1

A History of the Rubidium Frequency Standard

William J. Riley, Jr., Life Fellow, IEEE

Abstract--This paper is a history of the rubidium gas cell atomic military communications and now in widespread use by the
frequency standard, by far the most widely deployed type of telecom industry. Atomic clocks play a critical role in global
atomic clock. Since the early 1960’s, rubidium frequency stand- navigation satellite systems (GNSS). General atomic clock
ards (RFS) have offered an attractive combination of practicality references herein are [1] through [5].
and performance that spans a range of applications from low-
cost commercial devices for telecommunications timing, rugged- B. Requirements for an Atomic Frequency Standard
ized tactical units for military use, and high stability, high relia-
The fundamental basis of an atomic frequency standard
bility space clocks for global navigation systems. The paper de-
scribes their physical basis, the early scientific work that created (AFS) is utilization of an atomic resonance that is, in princi-
them, and their subsequent product development and widespread ple, an inherent and invariant natural property. In practice, to
commercialization from the perspective of an early contributor to implement an AFS, one must devise an apparatus that allows
this field. It emphasizes the author’s experiences at several or- one to observe an atomic resonance in a detectable manner
ganizations that continue to make RFS devices, and similar new and with minimal external disturbance.
atomic clock technology, smaller, better, and cheaper. That work
The requirements for implementing an atomic frequency
was carried out in the U.S. and elsewhere by a series of organiza-
tions and people following several distinct timelines and heritag- standard are means to confine the atoms, prepare them into a
es. The first complete RFS was designed in the late 1950’s at ITT particular state and then to interrogate them to probe their res-
in Nutley, NJ by Dr. Maurice Arditi with Dr. Thomas Carver of onant frequency2 [9].
Princeton, and that organization is now the largest user of space In the case of the common RFS, Rb atoms are confined in
Rb clocks for their GPS navigation payloads. The first commer- an inert buffer gas in a small glass “cell”, they are prepared in
cial RFS appeared in the early 1960’s from General Technology
a particular hyperfine atomic state by a process called optical
and Space Technology Laboratories. Some of the first commer-
cial RFS were built circa 1962 by Varian Associates in Palo Alto, pumping using an Rb spectral lamp, and their atomic reso-
CA, and that technological legacy continues today. nance is interrogated by means of RF energy applied to a mi-
crowave cavity. The Rb resonance at about 6.835 GHz is de-
Index Terms--Atomic clock, history, rubidium frequency tected optically by a photodiode that senses the light transmit-
standard. ted through the absorption cell.

I. INTRODUCTION C. Why Use An Atomic Clock?

T HE rubidium gas cell frequency standard has, over the last Relatively few atomic clocks are used as frequency stand-
60 years or so, become the dominant type of atomic clock ards per se but rather as sources of precise time. Maintaining
because of its combination of practicality and good perfor- time to the microsecond level (a rather large error for a fast
mance. While its design fundamentals have changed little over data stream) over a few days requires a frequency source error
that time period, there has been remarkable progress in adapt- well below the 1x10-11 level, which is a demanding require-
ing it to a wide range of demanding space-qualified require- ment for an ovenized quartz crystal oscillator 3. Thus an RFS
ments, rugged military applications and low cost commercial can be the most practical choice. Common RFS timing appli-
telecom usage. This paper will review some of that history. cations include cellular telecommunications, geophysical data
logging, GPS fast acquisition/robustness and frequency hop-
A. Atomic Clocks ping radios. High performance Rb clocks have become the
An atomic clock is a device that utilizes an atomic reso- device of choice onboard GPS satellites.
nance to produce a highly-stable frequency suitable as a fre-
D. Previous RFS Historical Reviews
quency reference and/or for precise timekeeping.
The rubidium gas cell atomic frequency standard (RFS or The first historical review of atomic frequency standards by
“Rb oscillator”1) is, by far, the most commonly-used form of Arthur McCoubrey appeared in 1966 [1]. That was quite early
atomic clock. Its scientific basis began in the 1950’s and its in the history of rubidium frequency standards, and it con-
commercialization was well underway by the early 1960’s, tained a brief description of their physical principles (which
first as a laboratory instrument, then primarily for tactical were fully in place by then), their performance, and several
examples of RFS products from General Technology and Var-
ian. McCoubrey followed that survey by another one 30 years
Rev. A Dec. 4, 2019. Citation: W. J. Riley, “A History of the Rubidium later at the 50th anniversary of the Frequency Control Sympo-
Frequency Standard”, IEEE UFFC-S History, http://ieee-uffc.org/about-
us/history/a-history-of-the-rubidium-frequency-standard.pdf, Dec. 2019.
sium in 1996 [16]. Roger Beehler wrote a historical review of
W. J. Riley is the proprietor of Hamilton Technical Services, Beaufort, SC
29907 USA (e-mail: bill@wriley.com). 2
1
The term “Rb oscillator” is a misnomer because an RFS is a passive AFS. As I recall, these fundamental requirements were first elucidated by Helmut
The Rb physics package is a passive discriminator not an active oscillator. Hellwig in various papers and tutorials, including [2].
Nevertheless, the term is reasonably descriptive and widely used. 3
For example, the LTE timing requirement is 1.5 s.
 2

atomic frequency standards in 1967 that contains an excellent Univ., 19536, [17]), optical detection (H. G. Dehmelt, Univ.
summary of the development of practical RFS devices [7]. He Washington, 1957), hyperfine filtration (T. R. Carver and C.
followed that with a progress report on commercial rubidium O. Alley, Princeton Univ., 1958)7, and optical detection, iso-
(and other) frequency standards written in 1973 [10]. In 1972, topic filtration, buffer gas effects, TC, light shifts, etc. (P. L.
Norman F. Ramsey published his first history of atomic clocks Bender, E. C. Beaty, NBS and A. R. Chi, NRL, 1958)8. Pho-
[8], which was followed by several others over the succeeding tographs of some of these key contributors to early RFS tech-
years [12], [15], [19]. Reference [5] contains a very brief RFS nology are shown in Fig. 1.
history. This paper updates those previous reviews, covering
only the passive rubidium gas cell atomic frequency standard
and the closely-related CPT and CSAC devices4.

II. PHYSICS
The history of rubidium frequency standards began as a
practical application of scientific research conducted mainly in
the U.S. following World War II.
Isidor I. Rabi (1898-1988) Alfred Kastler (1902-1984)
A. Rubidium as the Atom of Choice Father of the atomic clock Optical Pumping
Photo Credit: Wikipedia. Photo Credit: Wikipedia.
Rubidium is an alkali atom, one having a 37
single electron in its outer shell, and it is the Rb
energy state of that valance electron that is used Rubidium
as an atomic frequency reference. 85.4678
In particular, the so-called hyperfine resonance that repre-
sents the small energy difference between the two outer elec-
tron spin states is well-suited for that purpose, having a fre-
quency in the microwave range. Rb is uniquely suited for this
purpose because, as explained below, it exists as two common
isotopes, Rb-85 and Rb-87, which facilitate its state selection Robert H. Dicke (1916-1997) Thomas R. Carver (1929-1981)
Buffer Gas Linewidth Reduction. Isotopic Hyperfine Filtration. etc.
by optical pumping. etc. Photo Credit: Wikipedia. Photo Credit: Princeton Univ.
See Reference [234] for information about the physical
properties of Rb-87. References herein to general RFS theory
are [37] through [45].
B. Underlying Physical Research
The physics of atomic frequency standards is based on un-
derlying research, some conducted by Nobel laureates study-
ing the fundamental properties of atoms, whose apparatus
evolved into practical devices for frequency control. Much of
that evolution can be followed in papers presented at the Fre- Carroll O. Alley (1927-2016) Maurice Arditi (1913- )
quency Control Symposium (FCS) during the 1950’s. Atomic Isotopic Hyperfine Filtration Complete RFS
clocks are a prime example of scientific research leading to Photo Credit: Univ. of Maryland. Photo Credit: IEEE.
applied technology. References herein to early RFS scientific
papers are [23] through [36].
C. The First Atomic Clocks 6
Robert Dicke (1916-1997) predicted and experimentally showed that
The first type of atomic clock still being made today, the collisions that restrict the long-range motions of radiating atoms in a gas can
cesium beam tube device, was based on the early scientific suppress Doppler broadening, which is the basis for gas cell atomic frequency
standards. [28]. He invented the lock-in amplifier that uses synchronous
research by Stern and Gerlach and I. I. Rabi [14]. demodulation to detect weak signals, a process that is used in all passive
For rubidium gas cell devices, the seminal concepts in- atomic frequency standards, and was a founder of Princeton Applied Research
clude: optical pumping (A. Kastler, Paris, 19505, [6]), reso- (later, in 1977, an EG&G company), the first manufacturer of those devices in
1962. He also invented the radiometer and just missed discovering the cosmic
nance line narrowing by a buffer gas (R. H. Dicke, Princeton microwave background radiation.
7
I visited Dr. Alley at his home near the University of Maryland in September
2008, and asked him about the invention of the rubidium frequency standard.
He did not give a definitive answer, and it seems that a number of people in
4
This document is “a history” not “the history” of the rubidium frequency several groups brought it all together in the late 1950’s. The scientists moved
standard. I am not a historian but rather a participant in that endeavor, whose on to other things and the engineers took over from there.
8
complete story would require a broader perspective. Robert Dicke led the group at Princeton. Tom Carver was an instructor
5
Alfred Kastler (1902-1984) invented the method of optical pumping to excite starting in 1954 and became a full professor in 1967; he consulted with ITT
the energy levels of atoms in 1950, which is the basis of the optical state (e.g., Maurice Arditi). Carroll Alley was a grad student there when the
selection used in most Rb frequency standards. He received the Nobel Prize in seminal Rb groundwork was done.
Physics for this work in 1966.
 3

The first suggestion for building an ‘atomic clock” is gen-

erally credited to a January 1945 lecture by I. I. Rabi of Co-
lumbia who won the 1944 Nobel Prize for his work on nuclear
magnetic resonance. In 1949-1950, F. Bitter of MIT and A.
Kastler in Paris discovered the technique of optical pumping,
for which Kastler received the Nobel Prize in 1966. That was
followed by the 1953 suggestion by R.H Dicke of Princeton
that the Doppler broadening of atomic resonance lines could
Hans G. Dehmelt (1922-2017) Earl C. Beaty (1930-2018) be reduced by a buffer gas, the method of optical detection by
Optical Detection Buffer Gasses, H.G. Dehmelt in 1957 (recipient of the 1989 Nobel Prize), and
Photo Credit: Wikipedia. Isotopic Filtration, etc.
Photo Credit: JILA/Univ. CO. the use of isotopic filtration in 1958 by T. R. Carver and C. O
Alley of Princeton. Meanwhile, P. L. Bender, E. C. Beaty of
NBS and A. R. Chi of NRL13 were studying buffer gas and
light shift effects. By that time these and other investigators
had established the scientific basis of the rubidium gas cell
atomic frequency standard that continues to be utilized today.
That scientific work was quickly utilized to build prototype Rb
frequency standards, and by the early 1960’s commercial units
became available.
Several patents from this era describe arrangements and
A. R. Chi (1920-1982) Peter L. Bender (1930-) show block diagrams that are little different from RFS designs
Buffer Gasses, Buffer Gasses, today (see Figs 2 and 3)14. References herein for RFS patents
Isotopic Filtration, etc. Isotopic Filtration, etc.
Photo Credit: IEEE. Photo Credit: JILA/Univ. CO. are [235] through [322]15.
Fig. 1. Some Key Early Contributors to RFS Technology. These
investigators established the scientific basis for rubidium frequency
standards.

Rabi, Dehmelt and Kastler were Nobel laureates. Dicke

probably should have shared one for discovery of the cosmic
background radiation (CMB).
Many scientists and engineers contributed to the develop-
ment of the rubidium gas cell atomic frequency standard
(RFS), and it is not possible to attribute its invention to any
single person or organization9, 10, 11, 12 [13].

9
A case can be made that the first serious effort to develop an actual gas cell
atomic frequency standard was undertaken by Maurice Arditi at the ITT
Federal Telecommunications Laboratories in Nutley, NJ where he worked
until 1965. Dr. Arditi was the right person at the right place at the right time to
do that. He was an ITT employee in France who came to the U.S. shortly Fig. 2. Early Rb Physics Package. This figure from a circa 1960 Space
before World War II with degrees in both engineering and physics. ITT was Technology Laboratory patent shows all the ingredients of a modern RFS
then, and remains today (as Harris) at the same location, in the avionics physics package: Rb lamp 14 and exciter 16, Rb-85 filter cell 24, Rb-87
business, and was quite close to Princeton where Robert Dicke’s physics absorption cell 10, C-Field coil 11 and photodetector 32. It mentions
group (e.g., Tom Carver) had laid the theoretical and experimental applying the microwave excitation with a cavity per Dicke’s 2,940,050
groundwork for this technology. Arditi began with sodium, mainly because its patent.
lower hyperfine frequency made the microwave hardware easier. He then
moved on to cesium and then rubidium. One supposes (but I was unable to
confirm) that the final preference for Rb was (as it remains today) because of
isotopic filtration).
10
Dr. Alley’s Wikipedia page includes the following: “His PhD dissertation
developed new methods of detecting the ground state hyperfine transition in
optically-pumped rubidium vapor for the type of atomic clock invented by
Princeton professors Robert Henry Dicke—Alley’s thesis advisor—and
Thomas R. Carver. This type of atomic clock was later used in the U.S. Global was developed initially at Princeton, notably by T.R. Carver, with
Positioning System and the European Union’s Galileo Global Positioning contributions by others, including the present writer…”.
System.” 13
Andrew R. Chi joined NRL in 1957 and collaborated with physicists from
11
I find it ironic that ITT in Nutley, NJ, now part of Harris and the largest user NBS in developing gas cell frequency standards using optical pumping tech-
of high performance space Rb clocks, and one of the main spawning grounds niques. He and his colleagues were the first to measure the Rb-87 hyperfine
of that technology, announced in a 1958 news release a “lightweight atomic frequency, and the first to measure the pressure shift and TC of various buffer
clock” as an “aid to space navigation in the future” but then, in 1965, declared gasses in a gas cell frequency standard.
14
“Dr. Arditi’s atomic clock research has not been considered by ITT as worthy See U.S. Patent Nos. 2,836,722, 2,882,493, 2,884,524, and 2,927,278.
15
of continued financial support”. He was clearly far ahead of his time. While relatively few of these patents are referenced in the text, they help to
12
In [13], on p. 396, Dr. Alley says “… the rubidium vapor gas cell type using show the evolution of RFS technology and the persons and organizations
optical pumping and optical detection, which was proposed by R.H. Dicke and involved.
 4

Fig. 4. RFS Block Diagram. This block diagram applies to any passive
atomic frequency standard.

A more detailed diagram of a classic lamp pumped Rb

physics package with separate filter and absorption cells is
shown in Fig. 518. This arrangement is still in use today (e.g.,
in the Excelitas GPS RAFS).
Fig. 3. Early RFS Block Diagram. This figure from a circa 1960 Varian
patent shows all the ingredients of a modern RFS block diagram: Rb lamp Magnetic Shield

19 and oven assembly with isotopic filter 22, absorption cell 11 / Lamp Oven Filter Oven Cavity Oven
RF
Excitation
microwave cavity 12 assembly, C-field coil 17, magnetic shield 13, Lamp
Filter Absorption
photodetector 28, preamplifier, synchronous detector 42, VCXO 43 & 33, Lamp Rb-87
Rb-85 Rb-87
Photo- Signal
phase modulator 34, and RF multiplier chain 37-39. Arditi shows a similar Exciter Lamp Detector Out

Cell Cell
block diagram in an ITT patent 3,174,114 filed in 1959 for a Na-based gas Coil

cell clock. Varian patent 3,363,193 filed in 1966 shows a completely C-Field Coil
C-Field
Current
modern RFS block diagram that includes a synthesizer scheme.
(3) Oven Temperature Sensors and Heaters

III. TECHNOLOGY Fig. 5. Rb Physics Package. This Rb physics package diagram shows a
separate isotopic filter cell.
Let us now examine some of the technology used in rubidium
atomic frequency standards and the way it has evolved. B. Isotopic Filtration
The key property of rubidium that makes it uniquely suited
A. The First Rubidium Frequency Standards
for an optically pumped gas cell atomic frequency standard is
The burst of intensive work during the late 1950’s (e.g., see the fortuitous overlap in the hyperfine optical spectrum be-
the 1958 FCS Proceedings) makes it difficult to ascribe the tween the Rb-85 and Rb-87 isotopes that allows increased
first practical RFS to a single individual or group. Neverthe- optical pumping efficiency by isotopic filtration as shown in
less, it seems to me that one can credit that to Maurice Arditi Fig. 6.
of ITT Laboratories and Thomas Carver of Princeton Univer-
sity (a member of Robert Dicke’s group)16, 17. The January
1963 Proc. IEEE paper by M. Arditi and T. R. Carver [53] is a
good overall reference to their work. One can also cite similar
work by Carpenter, Packard & Swartz and others. Much of
that work was described in papers at the 1958 Frequency Con-
trol Symposium. Prototype RFS units soon followed in the
early 1960’s as reported by Arditi & Carver, Carpenter, Pack-
ard & Swartz and others. Commercial units became available
from Varian and General Technology soon thereafter. Refer-
ences herein to early RFS R&D are [46] through [58].
A basic block diagram of an RFS is shown in Fig.4. This
block diagram applies to any type of passive atomic clock
wherein a crystal oscillator is locked to an atomic reference.
Fig. 6. Rb Lamp D2 Line Optical Spectrum. This spectrum is for a mixed-
isotope Rb lamp. The Rb-87 lines are stronger and they are separated by
the 6.835 GHz hyperfine frequency. Notice the overlap between the two
Rb isotopes for the lower hyperfine level. Photo Credit: Efratom.
16
I contacted Dr. Arditi via e-mail in 2009 thanks to detective work by Tony
Lomnicki at ITT/Exelis, who traced him down in France via his ITT C. Discrete and Integrated Filter Cells
retirement records. Maurice Arditi was born on March 1, 1913 in Paris, and An RFS can use either a separate Rb-85 filter cell or an in-
was still living there at age 96 at that time, when he kindly shared some stories
about his work on Rb clocks. tegrated resonance cell containing both Rb-87 and Rb-85
17
I had the privilege sometime around 2005 to examine several of Dr. Arditi’s (likely as natural Rb) [275]. Both arrangements work well, but
notebooks in the ITT Nutley, NJ library. One could feel him struggling to do
something that had never been done before. I recall somewhat the same
18
experience at General Radio when Herb Stratemeyer and I searched for the Rb In early RFS, the physics package was often call the “optical assembly” or
resonance with our experimental apparatus, and we, of course, knew where to “optical microwave assembly” (OMA, but not a German grandmother).
look thanks to Arditi’s and others work. “Physics package” has nuclear weapons implications, but that’s another story.
 5

the discrete filter cell allows better absorption cell homogenei-

ty, high S/N, ease of light shift nulling, lower sensitivity to
microwave excitation power. Two examples of Rb resonance
cells are shown in Fig. 7.

