Banana Skins – March 2019 (#153-171)

Items 153-159 below are provided by Keith Armstrong of Cherry Clough Consultants, and come from various speakers at the “War Stories” forum held on the 17th August 2001 at the IEEE’s 2001 International EMC Symposium held in Montreal, Canada.

153.  Magnetic fields from set-top box interfere with TV

An early set-top box was found to interfere with the picture of the TV it was placed on top of, but only after 2 to 5 minutes. The designers had spent months making sure that the emissions from the product were very low. Then they found that the product caused the same interference when no cables were plugged into it, and then even when it was switched off!

It turned out to be due to the magnetic fields from the stereo speakers in the TV. Placing any metal object on top of the TV caused similar interference problems, after a few minutes.

154.  Furniture ESD crashes computers

A particular computer manufacturer had a software lab that checked the compatibility of their new products with a number of applications. With one new product the evaluation systems shut down when the testing staff left the room. The new product at that time consisted of a motherboard and HDD with no enclosure, plus the usual keyboard, monitor and mouse.

It turned out to be caused by the induction field developed by the static charge between the staff and their furniture when they stood up (not a spark, just a changing ‘static’ field). This field interfered with sensitive circuits on the exposed motherboard and caused the shutdown.

155.  EMC test results varied with cloud cover

Testing the shielding effectiveness (SE) of an aircraft on an OATS (Open Area Test Site). The engineers had shielded a wheel well using aluminium foil and a ground strap and were confused by random variations in the SE of up to 20dB. These variations were eventually noticed to correlate with the clouds passing in the sky, but at night there were no variations in the measured SE.

The problem was eventually found to be caused by corrosion between the grounding surfaces. Heating and cooling of the aircraft’s metalwork due to the sunlight and shade caused by the clouds caused the quality of the electrical bond at the corroded grounding surfaces to vary, causing corresponding variations in the SE of the wheel well. It proved possible to simulate the problem by banging the aircraft with a length of 2×4.

156.  Amplifier IC’s lead-frame and bond wires very susceptible around 950MHz

A new non-invasive blood pressure monitor was an electronic version of the old ‘cuff’ method. While testing it for RF immunity it would fail to measure at all between 950 and 1000MHz. It was found that its pressure sensor was outputting misleading signals during the test, despite being a standard part that had been used for many years without problems (or so claimed its salesperson). The pressure sensor had a 6-pin package, with 2 unused pins marked “do not connect”. Copper tape over the transducer and its pins made the problem go away. The problem was then isolated to just three of its pins, one of which was a compensation capacitor for the sensor’s internal amplifier. An engineer working for the Japanese company that made the sensor said that he had seen the problem before in an automotive application.

The N/C pins were connected to the inputs of the internal amplifier and used for performance checks during production testing. The pins were acting as antennas, picking-up the external RF field and injecting it into the internal amplifier at its most sensitive point, where it would be inevitably rectified (demodulated) by the semiconductor junctions in the amplifier’s IC and cause major shifts in DC operating points. Even with these pins cut off from the package the problem still remained – the amplifier was so sensitive that the internal leads and the bond wires to the IC still made effective antenna at 950MHz.

Eventually the sensor was modified by the manufacturer so that it did not have this problem. In the meantime one year’s worth of production of the new product suffered the additional cost of $20 per unit for a shielding can and its fitting.

157.  Tape recorder interferes with aircraft control system

We put a tape recorder into a drone aircraft used for surveillance. When the tape recorder came on, the drone nose-dived. The 10kHz bias oscillator for the tape recorder was exactly the same frequency as was used by the aircraft’s control system, and this caused the problem. The moral of this story is to avoid using standard frequencies.

158.  How many EMC engineers does it take to change a light bulb?  No. 1.

A 20 Amp 20 Volt power supply for a medical xenon lamp had to meet EU emissions standards. The PSU was to be fitted in various boxes, some of which could be plastic, so needed not to have to rely on any shielding from its enclosure. The output of the PSU had a 50 microHenry choke in series, to generate the high voltage which would ‘kick-start’ the discharge in the xenon lamp. Unfortunately, 20mA of RF common-mode current was measured the lamp cable!

