Cherry Clough House
Denshaw OL3 5UE UK
tel: +44 (0)1457 871 605
fax +44 (0)1457 820 145
"Banana Skins" Compendium
thLast updated 7 October 2005
This is a compendium of all the ―Banana Skins‖ that have been published to date in the EMC & Compliance Journal, published by Nutwood UK Ltd and available on-line at: http://www.compliance-club.com/.
Banana Skin items briefly describe reports of electrical/electromagnetic interference. Some are personal anecdotes, some come from research, and some are extracted from official documents and reports. Some of these interference incidents had harmless or even amusing results, some lost companies significant amounts of time and/or money, even to the point of bankruptcy, and some resulted (or could have resulted) in injury or death.
I hope that these anecdotes help you to identify interference problems that could happen to your designs, so that you deal with them in advance as part of the normal design/development procedure and don‘t have the embarrassment and cost of trying to correct poor EMC design after products have been shipped or systems installed. Designing EMC in from the first saves time and cost, reduces customer returns and warranty costs, and reduces the possibility of liability claims.
If you have any suitable anecdotes or know of any relevant research or reports, please tell me about them so they can be included in future Banana Skins (anonymously, if preferred).
1) To cope with increased North Sea oil production, two new pumping stations with 6 MW adjustable
speed induction motor drives were built and installed in Scotland, one in Netherly and one in Balbeggie.
Soon after commissioning the local power utility and the telephone company received a flood of
complaints. Geographically the complaints came from concentrated pockets spread over an area up to
12.5 miles away from the 33 kV overhead supply lines feeding the drives. A payphone over 4 miles
away from the power line was noisy enough to be almost unusable, whereas just across the street a
householder's telephone was relatively unaffected. Other symptoms included loss of synchronisation
on TV sets (rolling pictures) and ringing on the supply to fluorescent lighting circuits.
Although the drives had been designed to, and met the supply industry's G5/3 harmonic limits, the
problems turned out to be with higher order harmonics than it covered, up to the 100th in fact (i.e. 5
kHz). The problem became a public relations nightmare for all involved, and culminated in questions
being raised at Government level. Remedial EMC work was urgently required and was in fact
accomplished, although under extreme difficulties because the cost of any downtime of the oil pumping
stations was so high.
(Taken from: ―Harmonic filtering of large induction motor variable frequency drives‖ by M J V
Wimshurst of Hill Graham Controls, High Wycombe, U.K., and Allan Ludbrook of Ludbrook and
Associates, Ontario, Canada. Presented at the 7th International Conference on Harmonics and Quality
of Power (IEEE) at Las Vegas, October 16-18, 1996, pages 354-359 in the Proceedings. Also
presented at the "Sixth International Conference on Power Electronics and Variable Speed Drives",
Nottingham, UK, 23-25 September 1996, IEE Conference Publication No 429, pp 24-29,
2) A (CE marked) portable PC carried up the stairs in a domestic household whilst operating, reliably
caused the "power shower" in the bathroom to turn itself off if it was in use at the time. (Personal
communication in 1997)
3) Medical technicians taking a heart-attack victim to the hospital in 1992 attached her to a
monitor/defibrillator. Unfortunately, the heart machine shut down every time the technicians turned on
their radio transmitter to ask for advice, and as a result the woman died. Analysis showed that the
monitor unit had been exposed to exceptionally high fields because the ambulance roof had been
changed from metal to fibreglass and fitted with a long-range radio antenna. The reduced shielding
from the vehicle combined with the strong radiated signal proved to be too much for the equipment. ―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 1 of 90
(An article in the Wall Street Journal reported in Compliance Engineering Magazine's European edition
4) Computers used in a room close to a door fitted with a high-technology (magnetic) cat flap caused the
latches on the cat flaps to rattle continuously whenever Windows was loaded or a Windows application
run. (From the New Scientist magazine, 7th May 1997)
5) The Langley (USA) Air Force Base Rescue Co-ordination Centre reported that its search and rescue
satellite was receiving interference on its 121.5 and 243 MHz distress frequencies. The area over
which interference was a problem was around 8 square miles, which was significant because normal
emergency transmitters on these frequencies can only be detected at ground level for about one mile.
The problem was eventually traced to poor connections on an overhead power line. (From an FCC
Field Operations Bureau news release, 1994.)
6) An advertisement for engineers for "The HERO Project" quoted Rear Admiral Roland T Guilbalt,
Deputy Director, Electronic Warfare Division US Navy as saying that both Desert Shield and Desert
Storm suffered from serious and significant EMI problems. We have no more information on this at
present, but presume it was due to the very heavy use of high-tech civilian equipment used for the first
time in a military situation. (From EMC Technology magazine, 1993.)
