This topic has probably ignited more discussion among the air traveling public than lost luggage or the in-flight meal service. In our endlessly connected age, people demand to be spliced into the communication stream 24/7. How dare a flight attendant tell them you cannot use some device at any time they desire? Everybody knows this “problem” is a total farce, and nothing bad will ever happen if they call, text, surf or game with everyone around them, whenever they choose. Planes don’t spontaneously explode when they text, and don’t crash or hit a mountain when they use their cell phones, so it must be a non-issue.
In addition, the existence of planes now equipped with internal Wi-Fi clearly proves there is nothing to this whole personal electronic device (PED) thing to passengers, as does the permission to let users whip out their laptops in flight and click away. So, in the collective minds of flying public, this is just foolishness, and there isn’t a single piece of solid evidence to justify the ban, and frankly, it’s hard to blame them. The case has not been reported to change behavior, and frankly everyone tries to downplay even the likelihood that such a problem exists, for a variety of legal liability reasons. It’s a hugely frustrating environment to work in for everyone, especially cabin crews.
So, first things first, do PEDs actually cause in-flight aircraft problems? Luckily for our purposes, and so we don’t have to rely on apocryphal rumors, NASA maintains an extensive (but voluntary) flight report database called ASRS (Aviation Safety Reporting Database), and you can access it yourself from here, please don’t just rely on my data collection to satisfy yourself: http://asrs.arc.nasa.gov/search/database.html
 |
| All PED Events |
NASA maintains an extensive (but voluntary) flight report database called Aviation Safety Reporting Database, available at http://asrs.arc.nasa.gov/search/database.html. In a search for all PED in-flight events, the database reported 108 occurrences. The first 50 events reach back to 2002, which I think is as far as I want to go, because the very nature of the technology is changing enough that I prefer to have only timely data. Excluding the reports of passengers simply refusing to comply with legal cabin crew instructions to stop using their PED rules out 30 percent of the reported events. Also, if we remove issues of crew uncertainty as to whether some device is acceptable, and miscategorized events, we can then eliminate another 12 percent. That leaves 58 percent of the remaining NASA reports as genuine in-flight system failures, directly tied to PED operation. That is a significant number.
The simple fact that cell operation is even possible inside the aircraft is also unfortunate proof that potentially disruptive emission is possible into aircraft systems from the passenger area. If the windows, vents and composite structures produced a seamless RF shield, then cell operation should not be possible, and neither would interfering trans-hull emissions into antenna systems.
  |
| Relevant PED Events |
I was particularly concerned that the largest valid fault report percentage (26 percent) was tied to fire, smoke and explosions. Before studying the reports, I had not really considered this a meaningful PED threat (more like a deliberate terrorist act), yet the report data clearly shows this to be a very significant problem, especially when passengers attempt to charge any of their electronic devices from onboard power. What was even more astonishing was the number of events where passengers attempted to conceal their involvement in the smoke/fire/explosion event, clearly indicating the level of responsibility people demonstrate in flight even when they are creating a potentially and visibly dangerous problem is negligible.
Also unexpected were the events tied to TCAS, where flight crews got alarming messages of imminent airborne collision, and moved the aircraft in response. The frequencies involved are 1030Mhz for query, and 1090Mhz for reply, which matches no known PED frequency directly. These events are so disturbing I have copied the short synopsis of the three events here:
➤ B737-800 FLT CREW EXPERIENCED SEVERAL TCAS RA’S ALLEGEDLY GENERATED BY A WI-FI ENABLED LAPTOP COMPUTER.
➤ A FALSE TCASII RA SENDS A DC9 FLT INTO A CLB TO AVOID A POTENTIAL TARGET 5 MI SE OF BUNTS INTXN, PA.
