Commentary
Need a really stable portable clock? Think atomic
Steve Fossi, director of New Business Development, Symmetricom Inc.
3/1/2011 9:30 AM EST
As the new decade gets underway, designers of military hardware have a
new tool in their component toolbox— portable atomic clocks.
Ultra-stable timekeeping can be critical for battlefield communications,
IED jamming, navigation, and tactical UAVs (unmanned aerial vehicles),
among other applications. A truly portable atomic clock would be a real
plus in these applications.
Take communications. In order to compress ever-increasing amounts of data into limited spectrum, increasingly higher bandwidth waveforms are needed — with ever-decreasing tolerance for clock drift. That means that the clocks used in today’s manpack radios — oven controlled crystal oscillators (OCXOs) and temperature controlled crystal oscillators (TCXOs) — may no longer be suitable. But an atomic clock would be— provided its size, weight and power consumption (SWaP) were low enough to be truly called portable.
Power consumption is also an issue for IED jammers. Today’s IED jammers have power requirements that can only be met by a vehicle’s generator. To make an IED jammer dismounted, all of the components (including the oscillator) would need to be low-SWaP. Also, all of the dismounted jammers would need to be tightly synchronized to allow predefined time slots in the signals (“look windows”) where friendly force communications could still get through. Such high synchronization is another reason to use an atomic clock.
Handheld GPS units and tactical UAVs would also benefit from an atomic clock’s stability — but, again, only if the clock also satisfied the SWaP profile for portability. In the case of UAVs, for example, suppliers are increasingly under pressure to increase mission length, with no reduction in payload capability, which of course puts pressure on payload SWaP. And because UAVs often fly in areas where GPS signals are being jammed, there is a need for a low-SWaP clock that also has the stability to maintain synchronization, even during extended periods of GPS outage. The clock should also provide a 1 PPS output, just like most GPS receivers do. And the 1 PPS output should be disciplined to the GPS signal, when it is available.
A portability baseline
Starting in 2011, chip scale atomic clocks (CSACs) will become small enough to fit on a circuit board alongside “regular” chips. In fact, if you examine the very first commercially available CSAC (Symmetricom’s SA.45s), you will see what equipment designers can now expect as a portability baseline:
. 16cm3 volume
. 35g weight
. ±5.0E-11 accuracy at shipment
. σy < 5 x 10-12 at τ = 1 hour short-term stability (Allan Deviation)
. <3.0E-10/month aging rate
. 115mW power consumption
. 1PPS input and output

The accuracy, stability and aging are all characteristic of a true atomic clock, while the power consumption is a 40x improvement over other atomic clocks offered as “low power.” The power consumption is also 10x to 20x less than what an OCXO requires, thus offering the possibility of better stability at much less power.
The CSAC achieves this impressively low power consumption by being not just a smaller version of an atomic clock, but a true reinvention. The resonance cell itself is miniaturized through the use of MEMS (micro electro-mechanical systems) technology, and then interrogated by a specially engineered low-power vertical cavity surface emitting laser (VCSEL).
The entire physics package is then surrounded by a high-quality vacuum, and it is this combination of very small size and very high thermal isolation that enables the physics package to be extremely power-efficient. All of the clock circuitry that surrounds the physics package has also been engineered for the lowest possible power consumption.
What this all means
System designers may not always need to know how an atomic clock works, portable or otherwise. But now that the first portable atomic clock is available, designers who previously thought ultra-stable timing was beyond the reach of portable applications can now reconsider. The Chip Scale Atomic Clock is here.