7) Excessive mains harmonics in the London area, due mainly to the rapidly increasing use of personal
computers, are causing overheating problem in AC power cables (including those that run under the
Thames). In the offices where the computers are, it is increasingly common for the power-factor
correction capacitors normally fitted to fluorescent lamps to blow (the electricians usually just remove
the blown capacitors). Damaged and overheated neutrals, and damaged electrical switchgear is
increasingly seen as a result of harmonic mains pollution. In the US, fire insurance companies are
being urged not to take on any new policies unless they have had the size of the neutral cables in the
company concerned checked for their adequacy for the heating effects of harmonic currents. (Personal
communications, January 1998)
8) Hartman Products of Los Angeles, California, has agreed to pay a civil penalty of $60,000 to settle
allegations that it failed to file a report regarding a defect in the 1992 Hartman Pro1600 hair dryer. The
CPSC (a US consumer safety agency) believes that these hair dryers can turn themselves on even
when the on/off switch is in the "off" position. While the dryers' heaters start, their fans do not,
potentially causing internal components to overheat and cause fires. (Compliance Engineering
9) The AA and RAC estimate that around 9000 breakdowns they attended in 1996 were the result of
remote key fobs being blocked by RFI. An AA spokesman said: "The number of cars being produced
with radio-activated keys is standard now. If we're getting 9000 now, what will the problem be like later
on?". (Electronics Times 13th Oct. 1997)
10) We recently bought what looked like a fine new idea for an executive toy. It consisted of a very strong
magnetic base with lots of ball bearings attracted to it, which it was possible to form into beautiful
sculptures. What we did not realise at the time is that magnets and office desks are not cheerful
companions. But we soon found this out when the discs with our accounts on them were mysteriously
wiped, and the monitor screen went all the colours of the rainbow. It is now only possible to use our
office desk toy when not at our desks, and well away from the office. (Letter from Michael Fell in 29
November 97 issue of New Scientist.)
11) Wheelchairs have come in for special scrutiny by the FDA (the US Food and Drug Agency). A few
months ago, the agency ordered makers of wheelchairs to shield them and to educate users about the
potential hazards of interference. The FDA acted after receiving "many reports of erratic unintentional
powered wheelchair movements." These included sudden starts that caused wheelchairs to drive off
curbs and piers when nearby police, fire or CB transmitters were activated. Miraculously, no fatal
injuries have been reported. (But broken limbs have occurred as a result of such interference - editor.)
(Compliance Engineering - European Edition September/October 1994)
12) Around 1990 Alan Little leased a derelict arch under the railway line in Camberwell from British Rail.
He borrowed money to convert it into a two-level mix of recording and rehearsal studios. The total cost
was pushing ?50,000. Up until November 1991 it was popular with up-and-coming bands needing
somewhere to rehearse and record. Then, one fateful Saturday morning, with three bands booked for
the morning and three for the afternoon, disaster struck. All the studio equipment, and the bands'
amplifiers, started warbling. The bands and studio crew thought at first that they had an equipment
fault. Then other studios in other railway arches in the area began phoning each other. They all had ―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 2 of 90
the same problem. Alan Little phoned British Rail and on the Monday morning a BR engineer came round, listened and said the cause was a new signalling system installed by BR.
BR controls its track lights by feeding electric current through its rails. When a train runs over the rails it provides a short-circuit between them, triggering a red light behind the train. Recently BR has begun changing to the use of alternating current. The long rails act as a highly efficient aerial, radiating a powerful AC magnetic field (this was actually around 1 Amp/metre over much of the studio - editor).
The AC is at audio frequency, using tones of between 1 kHz and 4 kHz. The tones are complex warbles, to safeguard the system from outside interference.
The effect was heard through the mixing desk, with pick-up from mains and connecting leads. It was even heard through unpowered loudspeakers (even when they were disconnected from their cables
and their terminals shorted - editor). It was worst when an electric guitar is plugged into an amplifier. Guitar pick-ups are designed to convert their magnetic fields, modulated by the movement of the steel guitar strings, into sound. They cannot distinguish between magnetic fields from a BR signalling system and those from vibrating strings. (Extracted from an article by Barry Fox in Studio Sound
Magazine, June 1992)
13) It was reported in the Sunday Times (15/2/98) and New Scientist (7/3/98) that Sabena Belgian World Airlines had installed magnetic tray tables in its new fleet of A340 Airbuses, to prevent the nuisance of rattling trays on their flights, but that these tray tables were apt to cause loss of data on PC hard disc drives. New Scientist of 28 March reported that the story was untrue, but that tables of this sort had been discovered on a train from Frankfurt to Berlin. The conclusion seems to be that if you intend to use your PC in any kind of vehicle you should always carry a (steel!) paper clip and use it to check for magnetised tables.
14) Quote from an article in the IEE's Control and Computing Journal, April 1998 (page 52): "High intensity radiated fields (HIRF) guns and electromagnetic pulse transformer (EMPT) bombs are already easy to build from off-the-shelf components. The effects of even hand-built HIRF or EMPT weapons can damage microprocessors at ranges of hundreds of metres. Possibly, in a few years, a van equipped with suitable electronics could cruise down Wall Street (or through Canary Wharf - ed.) and disrupt the
information processing capability of thousands of computers without being detected by the local police."