➤ DC-9 FLT CREW RECEIVED A FALSE TCAS RA DURING DEP CLIMB AND INCREASED THEIR RATE OF CLB TO AVOID A FALSE TARGET APPARENTLY GENERATED BY A PAX LAPTOP COMPUTER.
 |
| Relevant PED Events |
I fully expected to see COM-related problems; my own lab experiments showed laptops by themselves could easily disturb VHF AM radio transmissions. This was in turn tied to specific software running as that excited the memory bus and dramatically increased emissions. Cell phones, two-way pagers and Wi-Fi devices all actively transmitting pose a much bigger RF interference threat to on-board avionics.
No doubt everybody who flies has noticed there is a cabin announcement to stop using all electronic devices as landing begins, and so the expectation should be that Localizer and Glide Slope problems should occur in the reports, and they do. This is a very critical flight phase, especially under IFR conditions, and the potential for an accident is significant if the situation is not remedied.
Magnetic compass errors are also reported, which should be no surprise, as the flux valves used to extract magnetic field data (usually wing or tail mounted, to avoid local magnetic field disturbance) are extremely sensitive, and easily upset by detected RF signals at the flux valve connections.
To me, the reports that are the most disturbing (after TCAS alerts) are instances where significant cockpit instrumentation failure occurs. The stress in the cockpit this generates is very significant, and the potential for flight mis-cues and errors is profound, not to mention being alarming from an electronic interference viewpoint. Here are the condensed reports:
➤ A B757-200’S L FUEL GAUGE BLANKED AFTER TKOF AND BECAME OPERABLE PRIOR TO LNDG. CREW SUSPECTS POSSIBLE PED INTERFERENCE.
➤ A B767 HAS MULTIPLE ELECTRICAL SYSTEM MALFUNCTIONS INCLUDING LOSS OF ENG AND FLT INSTRUMENTS AS WELL AS CTL OF AUTO FLT SYSTEMS FOR A SHORT PERIOD OF TIME DURING CRUISE FLT. PAX STATED IT WAS A NEW PHONE BY AT&T, MODEL SIEMENS CEO168, HE WAS INSTALLING NEW PHONE NUMBERS IN DATA BANK AND NOT TALKING ON PHONE.
The issue here is PED/TPED RF emissions are affecting non-RF systems, which means the pulse energy or data interference is also reaching into other system wiring within the cabin area, with serious repercussions. This suggests further shielding or conduit runs are also required to protect wiring not involved with shipboard radio systems per se, if they transit the passenger cabin area.
Let’s think about numbers and risks. The reluctance to file any optional and potentially inflammatory report is high, and there is a large disincentive for any airline or airframe maker to generate such reports or encourage their generation or discussion, since it highlights a potential problem. In addition, many flight crews do not have the time or inclination to try a cabin scan for PEDs as serious problems occur, so many events just get recorded (if at all) in the aircraft snag reports as intermittent problems, and get “resolved” as No Fault Found (NFF) subsequently.
If we ignore all flight except commercial airline flights, that is roughly 28,000 a day in the United States, out of 87,000 total per day. Worldwide there were roughly 93,000 commercial airline flights per day as of 2008. That is a lot of flying, and relatively few PED event reports, far less than even one a day. It is not unreasonable to assume only 10 percent or even less of PED events get voluntary reports, but even at 1,080 per 10 years that is still a vanishingly small flight percentage. The odds of a reported PED flight event seem to be roughly 1 in 100,000. Keep in mind, however, over this same period, there were more than 418,000 incident reports filed in the NASA database, and some of these are certainly PED related, but reported as intermittent problems, so the real incidence and correlation remains hard to establish. The only certainty is a real problem.
An objective person has to conclude that while aircraft systems are affected by PEDs, the incidence of events is low, frankly a testament to good avionics and airframe design, and the relentless announcements to passengers. Unfortunately for everyone, the landscape is changing for the far worse with regard to passenger equipment RF emissions, and it is a big mistake to assume the low voluntary report rates, and incidents of mild severity are going to remain the norm.
