Banana Skins – December 2018 (#132-141)

132.  Interference in the 2.4GHz band

Any time you have more sources of RF energy, the EMC design must accommodate with greater immunity. One area of growth is the use of the 2.45GHz band, where such activity as Bluetooth, cordless phones, HomeRF, new RF lighting, and other systems are all vying for use and must work with each other’s ambients. There have been claims of interference, but this situation is still coming to a boil. It is only a matter of time before products with lesser immunity in this band will not work together at all user locations.

(Taken from “EMC in a High-Frequency World” by Donald N. Heirman, Compliance Engineering, Jan/Feb 2001, page 30.)

133.  15 percent of all computer server crashes can be attributed to electromagnetic interference.

(Taken from “The EMC Building: design and construction strategies” by Jose M Rio, ITEM Update 2000, page 28,

134.  Problems with GPS reception caused by interference

It can clearly be seen from table 1 that GPS receivers are very sensitive compared to the others (GSM and Bluetooth) and not surprisingly, this means they are more prone to interference. For military applications this problem is largely solved by the use of controlled reception pattern antennas which electronically ‘point’ the antenna at the satellite, boosting the signal to interference ration by typically 30dB.

GPS signals can be interfered with by harmonic interference from commercial TV stations and mobile telephone base-stations. In some countries, including Germany, Austria and Hungary the GPS frequency band is also shared by local fixed radio services and GPS reception is impossible in some (small) areas of Hungary for this reason.

(Taken from “GSM, GPS and Bluetooth in an Automotive Environment” by Dr Peter Miller, Euro-EMC’s EMC seminars 3-5 April 2001.)

135.  Catastrophic emissions from cable TV in German aeronautical security bands

In spite of using coaxial cables in cable TV distribution systems there is a lot of shield leakage, based on technical imperfections and ageing. Catastrophic emissions in the aeronautical security bands are jamming Germany. LANs and WANs are growing increasingly, adding to this critical situation.

(Taken from “Megabits Per Second on 50Hz Power Lines” by Diethard Hansen, IEEE EMC Society Newsletter,
January 01,

136.  Interference possibilities between AM radio broad-casting and telcom xDSL

According to her study (Kate Harris, International Switching Symposium 2000, Birmingham, UK) measured ADSL data rates suffer as much as a 2000-b/sec drop when exposed to RFI, which can occur in the bands where DSL networks and AM broadcast share the same spectrum. In North America, this sharing occurs at the medium-wave AM broadcast band. The downstream bands of ADSL and ADSL.lite intersect with AM radio broadcasts in the 535-1104 khz and 535 – 552 kHz ranges, respectively. There is no overlap between their upstream bands and AM radio broadcast. In fact, the narrow overlap of ADSL.lite and AM radio enables minimum capacity loss in ADSL.lite services when combined with low-pass filtering and modern
RFI immunity.

In Europe, however, the upstream band of symmetric high-bit-rate DSL (SHDSL) does share spectrum with AM radio, at the long-wave AM band. And the downstream bands of ADSL and ADSL.lite overlap the long-wave AM band in addition to the medium-wave AM band. Therefore, DSL networks in Europe must contend with a broader RFI threat.

According to Eckert, regulatory agencies are more concerned about egress (interference with AM radio caused by the DSL systems) rather than ingress (interference experienced by the DSL system from AM radio transmissions). But because DSL services are ‘white’ and not concentrated in a carrier, interference produced by these services sounds like white noise, making it difficult for AM radio users to identify the problems as interference. Complaints received by a regulatory body would not necessarily pin-point the origin of the interference or the use of a DSL product as the culprit. So regulatory bodies have been slow to act, which has in turn slowed standards work.

(Taken from “Addressing the Risk of RFI to and from DSL Networks” in Compliance Engineering Jan/Feb 2001, pages 12 and 14.)

137.  Electronic article surveillance (EAS) can interfere with implanted medical devices

Additional areas being investigated by the FDA include electronic article surveillance (EAS) machines and their impact on implantable medical devices. The EAS machines are utilised as anti-theft devices in the exits of many retail stores. There have been instances caused by the interaction of these electromagnetic machines and implanted medical products.

(Taken from “Update on Medical Devices and EMC” by Daniel D Hoolihan, ITEM 2000, page 84,

138.  Unusual types of ESD

Two unusual forms of ESD, internal chair discharges and metal-to-metal discharges from “jingling change” have cause severe field problems in electronic equipment. These forms of ESD are not covered by any current standard.

  • When a person rises from a chair, charges are generated on both the surface of the chair seat and internally that can cause ESD events to occur inside of the chair. These discharges are between metal parts of the chair that are not electrically connected to each other. The discharges cause intense electromagnetic fields to be radiated from the metal parts of the chair, usually the legs. This radiation has been shown to be capable of disrupting the operation of nearby electronic equipment.