Fig. 8. Buffer Gas Linewidth Reduction. A lower relaxation rate narrows

the resonance line and improves its Q. There is an optimum buffer gas
pressure as a tradeoff between confinement and collisional broadening.

Using the same Rb-87 isotope as the absorption cell, it is

automatically at the right nominal wavelength, and its plasma
Fig. 7. Rb Resonance Cells. Left: Symmetricom 8130A resonance cell 1’’
diameter x 1” long. Right: Symmetricom X72 resonance cell about 0.3” broadens its spectrum sufficiently so that it is stably insensi-
long. Photo Credit: Symmetricom. tive to small variations in operating conditions. The lamp con-
D. Buffer Gas tributes negligible noise above that of the shot noise intrinsic
to the optical detection process.
Absorption/resonance cell buffer gas is used to narrow the
resonance line by confining the Rb atoms to reduce wall colli-
sions and Doppler shift, as shown in Fig. 819. Nitrogen is uni-
versally used to quench re-radiated pumping light. Various
buffer gas species have different pressure shift 20 and tempera-
ture coefficients thereby determining cell frequency offset and
TC. These buffer gas frequency shifts make the RFS a “sec-
ondary” (albeit stable) standard of frequency. Cell envelope
glass is selected for low microwave loss and low helium per-
meation21. Wall coatings are seldom used instead of buffer gas Fig. 9. Rb Lamp. This is a GPS RAFS lamp in its screw-in holder. Photo
in an RFS22. Credit: PerkinElmer.

E. Rb Lamps The classic problem with Rb lamps has been life and relia-
An electrodeless Rb discharge lamp (see Fig. 9) is a remarka- bility due to rubidium depletion as the excess Rb diffuses into
bly effective and low noise optical pumping source. the glass envelope. There is a tradeoff between having suffi-
cient Rb fill for satisfactory life and too much excess Rb that
increases lamp noise. That problem was solved by Tom Lynch
at EG&G circa 1980 when he invented a way to precisely
measure lamp Rb fill using a differential scanning calorimeter
(DSC) to measure the amount of heat energy required to melt
it, as shown in Fig. 1023, [115]. References herein for Rb
lamps are [111] through [117].
19
Buffer gases are commonly used to reduce the linewidth of microwave
transitions in alkali atoms. The lifetime of the atomic state determines the
width and Q of the resonance. To improve the atomic Q, around 10 Torr of an
inert gas such as Ne, Ar or N2 is added to the cell. The alkali atoms collide
frequently with the buffer gas and therefore diffuse slowly throughout the cell
before they hit the wall. The atomic Q is increased to about 10 8, which results
in a much better frequency stability than could be obtained with a pure Rb
vapor alone.
20
Pressure shift coefficients are positive for light gasses (e.g., He, Ne) and PerkinElmer
negative for heavy gasses (e.g., Kr, Xe). The pressure is usually measured at DSC
room temperature where the cell is filled and the frequency shift is measured
at the cell operating temperature. The values are quite similar for both large Fig. 10. Rb Lamp Calorimeter Record. The area under the Rb lamp
and small, rubidium and cesium cells. calorimeter record indicates the amount of heat energy required to melt it,
21
Helium is highly permeable through most glass. Atmospheric He can dif- which, by knowing the Rb heat of fusion, determines its mass. Photo
fuse into an absorpsion cell, changing its buffer gas mixture and raising its Credit: PerkinElmer, Plot Credit: EG&G.
frequency. In subsequent space (vacuum) operation, the frequency can age
lower. A copper cell with brazed sapphire windows and a pinchoff would
have excellent cleaning, bakeout and sealing properties, and could also serve
23
as a cavity/oven, but this has never been widely used (see U.S. Patent No. Zero in-orbit GPS RAFS Rb lamp failures have occurred since adopting
6,215,366). DSC lamp screening. A somewhat similar but less severe issue regarding lamp
22
A wall coated cell can have a narrow resonance line width but it also com- Xe buffer gas fill was resolved more recently by non-destructively measuring
monly has an unacceptably large temperature coefficient. A wall coated cell the Xe buffer gas pressure by operating the lamp as an absorption cell and
wthout buffer gas to confine the Rb vapor in a TE 111 microwave cavity must measuring its frequency offset.
have a septum to separate the two regions of opposite RF H field polarity.
 6

F. Lamp and Cell Processing cut-away physics package is shown in Fig. 12. A reference
Rb lamps and cells are processed on high vacuum systems herein for lamp exciters is [118].
like that shown in Fig. 11. The general procedure is to fabri-
cate the glass envelopes and their tubulations, chemically
clean them, attach them to a manifold on the vacuum system,
bake out, plasma clean and pump down the empty lamps or
cells, break an Rb ampoule inside the manifold and distill Rb
into the lamps or cells, introduce the appropriate buffer gas at
the correct pressure, pull the lamps or cells off the system, and
finally shorten their tubulations24, 25.


Fig. 12. Operating Rb Lamp in Cut-away Physics Package. Note the

excitation coil that surrounds the lamp in its oven. Photo Credit: Temex
Neuchatel Time.

H. Optical Filtration
An Rb lamp produces an output that contains spectral lines
from both its Rb fill and its buffer gas (typically Xe), all of
which contribute shot noise at the photodetector. An optical
filter can be used to remove much of the unwanted light and
thus improve the S/N ratio and short-term stability.
I. Pulsed Light
In principle, light shift could be avoided in an RFS by puls-


ing the pumping light and doing the interrogation in the dark.
Fig. 11. Rb Lamp Processing. A manifold of Rb lamps attached to a high
vacuum processing system. Photo Credit: EG&G. This technique has not, however, been widely used31.
G. Lamp Exciters J. C-Field
Rb lamp exciter circuits have been the subject of much de- A DC magnetic bias field or “C-Field” on the order of 50-
velopment effort to assure reliable lamp starting and stable 250 mG is necessary to provide a magnetization axis and to
operation. These circuits require about ½ watt of RF power at separate the Zeeman lines so that the frequency lock servo can
around 100 MHz26. The Rb lamp is usually excited inside a lock on to the central “field independent” line 32. It can also be
series-tuned coil, and a power oscillator has been found to used to adjust the RFS frequency, but, for best stability, it is
work better than a separate source and amplifier. Early RFS better to operate at a fixed minimum C-field and use another
designs used vacuum tube lamp exciters, and today’s rugged means (e.g., a synthesizer) for tuning [167]. It may be neces-
RF power transistors make these circuits very reliable. Starting sary to boost the C-field during lock acquisition to avoid lock-
requires developing a high resonant RF voltage across the up on a Zeeman line.
coil27, 28, 29, 30, while the lower-impedance lit lamp is driven By periodically commutating the C-field polarity it is pos-
mainly by the RF magnetic field. An operating Rb lamp in a sible to reduce RFS sensitivity to external DC magnetic fields
[168]33, 34. References herein for the RFS C-field are [167] and
24 [168].
The ultra-high vacuum cell processing systems at
EG&G/PerkinElmer/Excelitas originally came from General Radio/GenRad K. Magnetic Shielding
and thus date back to the mid-1960’s, although some of their pumps, valves,
gauges, etc. have, of course, been replaced. All RFS units require magnetic shielding to establish a re-
25
The need for ultra-high vacuum (say below 10-10 Torr) before lamps and gion of uniform C-field in the physics package and to reduce
cells are filled with Rb and buffer gas is debatable considering the finite purity
of the buffer gas and glass out-gassing during tipoff. The most important thing
their sensitivity to external magnetic field variations. Refer-
is probably the condition of the glass envelope surface, and the chemical and
ion bombardment processes used to clean it.
26 31
It is important to shield this RF energy to avoid unwanted conducted and The most serious attempt at pulsed light operation of a conventional lamp-
radiated emissions from the RFS. pumped RFS was probably that done by Tom English et al. at Efratom [113].
27
Cosmic radiation is thought to play a role in initial lamp ionization. But the Rb lamp isotopic mix and Rb-85 filter cell recipes can do an excellent
28
Lamp starting after prolonged storage has been a concern but never an job of light shift suppression. Pulsed light is a more attractive possibility
actual problem. Even if the rubidium forms a conductive film that inhibits when a diode laser is used as the pumping source.
32
starting, RF induction heating will clear it as long as the lamp exciter oscil- The term “C-Field” apparently originated with a label on Rabi’s NMR
lates. Experience with GPS Rb clocks has shown that they start OK after apparatus.
33
decade-long in-orbit storage. Periodic reversal of the C-field has been used to reduce RFS magnetic
29
A room full of Rb lamps can be lit with a single camera flash. sensitivity. The Collins AFS-81 used a Hall sensor to compensate for external
30
Use of a radioactive buffer gas to initiate a plasma discharge was considered magnetic fields.
34
for Rb lamps at EG&G in Salem, MA, which had a licensed Kr-85 facility. C-field commutation has been used by AccuBeat and by Stanford Research
But that was never tried, it being considered unnecessary and not worth the Systems in their PSR10 rubidium frequency standard. In the latter, clever
trouble. Bob Vessot alluded to using it to aid H-maser dissociator starting, but digital signal processing avoids problems with interference and achieves a x25
again it was probably never actually done. reduction in external magnetic field sensitivity.
 7

ences herein for RFS magnetic shielding are [169] through Any passive AFS must employ some means to acquire
[173]. lock-up on the atomic resonance. While that can happen by
itself if the active frequency source is highly accurate, in most
L. Rb Signal Modeling
cases it is necessary to search for lock by sweeping the crystal
The discriminator signal from an Rb physics package is oscillator frequency39.
well understood and can be accurately modeled using methods
devised by R.P. Kenschaft at General Radio in 1969 [164] and N. Radioactivity
later by Jim Camparo and Bernardo Jaduszliwer at the Aero- An RFS does not utilize radioactivity, and needs no special
space Corporation35. The former devised the so-called precautions for shipping or handling40, 41, 42.
Kenschaft model (see Fig. 13) and the latter a refined version
O. Physics Package Optimization
of it36, 37.
It is desirable to optimize RFS physics package perfor-
Bandpass Filter Lowpass Filter mance by minimizing the temperature coefficient of its vari-
2
Absorption
Depth +
s+1
__________________
(H1)
_____
Optical
Absorption
ous elements. In the case of an integrated resonance cell, the
Amax
(s+1)(s+2)+(H1)
2
(s+)
Signal
_
A(t, )
lamp oven TC can be nulled by adjusting the lamp Rb isotopic
Lowpass Filter

1 H1=Rabi Rate
ratio [277], and the cell oven TC can be nulled by adjusting
X
_____
X 1=Population Relaxation Rate the cell buffer gas mixture. In the case of a discrete filter cell
(s+2)
_
Frequency Offset, 
2=Coherence Relaxation Rate
that shares an oven with the absorption cell, the lamp oven TC
can be nulled with the filter cell length or temperature and the
Control
Voltage
+ Modulation
filter/absorption cell TC nulled with the absorption cell buffer
Block Diagram of Kenschaft Model of Rb Signal gas mixture [287]. If a separate discrete filter cell oven is used,
it will necessarily have a fairly large negative TC, but its set
One important application of the point temperature allows nulling of the light shift coefficient
Kenschaft model was to analyze [121].
the response the response of an
Rb physics package to servo P. RF Synthesis
upset such as can occur due to
transient radiation [165]. Nature has not provided us with an atomic resonance at a
standard 10 MHz related frequency, but most RFS units need
R.P Kenschaft to produce that output43 and many synthesis schemes have
Photo Credit: GR 1969 Picture Book. been used for that purpose44 [5]. The absorption cell buffer gas
Fig. 13. Kenschaft Model of Rb Signal.
frequency offset must match the synthesis scheme, and it is
References herein for Rb signal parameters are [160] desirable that the buffer gas pressure be optimized for a nar-
through [163] and for the Kenschaft model are [164] through row line width, and that it is relatively low to ease its fill toler-
[166]. ance and to reduce barometric sensitivity.
M. Lock Detection and Acquisition
It is important that there be an indication when an RFS fre- 39
Acquisition sweep is quite easy to perform by simply inserting a small
quency control loop is locked, and fortunately that quite easy offset at the input of the servo integrator. Without a discriminator signal, the
control voltage will sweep until it finds the atomic resonance; at that point the
to do by detecting the presence of the strong 2 nd harmonic sig- signal will overpower the sweep, the system will lock up at a small offset, and
nal that exists only when the frequency of the applied interro- the lock detector can then remove the sweep. The acquisition thus happens
gation RF is near the center of the resonance line and the fun- automatically under control of the lock detector as long as there is adequate
recovered signal. During warm-up the acquisition system must sweep up and
damental signal is nulled by the servo loop38. down until sufficient signal is produced. See Varian U.S. Patent No.
3,364,438 for an early RFS automatic search sweep scheme.
40
Natural rubidium contains two common isotopes, 72% stable Rb-85 and
35
Bernardo Jaduszliwer developed and verified Python code that implements 28% slightly radioactive Rb-87, a 0.28 MeV beta emitter whose 50 billion
an updated version of the Kenschaft model in Aerospace Report TOR-2017- year half-life is longer than the age of the universe. A rubidium frequency
02380 of September 2017 (Restricted Distribution). standard contains under a milligram of Rb-87, and its radioactivity of about 20
36
The discriminator slope is the magnitude of the fundamental signal response pCi is less than that of a banana. Plus, in an RFS, it is surrounded by metallic
per unit of fractional frequency detuning; it is what the physics package “gets oven material and magnetic shields.
41
paid” to produce, and is optimized as a function of the pumping light It can be a hassle to travel by air with an “atomic clock”.
42
intensity, filter and absorption cell (or integrated resonance cell) lengths and Another type of Rb clock is the one used for long-term geological dating by
temperatures, the RF interrogation power, and the modulation rate and comparing the ratio of Rb-87 to its Sr-87 decay product.
43
deviation. Excess light and/or RF power broadens the line. Cell temperatures RFS units are sometimes allowed to produce a “natural” frequency output.
determine the Rb vapor pressure and density. Absorption/resonance cell buffer For example, the EG&G/PerkinElmer/Excelitas GPS satellite Rb clock has a
gas types and partial pressures determine frequency offset. Filter cell 13.40134393 MHz output, exactly 1/510 of its atomic resonance, which
length/temperature determines light shift. simplifies its RF chain. A VCXO at a direct submultiple of the Rb frequency
37
Kenschaft’s model was implemented mathematically, in block diagram can also be used with a DDS (e.g., the PerkinElmer LPR10 commercial RFS).
44
form, and as C language software. Ken Lyon at EG&G devised analog circuit Between 1961 and 1972, atomic clocks were required to make rate
and PC application implementations of it. (frequency) adjustments so that their time stayed in agreement with the
38
Slow FM produces a squarewave recovered signal with transients as the Earth’s rotation. This was awkward from both a hardware and an operational
frequency changes, passing through the resonance peak twice per modulation standpoint and was abandoned in favor of today’s leap second scheme. To
cycle. As the FM rate is increased, this produces a strong 2nd harmonic comply, some atomic clocks of that era (e.g., the HP 5065A [59]) used
component with the fundamental component nulled at the center of the synthesizers as so-called “time-scale changers”.
resonance line.
 8