The most significant change to the cabin ecosystem is PEDs with accidental emissions are now being supplanted by Transmitting PEDs (T-PED), devices that deliberately emit high levels of RF. This means increased RF levels by 80-100dB, a huge increase. This includes cell phones (now in many bands, up to 2Ghz), both 2.4 and 5.8Ghz Wi-Fi within many devices not thought of specifically as radios, Bluetooth devices which can have significant range, FRS (Family Radio Service), CB radios, ISM band devices, two-way pagers, and all manner of handheld radios from low band VHF to over 1 Ghz, and many GPS-enabled devices which can act as secondary emitters, usually from local oscillator injection. Any or all of these could be operated in the cabin area either deliberately or accidentally. In fact, all the aviation communication and navigation bands are now bracketed both above and below by a sea of potential RF emissions, and are now clearly at risk.
Keep in mind, all RF emissions at any frequency also contain harmonics, unwanted spurious content, and produce mixer (sum and difference) products at the victim receivers. This synergistic unhappiness dramatically increases the chances of accidental “in-band” interference, even though no emitting source is nominally actually “on” that frequency intentionally. The potential is also there for highly elevated noise floor levels, which can compromise system operation, especially for GPS, which is an extremely low level service.
In addition, new aircraft have increased composite structures with no shielding, providing much easier RF contamination paths, and lower resistance to unwanted signal contamination. This trend is especially profound in helicopters, which not only have huge window areas and tightly packed cabins with antennas very close by, but now have minimal RF shielding in literally every direction.
Human Behavior
It’s important to look at passengers behavior while in flight. In 2006, the IEEE published a very useful study, in which researchers spent three months in 2003 scanning 37 domestic flights for PED emissions. They found passengers used cell phones on the average at least one to four times per flight, even during landing, and despite being told not to do so. They also found emissions could clearly interfere with GPS reception, echoing NASA studies. In addition, they detected signals showing at least one phone was on and attempting to attach to a network. As noted earlier, it is unrealistic to expect passengers to do anything to curtail PED use, even to maintain their own safety.
it is the combination of multiple sources, and their combined spurious emissions, that erode the safety margin of operating avionics systems.
The article also confirmed what I have seen in testing it is the combination of multiple sources, and their combined spurious emissions, that erode the safety margin of operating avionics systems. It is the gap between what FAA accepts as allowable signals and what the FCC accepts as allowable spurious emissions in approved devices that is the root of the problem. GPS, DME, NAV and COM interference were all observed though combinatorial interference using FCC approved devices. The key to this situation is to control the environment so that even passenger non-compliance does not threaten airworthiness.
You should review these NASA publications: NASA/TM-2004-213001
http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20040040193_2004035943.pdf for their analysis of specific mobile phones and GPS interference.
NASA/TP-2003-212438 http://ntrs.nasa.gov/archive/nasa/casi.ntrs.nasa.gov/20030067884_2003078491.pdf for Portable Wireless LAN Device and Two-Way Radio Threat Assessment for Aircraft Navigation Radios
Both the FCC and PED equipment makers are to blame for these potential PED problems. FCC permitted spurious emission levels are far too high and enforcement of compliance, especially on imported low end equipment is minimal, and largely self-policed. Many items from kid’s toys to music players have no real FCC oversight, yet are plastic-housed significant emitters loaded with noisy digital circuits, and often wireless links as T-PEDs.
Little or no effort is expended to make “airplane” mode easy to find on equipment like cell phones to iPads, and many items have no clearly marked power-off switch. Something as simple as a certification “flight safe” program to check equipment for unwanted emissions would totally change the dynamic created by the toothless government regulations in force. Making certification compliance a USER and MAKER effort issue, not a flight crew issue is a critical safety step we are currently missing.
It is very important to note that once spurious emissions standards are set (and especially if they are set poorly), then component makers and designers will use them as the de facto design levels. Filters and other related parts will be made to those specs, and no effort will ever be made to improve them by any manufacturer. There is simply no reason to do so, and a direct financial disadvantage to do it.
Lack of an incoming emissions test of passengers is the next practical problem to overcome. Self-check-in makes this much worse, as there is no opportunity to advise passengers or determine if they are cooperative enough for flight. A simple RF scan as people travel through the gangway, with a divert back if emissions are found, would be a hugely helpful tool.
