Global Navigation Satellites may revolve around the world, but the world of Global Positioning Satellites revolves around atomic clocks, and each satellite typically carries multiple atomic clocks. These are very sophisticated (and very costly) atomic clocks. They broadcast time (and clock bias information) all over the globe—they have become “the world’s clock.”
There are, perhaps by a hundredfold, more timing users of the GPS system than there are positioning users, and most do not realize just how critical such precise timing is, let alone that these critical timing services come from GPS satellites. Computer networks, radio transmissions, cellular systems, the whole internet for that matter all depend on the precise timing originating in those atomic clocks orbiting above.
In January of 2011, Symmetricom Inc. announced the first commercially available chip-scale atomic clocks (CSAC). This has been viewed as a kind of holy grail for many in the precise timing industry, and even by some in the GNSS industry. Others have reacted to the notion of a coin-sized atomic clock with shocked concern: “What is this about a pocket atomic device?” It’s not a James Bond gadget, and just like all atomic clocks there are no radioactive materials involved. This is not a nuclear process; rather, these clocks work on processes that excite certain particles at the atomic and sub-atomic level and derive precise oscillations from them. (That’s an oversimplification, but I could not go any deeper into the theory without developing a headache.) There is great potential to solve certain timing conundrums with innovations like this, as well as some for GNSS, but how? We have to go back to the foundation process of GNSS to find out.
Fundamental to the geometry being solved for global positioning is the distance between the satellite and your terrestrial receiver. If the position of the satellite is known (per tracked orbit data) and you know what time it was when the signal left the satellite (per the atomic clock), the signal speed (speed of light ±), and what time it was when the signal got to your receiver (onboard clock plus broadcast clock data and clock biases) and then you’ve got that distance. Do this with several satellites’ signals simultaneously, add some differential or other approach to mitigating the ionospheric and tropospheric delays in the transit of those signals, and voilà, you’ve got terrestrial positions from those satellites. There are a few weak links in the process, one being the relatively imprecise clock in your receiver … imprecise when compared to a high-end atomic clock.
Standard for most GNSS receivers are oven-controlled crystal clocks, OCXO for short. These use the piezoelectric effect of crystals—the mechanical vibration for an oscillating circuit—cool, but limited to a range of milliseconds and microseconds. High-end atomic clocks operate in the range of nanoseconds. It would be potentially wonderful if you could have the equivalent of one of the space-born atomic clocks on your rover, but not very practical. Though atomic clocks have shrunk since the first “fridge-sized” ones first operated in 1949 by the National Bureau of Standards (now the National Institute of Standards and Technology), a clock like those in the satellites would still be the size of a microwave oven, weigh over 150 pounds, and add many, many zeros to the cost of a rover.
Some wonder if a more precise clock is necessary; GNSS developers have had decades of success in resolving timing inequities and have used many innovative approaches to getting sub-centimeter positions in seconds. But there may be other reasons why a better clock on the receiver end might be a good thing, if not a game changer in some instances.
In researching CSACs I found out a few surprising things. First, my bubble was burst to some degree regarding the immediate future of high-precision GNSS and CSACs. However, judging by the evasive answers and reluctance to be quoted by developers (among them several big surveying equipment manufacturing companies) I questioned about this subject, there indeed must be a serious potential for implementation—no one wants to show their cards at this time.
The person willing to discuss CSACs and practical implementations thereof (even for many low-precision GNSS uses) was Steve Fossi, director of new business development at Symmetricom. “We made it through the third phase of the DARPA development program and decided to commit additional R&D into developing this as a commercial product,” he said.
Fossi was referring to the participation of Symmetricom, along with four other teams, in a program sponsored by the Defense Advanced Research Projects Agency (DARPA) that sought (from their website) “the development of CSAC enabled ultra-miniaturized and ultra-low power time and frequency references for high-security Ultra High Frequency (UHF) communication and jam-resistant GPS receivers.” And as with other DARPA-sponsored initiatives, many other military and commercial applications were made possible in the quest for such innovation.
On the military side, Fossi noted some examples. “You have soldiers in the field with GPS units ducking in and out of buildings and from under cover; they lose [GPS solution], and [this kind of chip] can greatly speed up the time to subsequent fix, no need for coarse acquisition.” There are many other military applications; having this precise timing source, independent of GPS time, means we can compare various signals to help detect and mitigate some forms of jamming, spoofing, and interference. Such solutions could be key in further developing indoor positioning solutions.
I had asked Fossi about something that a few of the GNSS developers had hinted at: the notion of using the CSAC as the “extra satellite” in some GNSS solutions. In high-precision GNSS, such as in ambiguity fixing for RTK/RTN, the clock of a fifth satellite is used along with the observations from a minimum of four others—this could be quite useful in poor-sky-view situations. He said, “Of course, and [even for] low-precision GPS where you use three satellites,” the CSAC could act “as the fourth, or if you didn’t care about the elevation, there could even be only two satellites with the CSAC [acting as] the third.”
The size, weight, and low power requirements of the CSACs have sparked a lot of interest in other positioning and timing applications as an alternative or supplement to satellite-based timing. It’s possible, “certainly in mining, with limited or no sky view, and in underwater applications,” said Fossi. “We have one client that is working on replacing [legacy] timing [chips] with CSAC in the design of deep sea sensors. The long life, low degradation in time, and very low power of the [CSACs] means this is a very practical and affordable solution—the whole unit uses a max of 120 milliwatts.” Cost is another challenge; a single one of these might cost as much as a laptop, but, if these were widely implemented (say, by the time they get to your rover), those costs would be greatly reduced.
The DARPA program called for an atomic clock to be 0.01 of the size of the common units and 0.001 of the power; Symmetricom succeeded. But how do you stuff an entire atomic clock into something where the main “atomic” component is smaller than a sugar cube? Note: Additional clock circuitry on a circuit board would also be needed to get time from the CSAC. “Our [CSAC] is a gas-cell-type atomic clock,” explains Fossi. “Those hydrogen-maser-type atomic clocks work on different principles.”
Fossi gave me a crash course in gas cell clocks. He described the gas cell (itself not a new idea) as a small chamber with typically rubidium in a gaseous state (in their unit they use cesium); a light source is introduced, and electrons go from a “base” state to an “excited” state. The shift from base to excited states is very sharp because only a specific amount of energy will enable the shift—a true quantum effect. From this shift they get very precise oscillations, and precise time from that.
But how practical would these be for high-precision GNSS receivers like the ones surveyors use? One trade off for miniaturization and low power (when compared to the atomic clocks on the satellites) is precision. The CSAC is not as precise as the atomic clocks on the GNSS satellites, by a factor of roughly 1000; however, the CSAC is probably 100 to 1000 times more precise than the clocks currently used in surveying equipment. It is more likely that we might see such CSACs implemented in some of the more “industrial strength” GNSS receivers such as those used for aviation and the military before we see them in RTK rovers. There may be implementation on terralites and pseudolites in the near future as well.
As CSACs eventually do find their way into rovers, you can be assured that they will not be just another “marketing feature” or acronym (“hey, press this hotkey and our rover emits a fresh lemony scent!”), but that the addition of CSACs (if/when) will have a significant impact on field capabilities. Perhaps it won’t be an improvement in positional precision per se, but more in the potential of being able to work better in sky-view challenged areas, to detect and mitigate spoofing, jamming, and interference, to enhance PPP-RTK, and to enable better synchronization of data and communications solutions that will soon be coming our way. We’ll have to examine those other potential solutions in future articles.