One very common RFS synthesis scheme (used by the magnetic bias field, but these do not have to be aligned with
Efratom RFS units and many others since the 1970’s, (see Fig. the optical axis, as they are in most cylindrical cavities. For
14) divides a 10 MHz output crystal oscillator by 2 to 5 MHz example, see the small so-called Jin resonator used in the
and again by 16 (both easy binary factors) to 312.5 kHz, and Symmetricom X-72 [178]. References herein for RFS micro-
mixes them additively to 5.3125 MHz. That signal is then sub- wave cavities are [174] through [182].
tracted from 6840 MHz from a harmonic and step recovery
S. Environmental Sensitivity
diode (SRD) multiplier chain to obtain Rb excitation at
6834.6875 MHz, where the SRD also serves as the final mix- AFS environmental sensitivity has been studied quite ex-
er. PM is performed near the bottom of the multiplier chain tensively46 [82]. The principal RFS environmental sensitivities
where minimal modulation index is required. A modern ver- are:
sion of this scheme substitutes a DDS for the 5.3125 MHz  Temperature and Temperature Slew
signal, providing tunability and low distortion FM [78], [80].  Thermal and Power Cycling
 Barometric Pressure
 Humidity
 Supply Voltage
 Acceleration, Vibration, G-Sensitivity & Orientation
 Mechanical Shock
 Magnetic Field
 EMI Susceptibility
 Radiation (Total Dose, Transient, SEU)
 Relativity

More details about these sensitivities are given below and

Fig. 14. Popular RFS Synthesizer Block Diagram. herein in references [82] and [83].
T. Baseplate Temperature Control
RF chain considerations include buffer gas offset, phase
noise, spectral purity, tunability, microwave power stability Active control of external RFS temperature, most often ap-
and modulation distortion. plied via its baseplate in a space application, is an effective
way to reduce its overall TC to essentially zero (at the expense
Q. Servo Modulation of additional power, and some increase in complexity, size and
Another distinguishing aspect of an RFS design is the way weight). A good example of this is the Excelitas GPS RAFS
low frequency FM is applied to its microwave interrogation in where an attempt was made by ITT to use software tempera-
order to support synchronous detection in its frequency lock ture compensation but that was found of limited effectiveness
loop. Classically, this was done by applying triangular audio because of the many small internal TC contributors and their
modulation to a varactor diode in an RC network at the lower differing time constants. Active baseplate temperature control
end of the RF chain to produce squarewave FM. Lower distor- is better. Figure 15 shows the thermal characteristics of a typi-
tion FM can be achieved using an all-pass network [121] since cal baseplate temperature controller (BTC).
modulation distortion is an important error factor. Nowadays,
40
perfect squarewave FM is applied via a direct digital synthe- 39 W

Baseplate Temperature Controller

sizer (DDS). Sinusoidal FM has also been, and still is, used 35
Thermal Characteristic

[81]. It is usually necessary to avoid modulating the RF out- 30 Slope=0.7 W/C

put45.
Max
BTC (0.6 W/C Insulator Thermal Conductance
Power Plus Radiative Heat Transfer)
25
25 W
Total Power, Watts

R. Microwave Cavities 20

A traditional RFS requires a microwave cavity to produce a 15

14 W
longitudinal RF magnetic field to interrogate the Rb atoms at 25 C Minimum Operating Range 11 C Margin
10

their 6.835 GHz hyperfine frequency. A classic unloaded Fixed

RAFS
Power

TE011 cavity at that frequency is the size of a coffee mug, 5 14 W

-11 14

which would in turn determine the size of the absorption cell 0

-20
C

-10 0 10
C

20 25 30 40 45 50

(say 1” diameter and length), the size of the cavity oven, its
BTC
S/V Panel Temperature, C Setpoint

thermal and magnetic shielding, and the size of the overall Fig. 15. BTC Thermal Characteristic. In this plot, the warm-up demand
unit. Somewhat smaller TE111 cavities are widely used, dielec- power at the left, the set point is at the right, and the operating power range
is between the two dots. The slope of the BTC characteristic indicates the
tric loading can make the cavity smaller, and other arrange- thermal conductance of its insulator plus radiative heat transfer.
ments have been tried, but it is often the case that the design of
an RFS starts with the cavity and proceeds outward from
there.
46
Space is a generally benign place for gas cell rubidium frequency standards
(e.g., no barometric pressure changes). After launch, except for radiation,
It is necessary that the RF H-field be aligned with the DC these clocks perform well in orbit, and radiation exposure can be managed by
proper component selection, screening and shielding. The Rb physics package
45
Servo modulation on the RF output may be tolerable for an RFS used only itself is not particularly sensitive, and zero-g operation is not a problem. Tin
as a clock. Another case is when the RFS output is used to drive a DDS, whisker growth is a concern for all space electronics. Magnetic torquer inter-
which has the effect of masking the discrete FM sidebands as wideband noise. ference is another concern for an Rb clock onboard a spacecraft.
 9

U. Vibration Sensitivity X. Transient Radiation

In general, an RFS is less sensitive to mechanical vibration Few RFS units have to contend with transient radiation, but
than a quartz crystal oscillator. Nevertheless, vibration can those that do require special design considerations because
affect the stability of an RFS, and this can be an important their large area photodetectors are excellent radiation sensors
consideration in tactical environments. That sensitivity is whose response can upset the frequency control servo. One
mainly associated with two mechanisms: (a) disturbance of the effective means of hardening disables the servo until the unit
physics package light beam when the vibration frequency is at can recover51, 52.
or near that of the servo modulation, and (b) disturbance of the
Y. High Temperature Operation
quartz crystal oscillator when the vibration is at or near twice
that of the servo modulation47. The former is reduced by a High temperature RFS operation requires that their oven set
rigid physics package construction, while the latter may re- points be raised sufficiently to maintain temperature control.
quire a crystal resonator with low acceleration sensitivity or That can reduce the Rb discriminator signal and compromise
active compensation. Both can be reduced by vibration isola- performance. A small (thin) absorption cell is advantageous.
tion, although that is awkward at low vibration frequencies. Raoult’s law can be utilized to depress the Rb vapor pressure
References herein for RFS acceleration, vibration and g sensi- by using a mixture of another alkali element such as potassi-
tivity are [186] and [187]. um. The Rb vapor pressure characteristic is fundamental to
RFS physics package design [234].
V. Fast Warm-Up
Z. Power Cycling Endurance
Fast warm-up can be a requirement for some RFS applica-
tions. That process involves lamp starting, oven warm-up, Power cycling endurance is a critical aspect for RFS units
frequency servo lockup and physics package thermal stabiliza- that must operate intermittently, especially if fast warm-up is
tion. Lamp starting can be immediate, particularly if aided by also required. Careful physics package thermal design is need-
boosted RF power, a HV electrode or even photoionization. ed to assure that fatigue stress is minimized. Design verifica-
Fast oven warm-up involves not only excess demand power tion and/or ESS testing is necessary to assure acceptable
but, more importantly, applying it directly to the internal phys- warm-up endurance, which applies especially to an RbXO
ics package elements. Servo lock acquisition can be speeded (See: EG&G RbXO and [156] through [159]).
up by special circuitry such as fast sweep and a fundamental AA. RFS Testing
signal detector. An EG&G fast warm-up missile borne RFS RFS testing has also been studied quite extensively and is
design was able to achieve lockup in only 7 seconds from described in Reference [84]. In general, the electrical and me-
+5°C [155]. chanical aspects are much the same as any electronic instru-
W. Radiation Hardening ment, but RFS testing requires high resolution frequency
Some applications require that an RFS be hardened again measuring equipment and attention to its specific environmen-
nuclear radiation, for example when the unit is intended for a tal sensitivities. References herein for RFS testing are [84]
space or military environment. This may have design impact through [110].
on both the physics package and the supporting electronics. It BB. Rubidium Masers
was a significant consideration for tactical RFS units in the
1980’s at the end of the cold war, and still is for GPS satellite Active rubidium masers were investigated in the early days
clocks48, 49, 50. References herein for RFS radiation hardening of Rb atomic clock research, notably by Jacques Vanier and
are [206] through [209]. his group at Laval University, and by W. Stern, P. Davidovits
and R. Novick. While excellent short-term stability was ob-
tained, these devices had all of the disadvantages of the H-
47
The effect of phase noise on the microwave interrogation signal at twice the maser (size, complexity) without any real advantages, and
servo modulation rate (and the related 2fmod EMI or vibrational modulation
effect) was elucidated by R. P. Kenschaft in a 1969 appendix to a General work on them ended without any commercial units ever being
Radio unpublished design report [135] but was not widely appreciated until a made53 [217].
1991 IEEE I&M paper by C. Audoin, et al. [160].
48
RFS radiation hardening generally involves total gamma dose, CC. RFS Evolution
neutron/proton fluence, transient radiation and single event upset
considerations. Their physics packages are quite immune except for the
The general evolution of rubidium frequency standards can
photodetector. Their electronic hardening is similar to most such circuitry, and
centers on the analog circuitry.
49 51
An RFS radiation hardening analysis has the very desirable side effect of The effect on clock time error is probabilistic, depending on when during
also performing a detailed worst case circuit analysis by a highly-skilled ana- the servo modulation cycle the event occurs. Thus it is necessary to perform
lyst (e.g., Terry Flanagan/IRT, Dave Swant/GE for GPS clocks). Modern many FXR shots to determine a worse case. Ken Lyon and I held the record
circuit analysis tools are remarkably effective. That modeling relates closely for the most shots in a day at the GE Valley Forge facility.
52
to initial and end of life (EOL) performance margins. Furthermore, this infor- An FXR machine is an impressive device. Its thick Ta target is deformed
mation is important for verifying component stress levels and making reliabil- by the energy delivered to it by the electron beam. The TRW machine we
ity predictions. used in El Segundo had (for a few nanoseconds) more power than all of LA.
50 53
Ken Lyon/EG&G is an amateur entomologist. During total gamma dose Active Rb masers were made using both buffer gasses and wall coatings
RFS tests, he would place a bug in a matchbox near the unit under test. After (e.g., paraflint wax). They require large high-Q cavities to self-oscillate (a
one test at U. Mass. Lowell, the poor thing turned a sickly grey (at about a passive RFS operates at only about 1% of the oscillation threshold). Jacques
human fatal 10k rads). But amazingly it recovered to apparent full health after Vanier worked on an Rb CPT maser at Kernco.
a few days.
 10

be roughly summarized in a timeline as follows: That is due primarily to the relatively large size of the early
 Scientific Research (1950’s)
physics packages which used large TE011 cavities, large and
 Commercialization (1960’s) cool absorption cells and discrete filter cells, along with high
 Miniaturization (1970’s) performance ovenized crystal oscillators. This was before the
 Tactical Military (Final Cold War Era 1980’s) emphasis shifted to small size and low cost. The early RFS
 Telecom (2000’s to Present) units were used as local frequency standards, where high sta-
This timeline is, of course, only approximate. The early pe- bility and low aging were paramount. Today most RFS are
riod of scientific research first spans investigations into atomic used as holdover devices in timing applications where they are
structure and then ways to utilize that structure in atomic locked to a GPS reference most of the time and simply need to
clocks. That research underwent commercialization as labora- maintain microsecond timing for a few days until the GPS
tory apparatus evolved into products. Miniaturization, simpli- reference is restored. The modest-performance RFS is simply
fication and cost reduction followed, which greatly expanded a more economical way to accomplish that than a very high
atomic clock applications. Meanwhile, a period of military performance ovenized crystal oscillator.
buildup resulted in a demand for environmentally-hardened A. Early RFS Manufacturers
RFS units. Today, while laboratory and military applications
Several (mainly aerospace) companies were the first to of-
still exist, the vast majority of Rb clocks are used for telecom
fer commercial RFS units, as listed below. Besides their use as
timing.
laboratory frequency references, RFS were used in the late
DD. RFS Technological Lines of Heritage 1960’s by TV networks to synchronize color broadcasts.
Before beginning a detailed look at some of the organiza-
tions involved with RFS technology, it seems wise to present 1) General Technology
General Technology of Los Angeles, CA developed one of
an overview of the leading current participants and their his-
the first commercial rubidium frequency standards, the Model
torical heritage. The following list shows the technology orig-
304 (see Fig. 17).
inators and their successor organizations and names:
 Varian, General Radio, GenRad, EG&G, PerkinElmer, Excelitas
 R&S, Efratom, Ball, Datum, Symmetricom, Microsemi. Microchip
 Litton, Frequency Electronics
 Stanford Research
 Neuchatel Observatory, Temex, Spectratime, Orolia
 Tadiran, Time & Frequency Ltd., Accubeat

A timeline of some of these RFS organizations is shown in

Fig. 16, and references herein for RFS products from some
from these organizations are [59] through [81].
Parent Organization

Stanford Research
(1980-Present) 1980 1995

1962 1985
FEI (1962-Present)

1953 1983
Litton
(1953-Present) Fig. 17. General Technology 304-B RFS. Photo Credit: Univ. of
Efratom
(1971-Present as Microsemi)
1971
1982
Ball Datum Symmetricom
1995 2002
Microsemi Microchip
2013 2018
Qeensland, Australia.
Rohde & Schwarz ?
(1933-Present)
1968

EG&G
(1947-Present)
PerkinElmer
1999
Excelitas
2010
2) Tracor
General Radio
(1915-1991)
1965
GenRad Tracor acquired General Technology in 1962 and continued
1975 1978 1980

Varian
1960
to make the Model 304 RFS and other T&F products (e.g., the
(1948-Present) 1967

Early Research
Princeton ITT Model 527E frequency difference meter). There was also a
Paris NBS
Etc.
Tracor Model 308 RFS (see Fig. 18).
1950

1960

1970

1980

1990

2000

2010

2020

Year

Commercial- Miniaturi-
Era Research
ization zation
Military Telecom

Fig. 16. RFS Organizational Timeline. For GPS rubidium clocks, there are
two timelines: Blocks I, II, IIA: Rohde & Schwartz, Efratom,
FTS/NRL/NTS-1, Efratom/Rockwell/Autonetics and Blocks IIR, IIR-M,
IIF, III: Varian/General Radio/GenRad, EG&G/PerkinElmer/Excelitas.

IV. ORGANIZATIONS AND PEOPLE

The history of rubidium frequency standard products can best Fig. 18. Tracor Model 308A RFS. Photo Credit: eBay.
be traced through the organizations and people that worked on
them. 3) Space Technology Laboratories
It is interesting that the emphasis of the early RFS manu- Space Technology Laboratories (STL) in Redondo Beach,
facturers was on performance, which in many respects ex- CA, a subsidiary of Thompson Ramo Wooldridge (TRW),
ceeded that specified for many of today’s commercial units. now Northrop, and a predecessor to the Aerospace Corpora-
 11

tion, made some of the first commercial RFS in the early The Varian R20 rubidium frequency standard was the suc-
1960’s. Especially noteworthy was their 1961 launch on Atlas cessor to the V-4700A circa 1965 (see Fig. 20).
rockets as a test for missile guidance. The principal RFS tech-
nologists at STL were J. M. Andres, D. J. Farmer, and G. T. (Left) February 1965 Varian Ad
Inouye [51]. in IEEE Spectrum for “World’s
Smallest Atomic Frequency
Standard”.
4) Clauser Technology
Little information is available about Clauser Technology
Corporation of Torrance, CA which was founded by Milton U.
Clauser (1913-1980) who was previously a VP at STL. The
company operated briefly between 1960 and 1962, and had
some RFS-related activity.

