Presented at RTCM
(Radio Technical Commission for Maritime Communications)
Offshore Systems Ltd.
a manufacturer of ECDIS (Electronic Chart Display and Informantion
that has installations around the world, including many ships in the
Lakes area. These systems usually incorporate a GPS (Global Positioning
System) receiver and a differential beacon receiver, which improves the
accuracy of the GPS position. Together, these two receivers are
to as Differential GPS, or DGPS. Typically, DGPS can provide real time
position information, accurate to within 10 metres, and often within 3
metres. Recently, technicians from OSL, and MLM Marine, of St.
Ontario, Canada, discovered the cause of an intermittant and unusual
of the DGPS receiver.
The "M.V. Manatoulin" is a 25,000 ton (approx) self-unloading bulk carrier, owned and operated by Canada Steamship Lines (CSL), of Montreal, Canada. This ship has operated on the Great Lakes for many years, carrying coal, iron ore, soybean, and other bulk cargoes. In the summer of 1994, OSL installed ECDIS on the CSL fleet, including the "Manatoulin." This ECDIS included the DGPS receivers, and was also interfaced to the LORAN-C, RADAR, ARPA, and other electronic devices on the bridge. Since that time, there have been a number of minor software and hardware upgrandes, and addtional sensors, such as wind and depth have been installed and connected to the ECPINS system. The GPS receiver was a Leica MX9212, and the diferential beacon receiver was a Leica MX-50R.
Over the years, the captain and crew of the "Manatoulin" have come to look favourably upon the ECDIS, and use it extensively for route planning and safe navigation. The captain believes that the system has not only made navigation safer, but more efficient, by following the best track in a given situation, and by allowing the ship to travel at the optimum safe speed. Training, and partiuclary regulary use of the system has allowed the captain and crew to become proficient in the use of the system. Unfortuneatly, in the early summer of 1996, repeated, apparently random failures of the DGPS receiver caused the captain and crew to lose faith in both the reliability and accuracy of the ECDIS.
These failures caused the position of the ship, as displayed on the electronic chart, to move erratically and dramatically, often accross large expanses of land. Usually the Speed Made Good display (SMG - an indication of the vessels speed, relative to the lake bottom or surrounding land) would start to increase, and sometimes indicate speeds of 500 knots or more. Sometimes, the problem would correct itself, and other times, the crew had to reset the system. While this type of failure was annoying and frustrating, it was very obvious. However, the crew became concerned that minor, less obvious inaccuracies were occuring, and this concern caused them to distrust the ECDIS more and more.
At first, the failures seemed to be completely random. Overtime, however, the crew began to notice that failures seemed to occur more frequently in particular places. Two of these places were in the St. Lawerence Seaway, south of Montreal, and about 6 NM off Pt. Weller Piers, in south-western Lake Ontario. (Pt. Weller is the upbound entrance to the Welland Canal.) While failure in these areas was not absolutely certain in these areas, they were very regular. Confusing the situation, was that most of the time, the DGPS receivers seems to work flawlessly. Even stranger, the failures near Pt. Weller seemed to only occur in the upbound direction.
These geographically localized failures cause OSL to suspect that there was some type of Radio Frequency Interference (RFI - sometimes also referred to as Electro-Magnetic Interference, or EMI) in that area. However, many other ships, several with equipment identical to that installed aboard the "Manatoulin," transitted those same areas regularly, and without incident or failures of any kind. This led us to the contradictory conclusion that the problem must be with the equipment aboard the ship. Eventually, we had to convince ourselves that both conclusions were true. This was not a very satisfing conclusion for several reasons, including the improbability of it, and the difficulty and expense of rectifing the problem.
OSL made many attempts to correct the problem, none of which were successful. These included relocating the GPS antenna away from the RADAR antenna and other sources of microwave and radio frequency (RF) radiation, replacing antennae, replacing receivers (several times,) reprograming receivers, testing cables, and so on. Often, an attempted repair would seem to have resolved the problem, only for the failure to reoccur shortly thereafter. Repeated inquiries were made to both Leica, Inc., the manufacturer of the DGPS equipment, and the US Coast Guard (USCG) GPS information lines. None of these inquiries provided a solution.
In November of 1996, OSL contacted Daryl Stewart, of MLM Marine, in St. Catherines. MLM was to install a replacement DGPS receiver aboard the "Manatoulin" for two purposes; first, to provide the ship with a reliable positioning receiver, and secondly, to demonstrate that the failures were confined to the original DGPS receiving equipment. The replacement DGPS receiver was an Star-Tech D-Link 212 integrated receiver, combining both the GPS and beacon receivers in the same housing.
Both OSL and MLM were suprised to discover that the replacement DGPS receiver began to exhibit the same times of failures in the same geograhpic areas. This led to the suprising conclusion that the original DGPS equipment was probably working correctly, and that a combination of geographic location, ship orientation, and the ship itself were responsible for the problems. This unlikely cause posed serious problems, not the least of which was convincing the owners of the ship.
In an all out effort to resolve the problem once and for all, technicians from both OSL and MLM boarded the "Manatoulin" in the early morning hours of December 10, 1996, at Stelco, in Hamilton, Ontario. Furthermore, MLM brought aboard two additional GPS recievers, a Trimble NT200, and a battery powered, portable, handheld unit, made by Magellan. The Trimble GPS receiver was temporarily installed on the bridge, with the GPS antenna installed on the opposite side of the ship. The Magellan handheld unit was to be used to determine if there was some unusual power problem on the ship, and possibly also determine if any interference was limited to the bridge. By the time the "Manatoulin" departed Stelco at about 0530, there were four independent GPS recievers, each made by a different manufacturer, all working correctly and giving positions in close approximation with each other.