15) More on radio activated key lock-out problems (banana skin No. 9 from the April issue of EMCJ): a quote from an Electronic Times article (29/9/97) says: "Most radio activated key-entry systems have a manual override. Unlocking the door can be as simple as inserting the mechanical key into the lock and trying the lock according to the instructions printed in the car manual." The trouble with this advice is that the manual will usually be locked inside the car (or are we supposed to carry it around with us at all times?).
16) More medical incidents:
The magnetic field caused by ground currents in a water pipe system made it impossible to use sensitive electronic instruments in part of a hospital.
A patient-coupled infusion pump was damaged by an electrostatic discharge, but thankfully the alarm system was not affected and a nurse was alerted.
An operation using a plastic welding machine caused interference with a patient monitoring and control system, causing failure to detect that the circulation had stopped in a patient's arm, which later had to be amputated.
(Taken from Compliance Engineering European Edition March/April 1998)
17) At 1:23 pm on Sunday, 24th July 1994 there was an explosion at the Texaco Refinery, Milford Haven. Its force was equivalent to 4 tonnes of high explosive and it started fires that took over two days to put out. Shops in Milford Haven 3km away had their windows blown in. 26 people sustained minor injuries, and the fact that it was Sunday lunchtime and the site was only partially occupied meant it could have been very much worse. Damage to the plant was substantial. Rebuilding costs were estimated at ?48 million. There was also a severe loss of production from the plant – enough to significantly affect UK
refining capacity. The incident was initiated by an electrical storm between 7:49 and 8:30 am on the Sunday morning which caused a variety of electrical and other disturbances across the whole site. (IEE Computing and Control Engineering Journal April 1998 pp 57 - 60). There is an HSE report on
this incident: "The explosions and fires at Texaco Refinery, Milford Haven, 24th July 1994" HSE Books,
―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 3 of 90
Comments: The author has not read the HSE report, but understands from private conversations with HSE experts that the large explosion was caused by the electrical storm giving rise to power surges which tripped out a number of pump motors whilst leaving others running. As there was a great deal of panic and confusion due to the information overload caused by the numerous small fires and equipment outages from the time of the storm, it was not noticed that flammable substances which should have been flared off were accumulating in pipework and vessels. After five hours something ignited the total accumulation, resulting in the large explosion.
The general incidence of surges in the UK's AC power distribution network is quite low, and this often leads people to believe that qualifying the power surge immunity of their products, systems, or installations is not important. This belief is often supported by the observation that neither of the generic immunity EMC standards included surge testing in their normative sections. But it only takes a single incident such as the above in the lifetime of even a very large plant to make an excellent economic case for a proper preventative strategy. Suitable basic test standards include IEC 61000-4-5 or IEC 61000-4-12 (ring wave), both of which are intended to simulate the indirect effects of electrical storms on power networks.
Engineers are always under pressure to save costs, and the costs of preventative measures are easy to quantify. However, many engineers are uncomfortable with estimating the risks of infrequent and unpredictable events such as thunderstorms so do not effectively communicate the actual risk/cost and safety implications to their managers.
As someone said recently: Doctors kill people in ones, but engineers do it in hundreds. Careers and personal liability are also at stake here too, so it is always best to make an informed cost/risk case and get a written decision from management. There is no shortage of advice and assistance on this sort of thing – sources include:
The Institute of Risk Management: phone 0171 709 9808, fax 0171 709 0716, or visit IRMG@aol.com The Hazards Forum; phone 0171 665 2158, fax 0171 233 1806, Email: email@example.com, or visit www.ice.org.uk
The Safety and Reliability Society: phone 0161 228 7824, fax 0161 236 6977, Email firstname.lastname@example.org, or visit www.sars.u-net.com
The British Safety Council: phone 0181 741 1231, fax: 0181 741 0835, Email
email@example.com, or visit www.britishsafetycouncil.co.uk
Health and Safety Executive: Infoline: 0541 545 500, fax 0114 289 2333, or HSE Books: phone 01787 88 11 65, fax 01787 313 995, or visit www.open.gov.uk/hse/hsehome.htm
The IEE Library: phone 0171 344 5449, fax 0171 497 3557, Email firstname.lastname@example.org, or visit www.iee.org.uk/Library/libgen.htm
IEE Publication Sales Helpline: phone 01438 76 7328, fax 01438 742 792, Email email@example.com, or visit www.iee.org.uk/pub/
The Engineering Council: phone 0171 240 7891, fax: 0171 240 7517, Email: firstname.lastname@example.org, or visit www.engc.org.uk
18) A very powerful (?8,000 Amps) DC drive was recently purchased and installed in an industrial plant. It was contractually agreed that it would meet and be declared compliant to the EMC Directive. A control room was also required (like most modern control rooms it was full of PCs and CRT-based VDUs) and the drive manufacturer said that it could be installed near their drive cabinets. When the drive was operated the images on the VDUs were squashed into 50% or less of the screen width. It was possible to tell the direction and loading of the drive directly from the movement of the VDU images, which of course were completely unreadable. The magnetic fields caused by the drive were of the order of 235;T, and most CRT-based VDUs show image movement at greater than 1;T (1;T is approximately
equal to 0.8Amp/metre and to 10 milligauss).