In addition, it is important to include warning information in the seat pocket hand-out, not to leave the RF defense to over-worked and harried flight attendants. It must spell out all electronic equipment is to be OFF during take off and landing, and communication from the plane during approach, take off and landing is an actionable offense according to Homeland Security. Some action by Sky Marshals is preferable periodically to clarify that it is not a joke.
Physical proximity is an important issue for interference, and small aircraft and helicopters are much more susceptible to PED/T-PED interference than larger airlines simply because the close proximity of the user to the system, and the lack of shielded barriers. Helicopters are an especially difficult problem because the use of unshielded composites is now common, it is not easy to even achieve good antenna performance, and inter-system interference is the norm rather than the exception. Adding the RF burden of PEDs and T-PEDs simply increases problems for every system because shielding between the user and the victim system is negligible in almost every direction.
The normal use of a cell phone or other T-PED is often by the user’s head, and thus close to the outer top skin of the airframe, this makes coupling to the coax cable runs, as well as through the windows to the antennas more likely. NASA studies on airliners showed a series of coupling models that clearly highlight windows, air ducts, composites and door seams as the principal emission pathways into external antennas. This coupling factor was about 80 to 85dB in a large aircraft, amazingly high when you consider there is no real line of sight possible.
The smaller the airframe is, the higher the coupling factor, and so the higher the chances of unwanted interference to onboard systems. In addition, roof access coax feedlines will be most susceptible to T-PEDs, a factor often easily forgotten.
Hardening the Airframe:
1. Windows seem to be the key to RF suppression in large ships, but simply may not be effective in helicopters due to their physical design. Use of either clear conductive coatings like ITO (Indium Tin Oxide) (up to ~40dB attenuation) or conductive micromesh (up to 60-80dB) can dramatically reduce both RF leakage out, as well as incoming signals to discourage users from attempting cell phone calls.
2. Use of extra shielding on all antenna feedlines, with the outer (non-current-carrying) shield bonded at each mounting clamp. This is in essence, a Triaxial cable with the outer braid treated for non-corrosion, and externally exposed and clamped. Use of an inner foil to achieve 100 percent coverage under the braid is recommended. A single cable conduit is also ideal if possible at airframe construction time. In essence, provide a well shielded path for any RF cables transiting the passenger cabin area.
3. Use of a well-grounded thin cover foil over all wiring exposed to the cabin area, so that all cabin exposed wire runs are shielded from gross signals, especially in overhead areas.
4. Use of passivated conductive seam tape over all mechanical body seams, and conductive mesh filters in air ducts. Add door RF finger seals if possible.
5. Install an RF tight quarantine area for problem devices found in flight. Use of a zip lock highly conductive bag may also work if the item is left with the passenger.
6. Install RF sniffer monitors or provide portable sniffers for crew in the cabin to look for problem RF transmissions.
7. Install Radiax (leaky cable) or GORE Cable-Based Antenna style internal Wi-Fi systems (probably floor run to avoid roof mounted GPS antenna interference), with demonstrated non-interference to ship’s systems. This provides the 2-way pathway for useful and safe passenger communication at low RF levels. Make it under pilot control, so it can be disabled during approach/landing and take-off. Use a 50 ohm attenuator to reduce system output, and stabilize the cable to reduce unwanted emissions due to ringing.
8. Modify any cabin wall sound insulation to also be RF absorbent, to dramatically reduce reflections within the cabin and emissions into wiring.
Part of the fix to this problem is somewhat counter-intuitive, namely, implement one effective Wi-Fi access method for all to use for free, which can be effectively controlled, and try and render other T-PEDS ineffective, to discourage their use. In this way, the RF environment can be made useful to passengers, but safe for on-board avionics. In addition, focus on this frequency allows direct frequency filtering to remove it, rather than some kind of very lossy wideband filter in every system as a control technique.