After a lot of work on the power supply, to no avail, someone tried a different type of lamp and found that the RF noise was 40dB less. Three other types of lamp were also found to be 40dB less noisy. Then other xenon lamps of the same type as the original noisy one were tried and found to be 40dB quieter too. So it seems that all xenon lamps are not created equal.

By the way, the answer was: four engineers.

159.  How many EMC engineers does it take to change a light bulb?  No. 2.

We had designed a shoe repair kit for use by the US army. The only electrical item in it was a standard domestic-type incandescent filament lamp that ran from 110V 60Hz, so the squaddies could repair their shoes at night. Unfortunately, our regular contact with the military was on an assignment elsewhere and we had to deal with a novice who didn’t understand EMC at all and insisted that we had to fully test the shoe repair kit to MIL-STD-461, the US military’s EMC standard.

He would not be moved by our arguments that the testing was a waste of time. He was following the procedure and it was more than his job’s worth to believe us when we said that it didn’t need testing as it was bound to pass. So we had to do the tests.

Imagine our surprise when our shoe repair kit failed its emissions test by a significant amount at 45MHz! We soon discovered, of course, that this was due to the light bulb. When we contacted Sylvania, its manufacturers, we eventually discovered that around 1% of all incandescent light bulbs (not just Sylvania types) had VHF oscillations, typically occurring between 28 and 45MHz and caused by a ‘monode’ gas plasma oscillator occurring in the very hot gas close to the coiled filament.

The emission frequency could not be predicted because there was no configuration control during manufacture for the aspect of the filament construction that caused the VHF oscillation. As far as we know, this 1% problem with incandescent filament light bulbs is still around.

We don’t remember what the answer was in number of engineers, but it was quite a few.


Items 160 – 169 have been very sent in by David Blake BSc CEng MIEE, Managing Director of Electronic Design Solutions Ltd – a compendium of interference problems and their solutions over 30 years.

160.  Interference with TV sound from unused set-top box

This is a supplement to Banana Skin no 153 in Issue 37. The stereo speakers on our television had been making rude reverberating noises for some time, particularly so when the sound was moderately loud in the bass. On reading in Banana Skin 153 about the metal chassis of an early set top box causing picture interference, I removed the old set top box, which we never use, from our T.V. and the sound is now OK. So it seems that interference from a metal plate above the television set can affect the sound as well as the picture.

161.  Noise from alternator in motor-generator set causes computer malfunctions

In the mid 1960s, a London bank was experiencing malfunctions in its new mainframe computer. On the assumption that the cause was mains borne interference from extraneous sources, a motor alternator set had been installed at the bank to isolate the computer from the mains. That done, the malfunctions continued. So they asked Eric Langham for help.

He found radio frequency ringing on the mains input to the computer, triggered by abrupt changes in alternator volt drop apparently brought about by sudden changes in load current during operation. The addition of R C snubbers across the three phase input to the computer eliminated the RF ringing and cured the problem.

(Told to David by Eric Langham, Chartered Electrical Engineer.)

162.  Mains power noise causes 250hp fan to vary speed

Small amplitude hunting of the speed of the DC thyristor drive of a 250 hp extraction fan was largely unaffected by experimenting with the values of the R C stabilising circuit around its speed error amplifier. Then, quite suddenly, the hunting stopped   coincidentally, it transpired, with the chief electrician switching on the automatic power factor correction system in the electrical substation as the factory load increased.

With hindsight, it is apparent now that cyclic voltage dips in the electric mains originating from commutation in other phase angle controlled thyristor power equipment had been delaying the latching of the thyristors in the fan drive (which were fired by trains of short pulses) until the introduction of the power factor correction capacitors reduced the amplitude of the dips.

(From when David was Senior Systems Engineer for E M Langham, Chartered Electrical Engineer, 1962 79.)

163.  Inadequate contact suppression caused guillotine control

Various guillotines cutting material to length were frequently making double strokes. The source of the trouble was found to be RFI generated when the output contact of the length counter switched the initiating relay in the guillotine control. The contact had originally been suppressed by an R C snubber comprising 100 ohms in series with 0.1 microfarads connected across it inside the counter. In some counters these components had blown up and were open circuit!