5) Varian
Varian Associates were a major contributor to early RFS
technology. The Model V-4700A Rubidium Vapor Frequency
Standard was introduced in early 1962 by the Instrument Divi-
sion of Varian Associates of Palo Alto, CA (see Fig. 19). It is
believed to be the first commercially available RFS on the
market. The Packard and Swartz December 1962 IRE Trans-
actions on Instrumentation paper about it had stability data
taken in October 1961 [34]. The design was remarkably simi-
lar to today’s high-performance (analog) units. It was all solid-
state except for the ≈ 100 MHz lamp exciter, which used a pair
of triodes in a push-pull oscillator configuration. The optical Fig. 20. Varian R-10 Rb Frequency Standard. Credit: Varian Associates.
package had an Rb-87 lamp, a discrete Rb-85 filter cell, and
an Rb-87 absorption cell with an N2-Ar buffer gas mixture. The Varian Quantum Electronics Division had operations
The unit had 5 MHz, 1 MHz, and 100 kHz outputs, the latter in Palo Alto, CA and at Bomac in Beverly. MA. Most of the
two obtained via regenerative dividers. The synthesizer drove Rb gas cell activity was in Palo Alto, with Cs beam tube and
a varactor diode multiplier with signals at 120 MHz and 5 6/19 active H maser work in Beverly (next to where
MHz, the latter produced by a divide-by-19 regenerative di- FTS/Datum/Symmetricom/Microsemi is still located54). When
vider and a balanced modulator. The ≈ 5.3158 MHz synthesiz- those activities ceased in 1967, Varian retained their Rb mag-
er output was subtracted from the 57 th harmonic of 120 MHz netometer product line and General Radio continued their
at 6840 MHz in the multiplier to produce an interrogation sig- Apollo era contract work on an Rb space clock for NASA.
nal at the ≈ 6834.6842 Rb resonance, corresponding to a rela- The principal RFS investigators at Varian were W. Earl
tively small buffer gas offset of about 1.6 kHz. The interroga- Bell, Arnold L. Bloom, Martin Packard, B.E. Swartz, Art
tion signal was modulated at 107 Hz by a phase modulator at McCoubrey and Al Helgesson (see Fig. 21)55, 56 [34].
the bottom of the RF multiplier chain, which had many tuned
circuits. The discriminator signal from the optical package was
amplified, filtered, synchronously demodulated with funda-
No
mental and 2nd harmonic detectors, and integrated to produce a Photo
control voltage for the 5 MHz crystal oscillator. All of this Available
circuitry is familiar today, although the modular chopper-
stabilized operational amplifiers are a far cry from today’s IC
op amps. All-in-all, it appears to be a very workmanlike de-
sign, and at least one unit may still be in use today [58].
W.E. Bell A.L. Bloom
(1921-1991) (1923-2018)
Photo Credit: Spectra Physics.

54
Microchip Technology acquired Microsemi in 2018.
55
Art McCoubrey later was associated with (a “co-founder” along with Bob
Kern in 1971) Frequency and Time Systems in Danvers and then Beverly.
MA. They did not pursue RFS technology until much later as Datum, but did
help to introduce the Efratom unit in the U.S. We at General Radio explored
Fig. 19. Varian V-4700A. The 1st commercial rubidium frequency
an integrated cell Rb physics package for him shortly after that.
standard. Photo Credit: Varian Associates 56
Al Helgesson was instrumental in the Rb technology transfer from Varian to
General Radio in 1967.
 12

Martin Packard Byron E. Swartz

(1921- ) (1931- )
Photo Credit: Stanford Univ. Photo Credit: IRE

Photograph of Physics Package

This physics package, with its TE011 cavity, resembles the first General
Radio designs.
Fig. 22. NBS Portable RFS [51]. Photo Credit: NBS.

B. Major RFS Instrument Manufacturers

Quartz crystal oscillator laboratory frequency standards
Art McCoubrey Al Helgesson
(1920-2019) (1933- ) were a part of the standard product line for electronic instru-
Photo Credit: Legacy.com. Photo Credit: IEEE. ment manufacturers such as Hewlett-Packard, General Radio
Fig. 21. Key RFS Investigators at Varian. and Rohde & Schwarz, so it was natural for them to enter the
field of atomic frequency standards in the early 1960’s when
6) NBS/Varian those instruments started to become feasible. This traditional
NBS started work in 1959 on a portable RFS for a satellite market was quite limited, and was, to some degree, based on
gravitational redshift experiment (see Fig. 22) [51]. The exper- the need to offer a complete product line.
iment was “deferred”, but three units were built using a Varian
physics package and a “natural frequency” 4.9961136 MHz 1) Hewlett-Packard
crystal oscillator (multiplied by 3·2·2·114 to the Rb hyperfine Hewlett-Packard, while emphasizing cesium beam tube
frequency, with a separate “translator” to 5 MHz). This unit atomic frequency standards, also developed and manufactured
illustrates how advanced this technology was at that early a high-performance laboratory RFS, the 5065A (see Fig. 23)57,
date. The HP Rb technology was independently developed with
some heritage from Varian. The principal contributors were
Darwin Throne (project leader), B. E. Swartz (electronics,
from Varian) and R. A. Baugh (optical package).

Rb Resonance Line

Fig. 23. HP 5065A Rubidium Frequency Standard. Photo Credit: Hewlett

Packard [62].

The HP 5065A had (and a few units still operating have)

excellent performance and it is probably the best laboratory
RFS ever made. HP also developed a miniature HP 10816
RFS in the late 1970's that was never produced58.

57
The Varian Cs technology (e.g., Len Cutler) was transferred to HP in
Block Diagram Beverly, MA (and later Santa Clara, CA) and the H-maser technology (e.g.,
Bob Vessot) was transferred to the SAO in Cambridge, MA.
58
Information about the HP mini RFS was kindly provided to me by R. K.
“Rick” Karlquist. He was the RF engineer on the project.
 13

2) General Radio/GenRad
General Radio (later GenRad, abbreviated GR) was, be-
tween 1915-1991, a leading manufacturer of electronic in-
struments, including a complete line of time and frequency
products. After many years of making quartz crystal frequency
standards, it established its own rubidium gas cell technology
circa 1965, later augmented by similar Varian technology.

Fig. 27. SATS Physics Package Lamp Assembly. Photo Credit: General
Radio.

Fig. 24. General Radio Concord MA Plant Circa 1965.

The GR RFS activities emphasized military and aerospace

designs rather than the company’s traditional commercial
products. Photographs of two of those RFS units are shown in
Figs. 25 and 29, and details of the SATS physics package are Fig. 28. SATS Physics Package Cavity Assembly. Photo Credit: General
Radio.
shown in Figs. 26-28. It used a combination of General Radio
and Varian technology59.
The Collins physics package was similar to that used in the
SATS but included a thermoelectrically cooled sealed outer
oven [145].

Fig. 29. Collins Radio AFS-81 Military RFS..The AFS-81 contained an

Rb physics package made by General Radio. Photo Credit: Collins Radio.

Fig. 25. Spacecraft Atomic Timing System (SATS). The SATS unit
The work on rubidium frequency standards began at the
comprised both a rubidium frequency standard and a digital clock and time General Radio Company in about 1965 as a logical successor
code generator. Photo Credit: General Radio. to its traditional quartz frequency standard product line when
it became clear that the last such quartz crystal oscillator, the
GR Model 1115, would be supplanted by the rapidly-emerging
atomic clocks. This work started at the Concord. MA head-
quarters and was subsequently moved to the new Bolton, MA
facility. The principal contributors toward rubidium frequency
standard technology at General Radio were as follows (see
Fig. 30):
 Richard W. Frank Engineering Group Leader
 Herbert P. Stratemeyer Engineering Project Manager
 William J. Riley, Jr. Electronic Engineer
Fig. 26. SATS Physics Package Cross Section. Photo Credit: General  Roland P. Kenschaft Physicist
Radio.  John E. Wilhelm Electronic Engineer
 Richard A. Mortenson Mechanical Engineer
 William J. DeFlorio Glassblower
59
Although the SATS lamp and cavity assemblies are basically the same as a
modern RFS unit, it interesting to note some of the differences between it and
the GPS RAFS that followed. The SATS filter and absorption cells shared the
same oven and had long tubulations. The SATS ovens used urethane foam
rather than vacuum insulation and used bifilar heater windings rather than
double layer foil heaters. The SATS lamp exciter circuit was inside the lamp
oven and used somewhat unreliable piston trimmers rather than ceramic chip
capacitors. The SATS cavity mode was TE011 rather than a smaller TE111 and
it had an unnecessary waveguide below cutoff at its photodetector.
 14

Model 1122 Rb physics package, developed in collaboration

with Marv Frerking and Don Johnson of the Collins Radio
Company. Initially intended for an airborne collision avoid-
ance system (CAS)60, it ended up being used in the U.S. Navy
Verdin "doomsday" VLF communication system. Collins (lat-
er Rockwell) built the complete AFS-81 militarized rack-
mounted rubidium frequency standard while GR made the
physics package (which included its power supply, oven con-
troller, RF multiplier and preamplifier supporting circuits).
This unit was extremely rugged, operated under harsh envi-
ronmental conditions (including thermoelectric cooling for
high temperatures and a sealed physics package to avoid bar-
Fig. 30. RFS Developers at General Radio. Photo Credit: 1969 GR
Personnel Book. ometric sensitivity in airborne use), and provided high perfor-
mance (even by today's standards).
The major rubidium frequency standard projects and prod- This work was carried out by an informal "aerospace"
ucts developed and produced at General Radio were as fol- group that operated somewhat independently from the com-
lows: mercial products because of its contract work and the need for
military parts, quality systems and testing. Those activities
 Model 1118 Commercial Rubidium Frequency Standard (never pro-
duced) were supplemented by other contract work such as quartz
 Spacecraft Atomic Timing System (SATS) for NASA (developed and crystal measuring equipment (Microelectronic Bridge and
qualified) Tracking Servobridge Detector) for the U.S. Army. It is inter-
 Model 1122 Military Rubidium Physics Package for Collins Radio
(about 300 produced)
esting that these early RFS products had relatively high per-
 Missileborne RFS for General Electric (design study) formance compared with today's miniature units. They were,
 Airborne RFS for Magnavox (design study) however, much more complex and expensive. Even commer-
 Model 1126 Space Rb Clock for Rockwell Global Positioning System cial units were the price of a car, while the ones now used in
(GPS) (brassboard unit)
large quantities for commercial telecom applications are only
The initial investigations of rubidium frequency standard about 5% of that relative cost.
technology were conducted with company R&D funding. Ear- The missileborne and airborne RFS design studies empha-
lier work by Maurice Arditi at ITT and others had established sized performance under vibration, and it was recognized that
the viability of this technology. The work proceeded quite vibrational modulation of the crystal oscillator that produced
independently of outside help or support, except for some ini- the interrogation signal was critical at frequencies related to
tial consultation by Peter Bender of NBS on Rb cell/buffer gas the servo modulation. R. P. Kenschaft contributed important
technology. General Radio had expertise in frequency control, theoretical insight into the Rb recovered signal and its pro-
quartz crystals, flash lamps, high vacuum processing and cessing.
RF/microwave electronics. The first cells were made at the The final major project was the development of a high-
EG&G Bedford laboratory before suitable glassworking and performance rubidium clock for the Rockwell Global Position-
cell processing facilities were established at GR. The core GR ing System (GPS). That effort resulted in a smaller physics
rubidium frequency standard technology used a pure Rb-87 package and a new electronic design intended for space. In the
lamp, a discrete Rb-85 isotopic filter cell, and an Rb-87 ab- meanwhile, Rockwell's own Autonetics division was adapting
sorption cell. Reasonably good lamp life was obtained by the an Efratom physics package for the same application. The GR
use of alkali-resistant glass and burn-in (calorimetric process GPS design was not pursued until several years later at
control did not exist). The lamps were generally operated in EG&G.
the intense, high-temperature red mode. Light shift was nulled The RFS work at GR wound down in around 1978 with the
by varying the length (or later) temperature of the filter cell. retirement of Dick Frank and Herb Stratemeyer and the depar-
Excellent aging was obtained by ultra-high (ion pump) absorp- ture of Bill Riley. John Wilhelm continued to support the
tion cell processing. All designs until the final GPS one used a Rockwell physics packages for a while, but no new projects
full-size TE011 microwave cavity that held both the filter and were undertaken. The company was in the process of transi-
absorption cells at the same operating temperature. The GPS tioning from electronic instruments to automated test equip-
unit used a smaller TE111 cavity and a separate filter cell oven. ment, and never regained its former leadership position. The
An important early milestone was the closing of the Varian main legacy of the General Radio rubidium frequency stand-
Quantum Electronics Division in Beverly, MA and the as- ard technology is the GPS RAFS units built by EG&G (later
sumption by General Radio of their NASA contract to develop PerkinElmer, now Excelitas). Those are a direct descendant of
a Spacecraft Atomic Timing System for the Apollo program. the GR Rockwell GPS RFS design, and were refined into the
It was a combination of a rubidium frequency standard, digital
clock and time-code generator. That unit, completely rede- 60
The 1970’s aircraft Collision Avoidance System (CAS) would have worked
signed at GR, was qualified but never produced or flown. somewhat like GPS with precisely synchronized clocks and pulses whose time
Overall, the most significant production activity was the of arrival indicates proximity.
 15

highest performance such units ever made. That work was neers involved were Hugo Fruehauf and Dale Ringer at
carried on at EG&G by W. J. Riley, with initial consultation Rockwell, Chuck Wheatley and Frank Kopek at Autonetics,
by H. P. Stratemeyer. and Werner Weidemann at Efratom.

3) Rohde and Schwarz

Rohde and Schwarz was, as still is, a major electronic in-
strument manufacturer, and, much like General Radio and
Hewlett-Packard, had a line of time and frequency products,
including laboratory rubidium frequency standards.
Their first rubidium frequency standard (I believe) was the
XSR laboratory RFS developed in 1968 and shown in Fig.
3161. The XSRM shown in Fig. 32 was a later R&S RFS.
Fig. 33. Rockwell GPS RFS. This atomic clock was built for the first GPS
satellites in the late 1970s [119]. Photo Credit: Smithsonian Museum.

D. EG&G/PerkinElmer/Excelitas Technologies
EG&G entered the rubidium frequency standard business in
1980 at the request of the U.S. Air Force and Rockwell Inter-
national. The EG&G facility in Salem, MA was a logical place
to pursue this work because of their expertise in glassworking
Fig. 31. R&S XSR Rubidium Frequency Standard. Photo Credit: R&S. and high vacuum technology, and their familiarity with high
reliability critical defense programs (see Fig. 34)63, 64, 65.

Fig. 32. Rohde & Schwarz XSRM RFS. Photo Credit: Rohde & Schwarz.

Fig. 34. Excelitas (right building) in Salem, MA. EG&G opened its
C. Aerospace/Electronics RFS Contractors Salem, MA plant at 35 Congress Street in the Shetland industrial park, the
old Naumkeag Steam Cotton works on the Salem waterfront, in 1961. The
Several major aerospace/defense contractors became in- Rb department began activities in 1980 on the development of a GPS
volved with rubidium frequency standard technology, mainly clock. That work continued as PerkinElmer (1999) and now Excelitas
because of their own requirements or those of their Govern- today. Photo Credit: Shetland Properties.
ment customers.
The major contributors to the RFS work at EG&G are
1) Rockwell International shown in Fig. 35.
Rockwell (in Seal Beach, CA, now Boeing) was the first
contractor for GPS satellites (1st contract awarded June 1974).
They, along with their Autonetics division (in Anaheim, CA)
needed Rb clocks for this program, and opted to design an
RFS based on an Rb physics package from Efratom in Mu-
nich, Germany, which set up operations in Irvine for that pur-
pose (see Fig. 33)62. The effort was quite successful (although
63
there were some early in-orbit failures), The principal engi- The author worked on RFS technology at EG&G and consulted for them
after they become PerkinElmer and then Excelitas. I am therefore able to
describe some of those activities in greater detail. I would welcome similar
61
The Rohde & Schwartz XSR RFS was developed in 1968 per a May 2007 detail by people associated with the work done at other organizations.
64
e-mail from Hans H. Jucker of R&S who kindly sent me a data sheet for it. The U.S. atomic clock business has traditionally been centered in the
Boston area and Southern California.
62 65
Excerpt from The Origins of GPS: “On realizing that the small Efratom Why were and are so many of the atomic clock companies in the Boston
company would be incapable of producing a radiation-hardened, space- area? Well I suppose it just started there (and Palo Alto) and has just kept
qualified rubidium oscillator, RI’s GPS satellite program manager Richard going. Bomac/Varian/FTS/Datum/ Symmetricom/Microsemi all in nearly the
Schwartz created a teaming relationship with them, which included his chief same place with spinoffs to General Radio/GenRad/EG&G/Perkin
engineer, Hugo Fruehauf, plus Dale Ringer, Dr. Chuck Wheatley of Elmer/Excelitas, as well as Kernco, National (Atomichron), Pickard & Burns,
Rockwell’s Autonetics Division, and Efratom’s Werner Weidemann. With and the SAO. There would sometimes be so many of us on the way to the
heroic efforts, this team built a space-qualified clock in time for the first GPS FCS or PTTI that we could have presented papers on the plane.
launch in February 1978.”
 16

higher performance than their smaller commercial counter-

EG&G Group Scientist Sy
Goldberg (1927-2014, at left) with
parts, e.g., a short-term stability better than 1x10-12-1/2, a
EG&G co- founder Harold “Doc” flicker floor below 1x10-15, a temperature coefficient on the
Edgerton examining an RFS-10 order of 1x10-14/ºC, and aging under 1x10-12/month. They easi-
rubidium frequency standard in ly maintain time to better than 1 nanosecond over a day
1983. Photo Credit: EG&G.
onboard a GPS satellite, which corresponds to a range error of
1 foot71. About 100 of these Rb clocks are currently in-orbit.72

Sy Goldberg

Fig. 36. EG&G GPS Block IIR RAFS. Photo Credit: EG&G

Bill Riley Tom Lynch The story begins at General Radio in the mid-1970’s with
the adaptation of an Rb physics package built for Collins Ra-
dio (later Rockwell) as part of the U.S Navy VERDUN VLF
communication system [145]. That design used a full-sized
TE011 microwave cavity and had separate lamp and cell (cavi-
ty) foamed-in-place ovens inside a sealed main oven which
desensitized it from pressure changes in aircraft use. The ini-
tial GPS design, performed for Rockwell as part of their first
GPS satellite program, eliminated the outer oven and substi-
Ken Lyon John Vaccaro tuted a smaller TE111 cavity, and a separate filter cell oven,
The EG&G Salem general and was successfully completed through a brassboard phase
managers included Walter [120]. The principle contributors to that design at GR were
Saldarini, Len Colasanti, Paul
Beech and Arnie Shuman. The
Herb Stratemeyer, Dick Frank and this author.
Rubidium department managers Later, in 1980, when Rockwell was experiencing Rb lamp
included Bob Stitt, Ralph reliability problems with their own GPS clock that used an
Carpenter, Ed Bryant, Mal adapted Efratom physics package [119], work on the GR-
Schwalje, Walt Zarris and Murray
Tysinger66. based RAFS was continued at EG&G under the leadership of
this author with major contributions from Sy Goldberg and
Doug Lowrie Tom Lynch, and early consultation by Herb Stratemeyer and
Fig. 35. Major RFS Contributors at EG&G.