The simple go-no-go test for a successfully hardened ship will be that reception of outside cell networks is not possible, but use of the internal Wi-Fi network is possible inside only. If this can be achieved, then the likelihood of PED/T-PED transmission as a fault trigger is dramatically reduced, and passengers can still have the link they insist on for outside communication.
It is important to understand that short of removing all PEDs and T-PEDs from passengers, totally safe cabin operation is not possible, but it is possible to mitigate most threats far better than we are doing now. This will have a weight and cost penalty that is not trivial, unfortunately, but the risks are increasing, so some effective plan needs to be implemented.
I would have had a lot more confidence in ship shielding as a viable tool if I had not done experiments. To really consider physical shielding as effective, I felt I should be able to show this effect using ambient signals over a 1Ghz range, an antenna linked to a spectrum analyzer, and a suitable shield.
In my test, I used a small broadband test Yagi antenna tied to my IFR A-7550, and a desktop 1Ghz TEM cell (shielded RF test chamber, which will come into play again later in our search for emissions).
  |
Here’s the local ambient RF around
my test bench, sampled over 1Ghz.
Lots of emissions, from broadcast
FM, TV and local SCADA to cell
phones. |
  |
Here’s the antenna mostly sealed
inside the TEM cell, and insulated
from it. That looks worse in some
ways, better a bit at 800MHz+
area, certainly not impressive. |
 |
The antenna outside the TEM cell
but common ground to the TEM
cell. Finally, we do get some useful
attenuation of the airborne signals,
but it depends on the shield being
at a ground potential relative to the
receiver end. Yes, the trace would
be flat. |
This really highlights the shielding problem of an aircraft. Relative to handheld PEDs and T-PEDs, the aircraft is NOT an effective shield, since the skin is a floating element, acting more like an antenna element than a shield. In fact, the aircraft, by virtue of its insulated tires, is never really grounded at any time, and so not hugely effective as a shield relative to floating internal items and external fields. Helicopters landing on metal skid tubes are probably the only aircraft to ever become truly grounded. The airframe skin will work for internal (airframe grounded) avionics, but may never be truly effective for external fields. This is a situation not unlike being able to operate your cell phone in an elevator high up, which despite being a sealed metal box, can still operate.
  |
Here’s the background TEM
cell signature around VHF
Com frequencies. |
  |
Here’s our test iPod Touch,
playing a video, with earbuds
connected. |
  |
Our iPod emission in the AM
Com band. |
  |
Here’s our reference wideband
1Ghz level. |
These are very low-level emissions, and frankly are impressively so. For reference, the signature from my iPad was similar, but my homemade larger TEM cell that it fits in is not clean enough to use as a public data source, but take comfort from the fact that the iPad is also a low (but not zero) emitter.
Please note though emissions exist, and certainly at frequencies not expected. This is the story of almost all PEDs and T-PEDs, they all add to the RF background, and T-PEDs dramatically so. Once you blend together 100 of these in the cabin area, the picture becomes a lot murkier as to how safe they really are going to be in flight.
  |
Here is the iPod emission signature
playing video. Notice the appearance
of small peaks at ~10-20, 50, 220,
300, 570 and 750Mhz. |
  |
Here is the iPod Web browsing via
its Wi-Fi hardware. Note the
addition of a new peak at 400Mhz.
This appears to be a spur from
the 2.4Ghz Wi-Fi transmitter. This
pulses on and off while the unit
tried to acquire a working Wi-Fi feed. |
Any plan going forward has to recognize the unwillingness of passengers to cooperate, and the increasing RF pollution we will face. I doubt there is a single solution for every airframe, but I believe it is possible to improve the situation if everyone makes some effort. Clearly the risks are not zero, so doing nothing, especially in the face of rising T-PED RF levels and widening frequency coverage is just not going to be a viable option.
Walter Shawlee 2 is the president of Sphere Research Corp. in West Kelowna, British Columbia, Canada, and a senior designer at Technisonic Industries. He can be reached at [email protected].
To see a complete archive of Shawlee’s System Design columns, visit www.aviationtoday.com/shawlee.