The drive manufacturer claimed that his drive did meet the EMC Directive despite the fact that it caused interference with the control room VDUs. What they in fact meant was that it met the industrial generic standards, which do not include any limits for low frequency magnetic field emissions. They forgot that their EMC Declaration of Conformity binds them to not causing interference of any kind, and
that compliance with a harmonised standard only gives a presumption of conformity.
―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 4 of 90
The situation has been remedied by the use of LCD screens, which have only recently been available with a specification suitable for the SCADA system that was used. "Dog kennel" magnetic shields and active field cancellation devices were also investigated. The delay in the use of the control room was several months, and this had an impact on productivity far beyond the cost of the remedial measures.
19) We've learned to live with the condition that if we get up in the middle of the night to use the bathroom, when we turn the light on the fan timer starts. The fan will keep running for twenty minutes, and when it turns off it causes interference that turns on the outside security light (infra-red triggered 500W halogen) which then runs for its time period (15 minutes) whilst shining through the bedroom window. Now you'll have difficulty believing this bit... Monty Python eat your heart out... before the 500W halogen lamp we had a high pressure sodium lamp with an inductive ballast. When this switched off it would cause interference which would sometimes start the bedside radio. So the scenario was this... Get up at 2:00 am, go to bathroom, turn on light, turn off light, go back to bed, and after twenty minutes a bright light would shine through the window and wake you up. If you slept through that (or went back to sleep), fifteen minutes later when the light switched off the radio would start, and then you would wake again.
The moral of this story? If you have bad EMC immunity make sure you use the bathroom before you go to sleep. (From Chris Dupres via email@example.com, 8/7/98)
20) While taking classes in the early 80's, my prof got involved with a terrible incident down in New Jersey. Seems a hospital had a high incidence of infant deaths in the intensive care section of the maternity ward. Late at night, the alarms on the babies' monitors would go off for no apparent reason. Annoyed, the nurses would turn them off and do the rounds on foot.
After some preliminary investigations, my prof found out that a nearby TV transmitter was allowed by their FCC license to increase their output wattage by some enormous amount after say midnight but had to reduce it prior to 6am, or some such arrangement. The cable interconnecting the nurses station to the various baby monitors sang like a lark with these frequencies and set off alarms with the induced voltages.
Not sure now of all the specifics except what I have related above nor the name of the hospital, but they lost something like 6 kids before fixing it. (From Doug Mckean, via firstname.lastname@example.org, 29/7/98)
21) Undervoltage AC supplies (brown-outs) are common in underdeveloped countries, or where the AC supply network is incorrectly configured. Parts of Spain are known to experience around 150Vac for lengthy periods during the day, apparently due to network loading, despite an officially-specified mains supply of 230Vac ?6%. I had never experienced a brown-out in the UK, except maybe for a second or so prior to a complete supply failure during a thunderstorm, and I used to think that it must not be possible because of the way the supply network is operated here.
On Sunday 26/7/98 around 5pm in Denshaw village, Lancashire, U.K., the supply dropped to around 140VacRMS (40% below nominal), and stayed at that level for about three hours before shutting down completely as the engineers arrived to fix the problem. We switched off our fridge and other motor-driven appliances, mainly because they were making very strange noises. Our computers kept running, but the CRT screens blanked, making us concerned about what was happening to our data, so we switched them all off as well.
I am aware of electronic control equipment that can misoperate when operated considerably outside its specified AC supply range, and also understand that undervoltages can damage coils and motors. Apparently the motors can stall due to the low voltage, so they don't generate back-emf, so they draw excessive currents and overheat, damaging their insulation and suffering premature failure (if not electric shocks and fires).
Manufacturers of products for the developed world, and their safety test laboratories, usually do not test at supply voltages outside ?10% (sometimes ?15%). Until Sunday 26th July I had not thought this important. So far we have not discovered any damage to appliances or to data. (From Keith Armstrong,
Cherry Clough Consultants)
22) There was film footage on TV some years ago of a British in-flight re-fuelling exercise where the tanker aircraft was hit by lightning, but there was no on-going discharge downwards, the implication being that the plane was left charged to 100kV - 1MV or whatever. A few seconds later great balls of glowing gas came off the back of the fuselage and wing tips (where the fuel hose was) into the slip stream,
―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 5 of 90
presumably taking away a whole load of surplus electrons, or holes - whatever it was - as "ball lightning". (from Chris Dupres via emc-pstc)
Incidentally, New Scientist magazine recently reported that a theoretical basis for ball lightning may have been found.