At ITT: Tony Baker, Gerry Freed, Marv Epstein, Jay Phelan. At Aerospace:
The EG&G Electronics Components Division and Fre- Phil Talley, Charlie Volk, Jim Camparo.
quency Products Division in Salem, MA (as it was called dur- 68
A significant amount of credit for the success of the EG&G GPS RAFS can
ing the 1990’s) was managed by a variety of persons under be attributed to its long gestation period, and to the cordial relationship that
developed between the technical staffs at EG&G, ITT and Aerospace
several structures, while the technical staff remained largely Corporation. There was a strong belief in its worth, and a genuine team effort
stable and self-sufficient. to make it the best it could be.
69
The contract with ITT for the GPS Block IIR Rb clocks was extremely
1) EG&G GPS RAFS important to the EG&G Rb department. It led to a very mutually beneficial
The history of the exceptionally high performance EG&G relationship that still continues today. The RAFS has proven to be an
excellent clock, and close personal ties have developed between the
(later PerkinElmer and Excelitas) GPS Rubidium Atomic Fre- technologists in both organizations.
quency Standard (RAFS) is a prominent part of this history 70
It’s impossible to overstate the importance of GPS to the RFS business.
paper67, 68, 69, 70. These clocks (see Fig. 36) have significantly Besides the satellite clocks themselves, it spawned a huge market for com-
mercial RFS units in telecom cell sites. Of course, the opposite is also true:
Rb clocks have been the most critical devices onboard the GPS satellites.
71
It was originally planned to deploy a mix of two rubidium and one cesium
66
The first manager of the EG&G Rb department was Bob Stitt. Murray clock on the Block IIR GPS satellites. When the Kernco/SCI Cs team found-
“Butch” Tysinger was the manager of EG&G Frequency Products when, at its ered, the decision was made to go with three EG&G RAFS. This, of course,
largest, it included a crystal company called CINOX. Len Colasanti was the increased the pressure on us to achieve the performance and reliability goals.
EG&G Salem division manger during the critical years when the GPS RAFS For added assurance, two units were placed on life test at NRL [168]. Fortu-
entered production. nately, those RAFS met and exceeded all requirements.
67 72
Major contributors to the EG&G GPS RAFS design: At EG&G: Herb 60 GPS IIR, 24 GPS IIF, 3 GPS III, 11 QZSS as of June 2019. Many have
Stratemeyer, Bill Riley, Sy Goldberg, Tom Lynch, Ken Lyon, John Vaccaro. operated for 10-15 years, one for 20+ years.
 17

Jacques Vanier73, 74, along with close cooperation with the exploded. The three on-board RAFS were recovered largely
Aerospace Corporation. intact but in non-operable condition as shown in Fig. 37.
To the credit of Rockwell and their U.S. Air Force spon-
sors, the program was set up with sufficient time and resources
to do a careful bottoms-up design of a GPS-specific advanced
Rb clock. A principal immediate goal was to solve the lamp
reliability problem. The physics package was tailored for high-
performance space operation (e.g., the ovens used vacuum
insulation, had low radiative thermal loss, and high reliability
low residual magnetic field 2-layer foil heaters with high gain
temperature controllers). The EG&G RAFS design retained
the performance advantages of a discrete filter cell (e.g., high Fig. 37. GPS Block IIR RAFS S/N 004. Photo Credit: USAF.
S/N, absorption cell homogeneity, low RF sensitivity, ease of
3) Later GPS RAFS Units
light shift nulling) while overcoming its main disadvantage
Improvements were made in the GPS Block IIF RAFS, in-
(large negative filter cell TC) by tight oven temperature con-
cluding Xe lamp buffer gas, optical filtration and tighter BTC
trol.
thermal control, which enhanced its stability77, 78, 79. Figure 38
The lamp Rb depletion problem on the Rockwell/Efratom
shows a cross sectional diagram of the PerkinElmer GPS
GPS clocks was aggregated by using low Rb fill to reduce
Block IIF RAFS physics package. That clock included a sec-
noise, but this was hard to control “by eye” and it resulted in
ondary loop that produced a 10.23 MHz output.
short (e.g., 1 year) lifetime. Aerospace scientists Jim Camparo
and Charlie Volk established that the likely depletion mecha-
nism was diffusion of Rb into the glass lamp envelope. Tom
Lynch at EG&G invented a non-destructive way to measure
lamp Rb fill by measuring the energy required to melt it using
a differential scanning calorimeter (DSC). Employing that for
initial process control and subsequent lamp screening, the Rb
fill and diffusion rate were quantified, and fill tolerances es-
tablished to assure both low noise and long life 75. That meth-
odology is now in widespread use throughout the RFS indus-
try.
In summary, Rockwell opted to continue using its own in-
Fig. 38. PerkinElmer GPS Block IIF Physics Package Cross Section.
house (Autonetics in Irvine) GPS RFS design using the
Credit: PerkinElmer.
Efratom physics packages rather than further develop a 2 nd
source at GR. The EG&G GPS RFS concept lay dormant until The Rockwell Block IIF GPS satellites contain both Rb and
the Lockheed Martin/ITT GPS Block IIR program. Design Cs clocks, allowing their performance to be compared, as
work at EG&G began in 1990, a production contract was shown in Fig. 39. While both types of clock easily meet the
awarded in 1992, and clock deliveries (60 clocks, 3 per satel- IIF specifications, the EG&G RAFS contribute significantly
lite) occurred between 1994 through 1997. GPS Rb clock ref- lower range error.
erences herein are [119] through [132]76.

2) First GPS IIR Launch 77

An RFS radiation hardening analysis has the very desirable side effect of
The first GPS IIR satellite launch in January 1997 failed also performing a detailed worst case circuit analysis by a highly-skilled ana-
shortly after liftoff when a damaged Delta solid rocket booster lyst (e.g., Terry Flanagan/IRT, Dave Swant/GE for GPS clocks). Modern
circuit analysis tools are remarkably effective. That modeling relates closely
to initial and end of life (EOL) performance margins. Furthermore, this infor-
73
Jacques Vanier and his atomic clock group at Laval University researched mation is important for verifying component stress levels and making reliabil-
absorption cell buffer gasses and other aspects of the initial EG&G GPS ity predictions.
78
RAFS design. GPS clocks are hand made by dedicated craftspersons whose workmanship,
74
Jacques is one of the giants of this field. He (along with Claude Audoin) more than anything else, is responsible for their excellent reliability. Each
literally “wrote the book” about atomic clocks [3]. He was one of the princi- clock is individually observed and adjusted during testing to perform the best
pals at the Varian Quantum Electronics Division in Beverly, MA. He estab- it can.
79
lished a quantum electronics laboratory at Laval University, pioneered active Excerpt frpm GPS World 4/26/17: “GPS III Rubidium Atomic Frequency
and passive Rb frequency standards, researched wall coatings and buffer Standards (RAFS) have evolved from GPS IIR and IIR-M RAFS, which have
gases, analyzed RFS electronics, modeled the Rb signal, and then later largely collectively and reliably provided more than 250 years of on-orbit service,
invented the CPT interrogation technique, implementing it in ordinary and including significant time beyond their intended design lives. Our GPS III
“maser’ form at Kernco. RAFS clocks undergo rigorous environmental qualification and life tests to
75
Zero in-orbit GPS RAFS Rb lamp failures have occurred since adopting assure performance over this next generation satellite’s 15-year design life. In
DSC lamp screening. A somewhat similar but less severe issue regarding addition, each GPS III SV includes multiple RAFS for redundancy. GPS III
lamp Xe buffer gas fill was resolved more recently by non-destructively continually monitors the active RAFS to detect and mitigate clock anomalies.
measuring the Xe buffer gas pressure by operating the lamp as an absorption This is just one way that GPS III provides increased signal integrity for GPS
cell and measuring its frequency offset. users. Galileo clocks utilize different suppliers than GPS III clocks. The GPS
76
EG&G in Salem also made Rb magnetometer lamps, cells, lamp exciters III clock supplier has produced reliable RAFS clocks for GPS satellites over
and physics packages for a sister division, Geometrics. the past several decades.”
 18

days and has a modeled white FM noise level of only about 1x10-12-1/2
which corresponds to a time deviation of only about 0.1 ns at 1 day. Plot
Plot Credit: Excelitas.

4) EG&G GPS RAFS Aging

Mature GPS RAFS units show aging in the low pp1014/day
range. Where effective measures are taken to minimize light
shift and other factors, the most important contributor to aging
is likely diffusion of absorption cell N2 buffer gas into the
glass envelope and/or Rb surface film. This theory is support-
ed by their consistent negative aging direction and excellent fit
to a diffusion (√t) law, as shown in Fig. 42 [231]. RFS aging
Fig. 39. Estimated GPS IIF Clock Range Error. Plot Credit: Wu & Feess, references herein are [230] through [233].
2000 PTTI [170].

In preparation for GPS III, a program of further improve-

ments called RAFSMOD was initiated in 2005, which resulted
in even better stability, as shown in Fig. 40 where the flicker
floor extends below 1x10-15 at averaging times longer than
several days80.

Fig. 42. RAFS Aging Plot with Diffusion Law Fit. Plot Credit: EG&G.

5) EG&G Military RFS

As the EG&G GPS Rb clock design was progressing, the
same group developed a small, ruggedized unit directed at
military applications. That effort was driven by competitive
U.S. Air Force RADC programs at EG&G, Efratom, and Lit-
Fig. 40. Excelitas RAFSMOD EDU Stability. Credit: Dupuis, Lynch & ton to develop a tactical RFS (TRFS) for the Hazeltine Seek-
Vaccaro, 2008 FCS [132].
Talk anti-jam radios81. The main challenge was a very severe
operating environment that included fast warm-up, mechanical
The Excelitas GPS III RAFS displays exceptional stability,
vibration and high temperatures. All three organizations suc-
as shown in Fig. 41.
cessfully built prototype units with EG&G and Efratom gain-
ing new core products82, 83 and Litton launching a new techno-
logical base that was later sold to FEI. References herein to
MIL/tactical RFS history are [141] through [155].

6) The EG&G TRFS

The EG&G TRFS (and similar units from Efratom and Litton,
later FEI, see Fig. 48) were the most rugged tactical RFS units
ever made. Although the Seek-Talk radios were never pro-
duced, the concepts learned were utilized later in the EG&G
RFS-10 family (see Figs 43 and 44). References to RADC
Technical Reports and papers re these tactical RFS units for

81
The overseers of the three TRFS contracts at RADC were Nick Yanoni, Al
Fig. 41. Excelitas GPS III RAFS S/N 324 30-Day Drift-Removed ATP Kahan and Ferdy Euler (the latter for EG&G). Although little came of the
HDEV Stability Plot. This GPS RAFS reaches the pp1016 range for ≥ 5 Seek-Talk radio program, the TRFS work had important effects on the three
contractors. Efratom got a new product (the M-1000 evolved into the M-3000
with improved reliability). Litton got a new technology, which they soon sold
80
PerkinElmer received a contract in June 2009 for RAFS for the 1st two GPS to FEI. EG&G got an entry into the military RFS business to supplement their
III spacecraft, "Having built more than 100 atomic clocks for space, far GPS Rb clock work.
82
exceeding that of any other company, we are confident that our dedicated The EG&G TRFS and Efratom M-1000.
83
design and manufacturing capabilities and 30 years of experience will serve to Management seems to always push end-of-the-year sales. I delivered a
enhance GPS satellites for decades to come," said John Vaccaro, Technical critical contractual milestone test report on the TRFS to Ferdie Euler (along
Director of RAFS at PerkinElmer. with a box of wine) at the gate of Hanscom AFB on New Year’s Eve in 1982.
 19

the three participants will be found in References [146] 8) The EG&G RFS-10-7 and RFS-10-102
through [154]. The EG&G RFS-10-7 and RFS-10-102 were versions of
the RFS-10 in a larger box compatible with the popular
7) The EG&G RFS-10 Efratom M-100 militarized Rb frequency standard. One appli-
The EG&G RFS-10 was designed as an equivalent to the cation for these units was the 1976-1995 Raytheon (Wayland
Efratom M-100 MIL RFS in a smaller package. However that and Bedford, MA), and Unisys (Salt Lake City, UT) 2nd source
turned out to be a marketing mistake as the larger M-100 for- U.S. Air Force AN/TRC-170 troposcatter radio. Although its
mat had become an industry standard of sorts, especially in procurement ended in 1995, as of 2017 this radio system was
military procurements where multiple sources were encour- still in use.
aged. Thus most RFS-10 sales were for units where the small- Military anti-jam frequency hopping HF radios were anoth-
er RFS-10 core was put into a larger package. In later years, er common RFS application during the 1980’s, and there was
the smaller size became more widely used, was duplicated by stiff competition between EG&G and Efratom for that busi-
a Symmetricom M-100 replacement (see Fig. 50), and contin- ness. One such program was the 1983-1992 Magnavox (Ft.
ued to be produced well into the 2000’s. Wayne, IN) AN/TRC-179 Regency Net U.S. Army HF fre-
quency hopping radios, with most RFS procurement circa
198686.
Other users of the RFS-10-7 were GTE Sylvania Commu-
nications Systems Division in Tauton, MA and Brandywine
Communications.
References to RFS-10-7 and RFS-10-102 Technical Re-
ports will be found in [141] through [143].

9) The EG&G RbXO

Fig. 43. EG&G RFS-10. Photo Credit: EG&G. The RbXO (see Fig. 45) is a simple concept whereby a con-
tinuously-running crystal oscillator is occasionally syntonized
by a rubidium reference. The idea is to save power (really en-
ergy) in a timing application by activating the Rb reference
only as needed to maintain XO accuracy. This arrangement,
championed by John Vig, makes most sense when a very low
power ovenized crystal oscillator is used, and he oversaw a
parallel program at Bendix and then Piezo Crystal for the Tac-
tical Miniature Crystal Oscillator (TMXO)87.
EG&G and Efratom had RbXO contracts with the U.S.
Army, but only EG&G successfully completed the work. The
challenge was mainly to develop an Rb reference that could
quickly warm-up and syntonize the OCVCXO reliability thou-
sands of time with minimal energy consumption88.

Fig. 44. Seek-Talk Rb Physics Package. The EG&G Seek-Talk Rb

physics package utilized a discrete filter cell co-located in the microwave
cavity/oven with the absorption cell84. The crystal oscillator shared the
same thermal environment. It was rigidly attached to the Rb lamp
assembly. See U.S. Patent 4,494,085. Figure Credit: EG&G.
Fig. 45. EG&G RbXO with TMXO (left). Photo Credit: EG&G.
Thus the largest number of EG&G military RFS units were
RFS-10-7 equivalents to the Efratom M-100 (see below), 86
Mal Schwalje was the sales manager for Rb products at EG&G in Salem,
which were used in considerable quantities in such programs MA, and his standing joke was things were so tight that he had to take the bus
as the Raytheon TRC-170 troposcatter radio and the Mag- to Ft. Wayne, IN. EG&G never achieved much success in the commercial
navox Have-Quick aircraft radios85. RFS business and, at best, shared less than half the military business with
Efratom.
87
The TMXO per se never became a product, but to some extent led to the
many very low power OCVCXO’s available today.
84 88
There is a pitfall associated with placing the filter cell inside the microwave On-off power cycling endurance, especially at low temperatures with fast
cavity. If it contains any Rb-87 there can be a spurious resonance at a warm-up, is a difficult design challenge but essential to the RbXO concept.
frequency determined by the filter cell buffer gas. The EG&G RbXO was able to pass a tough design verification test (DVT) but
85
As I recall, in the largest single order for several hundred units, the a competitive unit was not. The DVT subjected 4 RbXO units to 180 days of -
procuring office made a mistake on the quantity; the excess units are probably 62C to +68C temperature cycling with 20 on-off power cycles per day.
still in some huge Government warehouse next to the Ark of the Covenant.
 20

Later, another program (circa 1985) called Modular Intelli- rubidium frequency standards at Rohde & Schwarz. Efratom
gent Frequency, Time and Time Interval (MIFTI) utilized the developed and produced small RFS units (FRK, M-100, etc.)
EG&G RbXO along with a microcontroller (with algorithms employing a then-innovative integrated resonance cell. The
from Sam Stein of Timing Solutions) to discipline the system company expanded to Irvine, CA in 1973 where it became the
time and frequency. That work was a bit ahead of its time and largest manufacturer of RFS. It became part of Ball Corpora-
was never completed89, 90, 91. But before long, there appeared a tion in 1982, Datum in 1995, then Symmetricom in 2002 (and
“Smart Clock” from HP and then a host of GPS-based disci- moved to Beverly, MA), Microsemi in 2013, and then Micro-
plined OCVCXO and Rb products (although those did not chip Technology in 2018.
emphasize low power). Major contributors at Efratom included president Hugo
Figure 46 shows an RbXO temperature/power cycling re- Fruehauf, physicist Tom English, engineer Werner Weideman
trace characteristic as a function of air temperature with dis- (see Fig. 53) plus engineers Bill Cashin and Jeff Crockett and
tinct hysteresis versus the direction of the temperature change. physicist Jin Deng.
The RbXO would follow this characteristic over and over Noteworthy accomplishments included the first atomic
again during successive power and temperature cycles. RbXO clock in space (a modified FRK on NTS-1), the 1st generation
references herein are [156] through [159]. Rb physics packages for GPS, rugged militarized units (M-00,
M-1000, M-3000) and thousands of commercial RFS units
(FRK, FRS, LPRO, X-72, etc.), see Figs 47-49 for examples.

