By Rich Keegan
LightSquared is currently conducting a public campaign intended to persuade federal regulators to approve a nationwide broadband service that would be detrimental to users and applications that depend on GPS. The campaign relies on misinformation, revisionist history, half-truths, and clear misstatements of fact. To understand the effort to convince regulators and legislators that the experts are wrong, one must consider 10 basic truths.
1: The MSS Band Was Not Meant for High-Powered Terrestrial Use. The FCC authorized use of ancillary terrestrial component (ATC) ground transmitters many years ago within the mobile satellite services (MSS) band. The LightSquared campaign claims that this proves the band was intended for primary high-powered terrestrial use. But note ATC means ancillary terrestrial component, not primary. The FCC allowed this use only to fill in small holes in coverage from satellites. The term MSS recognizes that the band was for use by low-powered satellites, not high-powered land transmitters.
The FCC conditional waiver given to LightSquared, if allowed to stand, would completely change the nature of the band, converting it to primary terrestrial use by 40,000 or more high-powered ground transmitters. Many FCC statements preceding the conditional waiver make it clear that the LightSquared effort is precisely what the FCC said would not be permitted.
2: Interference to GPS Has Not Been Resolved. LightSquared assured the GPS community when the conditional waiver was announced that all interference issues had been addressed, and its system would not interfere with GPS. It was immediately clear to GPS engineers that this was wrong, and subsequent testing ordered by the FCC, along with that done by manufacturers, federal agencies, and independent organizations, confirmed that the original LightSquared system would cause massive interference with all classes of GPS receivers.
Faced with irrefutable evidence of massive interference, LightSquared revised its system design to propose initial use of only 10 MHz of spectrum farthest from the GPS band (Low 10) for an unspecified period of time, after which it would be allowed to add the closer 10 MHz (High 10). While it may be feasible in the future to develop GPS receivers that could tolerate Low 10, several things are reasonably clear:
- High-precision receivers that can tolerate High 10 and work as well as the ones we now use can’t be built, now or in the foreseeable future. LightSquared’s claims that “we can innovate our way out of this” are wrong with respect to High 10. Filters that LightSquared presently touts to allow Low 10 would not work in the High 10 environment.
- Based on limited testing and analysis, Low 10 causes less interference than the original plan of Low 10+High 10, but the Low 10 effects on many receivers, particularly high-precision receivers in many high-value applications, remains substantial.
With this plan, LightSquared claims that 99 percent of existing GPS receivers would not suffer harmful interference. This conclusion relies on a definition of harmful interference of C/N0 degradation of 6 dB for general navigation devices (the GPS industry and FCC precedent require only 1 dB), and on testing cell-phone GPS with a simple pass/fail criterion, ignoring performance degradation and the fact that modern cell phones are much more like general navigation devices and PNDs than older cell phones. Slanted and unorthodox analytical parameters produced this rosy assessment.
Based on evidence of Low-10 interference, the NTIA and FCC ordered more testing specifically focused on Low 10. In response to mounting evidence of interference at this level also, LightSquared has now offered a third version of its system architecture, using Low 10 and limiting power on the ground. From a GPS interference perspective, this power reduction is useful. However, the latest LightSquared plan does not fully address three key problems:
- There has been no renunciation of High 10. LightSquared says that in 5–6 years it will need spectrum capacity beyond Low 10. It would be irrational to design receivers now that tolerated Low 10, only to find in a few years that the requirements had changed to require tolerance to High 10 also (which is not possible).
- There will still be interference with GPS receivers of various important classes in the power-limited environment of the latest plan.
- None of the evolving plans deals with the massive installed base of GPS receivers.
3: The GPS Industry Did Not Know of a Spectrum Conversion. LightSquared claims that for many years GPS manufacturers were aware of the proposed ground transmitters and should have designed receivers to avoid picking up strong signals in this neighboring band. These claims of foreknowledge of a recent fundamental change in proposed use of the MSS band are fallacious.
The U.S. GPS Industry Council at the time of the limited conditional approval of ATC transmitters (circa 2003) consisted of only two or three GPS manufacturers. It is clear from USGIC statements at the time that it did not anticipate a spectrum reallocation. In any case, it is a huge stretch to claim that USGIC represented all GPS manufacturers, let alone the entire GPS industry and users. The GPS industry had no indication that the FCC would ever radically reallocate MSS band for a stand-alone high-powered terrestrial network, prior to November 2010.
As [GPS World survey editor] Eric Gakstatter has pointed out, a major change with the potential to affect all GPS users should follow certain guidelines. The Air Force GPS Directorate demonstrated this in handling a much less important change to GPS signals: discontinuing support for the semi-codeless technique used in most high-performance receivers. In 2008, it hired consultants to question all manufacturers and many users of GPS about the potential impact. It then proposed that the signal change would occur on December 31, 2020, giving more than 12 years to prepare for the change.
Should we ask anything less from LightSquared’s far more radical proposal?
The FCC has a process that would have been much more appropriate for a proposal to reallocate the MSS L-band to high-powered terrestrial use: Notice of Proposed Rulemaking. Had it followed this process, we might be having a productive discussion of technical aspects.
4: GPS Receivers Properly Use the MSS L-Band. LightSquared asserts that GPS receivers intrude into LightSquared’s spectrum— a misleading claim. Many GPS receivers in fact have filters that do not block signals from the MSS band. There are several reasons for this:
- So long as the MSS band was a satellite band for signals from space to Earth, the signals from other systems in that band were low-power and not harmful to GPS reception. GPS receiver designers relied on this and assumed this allocation of the band would continue. The ability to use filters that overlap into the MSS band has enabled both low-cost and high-precision GPS receivers.
