EMI Causes Wireless Keyboard Cursor to Move Mysteriously

DN Staff

July 31, 2009

3 Min Read
EMI Causes Wireless Keyboard Cursor to Move Mysteriously

Responding to our blog post on EMI causing computers to spontaneously reboot, Gupta Harshwardhan (a dyed-in-the-wool mechanical engineer) wrote in wondering whether the HP wireless keyboard+mouse on his office desktop was suffering from the same problem. Every now and then, he said, the cursor would creep towards the left as if directed by a spirit hand.

(Okay, he didn’t add that last bit in — we’re taking a little editorial license because we like ghost stories.)

Though the clues were scant, we asked Daryl Gerke, an EMI expert who is co-owner of the consulting firm Kimmel Gerke Associates, to speculate on the case:

“EMI is always a possibility, but there may be many other causes such as a failing component or thermal issues. The symptoms, however, do match with an RF source causing upsets. If EMI related, I doubt it is due to other typical EMI sources, such as power transients or electrostatic discharge. These usually result in resets, lockups, memory corruption, etc.

One possible scenario is a nearby cell phone, making supervisory transmissions to the local cell tower. I’ve seen that problem in my own office. If the cell phone is next to my “expensive” $8 speakers, I can hear a ticking sound for a few seconds when the cell phone transmits. The insidious thing is that you don’t need to be talking on the cell phone — the supervisory transmissions are short, but they occur on a regular basis (3-30 minutes typical.)

Another similar source would be a mobile RF transmitter, such as a VHF/UHF “walkie-talkie.” I’ve seen that problem in both hospitals (fireman, paramedics) and industrial facilities (security guards.)

Another EMI possibility is electromagnetic radiation from transients, rather than regular radio transmitters. This can be a bit of a sleeper, since it is often unexpected. Nevertheless, we’ve seen upsets to electronic equipment 20-30 feet away from human electrostatic discharge (or during ESD testing.) In one case, a variation on this theme was resets due to an X-ray machine one floor below the vulnerable system. When the X-ray fires, it can also cause a big electromagnetic transient.

Just two weeks ago, I chased down a problem due to radiated transients from the inverters in a large UPS system. Whenever the vulnerable equipment was brought close to the UPS, it would beep and go wild. The good news there was that a 0.1 uF capacitor on a critical line fixed the problem. Lest I make it sound too simple, diagnosing the problem and figuring out just where to place the critical cap took some effort. Actually, we put in a whole bunch of filtering, and then backed stuff out. When the dust cleared, we found only the one cap was actually needed. Ah, if only they all ended up so simple…

If all else fails, we can always blame the software. (I’m a hardware guy, or course…)

Follow-up: We emailed Gerke’s response to Harshwardhan. He responded: “A hint is enough for a good engineer, I suppose. The mouse + keyboard transceiver was sitting next to my Wi-Fi transceiver. I have moved it away. Seems okay now.”

Case closed.

Sign up for the Design News Daily newsletter.

You May Also Like