Fig. 47. Efratom FRK Commercial RFS. Quoting from the Smithsonian
National Museum of American History web site for this item: “This
compact rubidium frequency standard is the commercial Model FRK, first
Fig. 46. RbXO Retrace Characteristic [210]. Plot Credit: EG&G. made by Efratom Elektronik, Munich, Germany, and later by Efratom
California in Irvine, Ca. Gerhard Hübner and Ernst Jechart established the
10) The EG&G SCOTT RFS-20 firm in 1971 and a year later supplied examples of the clock to the Naval
The EG&G RFS-20 was an RbXO intended for use in the Research Laboratory (NRL), Washington, D.C., for inclusion on NTS-1,
the first of the Navigation Technology Satellites (NTS) launched in 1974
U.S. Army TSC-124 Single Channel Objective Tactical Ter-
to validate the key concepts and hardware for the Global Positioning
minal (SCOTT) EHF MILSTAR ground terminal, a circa System (GPS). Relatively large rubidium frequency standards had been
1985-1990 Magnavox (Ashburn, VA) program initiated at developed in the 1950s, but the FRK—weighing roughly three pounds and
MIT Lincoln Lab in 1982. The RFS-20 was designed as the measuring about four inches on a side—were the smallest atomic
frequency standard of any type available. Efratom established a branch in
timing and frequency reference, with precise timing main- Irvine, California, in 1973 and manufactured compact rubidium frequency
tained for 10 days at low power during transport 92. Based on standards there for a variety of customers. The firm became a division of
the RFS-10, it was a very attractive design, but unfortunately Ball Aerospace in 1982 and then part of Datum in 1995. Symmetricom
acquired Datum in 2002.” Photo Credit: Ball Efratom (More photos on the
the SCOTT program was terminated in 1993 before any pro- Smithsonian web site).
duction. References to the SCOTT program and RFS-20
Technical Reports will be found in [144].
E. Efratom/Ball/Datum/Symmetricom/Microsemi/Microchip
Efratom was founded in Munich Germany in 1971 by Ernst
Jechart (along with Gerhard Hübner) 93 after his working on

89
Some of the ideas like taking advantage of environmental changes to learn
OCVCXO compensation versus the Rb reference are still not commonplace.
90
The RbXO was also a precursor to the U.S. Army SCOTT time and Fig. 48. Efratom M-1000 Tactical RFS. Photo Credit: Ball Efratom.
frequency reference (that was also never produced).
91
A personal side benefit is that the author was exposed to the C programming
language and PC programming in general, which led to other productive
results.
92
An RbXO depends on having a low power OCXO. Bendix and Piezo
Technology were unable to successfully produce the U.S. Army TMXO
needed for the SCOTT RFS. Sy Goldberg and Tom Lynch were granted U.S.
Patent 4,845,337 for a novel low power ovenized crystal oscillator assembly.
Efratom developed their EMXO as a low power OCVCXO.
93
Jechart came to the U.S. in 1973 and Huebner remained in Munich to run
the business there.
 21

Fig. 52. Symmetricom 8122 RFS Module. Photo Credit: Symmetricom.

Fig. 49. Datum LPRO RFS. Photo Credits: Datum.
An Efratom FRK rubidium oscillator was part of the atomic
The Symmetricom 8130A (Fig. 50) was designed circa clock exhibit at the Smithsonian Institution in Washington,
2002 as a modernized equivalent to the Efratom M-100 with DC between 1982-1988. Two FRK units flew on the NRL
the smaller size of the EG&G RFS-10 (see Fig. 50). M-100 era NTS-1 spacecraft, becoming the first atomic clocks in space.
tactical RFS units used MIL-spec parts that were increasingly Efratom supplied rubidium physics packages to Rockwell for
hard to obtain, while modern electronic parts had achieved use on the Blocks I, II and IIA GPS navigation satellites95, and
comparable reliability and offered more functionality (e.g., for a rubidium oscillator that was on the Huygens probe that
DDS synthesis). The semiconductors in the 8130A were, how- landed on Saturn’s moon Titan in 2005 [137].
ever, subjected to a custom screening process. Even that is
seldom required for today’s military RFS applications where
commercial units have been successfully adapted for use in
harsh environments94.

Ernst Jechart (1936-1991) Hugo Fruehauf

Co-Founder of Efratom President of Efratom
Photo Credit: GPS World. Photo Credit: GPS World.

Fig. 50 Symmetricom Model 8130A RFS. Photo Credit: Symmetricom.

The Symmetricom X72 is a very small RFS for commercial

telecom applications (see Fig. 51). It uses a conventional Rb
lamp and a tiny cell inside a unique microwave resonator in-
vented by Jinquan Deng [178].

Werner Weidemann (1943-2008) Tom English

Photo Credit: IEEE UFFC. Photo Credit: Logos Research.
Fig. 53. Major Efratom RFS Contributors.

Hugo Fruehauf, the president at Efratom during most of its

time in Irvine, CA, came from Rockwell when they began to
Fig. 51. Symmetricom X72 RFS. Photo Credit: Symmetricom. use the Efratom Rb technology for their GPS satellite pro-
More recently, Symmetricom has designed and manufac- gram.96 He was a charismatic leader with the highest profes-
tured RFS modules for the F-22 and F-35 aircraft based on the sional and personal standards.97 Later, he went to work for FEI
X72, as shown in Fig. 52. and is now a consultant.
As of 1993, more than 35,000 Efratom rubidium oscillators
F. Litton
had been made for applications such as commercial and mili-
tary communications systems, navigation systems, and me- Litton Guidance & Control Systems in Woodland Hills, CA
trology. By 1996, annual production exceeded 100k units, entered the field of Rb frequency standard technology as an
mainly for telecom applications, and there are now more than adjunct of their work on rubidium NMR gyros in response to a
a half-million units in use. 1983 USAF RADC contract for a tactical RFS (in which

95
There were a total of 17 Rockwell/Efratom RFS on the 11 Block I GPS
94 The Symmetricom 8130A MIL RFS was a nice design, especially the low S/Vs.
phase noise version, but reached the market too late to have any impact. 96
Hugo relates a story about the U.S. Government threatening to shut down
Besides the author, its designers were Ken Lyon and Larry Zanca, the GPS RFS work and deport Jechart because he was not a citizen.
97
He received the 2019 Queen Elizabeth Prize for Engineering.
 22

EG&G and Efratom also participated separately). We at

EG&G were fascinated to follow their (public) reports as they
independently and quickly rediscovered the basics of RFS
physics. The principal contributors to the Litton work were
Tae Kwon98, Tom McClellan and Bruce Grover [149]. The
Litton RFS technology was sold to Frequency Electronics in
or about 1985.
G. Frequency Electronics
Frequency Electronics (FEI) was founded by Martin Bloch
in 1962, and has become a major player in the T&F field, es-
pecially for ultrastable space crystal oscillators and custom Fig. 56. FEI FE-5680A Commercial Telecom RFS. Photo Credit:
Frequency Electronics.
designs for Government programs. It enjoys a good reputation
for delivering promised performance, and Marty Bloch grew
FEI into a very successful company99. Frequency Electronics
acquired the Litton Rb technology in circa 1985, and Tae
Kwon and Tom McClelland went to FEI in Mitchell Field,
NY. Figure 54 shows the two major RFS contributors at FEI.

Fig. 57. FEI FE-5680A Commercial Telecom RFS. Photo Credit:

Frequency Electronics.

H. AccuBeat
AccuBeat is a relatively small but well-established compa-
ny in Israel founded by Avi Stern that designs and manufac-
Martin Bloch Tom McClellan turers time and frequency products including rubidium fre-
Photo Credit: ION. Photo Credit: LinkedIn. quency standards. AccuBeat predecessors include FEI,
Fig. 54. Major FEI RFS Contributors. Tadiran and Time & Frequency, Ltd. (TFL, e.g., the TF-4000,
TF-4020 and TF-4030 RFS). Some of its products have been
FEI utilized that technology in an RFS subsystem for the sold through PerkinElmer/Excelitas in the U.S., and apparent-
MILSTAR and Advanced EHF communications satellites (see ly they now partner with Bliley. An AccuBeat commercial
Fig. 55), and later for several commercial RFS products and RFS is shown in Fig. 58.
also a 2nd source Rb clock for GPS. 19 FEI MILSTAR Rb
systems were delivered as of 2004. Examples of FEI commer-
cial RFS products are shown in Figs 56 and 57. The FE-
5650A was used in Lucent cell phone towers.

Fig. 58. AccuBeat AR40A RFS. Photo Credit: AccuBeat.

I. Stanford Research Systems

Stanford Research Systems (SRS) has an interesting and
Fig. 55. FEI MILSTAR Rubidium Master Oscillator. Photo Credit: diverse product line that includes rubidium frequency stand-
Frequency Electronics. ards and related instruments (e.g., the SRS620 counter, for
which it was designed as a time base). Their PRS10 Rubidium
Oscillator (see Fig. 59) has good specifications and a reputa-
tion for excellent performance. It utilizes a low noise
OCVCXO, a lamp with a large Rb reservoir, digital control,
and sinewave servo modulation.
98
Tae Kwon left Efratom in 1978 to join Litton.
99
FEI encountered legal problems when it was suspended from eligibility for
U.S. Government prime contracts in December 1993 based on an indictment
of four of its officers for conspiracy to defraud the Government. That suspen-
sion lasted through June 1998, when the charges were dismissed.
 23

development of rubidium frequency standard technology.

At the beginning, the National Bureau of Standards (NBS,
now NIST) and the Naval Research Laboratory (NRL) were
involved in the applied research that led to RFS instruments.
They have continued to stay involved, with NIST recently
doing early work on the Chip Scale Atomic Clocks (CSAC)
and NRL serving as a ground life test facility for GPS Rb
Fig. 59. Stanford Research PRS-10 RFS. Photo Credit: Stanford Research
clocks [130]. The U.S. Naval Observatory continues to closely
Systems. monitor those clocks in-orbit. The U.S. Army supported sev-
eral RFS-related programs (e.g., RbXO, MIFTTI, SCOTT)
The PRS10 includes hardware and firmware that allows it during the time Ft. Monmouth was still open. The U.S. Air
to be locked to an external 1 PPS signal from, for example, a Force has, in the past, supported several tactical communica-
GPS timing receiver, as shown in Fig. 60. The principal de- tions programs that employ RFS units (e.g., Have-Quick,
signer of the PRS-10 was John Willison. Seek-Talk, TRC-170), and it, of course has the major role in
the GPS program. NASA has occasionally supported RFS
work (e.g., SATS) and continues to support the JPL efforts on
the Deep Space Atomic Clock (DSAC) Hg ion clock which
has some similarities to an RFS100.
The quasi-government Aerospace Corporation has played a
significant role in RFS research and has provided important
support to the MILSTAR and GPS programs.
M. International Organizations
Several private and governmental organizations are active
in rubidium frequency standard technology outside the United
Fig. 60. PRS-10 ADEV Plot with 1 PPS Reference. Photo Credit: Stanford States. The former manufacture a line of RFS products while
Research Systems.
the latter are mainly associated with GNSS applications.
J. Quartzlock
N. Switzerland
Quartzlock in the UK has offered rubidium frequency
The atomic clock activity in Switzerland has been centered
standards as part of its product line for quite some time (e.g.,
at the Neuchâtel Observatory and various companies located
circa 1986 or earlier). These units were apparently first associ-
there.
ated with Dartington Frequency Standards (also located in
Devon, UK). The Quartzlock A10 Bench RFS (see Fig. 61) 1) Orolia
data sheet of 2002 includes a brief RFS history [79]. Orolia has a product line that includes commercial and
Quartzlock also developed (circa 2015) an E10-SPC space space rubidium frequency standards (see Fig. 62). Spectracom
qualified Rb clock. and SpectraTime (previously Temex Neuchâtel Time – found-
ed in 1995 - and Temex Time) are Orolia subsidiaries. Their
principal technologist is Pascal Rochat, managing director of
SpectraTime, Inc.
Orolia built the RAFS for the both the Galileo and Indian
IRNSS navigation satellites, which have both experienced
Fig. 61. Quartzlock A10 Bench RFS. Photo Credit: Quartzlock.
early in-orbit failures due to a “faulty component”, a problem
now believed to be resolved. These Rb clocks may also have
K. Small Companies been used by China in the BeiDou GNSS system. Their stabil-
It has always been difficult for a small startup company to ity (see Fig. 63), although excellent, is not quite as good as the
succeed in the RFS business. The technology is fairly complex Excelitas GPS RAFS (see Figs 40 and 41).
and customers tend to be large organizations that expect in-
depth support.
Aqtron was a small company that briefly manufactured and
sold rubidium frequency standards circa 1982. It was owned
by William R. Fowks who had previously worked on those
devices at Rockwell.
CJI Technology was another name from the past. IQD is a
current RFS name in the UK.
L. U.S. Government Fig. 62. Orolia LPFRS and Galileo RAFS. PhotoCredit: Orolia.
The U.S. Government has played a significant role in the
100
A DSAC demonstrator unit was launched in June 2019.
 24