- High-precision receivers cannot produce accurate measurements without using wideband GPS signals that occupy most or all of the GPS band. “Brick wall” filters that could capture all the energy in the GPS band and none of the energy in the adjacent MSS band do not exist.
- Lightsquared ignores hundreds of thousands of high-accuracy, high-value GPS receivers that receive signals from the MSS band, using it for its intended purpose — satellite to ground communication. Deere receivers use the StarFire system leasing use of transmitters on MSS band Inmarsat satellites. Trimble leases use of MSS band on LightSquared’s own satellites for OmniSTAR correction signals.
- GNSSs worldwide are modernizing their signals; many of these new signals are wideband. To take advantage of them, modern receivers of all classes will be wideband, as high-precision receivers are now, and will suffer interference similar to that of high-precision receivers now.
5: GPS Receivers Do Not Ignore Government Design Standards. LightSquared asserts that the fundamental GPS L1 signal specification mandates receiver design standards that the GPS industry has ignored, to save a few cents of cost. These claims are false. The GPS specification defines the signal-in-space and explicitly says that it is not a receiver design standard; it simply uses a nominal receiver design to be able to translate signal-in-space specification into navigation performance effects.
6: Receiver Replacement Costs and Schedules Are Large. LightSquared has offered $50 million to fund retrofit or replacement of legacy government receivers impacted by its signals. General Shelton of the Air Force Space Command testified to Congress that it would take billions of dollars to replace or retrofit the government receivers. He also estimated a 10-year time frame to test and validate replacement receivers.
LightSquared says it will not bear the costs of replacing commercial receivers since, it claims, manufacturers are responsible for the improper design of those receivers. This is wrong, as shown earlier. LightSquared should bear the cost of replacing commercial receivers, if allowed to proceed. A realistic time frame needed to replace high-accuracy, high-value commercial receivers is also about 10 years.
LightSquared argues that in five years, most current GPS receivers will be obsolete. This is clearly not true. Many current high-precision receivers are already prepared to use modernized signals from GNSS constellations. The L1C GPS civil signal, for instance, will not be available on any satellite until 2014, and the full constellation of satellites with L1C will not be available until 2026. Therefore, many receivers in use now will continue to be in use for many more years than five.
7: Other GNSS Are Also Affected. Because Galileo, Compass, and GLONASS use or will use signals similar to GPS, in the same band as GPS, they will suffer interference very similar to that suffered by GPS. Users will lose the benefits of these other constellations, as well as GPS.
The United States has entered into formal obligations to protect some other GNSS signals; LightSquared signals are not compatible with these U.S. obligations.
8: Handset Interference is a Serious Concern. LightSquared handsets do not yet exist, but testing to date makes it clear that the handset signals to communicate with LightSquared base stations also interfere with GPS receivers when they are nearby (a few meters). The interference to GLONASS reception is also likely to be harmful. The interference effects of a group of LightSquared handsets has not been fully evaluated, but will certainly create more interference for nearby receivers.
Out-of-band emissions from LightSquared handsets, if as high as FCC power masks currently permit, would substantially interfere with all GPS receivers, possibly more than LightSquared base stations.
9: The Solution Is Not a $6 Filter. LightSquared displayed a Deere high-precision receiver with a “$6 filter” and told Congress this proved it could be done inexpensively and quickly. The claim is based on half-truths.
- The Low 10 signal can be filtered out using low-cost parts, but the effect on performance is not known. There is good technical reason to be concerned about degraded performance from this filtering.
- The Deere receiver displayed is not capable of readily being retrofitted with LightSquared’s or any other filter. Like many high-precision units, it is an integrated, hermetically sealed device. Retrofitting would entail returning the unit to the factory, cutting open and discarding the case, replacing the antenna/preamp assembly with a redesigned antenna/preamp assembly, inserting the unit into a new case and sealing it, re-testing the unit, and returning it to the customer. A costly process.
- Filtering is one element of a design, usually distributed across several stages of the receiver. Changing filtering requires a redesign that may stretch across the entire RF front end, and cannot be done casually.
- The displayed filter’s specified insertion loss is 3 dB, well above what GPS designers normally accept, and would result in about 2 dB more loss of sensitivity than with current filters.
- LightSquared has suggested moving StarFire and OmniSTAR augmentation signals to the top of the MSS band, very close to the GPS band, so that filters that included GPS could include them. This is a reasonable approach, but the “$6 filter” might not permit that, as it would excessively attenuate at least the StarFire signal.
10: The GPS Industry Supports National Broadband. The GPS industry broadly supports the goal of extensive and pervasive national broadband, and of strong competition among providers. Pervasive broadband would be helpful for applications such as real-time kinematic (RTK) positioning. It would be beneficial to GNSS users to have broadband services available everywhere, but not if the cost is to degrade or deny GNSS service.
LightSquared’s broadband services require terrestrial base stations and cannot be done with the LightSquared satellites. It is unlikely that low-population areas will be covered with terrestrial base stations due to the economics involved, but if broadband coverage is nationwide, then so too will be GPS interference.
Rich Keegan is a senior principal engineer at NavCom Technology, Inc., a wholly owned subsidiary of Deere and Company.