fast Transient, or EFT. Failures include both damage (power interface) and upset (digital circuits).
- Self Compatibility. Often these are internal problems due to mixed technologies. For example, digital circuits are often upset by transients from power circuits (power supplies, motors, contactors, etc.). Low level analog circuits are often affected by digital circuits. A rapidly increasing problem is on board radio receivers (wireless, cell phone, and GPS) being jammed by digital circuits.
Gerke stresses that to avoid trouble with their designs, engineers need to pay attention to the details. “We like to say that 95% of the problems are caused by 5% of the circuits.” He exhorts engineers to start by identifying the critical circuits--clocks, resets, power regulators, and I/O--and make sure they are adequately filtered and decoupled. “Keep the leads short!” he warns. “Next, check out the placement and routing for those critical circuits.”
At the board level, he advocates a quick EMI review right after the boards are routed. And at the box level, he recommends EMI reviews of the packaging, grounding, and interfaces. Finally, he recommends that engineers consider "precompliance” testing during the design phase. Though it may seem like just so much drudgework, he says that design engineers shouldn’t wait until the end to do the EMI testing.
Those who wait run the risk of, well, a most unwelcome visitor.