Bench testing of the several types of relay used in the guillotines showed that the original capacitor was too small, in some cases, to limit the peak voltage transient to within its own voltage rating. The shortcoming was cured by connecting appropriate R C snubbers in parallel with the coils of the guillotine relays.

(Yet another from when David was with E M Langham, Chartered Electrical Engineer.)

164.  Missing low-cost capacitor causes costly machine shutdowns

In the early 1980s, an electronically controlled flying saw occasionally (perhaps once or twice during an eight hour run) switched itself off whilst cutting slowly moving heavy density material into short lengths. Each event cost some £1000 in lost output whilst the scrap material was cleared and production restored. At the end of an all night vigil, the user’s systems control manager traced the cause to be the false triggering of an integrated circuit monostable in the 24 volt DC control sequencing logic, coincident with a brief period of heavy regeneration of the thyristor controlled main DC drive during its operating cycle.

Bursts of thyristor commutation current generated disturbances which caused radio frequency ringing at the output of an autotransformer which had been installed to match the 430 volt factory supply to the 380 volt rated German electrics.

RFI was breaking through into the 24 volt DC supply because a small capacitor, shown on the circuit diagram, had been omitted by the manufacturer. Critically damping the autotransformer output leakage inductance by R C snubbers and installing the missing capacitor effected a complete cure.

(From when David was Systems Control Manager at the Cape Insulation Rocksil Works in Stirling, 1979 to 1987.)

165.  Furnace heat control interferes with oxygen sensor

Variations in the analogue output signal of a flue gas oxygen monitor operating from a probe in the waste gas duct of a glassmaker’s furnace were traced to common mode interference from the furnace heating phase angle controlled electric boost. The coupling was found to be directly conductive and its effect dependent upon probe temperature.

Operation adequate for using the probe signal to control the combustion air/oil ratio was achieved by re siting the probe further away from the furnace and its boost electrodes in a less hot part of the flue.

(Another from when David was at the Cape Insulation Rocksil Works.)

166.  Defective contactor interfered with thyristor heating controller

The cause of sporadic false firing of its thyristors which, when it occurred, switched the output of a single phase heating controller to full power was under investigation. Then the problem suddenly ceased when the firm’s electrical engineer, acting quite independently, discovered and disabled a defective electrical contactor in the mains power factor control system.

(Another from when David was at the  Cape Insulation Rocksil Works.)

167.  Incorrect installation of screened cables causes problems with machine

The output thyristors in photocell detectors in the rolling unit of a multi section pipe making machine were failing to latch on immediately and delaying the response of the control system. Cure was effected by re routing their screened cables directly back to the control cubicle through dedicated individual steel conduits. Then the same defect was found elsewhere on the machine. Detailed investigation revealed that the manufacturer’s technician had installed and earthed all the screened cabling in a manner contravening his firm’s explicit documented instructions.

So, during a fortnight’s shutdown brought about by the need to carry out other remedial work, the user’s electricians revamped the screened cable terminations and cured all the associated malfunctions.

(Yet another from when David was at the Cape Insulation Rocksil Works.)

168.  Switch interferes with crane control system

The inverter powering the hoist drive of an overhead traveling crane frequently tripped during long travel, coincidentally with the operation of a limit switch which directly switched a tungsten filament indicator bulb in the driver’s cabin. RFI from the switch was found to be injecting a false motor speed feedback pulse train into the control system and transiently grossly mismatching the inverter voltage and frequency output to the needs of the motor.

The malfunction was cured by removing the lamp bulb as the crane driver did not need the indication. (From when David was Managing Director of Pace (Stirling) Ltd   consulting engineers – from 1987 to 1998.)

169.  Walkie talkie interferes with crane hoist’s load cell

When checking the calibration of load cell equipment in the hoist mechanism of an overhead travelling crane, it was found that its electronic signal converter could be made to give any value of analogue output voltage between zero and full scale (10 volts) dependent upon the proximity and orientation of the tester’s walkie talkie radio. No remedial investigation was undertaken because no personnel would be on the crane during normal operations and maintenance staff now knew not to use a walkie talkie there.

Possible effects from other sources of RFI, such as inverter drives, were not looked for because the equipment performed reasonably enough in normal circumstances.