23) Helicopter blades and bodies tribocharge as they swish through the air, and they don't have a convenient green/yellow wire handy. There are reports of an oil rig computer system crashing whenever a helicopter landed, due to its sudden electrical discharge into the metal decking, and there is a film which shows a crew-member getting an awful shock when he reached up to touch the skids of a hovering helicopter. Here is another item from Chris Dupres via emc-pstc:
A US Coastguard Chief related to me that he had seen "arcs as bright as a welding stick" when an emergency pump was delivered to the deck of a freighter one night. He also told me that the Coast Guards' standards practice calls for NEVER lowering a flotation ring or sling to a person in the water; the person will automatically reach for the line! Instead, they dip the line into the water and drag it to the person.
24) Crane incidents.....
A new CNC machine being installed in a factory had a spindle controller which was a small invertor drive in a plastic case. When the spindle was first operated emissions from the invertor caused the overhead travelling crane to start up and drag its chains down the length of the factory. Luckily, the 18 ton casting the chains had been attached to had just that minute been released. (from Phil Hampton)
There was the famous case reported by the DTI in the early days of their EMC Awareness Campaign of the guy who was standing under his crane's load using his radio-control pendant when interference caused it to release its load, crushing him to death. Many recent crane incidents are due to the use of radio control, especially where crane radio-control systems share the same frequency bands as amateur radio and/or car radio-keyfobs. The soon-to-be-introduced TETRA system also shares some of these bands, and use 25W transmitters – so expect more wild cranes.
Just to prove that modern technology can't teach old technology any interference lessons, I once worked for a company that I was told had made the controls and drives for the first large-scale hovercraft testing tank in the late 1960's. It was in effect a sophisticated travelling overhead crane, which ran a gantry along overhead rails and towed a hovercraft shape along a large pool of water in an even larger building. In those days they used resistor-transistor logic which ran on a 40V rail to provide noise immunity. During commissioning the machine suddenly started up by itself and proceeded towards the far end of the pool – it had been set off by "some sort of mains transient". All the personnel on the site were standing by the access ladder to its gantry, but the only emergency stop button was on the gantry – but it and its ladder were moving just faster than running speed and they couldn't get to it. Since it was not operating according to its (hard-wired) programming, the crane ignored its limit switches and crashed clear through the end wall of the building. Luckily nobody was hurt. The next version had E-stops all around the building. (From Keith Armstrong)
26) A foetal heart monitor in a clinic in the UK in June 1998 picked up a cellphone conversation from elsewhere on the premises quite clearly. The visual output of the monitor was unaffected, but the staff tend to use the audio output, and the cellphone conversation was so loud that it swamped the heart signal they were listening for. It must have been an analogue cellphone, and it must have been getting in via the audio stages, or else the visual output would have been distorted. Even slow opamps will demodulate 900MHz signals (as hearing aid wearers are only too aware!). In common with many healthcare promises, the use of cellphones on the premises was banned, but you can't rely on people to read or follow signs. (from Ian Ball)
27) A 40kW RF welder (a dielectric welder for plastic materials) in use in a factory caused a mattress in a bed manufacturer's factory 60 yards away to catch alight. The bed springs must have just been the right length to make an efficient antenna at the frequency the welder was using.
28) Digital TV is more likely not to deliver a programme to the viewer than the analogue TV services it replaces. It appears that this newer technology is less robust, and that its users will on average suffer a higher loss of service than they may have become used to.
Broadcast digital TV, which can be picked up with existing TV antennae, has a sharper cut-off in performance as signal strength declines. R.S.Sandell, a Fellow of the IEE writing in IEE Review November 98, is concerned that: ―Whereas analogue viewers can live reluctantly with a picture that has to be viewed through varying angles of ‗venetian blind‘ and alternating densities of ‗boiling
―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 6 of 90
porridge‘, they can still follow the programme plot for most of the time. This dubious advantage may
not be available for some members of the digital generation, who will be confronted by a blank screen.
In particularly unfortunate reception location this condition may come and go with time as field
strengths vacillate‖. Viewers using indoor aerials (the TV with the rabbit ear antenna in the kid‘s
bedroom?) may find they need to invest in new external aerials or aerial amplifiers and splitters.
Satellite-delivered digital TV is very susceptible to lightning storms, both at the uplink and downlink
ends. This leads to the odd situation, when watching digital satellite TV in South Africa during very fine
clear weather that thunderstorms near the uplink in Europe can cause all 100 channels (or however
many there are) to disappear all together for periods of several minutes. This phenomenon was well
understood by the satellite broadcasters, who broadcast a little presentation on this topic every now
and again. (From Keith Armstrong of Cherry Clough Consultants)
29) GPS is another example of an advanced technology that everyone wants to use, but which has
important susceptibility problems. The signals from the GPS satellites are very weak, so the receivers
have to be correspondingly sensitive, which means they are readily swamped by interference from
industrial sites. Even though they are at microwave frequencies, interference with satellite
communications caused by such commonplace things as poor quality power line connections has
been observed several times. Added to this, the need to ―see‖ several satellites at once means that
GPS is unreliable in the urban canyons of cities.