About the author:
Steve Fossi, director of New Business Development, Symmetricom Inc. (San Jose, Calif.)
Steve Fossi has been with Symmetricom for 3 years. He started as the Sales Manager for Symmetricom’s Space, Defense, and Avionics product line, and then moved to business development, where he has focused on introducing the Chip Scale Atomic Clock. He has been working closely with early adopters of the CSAC since the spring of 2009, and has also led the market introduction campaign for the CSAC.
Prior to Symmetricom, Steve was at Hewlett-Packard, and its spinoff Agilent Technologies, for 27 years. While at HP/Agilent, he served in marketing manager, R&D manager, and general manager roles. During his management career he was part of an HP management team that was the subject of a Harvard Business School case study, which is still taught at HBS today. He also lived in the Netherlands
for two years during one of his marketing assignments.
Steve’s technical background is in RF and microwave measurements, and in semiconductor test. He brings a great deal of experience to the real-world applications of Symmetricom’s oscillators and atomic clocks.
Take communications. In order to compress ever-increasing amounts of data into limited spectrum, increasingly higher bandwidth waveforms are needed — with ever-decreasing tolerance for clock drift. That means that the clocks used in today’s manpack radios — oven controlled crystal oscillators (OCXOs) and temperature controlled crystal oscillators (TCXOs) — may no longer be suitable. But an atomic clock would be— provided its size, weight and power consumption (SWaP) were low enough to be truly called portable.
Power consumption is also an issue for IED jammers. Today’s IED jammers have power requirements that can only be met by a vehicle’s generator. To make an IED jammer dismounted, all of the components (including the oscillator) would need to be low-SWaP. Also, all of the dismounted jammers would need to be tightly synchronized to allow predefined time slots in the signals (“look windows”) where friendly force communications could still get through. Such high synchronization is another reason to use an atomic clock.
Handheld GPS units and tactical UAVs would also benefit from an atomic clock’s stability — but, again, only if the clock also satisfied the SWaP profile for portability. In the case of UAVs, for example, suppliers are increasingly under pressure to increase mission length, with no reduction in payload capability, which of course puts pressure on payload SWaP. And because UAVs often fly in areas where GPS signals are being jammed, there is a need for a low-SWaP clock that also has the stability to maintain synchronization, even during extended periods of GPS outage. The clock should also provide a 1 PPS output, just like most GPS receivers do. And the 1 PPS output should be disciplined to the GPS signal, when it is available.
A portability baseline
Starting in 2011, chip scale atomic clocks (CSACs) will become small enough to fit on a circuit board alongside “regular” chips. In fact, if you examine the very first commercially available CSAC (Symmetricom’s SA.45s), you will see what equipment designers can now expect as a portability baseline:
. 16cm3 volume
. 35g weight
. ±5.0E-11 accuracy at shipment
. σy < 5 x 10-12 at τ = 1 hour short-term stability (Allan Deviation)
. <3.0E-10/month aging rate
. 115mW power consumption
. 1PPS input and output
The accuracy, stability and aging are all characteristic of a true atomic clock, while the power consumption is a 40x improvement over other atomic clocks offered as “low power.” The power consumption is also 10x to 20x less than what an OCXO requires, thus offering the possibility of better stability at much less power.
The CSAC achieves this impressively low power consumption by being not just a smaller version of an atomic clock, but a true reinvention. The resonance cell itself is miniaturized through the use of MEMS (micro electro-mechanical systems) technology, and then interrogated by a specially engineered low-power vertical cavity surface emitting laser (VCSEL).
The entire physics package is then surrounded by a high-quality vacuum, and it is this combination of very small size and very high thermal isolation that enables the physics package to be extremely power-efficient. All of the clock circuitry that surrounds the physics package has also been engineered for the lowest possible power consumption.
What this all means
System designers may not always need to know how an atomic clock works, portable or otherwise. But now that the first portable atomic clock is available, designers who previously thought ultra-stable timing was beyond the reach of portable applications can now reconsider. The Chip Scale Atomic Clock is here.

Steve Fossi, director of New Business Development, Symmetricom Inc. (San Jose, Calif.)Steve Fossi has been with Symmetricom for 3 years. He started as the Sales Manager for Symmetricom’s Space, Defense, and Avionics product line, and then moved to business development, where he has focused on introducing the Chip Scale Atomic Clock. He has been working closely with early adopters of the CSAC since the spring of 2009, and has also led the market introduction campaign for the CSAC.
Prior to Symmetricom, Steve was at Hewlett-Packard, and its spinoff Agilent Technologies, for 27 years. While at HP/Agilent, he served in marketing manager, R&D manager, and general manager roles. During his management career he was part of an HP management team that was the subject of a Harvard Business School case study, which is still taught at HBS today. He also lived in the Netherlands
for two years during one of his marketing assignments.
Steve’s technical background is in RF and microwave measurements, and in semiconductor test. He brings a great deal of experience to the real-world applications of Symmetricom’s oscillators and atomic clocks.
Navigate to related information



jcog123
3/2/2011 3:33 PM EST
Design is from a NIST project several years ago.
Sign in to Reply
Bear1959
3/4/2011 9:08 AM EST
This is so cool!
How much does one cost?
Sign in to Reply
Roncalli
3/5/2011 8:16 AM EST
Interesting product. Quick search shows low volume price of 1500 lead 20 weeks-- new product so one presumes it is early on the curve.
Now if just they could lower the height to fit on a one wide PCIe card in a blind swap 1U...
Sign in to Reply