This effect was first reported in 1993 by Honda and Smith. Most chairs I have observed with this effect produce about a dozen discharges over the first 10 to 15 seconds after a person rises from the chair. However, some office chairs are capable of producing several hundreds of discharges over as much as a minute. Just purchasing “ESD safe” chairs alone will not eliminate the problem.

I have personally observed an “ESD safe” chair in a factory emitting this type of interference. Since 1993, many types of equipment have been affected by this phenomenon, including communications equipment, computer equipment, even critical equipment in the field of aviation.

  • When small pieces of metal, such a pocket change, move around inside of an insulating pouch such a pocket or plastic bag, they generate different charges. When they touch, small ESD events are generated, for the most part too small to be seen. I have measured risetimes of the fields to be smaller than 100 picoseconds, with sub-nanosecond pulse widths.

With the increasing speed of electronic circuits, many types of circuits have become susceptible to this form of interference. I have caused upset by shaking a plastic sandwich bag with a handful of pocket change near communications equipment, a 100MHz PC, and some consumer electronics. In one case, shaking a bag of coins 3 feet from a rack of equipment caused dozens of red LEDs to light!

(Taken from “Unusual Forms of ESD and Their Effects” by Doug Smith, Conformity 2001, page 203. The article originally appeared in the 1999 EOS/ESD Symposium Handbook, and can be downloaded from

139.  Intercepting and reconstructing VDU monitor signals at 1km or more

You might be interested to note that it is possible to intercept VDU emanations at 1km for monochrome and more for RGB. Both figures are likely to be greater using sophisticated technical means.

(Taken from a discussion about TV detector vans and TEMPEST in The Register,, 4th March 2001 by Andrew Orlowksi, sent in by Graham Eckersall.)

140.  Computers and earthing/grounding impedances at high frequencies

In a newly-constructed financial dealing room, the earthing was done as per IEE Protective Earthing (BS7671). After 2 weeks the US made computer equipment failed. US engineers said it was due to leakage currents from their ITE. Protective bonding for safety is only concerned with 50Hz currents, and is not adequate for modern computers which need a lower earth impedance at higher frequencies because of earth leakage from their mains filters.

In a large computer installation earth leakages (alone) of 70A have been measured at the main earthing terminal, and they are rarely less than 10A.

(From Peter Smith’s presentation of his paper “Protective or Clean Earthing – a Potential Difference” at ERA’s Earthing 2000 Conference 2000, Solihull, 21/22 June 2000.)

141.  Experiences with filtering RS232

We have a product used in-vehicle for vehicle handling testing (see the SR30 robot stuff on our website if you’re interested). One version has a control box that incorporates the closed-loop position controller, servo-amp, interlocks etc and communicates with a remote lap-top, usually part of the end-users data capture system to select the test type and set-up and to upload any test information captured in the controller card.

Early experience suggested that separating the ground references of the remote equipment (normally cigarette-lighter powered) and our control system (direct battery powered; it takes up to 100A instantaneously but has a tare drain of 5A) were a good idea, to prevent the inevitable fighting over the apparent vehicle chassis ‘0V’ reference. We therefore installed the proprietary opto-isolated serial comm’s card supplied by the controller manufacturer and used a shielded comm’s lead (RS232) with the shield connected at one end only – our end in fact.

Six of these units had performed quite happily in the field (and flew through the fairly arduous 30V/m tests used in automotive EMC testing) but the seventh seemed to be very sensitive to the particular laptop/power unit combination used. Very regularly the serial comm’s would lock up when the PWM servo amp was enabled.

With the unit back at our base the symptoms were all too readily reproduced on the end user’s laptop (but not with the newish Dell used in all our testing). Looking at the serial comm’s lead I noticed something different about the 9-pin D-type connector at our end; it had been fitted with a flexible push-in capacitive filter ‘thingy’, of the type available from RS and used on some of our equipment.

As this was not (and had not been) part of the standard assembly I removed it, and the serial comm’s problem was instantly solved!

I can only imagine that this ‘thingy’ was not acting as a filter but was acting as a convenient means for noise to be injected onto the cores of the serial comm’s lead. The connections entering our control case are on the isolated side of the opto card, referenced to the 0V potential of the remote laptop. The connector shell is referenced to our case and its control 0V potential, therefore if our case potential moves relative to the RS232 cores (as I can only imagine it must be when the amplifier is enabled) the capacitors in the filter will act as convenient low impedances directly injecting noise onto the cores.

Fitting the filter ‘thingy’ at the laptop end is, of course, the thing to do …… and moving the shield connection to that end should also help, otherwise the transfer capacitance twixt shield and cores could also be a problem in much the same way as the capacitive filter!

(From Dave Bethell, Anthony Best Dynamics Ltd,, 25th June 01)

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:, 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.