thI was intrigued to see two items on GPS in the New Scientist magazine dated 10 January 1998. The
first was an article about the concern of the US military about a Russian GPS jammer. With only 4W of
power (about the same as a hand-held CB or security guard walkie-talkie) this device is claimed to
prevent GPS systems from working over a 200km radius (yes, 200 kilometres!). Apparently any
competent electronic engineer could build such devices from readily available components. The
second item in the same issue was an advertisement from BT for their MoBIC mobility system for blind
people. This uses a computerised map, speech simulator, and GPS to guide blind people to their
destination. Quote: ―Getting around the shops is much easier since I started using the US Military‘s
Satellite Guidance System.‖
Designers building GPS into their products, especially where these are used for critical purposes,
might like to consider the lack of robustness and ease of jamming of this system. I have visions of
hordes of planes, cars, and pedestrians all milling around a factory until a certain machine is switched
off, because their satellite navigation systems are blocked by its microwave noise. (From Keith
Armstrong of Cherry Clough Consultants)
30) Interference in the Amateur Radio 144MHz band traced to an ultrasonic rodent repeller. (Brad
Thomson, Editor of Test and Measurement World, Feb 95)
31) The operation of a domestic appliance used to reset a surround-sound processor, causing a ? second
gap in the audio (From Neil Gardner, Plantronics, August 98)
32) An advanced hi-fi system would change input selection due to taxicab radio transmitters when they
called at a public house 100yards away. (From a Technical Director of Lumonics, 1996).
33) A Tissot Two-Timer digital/analogue wrist-watch went into time-travel mode (about x 60) whenever a
particular Motorola Micro-Tac portable phone nearby had someone actually speaking into the
mouthpiece. (Chris Duprés, 7/7/98. It was his watch!)
34) With over 18 years in EMC I could go on listing interference incidents for a long time. Some examples
this year already: My computer (FCC Class B) interferes with my cordless phones, to some degree on
all 10 channels. My fax machine (FCC Class B) interferes with my TV and some channels of my
cordless phones. My garbage disposal unit interfered with everything! My small personal fan destroys thmy monitor picture. (From Derek at LF Research, 6 July 98)
35) Domestic microwave ovens can activate the microwave security sensors fitted to some vehicles. (From thTerry Beadman, MIRA, 6 November 1998).
36) The Canadian Centre for Marine Communications claim there is evidence that EMI may have
contributed to two boat capsizes, via autopilot malfunctions. One was the 16metre fishing vessel the th―Dalewood Provider‖ on August 17 1989, the other was the 64 tons ―Martin N‖ on April 25 1987. In the
latter case three lives were lost. In both cases the concern is that the on-board VHF radiotelephone
system interfered with the autopilot sufficiently to turn the rudder hard over. Staff at the Centre report
that erratic alterations in a boat‘s course when autopilot is engaged and VHF radio used is
commonplace, generally due to insufficient EMI suppression at the autopilot‘s interface and control ―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 7 of 90
cables. (Extracted from: ―Need for EMI/EMC standards and regulations on small boats: a Canadian perspective‖ by Byron R Dawe and Albert Senior of the Canadian Centre for Marine Communications, and Peter Ryan of Fisheries and Oceans Canada, EMC Technology magazine, Nov/Dec 1998, pages 17-19. )
37) The HSE recently prosecuted the supplier of an item of equipment which led to a release of chlorine in a semiconductor plant. The equipment was not sufficiently immune to mains transients (and proven to be so by the HSE‘s own labs). They were prosecuted under section 6 of the Health and Safety at Work Act because the supplier, though aware of the problem, did not inform the users of the equipment. The thcompany pleaded guilty. (From Simon Brown of the HSE, 13 January 1999)
38) The Therapeutic Goods Administration of Australia (TGA) continues to review findings of clinical and laboratory research indicating a potential for temporary interaction or interference between mobile phones and the operation of pacemakers and implantable defibrillators. The findings have indicated that interference may be caused by holding the phone within about 150mm of the implanted device, or in direct contact between the phone antenna and the user‘s skin. Interference can occur with the phone in standby mode, as well as in use. Some phones incorporate magnets, at least in their loudspeakers, and while held close to the implanted device these can cause them to go into their ―magnet‖ mode, which for a pacemaker is a fixed pace.
Based on the most recent testing, simply moving the phone away from the implanted devices will return it to its correct state of operation. Recommendations for users of implanted pacemakers or defibrillators include: not keeping the phone in a pocket over the site of an implant; using the ear that is furthest away from the site of the implant when using the phone; and not allowing the phone antenna to touch any part of the body. (From Compliance Engineering‘s European edition Jan/Feb 1998)
39) There is a story of how something was causing havoc with the emergency services‘ two-way radio
communications in Nevada, i.e. police, fire, and ambulances. An exhaustive investigation led to one or more really noisy pinball machines at a roadside pub (editors note: I thought they had bars in Nevada
instead of pubs). The owner was ordered to get rid of them. He got rid of them and the problem went away. However, it soon reappeared, as another pub owner wound up with the same machines. (From thGeorge Alspaugh of Lexmark International, 7 July 1998. )
40) I have a security floodlamp system for my backyard, equipped with a thermal motion sensor. I have found that I have a reliable, though unintentional, remote control capability simply by flicking the kitchen range fan on and off a couple of times. I told my wife that it‘s a special purpose, hard-wired, thdigital controller. (From Ed Price of Cubic Defense Systems, San Diego, 8 July 1998.)