R. India
India has recently developed an indigenous Rb clock for
their Indian Regional Navigation Satellite System (IRNSS). It
may use either an Rb clock from SpectraTime or one devel-
oped locally. Information about the latter will be found in Ref-
erence [140].
S. Italy and France
The French telecom company Thomson-CSF (now Thales)
made a HAM-111 laboratory RFS in the early 1970’s. Most of
the current atomic clock activity in these countries is research
oriented at organizations such as IEN/INRIM and CNRS.
T. Patents
Fig. 63. Typical Stability of Galileo RAFS. Plot Credit: Spectratime.
Patents have never been a significant factor in RFS com-
O. Japan merce and very few have resulted in royalty payments 101.
Japan has had for many years indigenous RFS technology Nevertheless, they can make for interesting reading (see patent
that has been applied mainly by telecom companies for inter- references herein), and some are quite informative. Trade se-
nal use. RFS manufacturers have included Fujitsu, Nippon cret protection regarding Rb physics package processes is
Electric Company (NEC, e.g., Neatomic Rb-1003) and Japan much more commonly used. Recent devices often contain
Radio Company (JRC). firmware that is generally tightly controlled. 
The Japanese Quasi-Zenith Satellite System (QZSS) is a
communication, broadcasting and positioning satellite system U. RFS Formats
that offers a complementary service for GPS. It uses an RFS devices are available in a number of formats to serve a
Excelitas Rb clock similar to that of GPS-IIF [139]. variety of applications:
 Components (e.g., CSAC – though mainly Cs)
P. Russia  Boards (with CSAC or small RFS)
In general, I know little about Russian RFS technology and  Modules (most small RFS units)
products, but they have a long history with Rb, Cs and H-  Black Boxes (military, avionics and aerospace units)
 Instruments (with power supply & controls)
maser atomic clock technology.
 Systems (with time/frequency distribution, etc.)
The Russian Global Navigation Satellite System
(GLONASS) space vehicles use a mixture of Russian-made Quite a few companies bundle RFS units into more com-
Rb and Cs atomic clocks [139]. plete assemblies and the core Rb manufacturer sometimes
There also appear to be Russian manufacturers of small isn’t clear.
commercial RFS units, including Ruknar, which was founded
in 1997 (see Fig. 64). V. RFS Applications and Markets
RFS applications gradually evolved from laboratory stand-
ards to military communications and avionics equipment and
then to commercial telecommunications with yearly manufac-
turing volumes increasing from hundreds to tens of thousands,
influenced greatly by the advent of GPS. The combination of a
GPS time reference and a small Rb oscillator for short-term
stability and holdover is essential to today’s telecom networks.
It’s interesting that the trend toward smaller and lower cost
Fig. 64. Ruknar CH1-1012 RFS. Photo Credit: Ruknar JSC . RFS has been accompanied by generally lower performance
except for the GPS/GNSS clocks where performance has been
Q. China significantly improved. One report puts the global RFS market
Little general information is known about Chinese RFS at $130 million in 2019 with an annual growth rate of 6.8%
technology and products. and about half produced in the U.S.
Chengdu Spaceon Electronics Co., Ltd. offers RFS prod-
ucts, as may the China Aerospace Science and Industry Com- Over the years, high-end RFS performance has improved
pany (CASIC). remarkably by 2 to 3 order of magnitude with flicker floors
Rb clocks are used onboard the BeiDou GNSS system
[139]. In late 2018, the BeiDou-2 constellation consisted of 14 101
For example, Jechart’s U.S Patents 3,798,565 and 3,903,481 on the
active satellites all of which use Rb clocks. Per a 2018 paper, integrated resonance cell and mixed lamp isotopes for light shift reduction did
not bring in revenue as far as I know, nor did Goldberg’s 4,494,085 which
the stability of the improved Rb AFS on BeiDou-3 is compa- was infringed but never seriously enforced. The legal costs of enforcement
rable to that of the Galileo RAFS [138]. are higher than any returns. Some, like Riley’s 4,721,890 have been honored
and worked around by competitors.
 25

dropping from pp1011 to pp1014. Similarly, RFS volumes have devices until the mid-1990’s. Discrete semiconductors were
been reduced from the size of a microwave oven to a deck of gradually replaced by ICs, and better devices (e.g., op amps)
cards, power has been reduces from 10’s of watts to fractions improved RFS performance and reliability. Some of today’s
thereof, and the cost of commercial units has gone from the designs have become largely digital, as DDS/PLL synthesizers
price of a car to 5% of that. enabled high-resolution tuning, perfect squarewave FM, se-
lectable output frequency and eased absorption cell buffer gas
W. Surplus RFS Units
fill tolerances. Microprocessors, DSP and firmware have ena-
Used RFS units, particularly Efratom LPROs (see Fig. 65), bled digital frequency lock servos, fast lock acquisition, better
are currently available on the surplus market for around $250 oven control102, temperature compensation, C-field commuta-
or even less. These are apparently some of the thousands re- tion, digital monitors, user interfaces, external locking and
moved from telecon cell sites for preventive maintenance or self-contained calibration, test and aging measurements.
technical obsolescence. Most still operate fine and have years
of life left, and are used by radio amateurs and hobbyists, of- B. New Technologies
ten with GPS disciplining. These days, there is little reason for The classic rubidium frequency standard hasn’t changed
anyone who needs or wants an atomic clock to not have one. drastically since its beginnings (see early patents), but rather
has enjoyed gradual refinement including improved perfor-
mance along with reduced size, weight, power and cost as
advances have been made in their physics packages and elec-
tronics. In particular, digital synthesis and signal processing
have replaced analog circuitry. The three most significant re-
cent technical innovations, laser pumping, CPT interrogation
and MEMS micro-fabricated physics packages, are based on
the availability of semiconductor diode lasers.
Fig. 65. Surplus Efratom LPRO RFS. Photo Credit: E-Bay.

X. GPS and RFS 1) Diode Laser Optical Pumping Source

Replacing the conventional Rb spectral lamp and lamp ex-
It can be argued that GPS is the single most important thing
citer (and eliminating the Rb-85 isotopic filter cell) with a
affecting RFS usage since that system was invented. High
single mode diode laser (e.g., VCSEL, DBR or ECL) with its
performance Rb clocks have turned out to be the best match
current source (and likely a cooler/temperature controller) is a
for the time reference requirements on board a GPS satellite
potentially attractive choice for an RFS optical pumping
thanks to their combination of short and medium term stability
source that offers more efficient pumping, a larger signal and
(out to several days) along with practical factors like size,
higher S/N ratio. The main challenges are laser noise, wave-
weight, power and (especially) life and reliability. On the
length/light shift control, and device availability and reliabil-
ground, the combination of a GPS reference and an inexpen-
ity. This approach has been applied to high performance units
sive modest performance RFS provides an ideal time reference
with complex laser systems that have achieved order-of-
for telcom applications at thousands of cell sites. The fit be-
magnitude better short term stability, to small Rb and Cs de-
tween RFS capabilities and those market needs has resulted in
vices using conventional microwave interrogation, and to
Rb clocks with exceptional performance on one hand and re-
small RFS units using CPT interrogation.
markably low cost on the other.
Y. T&F Industry Business Relations The evolution of ultra-miniature gas cell physics packages
With very few exceptions, the author has found business re- arguably began with a laser-pumped Cs device using a very
lations within the time and frequency community to be cordial small 0.2 cm3 cell in a heavily dielectrically loaded cavity at
and collegial, even between competitors. Business is conduct- Westinghouse (later Northrop-Grumman) in the early 1990's
ed fairly and information is shared quite freely, with an em- by P. J. Chantry, I. Liberman, et al., see Figure 66 [194],
phasis on program success by all participants. It is a diverse [296], [297].
international group from academia, government, and industry. Fig. 66.
Personal friendships and organizational ties are fostered at Ultra-Miniature Cs Resonator
meetings and conferences. Similarly, the metrology, military, Photo Credit: Northrop-Grumman
telecom and aerospace customers for T&F products in gov-
ernment and industry are, almost without exception, knowl-
edgeable, competent and a pleasure to deal with.

V. NEW TECHNOLOGIES
A. Analog to Digital
Even the earliest RFS units in the mid-1960’s were solid 102
For example, digital oven controllers use A/D and D/A converters and DSP
state (except for the first lamp exciters which used vacuum between their thermistor sensors and power transistors (instead of just a wire)
but gain programmable set point and demand power along with improved
tubes), but Rb frequency standards remained mostly analog control dynamics.
 26

2) Coherent Population Trapping 3) Chip Scale Atomic Clocks

Coherent population Trapping (CPT) is an alternative interro- The Chip Scale Atomic Clock (CSAC, see Figs 70 through 71)
gation method that has been successfully used to implement is a promising newer technology related to the RFS 106, 107, 108.
both Rb and Cs gas cell atomic clocks 103, 104. Its implementa- Although mainly using Cs rather than Rb gas cells, and gener-
tion uses wideband modulation of a VCSEL diode laser at ally having lower performance, a CSAC offers significantly
one-half the atomic hyperfine frequency to excite the atoms in smaller size, lower power and potentially lower cost than a
a gas cell with the resulting optical sidebands, thereby elimi- traditional RFS.
nating the need for a bulky microwave cavity.
A photograph of a 2004 NIST CSAC physics package is
Jacques Vanier traces the history shown in Figure 69. The principal investigators for this work
of CPT from the 1976
were John Kitching and Leo Hollberg with Svenja Knappe.
experimental work of Alzetto et al
with sodium in [324] and the 1979 Quoting from NIST “The physics
theory of Orriols in [325], later package of the NIST chip-scale
interpreted by him in [3] and atomic clock includes (from the
[329]. The use of CPT in a bottom) a laser, a lens, an optical
attenuator to reduce the laser power,
frequency standard was proposed a waveplate that changes the
by Cyr, Têtu and Breton in [326] polarization of the light, a cell
Fig. 67. Jacques Vanier. Photo
and followed up by Levi, Godone containing a vapor of cesium atoms,
and Vanier in [327] and [328]. and (on top) a photodiode to detect
Credit: IEEE
the laser light transmitted through the
cell. The tiny gold wires provide
An early CPT clock was developed at Kernco in 2003 with Figure 70. NIST CSAC Physics electrical connections to the
Package. Photo Credit: NIST
ONR sponsorship105. CPT atomic clocks have found a sub- electronics for the clock.”
stantial place in the atomic clock market for small, low power
devices with modest performance (see Figs 68 and 69 for ex- Commercial CSAC products are shown in Figs. 71 and 72.
amples). References herein for CPT interrogation are [229]
and [323] through [328]. A CPT clock is nicely described by
Vanier in U.S. Patent 6 320 472 [323].

Fig. 71. Microsemi CSAC & Physics Package. Photo Credit: Microsemi.
& Symmetricom.

Fig. 72. The Symmetricom CSAC

Fig. 68. Microsemi SA.3Xm CPT Rb Miniature Atomic Clock. Reference is used in products made by
[229] gives an excellent description of a small, low power Rb clock (the Jackson Labs. Photo Credit:
Symmetricom, now Microsemi SA.3Xm) employing CPT interrogation. Jackson Labs.
This technique provides performance equivalent to a conventional RFS
unit in a much smaller package, and seems to represent the future of those
devices. Photo Credit: Microsemi.

106
John Vig deserves much credit for spurring CSAC development by
bringing together the MEMS and frequency control communities, followed by
several DARPA-sponsored programs that led to the development of working
devices by NIST, Symmetricom/Microsemi and others. Work by Jacques
Vanier and others on CPT interrogation was an important precursor. At NIST,
John Kitching showed theoretically that tiny gas cells can have usefully-small
resonant linewidths. NIST pioneered the design and fabrication of CSAC cells
Fig. 69. AccuBeat NAC-1 CPT Rb Atomic Clock. Photo Credit: and devices [Ref, Ref, Ref] At Symmetricom, the CSAC work was led by
AccuBeat. Mike Garvey, with Robert Lutwak devising the physics with electronics
support by the author. Draper Labs MEMS technology was crucial, as were
VCSEL devices from Sandia Labs.
107
The majority of early CSAC devices have used Cs gas cells rather than Rb.
103
CPT interrogation was first extensively investigated by Jacques Vanier. At Symmetricom, this was somewhat due to familiarity with optically-pumped
104
I recall sitting with Jacques Vanier in a Neuchătel crĕps shop during an Cs beam tube technology, and with more readily available VCSELs at the Cs
EFTF conference as he tried to explain to me (without much success) CPT wavelength.
108
and quantum superposition states. CSAC devices are increasingly finding a place in the commercial atomic
105
Contributors at Kernco included Bob Kern, Marty Levine, Mike Delaney, clock market, and low noise and rad-hard space versions are available. As of
Dan Janssen, and Cam Everson, with consultation by Jacques Vanier and Sy 2018, over 95,000 units had been sold by Microsemi [Ref].
Goldberg.
 27

The most important CSAC attribute is arguably its low power Tom English, Hugo Fruehauf, John Kitching, Jacques Vanier
consumption compared with other atomic clocks as shown in and John Vig. I thank you all very much.
Fig. 73. CSAC references herein are [220] through [228].
Rev. A of this paper mainly adds additional information about
CPT compared with the original version of July 2019.
B. Professional Organizations
The three professional organizations listed below have pro-
vided for the exchange of information in the form of confer-
ences, tutorials, standards and other committees and personal
contacts within the time and frequency community:
 The IEEE Ultrasonics, Ferroelectrics and Frequency Control Society
(UFFC) and the IEEE International Frequency Control Symposium
Fig. 73. Atomic Clock Power Consumption. Fig. Credit: J. Kitching [228]. (FCS). Access to abstracts from the annual FCS Proceedings is open, but
Note: This is an excellent review of CSAC devices. UFFC membership or an IEEE Xplore subscription is required to obtain
the full text.
 The annual Precise Time and Time Interval (PTTI) conference. An index
VI. COMMENTS AND CONCLUSIONS to the full PTTI papers between 1969 and 2012 is available. After that,
A. Comments one must go to the Institute of Navigation (ION) web site, where mem-
bership is required for full paper download.
It is quite remarkable that the three main atomic clock  The annual European Frequency and Time Forum (EFTF). Complete
technologies today (Cs beam, Rb gas cell and H maser) remain proceedings are available for free download since 1987.
basically the same as those in the 1960’s. While there have
Early RFS work was reported mainly at the FCS and the
been numerous additions to the types of laboratory/primary
scientific literature. Later, FCS papers tend more toward de-
standards, few fundamentally new types of commercial atomic
vices and PTTI papers toward applications (e.g., GPS), while
clocks have found acceptance, only variants like CPT interro-
the EFTF emphasizes European work.
gation, laser pumping and CSAC fabrication.
Also noteworthy are (a) the huge number of medium-
ABBREVIATIONS AND ACRONYMS
performance low-cost commercial Rb clocks that are utilized
by the telecommunications industry, (b) the exceptionally high The following abbreviations and acronyms are used herein.
performance and reliability that has been achieved by the GPS ADEV Allan Deviation
Rb clocks, and (c) the general improvement in electronic AFB Air Force Base
AFS Atomic Frequency Standard
component reliability that allows use of commercial Rb de- Ar Argon
signs in rugged military environments. ATP Acceptance Test Procedure
DBR Distributed Bragg Reflector
B. Conclusions BTC Baseplate Temperature Controller
The rubidium frequency standard has been remarkably suc- CAS Collision Avoidance System
CMB Cosmic Microwave Background
cessful and can be expected to continue to serve as the most CNRS National Center for Scientific Research (France)
widely used atomic clock. Change comes slowly to this field, CPT Coherent Population Trapping
but it likely that small laser-pumped, CPT-interrogated units Cs Cesium
CSAC Chip Scale Atomic Clock
will become increasingly used. The originators of RFS tech- DARPA Defense Advanced Research Projects Agency
nology are disappearing from the scene and one hopes that the DDS Direct Digital Synthesizer
next generation of atomic clock technologists will continue to DSP Digital Signal Processing
DSC Differential Scanning Calorimeter
follow with new ideas and accomplishments 109. DSAC Deep Space Atomic Clock
DVT Design Verification Test
ACKNOWLEDGEMENTS ECL External Cavity Laser
EFTF European Frequency and Time Forum
A. Credits EG&G Edgerton, Germeshausen and Grier
EHF Extremely High Frequency
The figures herein without explicit credit are from the au- EMI Electromagnetic Interference
thor’s personal papers and archives, and/or Reference [5]. ESS Environmental Stress Screening
Product photographs are generally from their manufacturer’s FCS Frequency Control Symposium
FEI Frequency Electronics Incorporated
literature or web site. Thanks are due to the Princeton physics FM Frequency Modulation
department for the picture of Tom Carver. FTS Frequency and Time Systems
FXR Flash X-Ray
GE General Electric
Several colleagues have reviewed and provided construc- GNSS Global Navigation Satellite System
tive comments about this history paper, including Peter Cash, GPS Global Positioning System
GR General Radio, GenRad
109
The history of the rubidium gas cell atomic frequency standards extends H Hydrogen
back some 60 years. Looking forward, it seems clear that there will be a con- HDEV Hadamard Deviation
tinuing need for precise time and frequency using these and similar devices. He Helium
HP Hewlett Packard
 28

IEEE Institute of Electrical and Electronic Engineers HISTORICAL OBSERVATIONS (by Tom English)
IEN National Electrotechnical Institute (Italy)
INRIM National Institute for Metrological Research (Italy)
IRNSS Indian Regional Navigation Satellite System
I&M Instrumentation & Measurement
ION Institute of Navigation
IRE Institute of Radio Engineers
ITT International Telephone and Telegraph Company
JPL Jet Propulsion Laboratory
JRC Japan Radio Company
Kr Krypton
LTE Long-Term Evolution
MEMS Micro Electro-Mechanical System
MIFTTI Modular Intelligent Frequency Time Time Interval
MIT Massachusetts Institute of Technology
NASA National Aeronautics and Space Administration
NBS National Bureau of Standards
Ne Neon
NEC Nippon Electric Company
NIST National Institute of Standards and Technology
NMR Nuclear Magnetic Resonance
NRL Naval Research Laboratory
OCXO Oven Controlled Crystal Oscillator
OCVCXO Oven Controlled Voltage Controlled Crystal Oscillator
OMA Optical Microwave Assembly
ONR Office of Naval Research
PLL Phase Locked Loop
PM Phase Modulation
PTTI Precise Time and Time Interval
QZSS Quasi-Zenith Satellite System
RADC Rome Air Development Center
RAFS Rubidium Atomic Frequency Standard
R&S Rohde & Schwartz
Rb Rubidium
RbXO Rubidium Crystal Oscillator
RFS Rubidium Frequency Standard REFERENCES
SAO Smithsonian Astrophysical Observatory
SATS Spacecraft Atomic Timing System The following rubidium frequency standard references are
SCI Space Craft Incorporated listed chronologically according to their type. Note that a sub-
SCOTT Single Channel Objective Tactical Terminal
S/N Signal to Noise
scription to the IEEE Xplore Digital Library is recommended
SRD Step Recovery Diode for access to many of these references.
STL Space Technology Laboratories
S/V Space Vehicle General Atomic Clocks
SWP Size, Weight and Power [1] A. O. McCoubrey, “A Survey of Atomic Frequency Standards,” Proc.
TC Temperature Coefficient IEEE, vol. 14, no. 2, pp. 116-135, Feb. 1966.
T&F Time and Frequency [2] H. Hellwig, “Areas of Promise for the Development of Future Primary
TMXO Tactical Miniature Crystal Oscillator Frequency Standards,” in Proc. 24th Annu. Symp. Freq. Contr., Atlantic
TRFS Tactical Rubidium Frequency Standard City, NJ, USA, Apr. 1970, pp. 246-258.
TRW Thompson Ramo Wooldridge [3] J. Vanier and C. Audoin, The Quantum Physics of Atomic Frequency
UFFC Ultrasonics Ferroelectrics and Frequency Control Standards, vol. 2, Bristol, UK and Philadelphia, PA, USA: Adam Hilger,
VCSEL Vertical Cavity Surface Emitting Laser 1989.
VCXO Voltage Controlled Crystal Oscillator [4] F. G. Major, The Quantum Beat: Principles and Applications of Atomic
Xe Xenon Clocks, 2nd ed., New York, NY, USA: Springer-Science, 2007.
XO Crystal Oscillator [5] W. J. Riley, Rubidium Frequency Standard Primer, Morrisville, NC,
USA: LuLu Press, 2011.