Departing Hamilton harbour requires passing under a large steel bridge, between two concrete piers, known as Burlington Piers. While passing under the bridge, each of the GPS receivers momentarily lost the satellite signals - this is a normal occurance - but each regained the signals after a few seconds. A few seconds after that, just as the ship was clearing the piers, all four GPS receivers quickly lost almost all satellites. Different receivers would occaisionally pickup one or two satellites at different times, but none of the four receivers was producing a usable position. In an attempt to determine the cause of this complete failure of the four units, the captain made a sharp turn to port, turning from our original course of 087°T to about 005°T, and then a sharp turn starboard coming to about 170°T. We then returned to our original course, which was to take us within a few miles of Pt. Weller piers, where regular GPS failures were common. As well, the ship's speed was decreased and increased, main engines were taken offline, and put back on, as were generators.
During the several minutes it took to carry out these manouveres, the GPS satellite signals slowly returned to normal, but we were unable to ascertain whether the course changes had any influence. While the ship was making the turns, Daryl Stewart, of MLM Marine, took the portable GPS outside to the main deck, where he walked to the aft end of the ship, about 600 feet away. He walked down one side of the ship, and then returned on the other side of the ship. During this period of about 15 minutes, he was unable to receive a position on the handheld GPS unit. Shortly after the three GPS units on the bridge began to receive again, the handheld was also able to obtain a position. The ship then continued on course, which was modified slightly, to take the ship near Pt. Weller Piers.
On approaching Pt. Weller Piers, nothing happened. This was not completely unexpected, as we were heading in the downbound direction, and the failures usually occurred in the upbound direction. We maintained a close watch on the GPS receivers for the next 30 hours, but all GPS receivers continued to operate normally. This did not surprise us, as we did not expect any further failures until we approached Montreal, the next day.
While in the St. Lawerence Seaway, approaching Montreal, the captain remarked that "just up around this bend we should start to lose it again." As we approached the turn, the captain gave the wheelsman the new steering order: 087°. This was the same course which the ship was on the previous day when all four GPS receivers lost the satellite signals. As we passed through about 080° the GPS receivers again started losing satellites. By the time we had come around to 087°, all GPS receivers had once again lost all satellites.
In an attempt to determine what was causing this problem we immediately began shutting down all equipment on the bridge: RADARs, depth sounders, LORAN-C, Nav Text receivers, M-Sat phone, VHF radios, including portables, and so on. In fact, the only devices left running were the gyro compass and the GPS receivers. Even after "blacking out the bridge" all four GPS receivers were still not receiving.
There was, however, one other antenna on the wheelhouse top, that we had earlier identified, and ruled out as a likely source of interference - the television antenna for the captain's quarters. But in an attempt to demostrate that it was not the problem, and to cover all bases, we unplugged it anyway. Astonishingly, within about 20 seconds all of the GPS receivers again begain to receive the satellite signals, and display a good position. To insure that this was not a coincidence, one of the technicians took a portable VHF radio to the captains quarters, to again disconnect the TV antenna. Within one or two seconds of the antenna being reconnected, the GPS receivers again lost the satellite signals. And again, within a few seconds of disconnecting the TV antenna, the GPS receivers began to work properly. This connect disconnect cycle was repeated several time over the next few minutes, with the same results.
Needless to say, we left theTV antenna disconnected.
Upon examining the ship's log books, we determined that the TV antenna had been installed in the late spring of 1996, shortly before the problems with the GPS receiver had begun. It appears that this antenna was purchased at a consumer electronics store, and installed by the ship's crew. As we were unable to locate the documentation for the antenna, our information about it is sparse. We believe it has an RF amplifier built into the antenna, and is also directional, with a control located near the television that allowed the antenna to be pointed in different directions. (The actual antenna was enclosed in a round placic cylinder.) The only identifing marks we cound find on the control indicated that it was distributed by TDP Electronics, of NC. We suspect it was originally designed for Recreational Vehicles.
We do not have the antenna in our posession, nor do we have the test equipment to determine what excatly was happening. Based on what we have observed, however, we surmise the following:
The antenna had an inexpensive,noisy, RF amplifier built in, that was easily overloaded,thus causing the interference with the GPS signals.
The directional nature of the antenna exacerbated the RF noise problem when pointed towards a strong television signal, and reduced the problem when pointed away from a strong signal. This is probably why the interference only occurred when the ship was heading in certain directions.
When the antenna was far from TV transmitters, it posed no problems. However, when it was near large cities (Montreal, Toronto, Hamilton) where there were powerful TV transmitters, it was subject to overloading. This is probably why the interference only occurred in certain geographic areas.
When overloaded, the antenna appeared to generate enough RFI to "jam" the GPS satellite signals at distances of up to 600 feet or more.
It should be noted that at no time during this trip was there any interference to the low frequency (»300 kHz) differential receivers.
Marine TV Antenna Systems
Naval Electronics, Inc.
7028 West Waters Ave Suite 393
Tampa, Florida USA 33634
phone (813) 885-6091 fax (813) 885-3601