41) Eurostar and Railtrack officials admitted this week the threat of EMI causing signal failures is delaying the introduction of European rail services north of London. EMI generated by overhead power lines can affect the trackside signals such that red lights are forced to green. A Eurostar spokesperson said: ―In electrical terms, we have found with new trains, such as Eurostar, there tends to be a degree of stray electrical current. This can cause an interference with signalling and effect the integrity cause a signal to go from red to green.‖ Railtrack, responsible for the track and signalling systems, is refusing to allow the trains to run commercially until Eurostar can demonstrate their safety. ―We are working hand-in-hand with to solve this problem as quickly as possible,‖ Railtrack said. Eurostar engineers have designed an interference current monitoring unit. When it senses EMI, the motor is stopped and the train coasts to a stop. However, for the highest safety the unit must be set to maximum sensitivity. This could cause the train to stop every few miles. (From Electronics Weekly rdOctober 23 1996) (Editors note: has anybody seen a Eurostar north of Watford Junction yet? How much has this cost our national economy, especially northern companies? I understand that all Eurostar trains have had TCFs done for them under the EMC Directive and that traditionally both British Rail and Railtrack always imposed stringent EMC immunity standards on their signalling equipment, using the RIA series of standards.)
thAlso see the Lords Hansard text for 14 July 1998 (180714-02) in which Baroness Hayman,
Parliamentary Under-Secretary of State, Department of the Environment, Transport and the Regions stated:
―My Lords, the technical issues which prevent the operation of regional Eurostar rolling stock on parts of the existing network relate to electrical interference associated with track circuits. These matter shave prevented the issue of safety clearances which are required before passenger services can be operated.‖
―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 8 of 90
In October 2003 the interference problems still do not seem to have been solved.
42) There are a number of railway trains which have been unable to be taken into service because they interfere with signalling. (From Ray Garner of Datel Defence Ltd, November 98, quoting an earlier
article in a national newspaper.)
43) A 1.5 MW induction furnace controlled in on/off time-proportioning mode (using large contactors to switch the current) interfered with the computers in a Marks and Spencer‘s store ? mile away. (From
Laidler Associates Consulting Services, June 1998)
44) A new large turbogenerator in a UK power station was designed to have its 20kA three-phase output busbars split either side of one of its support pillars, because of a lack of space. The support pillars were steel, part of a steel framework, and created a single-shorted turn around one of the busbars. In operation, the pillar (made of 2 inch thick steel members) got hot enough to blister its paint, and increased in height by 5mm, putting a bearing out of alignment and causing a terrific noise which caused the station workers to run for their lives. (Editor‘s note: a large turbogenerator up to speed and
adrift from its bearings is a fearsome object!) The cure was another shorted turn, this time around the thpillar and made of ? thick aluminium. (Conversation at Mersey and District Club Européen, 28
45) Further information on No. 37: The case referred to was the prosecution of Fluid Systems International th(t/a Cambridge Fluid Systems) at Swindon Magistrates Court on November 25 1998. The defendant
entered a plea of guilty to a charge brought under S6 (1) (d) of the Health & Safety at work etc. act , 1974. The magistrates imposed a fine of ?5,000 and made an order for the defendants to contribute ?7,000 towards HSE‘s costs of ?9,482. The case concerned a microprocessor based valve control panel used to control the flows of chlorine and nitrogen in a semiconductor plant. There had been a release of chlorine resulting from all of the valves in the control cabinet being set to an open position. Investigation by the HSE found that the unit was susceptible to conducted transients on the mains supply. There were no precautions against electrical interference in the power supply and the microprocessor watchdog was not effective in ensuring a safe state following detection of a fault. The HSE inspector who dealt with this case was Eifion Davies in our Cardiff office. (From Simon Brown of rdthe HSE, 3 March 1999)
46) Scenario: Large open-plan office in a publishing company. Lots of eager beavers with 21 inch displays on their MACs, doing all sorts of clever graphics things for page make-up and other arcane processes.
Problem: The displays on only some of the monitors oscillate sideways about 0.5 mm at most, at about 1 Hz.
Diagnosis (partial): The combined magnetic fields of mains cables under the floor and a power transformer on the floor below are sufficient to cause this very small effect. Unfortunately, once you notice it, it keeps catching your eye and it eventually drives you mad! The 1 Hz is due to a beat between the third harmonic of 50 Hz and the second harmonic of the 75 Hz frame rate of the displays.
Solution: Move the transformer. Replace the large feeder cable to it by individual lower-current feeds to the loads served from it, spread out across the void below the office floor.