(Another from when David was Managing Director of Pace (Stirling) Ltd.)


170.  Some thoughts on EMC and safety, and the security of bank accounts

Personally, I could list a ton of stuff that would instil fear and loathing amongst the faintest of EMC hearts. Sitting in a jet airliner at the end of the runway readying for take-off and watching the cabin lights dim slightly in sync with the sweep of the main radar dish just a couple of hundred yards away. ESD events in the kitchen area of the airliner causing the phone in the cockpit at the other end of the plane to ring making the pilot pickup to answer. ESD events in the control tower of an airport causing the computer and other essential equipment to crash. Enough spurious radiation events to require laptops and cell phones to be turned off upon takeoff or landing.

Why am I and hundreds of others trusting out lives on something so … sensitive? But we think nothing of dialling up the cell phone inside a car packed with digital controls for things like the brakes, the accelerator, gas control … The automobile industry does its best to test for the severest of electrical events with lightning simulations. But what about internal to the car less than a meter away? And by the way, do they allow cell phones and laptops in that airport control tower? Do they have conductive floors and require people to wear ESD proof shoes? Define safety related issues? Does it necessarily have to do with physical safety? How about the spurious radiation from an ATM being decoded by someone nearby to gain access to your bank account to drain it?

(Posted by Doug McKean on emc-pstc@ieee.org, 2nd Jan 2002)

171.  Some examples of interference in residential environments

EMC? Ha! I’ve replaced the incandescent lamp on my bedside table with a new energy-saving compact fluorescent lamp. With the lamp on, I cannot listen to even the strongest AM radio station on my clock radio (on the same bedside table) due to the lamp interference. This must not be the usage contemplated by EMC requirements.

My TV and stereo are more-or-less integrated (they are in close proximity). On New Year’s Day, I wanted to listen to the radio version of the football game description while watching the TV. With the TV on, I cannot listen to even the strongest AM radio station due to the TV interference. This must not be the usage contemplated by EMC requirements.

I take my Grundig portable radio with me when I travel. Most hotels have sufficient interference sources that I cannot listen to AM radio, and sometimes not even FM radio (with lights and TV off!). This must not be the usage contemplated by EMC requirements. EMC? Ha!

(Posted by Rich Nute on emc-pstc@ieee.org, 3rd Jan 2002. Rich is based in the USA.)

I have it from a message on the RFI@contesting.com list that Phillips bulbs produce less RF noise than others. I can’t vouch for that, however.

(One of the replies to the above, from Cortland Richmond, 3rd Jan 02)


The regular “Banana Skins” column was published in the EMC Journal, starting in January 1998. Alan E. Hutley, a prominent member of the electronics community, distinguished publisher of the EMC Journal, founder of the EMCIA EMC Industry Association and the EMCUK Exhibition & Conference, has graciously given his permission for In Compliance to republish this reader-favorite column.

The Banana Skin columns were compiled by Keith Armstrong, of Cherry Clough Consultants Ltd, from items he found in various publications, and anecdotes and links sent in by the many fans of the column. All of the EMC Journal columns are available at: https://www.emcstandards.co.uk/emi-stories, indexed both by application and type of EM disturbance, and new ones have recently begun being added. Keith has also given his permission for these stories to be shared through In Compliance as a service to the worldwide EMC community.

We are proud to carry on the tradition of sharing Banana Skins for the purpose of promoting education for EMI/EMC engineers.

About The Author

Keith Armstrong

After working as an electronic designer, then project manager and design department manager, Keith started Cherry Clough Consultants in 1990 to help companies reduce financial risks and project timescales through the use of proven good EMC engineering practices. Over the last 20 years, Keith has presented many papers, demonstrations, and training courses on good EMC engineering techniques and on EMC for Functional Safety, worldwide, and also written very many articles on these topics. He chairs the IET’s Working Group on EMC for Functional Safety, and is the UK Government’s appointed expert to the IEC committees working on 61000-1-2 (EMC & Functional Safety), 60601-1-2 (EMC for Medical Devices), and 61000-6-7 (Generic standard on EMC & Functional Safety).

Related Posts

Leave a Reply

Your email address will not be published.

X