Continuing problem: Now that the sideways movement has been eliminated, an even more subtle
―vertical‖ movement of the displays is discovered. Again, it's difficult to see, but once you see it, you can't ignore it. This effect is not continuous: it occurs for a few minutes and then disappears for about ten minutes or more.
Diagnosis: An air-conditioning unit is found to have an intermittent fault to earth, resulting in some 3 A flowing in the armour of the cable feeding it. This current is not balanced by currents flowing in the conductors of the cable, and creates a 50 Hz magnetic field with a horizontal component sufficient to cause the effect.
Solution: By turning a monitor through a right-angle, so that the strong horizontal component of the field is parallel to the electron gun axis, the movement disappears. However, it is obviously necessary to correct the potentially hazardous fault in the air-conditioner.
The main point here is that the tolerable amount of display movement is ―very small indeed‖ when thpeople are working on complex artworks on large-screen displays. (From John Woodgate, 8 March
―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 9 of 90
47) ―When I was a lad‖…. reminiscences from Keith Wilson: The relaxed attitudes of those times did not
always pay off, however. Slightly later in my career, I moved to a company where engineering
standards were, lets put it politely, a little lacking. For example, interlocking between contactors in
reversing pairs was considered an unnecessary expense, and no one would ever consider such
niceties as interrupting capacity when selecting a fuse. If the current rating was right, the fuse was
good enough for the job.
The error of these ways wasn‘t long in revealing itself, however. In one mechanical handling job, we
had around a dozen reversing starters, all protected with totally inadequate fuses. Even worse, the
contactors were controlled by some very dodgy solid-state switches which had been ―designed‖ in-
house. Now, in those days, EMC hadn‘t even been invented. The result was that even the slightest
spike on the supply made these solid-state switches turn on – just for an instant – but long enough for
all of the contactors to jitter. Frequently, both contactors in a reversing pair would close for an instant,
placing a short-circuit across the supply. This meant a mighty bang as the inadequate fuses shattered
and spilled their silica contents all over the floor of the enclosure.
After a lot of time on site, during which much wiring was re-arranged and many capacitors were added
to the system, we managed to get the equipment working after a fashion but, ever since, I‘ve been
suspicious of control panels with a layer of silica sand in the bottom! (Taken from Panel Building
Magazine, February 1999, page 17)
48) I was testing an item of IT based instrumentation the other day that failed conducted emissions. We
replaced its 3 metre long screened 25-way D-type lead, which had been purchased as a ―fully
screened cable‖ from a well-known distributor, with my own home-made 15 metre long 25-way D-type
lead, which simply used a single braid cable and metallised plastic backshells. The conducted
emissions problem (on the mains lead) went away. My customer is now trying to source cables which
really are screened. So caveat emptor, even when buying from large distributors. (From Ian Ball of A.
D. Compliance Services Ltd, which used to be Dedicated Micros EMC Test Centre.) Items 49-52 below are four real-life case-histories of industrial projects that failed in a big way, due a failure to correctly appreciate EMC. The names and details that might allow identification of the companies concerned have been suppressed for confidentiality
These examples have been extracted from the paper ―The real engineering need for EMC‖ by John Whaley, General Manager of SGS International Electrical Approvals (UK), presented at the IEE event rd―Electromagnetic compatibility in heavy power installations‖, Teesside, UK, 23 February 1999.
The other papers from this event will also be of value to anyone involved with industrial products and installations (not just heavy power applications). Contact IEE Sales and ask them to send you digest reference 99/066. These cost ?20 each for delivery in the UK, and they normally require a cheque for the full amount before posting. An extra postage charge may be made for overseas customers. Phone +44 1438 313 311, fax +44 1438 313 465, or e-mail: email@example.com.
49) Failure to correctly specify EMC performance. A large manufacturer of industrial fasteners,
negotiating with a major customer, agreed to install a packaging cell containing an automatic weighing
machine which filled plastic packets with fasteners and an RF welding machine to seal the packets.
For cost reasons the two machines were purchased separately. No assessment of the electromagnetic
environment took place, and the machine contract specifications included no EMC requirements other
than ―shall meet all legal requirements―.
Both machines were supplied, installed, and tested successfully. Unfortunately when both were
operated together the weighing machine suffered >25 % errors due to interference from the RF used
by the welder (not an uncommon problem). In an 8 hour shift the cell should have packaged ?20,000
of fasteners, but could have given away up to ?4,000 of product in incorrect weights.
There was no comeback on the machine suppliers, whose products met specification. Both suppliers
appeared willing to help, but when pressed blamed each other. Expert technical assistance was
brought in and solved the problem. The fastening manufacturer lost 6 weeks production, suffered
additional costs, and lost credibility with their major customer.
50) Over-specification of EMC. A machinery manufacturer needed a special invertor drive for a new
range of machines, and out them out to tender. A manufacturer of AC inverters won the contract for
this large project against stiff competition, but didn‘t notice that the specification required meeting
military EMC standards. Their normal inverter designs failed the EMC tests, and their customer ―Banana Skins‖ compendium ? Cherry Clough Consultants Oct 2005 Page 10 of 90