History
[6] A 1966 Nobel Prize biography for Alfred Kastler.
[7] R. H. Beehler, “A Historical Review of Atomic Frequency Standards,”
Proc. IEEE, vol. 55, no. 6, pp. 792-805, June 1967.
[8] N. F. Ramsey, “History of Atomic and Molecular Standards of Frequen-
cy and Time,” IEEE Trans. Instrum. Meas., vol. IM-21, pp. 90-99, May
1972.
[9] H. Hellwig, Frequency Standards and Clocks – A Tutorial Introduction,
NBS Technical Note 616, Apr. 1972.
[10] R. H. Beehler, “Recent Progress on Atomic Frequency Standards,” Time
and Frequency: Theory and Fundamentals, NBS Monograph 140, U.S
Department of Commerce/National Bureau of Standards, pp.101-109,
May 1974.
[11] P. Forman, “Atomic Clocks: Preview of an Exhibit at the Smithsonian,”
in Proc. 36th Annu. Symp. Freq. Contr., Philadelphia, PA, USA, June
1982, pp. 220-222.
[12] N. F. Ramsey, “History of Atomic Clocks,” Journal of Research of the
National Bureau of Standards, vol. 88, no. 5, Sep.-Oct. 1983.
 29

[13] C. O. Alley, “Proper Time Experiments in Gravitational Fields with [39] A. Kastler, “Displacement of Energy Levels of Atoms by Light,” J. Op.
Atomic Clocks, Aircraft, and Laser Light Pulses,” in Quantum Optics, Soc. Am., vol. 53, no. 8, pp. 902-910, Aug. 1963.
Experimental Gravity, and Measurement Theory, Plenum Publishing [40] M. Arditi and T. R. Carver, “Hyperfine Relaxation of Optically Pumped
Corp., 1983. Rb87 Atoms in Buffer Gases,” Physical Review, vol. 136, pp. A643-
[14] J. S. Rigden, Rabi, Scientist and Citizen, New York, NY, USA: Basic A649, Nov. 1964.
Books, 1987. [41] R. Bernheim, Optical Pumping, New York, NY, USA: W.A. Benjamin,
[15] N. F. Ramsey, “The Past, Present, and Future of Atomic Time and Fre- 1965.
quency,” Proc. IEEE, vol. 79, no. 7, pp. 921-926, July 1991. [42] R.W. Frank, et al., “Design Concepts Study for an Aerospace Rubidium
[16] A. O. McCoubrey, “History of Atomic Frequency Standards: A Trip Standard,” General Radio Company, Bolton, MA, USA, unpublished re-
Through 20th Century Physics,” in Proc. 1996 IEEE Int. Freq. Contr. port, Oct. 1969.
Symp, Honolulu, HI, USA, June, 1996, pp. 1225-1241. [43] B. E. Blair, Editor, Time and Frequency: Theory and Fundamentals,
[17] W. Happer, P. J. Peebles, and D. T. Wilkinson, “Robert H. Dicke, a NBS Monograph 140, U.S. Department of Commerce, May 1974.
Biographical Memoir”, Nat. Acad. Press, Washington, DC, USA, 1999. [44] G. Missout and J. Vanier, “Some Aspects of the Theory of Passive Ru-
[18] C. Audoin and B. Guinot, The Measurement of Time, Cambridge, U.K.: bidium Frequency Standards,” Canadian Journal of Physics, vol. 53, no.
Cambridge University Press, 2001. 11, pp. 1030-1043, 1975.
[19] N. F. Ramsey, “History of Early Atomic Clocks,” Metrologia, vol. 42, [45] J. Vanier and C. Audoin, The Quantum Physics of Atomic Frequency
Issue 3, Article S01, June 2005. Standards, Adam Hilger, 1989, ISBN 0-85274-433-1.
[20] L. Mallette, P. Rochat, and J. White, "Historical Review of Atomic
Frequency Standards Used in Satellite Based Navigation Systems,” in Early RFS Research and Development
Proc. 63rd Annu. Meeting of the Institute of Navigation, Cambridge, [46] T. R. Carver, “Rubidium Oscillator Experiments,” in Proc. 11th Annu.
MA, Apr. 2007, pp. 40-48. Symp. Freq. Contr., Asbury Park, NJ, USA, May 1957, pp. 307-317.
[21] J. Camparo, “The Rubidium Atomic Clock and Basic Research,” [47] M. Arditi, “Gas Cell Atomic Clocks Using Buffer Gases and Optical
Physics Today, vol. 60, No, 11, pp. 33-39, Nov. 2007. Orientation,” in Proc. 12th Annu. Symp. Freq. Contr., Asbury Park, NJ,
[22] J. Levine, “The History of Time and Frequency from Antiquity to the USA, Apr. 1958, pp. 606-622.
Present Day,” Eur. Phys. J. H, Mar. 2016, doi: 10.1140/epjh/e2016- [48] M. Arditi, “Evaluation of ITT Breadboard Gas Cell Frequency Stand-
70004-3. ard,” in Proc. 13th Annu. Symp. Freq. Contr., Asbury Park, NJ, USA,
May 1959, pp. 655-667.
Early Scientific Papers [49] J. M. Andres, D.J. Farmer, and G.T. Inouye, “Design Studies for a Ru-
[23] J. Brossel and A. Kastler, “Résonance Optique,” C.R. Acad. Sci. (Paris) bidium Gas Cell Frequency Standard,” IRE Trans. Mil. Electron., MIL-
vol. 229, pp. 1213-1215, Dec. 1949. 3, pp. 178-183, Oct. 1959.
[24] F. Bitter, “The Optical Detection of Radiofrequency Resonance,” Phys. [50] M. Arditi, “A Gas Cell “Atomic Clock as a High-Stability Frequency
Rev., vol. 76, no. 6, pp. 883-885, Sep. 1949. Standard”. IRE Trans. Mil. Electron., vol. MIL-4, no. 1, pp. 22-28, Jan.
[25] A. Kastler, “Quelques suggestions concernant la production optique et la 1960.
détection optique d'une inégalité de population des niveaux de [51] R. J. Carpenter, et al., “A Prototype Rubidium Vapor Frequency Stand-
quantifigation spatiale des atomes. Application à l'expérience de Stern et ard,” IRE Trans. Instrum., vol. I-9, no. 2, pp. 132-135, Sep. 1960.
Gerlach et à la résonance magnétique,” [Some suggestions concerning [52] M. Arditi, “Frequency Control by Gas Cell Standards – Fundamental
the optical production and the optical detection of a population Problems in the Light of Recent Experimental Results,” in Proc. 15th
inequality of the levels of spatial quantification of atoms. Application to Annu. Symp. Freq. Contr., Atlantic City, NJ, USA, Apr. 1961, pp. 181-
the Stern and Gerlach experiment and magnetic resonance], (in French), 202.
J. Phys. Radium, vol. 11, no. 6, , pp. 255-265, June 1950. [53] M. Arditi and T. R. Carver, “The Principles of the Double Resonance
[26] A. Kastler, “Méthodes optiques d'étude de la résonance magnétique,” Method Applied to Gas Cell Frequency Standards,” Proc. IEEE, vol. 51,
[Optical methods of studying magnetic resonance], (in French), Physica, no. 1, pp. 190-202, Jan. 1963.
vol. 17, no. 3-4, pp. 191-204, Mar.-Apr. 1951.. [54] D. J. Farmer, “Performance and Application of Gas Cell frequency
[27] H. G. Dehmelt, “Slow Spin Relaxation of Optically Polarized Sodium Standards,” in Proc. 17th Annu. Symp. Freq. Contr., Atlantic City, NJ,
Atoms,” Phys. Rev., vol. 105, no. 5, p. 1487, Mar. 1957. USA, Apr. 1963, pp. 449-461.
[28] R. H Dicke, “The Effect of Collisions Upon the Doppler Width of Spec- [55] R. J. Carpenter, “A Portable Rubidium Vapor Frequency Standard,”
tral Lines,” Phys. Rev., vol 89, no. 2, pp. 472-473, Jan. 1953. NBS Technical Note 235, Apr. 1965.
[29] R. H. Dicke, “Atomic and Molecular Frequency Standards,” in Proc. [56] Technical Information, R-20 Rubidium Frequency Standard, Varian
10th Annu. Symp. Freq. Contr., Asbury Park, NJ, USA, May 1956, p. Quantum Electronics Division, Varian Associates, Palo Alto, CA, Circa
259-267. 1966.
[30] T. R. Carver, “Rubidium Oscillator Experiments,” in Proc. 11th Annu. . [57] A. R. Chi, et al., “Long-Term Stability Frequency Stability Measure-
Symp. Freq. Ctrl., Asbury Park, NJ, USA, May 1957, pp. 307-317. ment of Rubidium Gas Cell Frequency Standards,” in Proc. 22nd Annu.
[31] W. E. Bell and A. L. Bloom, “Optical Detection of Magnetic Resonance Symp. Freq. Contr, Atlantic City, NJ, USA, Apr. 1968, pp. 592-604.
in Alkali Metal Vapor,” Phys. Rev. 107, 1559, Sep. 1957. [58] R. Lash and M. Fremont, BAMBI SETI Site Varian V-4700A Rubidium
[32] P. L. Bender, E. C. Beaty, and A. R. Chi, “Optical Detection of Narrow Atomic Frequency Standard, Oct. 1998.
Rb87 Hyperfine Absorption Lines,” Physical Review Letters, vol. 1, no.
9, pp. 311-313, Nov. 1958. RFS Products
[33] R. J. Carpenter, el al., “A Prototype Rubidium Vapor Frequency Stand- [59] D. H. Throne, “A Rubidium Vapor Frequency Standard for Systems
ard,” IRE Trans. Instrum., vol. I-9, no. 2, pp. 132-135, Sep. 1960. Requiring Superior Frequency Stability,” Hewlett-Packard J., vol. 19,
[34] M. E. Packard and B. E. Swartz, “The Optically Pumped Rubidium no. 11, pp. 8-16, July 1968.
Vapor Frequency Standard,” IRE Trans. Instru., vol. I-11, no. 3 & 4, pp. [60] D. H. Throne, “A Report on the Performance of a New Rubidium Vapor
215-223, Dec. 1962. Frequency Standard,” in Proc. 23rd Annu. Symp. Freq. Contr, Atlantic
[35] A. Kastler, “Optical Methods for Studying Hertzian Resonances,” Sci- City, NJ, USA, May 1969, pp. 274-278.
ence 158 (3798): 214-221, 1967. [61] K. Chiba and T. Hashi, “An Ultra-Miniature Rubidium Frequency
[36] N. D. Bhaskar, et al., in “A Historic Review of Atomic Frequency Standard,” in Proc. 35th Annu. Symp. Freq. Contr, Philadelphia, PA,
Standards Used in Space Systems,” in Proc. 1996 IEEE Int. Freq. Contr. USA, May 1981, pp. 646-650.
Symp., Honolulu, HI, USA, June 1996, pp. 24-32. [62] T. Hashi, K. Chiba, and C. Takeuchi, “A Miniature, High-Performance
Rubidium Frequency Standard,” in Proc. 39th Annu. Symp. Freq.
General RFS Theory Contr., Philadelphia, PA, USA, May 1985, pp. 54-58.
[37] A. L. Bloom, “Optical Pumping,” Scientific American, pp. 72-80, Oct. [63] G. M. Saxena, et al., “Evaluation of the Rubidium Atomic Frequency
1960. Standard Developed in India,” in Proc. 39th Annu. Symp. Freq. Contr.,
[38] M. Arditi and T. R. Carver, “Pressure, Light and Temperature Shifts in Philadelphia, PA, USA, May 1985, pp. 59-63.
Optical Detection of 0-0 Hyperfine Resonance of Alkali Metals,” Physi- [64] W. Weidemann, “Subminiature Rubidium Oscillator Model FRS,” in
cal Review, vol. 124, no. 3, pp. 800-809. Nov. 1961. Proc. 40th Annu. Symp. Freq. Contr., Philadelphia, PA, USA, May
1986, pp. 470-473.
 30

[65] A. Stern, et al., “TF-4000A, TFL’s High Performance Ruggedized Ru- [91] Electronic Equipment, Ground, General Requirements For, MIL-E-
bidium Frequency Standard,” in Proc. 43rd Annu. Symp. Freq. Contr., 4158.
Denver, CO, USA, May-June 1989, pp. 124-129. [92] Electronic Equipment, Aerospace, Extended Space Environment, Gen-
[66] A. Stern, et al., “Rubidium Frequency Standard with a High Resolution eral Specification For, DOD-E-8983.
Digital Synthesizer,” in Proc. 1992 IEEE Freq. Contr. Symp., Hershey, [93] Test Requirements for Space Vehicles, MIL-STD-1540.
PA, USA, May 1992, pp. 108-113. [94] Electromagnetic Compatibility Requirements for Space Systems, MIL-
[67] J. Vanier, et. al., “Aging, Warm-up Time and Retrace: Important Char- STD-1541.
acteristics of Standard Frequency Generators,” in Proc. 1992 IEEE [95] Calibration System Requirements, MIL-C-45662.
Freq. Contr. Symp., Hershey, PA, USA, May 1992, pp. 807-815. [96] Military Specification, Oscillators, Crystal, General Specification For,
[68] H. Fruehauf, Precision Time and Frequency Handbook, 9th ed., Irvine, MIL-O-55310B.
CA, USA: Ball Corporation, Efratom Time and Frequency Products, [97] Test Methods for Electronic and Electrical Component Parts, MIL-STD-
1993. 202.
[69] M. Bloch, et al., “Subminiature Rubidium Frequency Standard for [98] Electromagnetic Interference Characteristics, Measurement of, MIL-
Commercial Applications,” in Proc. 1993 IEEE Int. Freq. Contr. Symp., STD-462.
Salt Lake City, UT, USA, June 1993, pp. 164-177. [99] Platform Electrical Power Characteristics, MIL-STD-704.
[70] P. Rochat, et al., “Developments of Rubidium Frequency Standards at [100] Characteristics of 28VDC Electrical System in Military Vehicles, MIL-
Neuchatel Observatory,” in Proc. 1994 IEEE Int. Freq. Contr. Symp., STD-1275.
Boston, MA, USA, June 1994, pp. 716-723. [101] Environmental Stress Screening Process for Electronic Equipment,
[71] Y. Koyama, et al., “An Ultra-Miniature Rubidium Frequency Standard MIL-STD-2164(EC).
with Two-Cell Scheme,” in Proc. 1995 IEEE Int. Freq. Contr. Symp., [102] Standard Methods of Test of Magnetic Shielding, ASTM 346-64, Ameri-
San Francisco, CA, USA, May-June 1995, pp. 33-38. can Society for Testing and Materials.
[72] T. McClelland, et al., “Subminiature Rubidium Frequency Standard: [103] “Guidance for Aircraft and Electrical Power Utilization and Transient
Manufacturability and Performance Results from Production Units,” in Protection,” Aeronautical Radio, Inc., ARNIC Report 413, Dec. 1976.
Proc. of the 1995 IEEE Int. Freq. Contr. Symp., San Francisco, CA, [104] Navy Manufacturing Screening Program”. NAVMAT P-9492.
USA, May-June 1995, pp. 39-52. [105] R. G. Lambert, “Case Histories of Selection Criteria for Random Vibra-
[73] A. Couplet, et al., “Miniaturized Rubidium Clocks for Space and Indus- tion Screening,” Inst. Env. Sci. Workshop Proc., Sep. 1984.
trial Applications,” in Proc. 1995 IEEE Int. Freq. Contr. Symp., San [106] W. Riley and J. Vaccaro, “Rubidium-Crystal Oscillator (RbXO) Devel-
Francisco, CA, USA, May-June 1995, pp. 53-59. opment Program,” U.S. Army LABCOM R&D Technical Report
[74] T. McClelland, et al., “Subminiature Rubidium Frequency Standard: SLCET-TR-84-0410F, Apr. 1986.
Performance Improvements,” in Proc. 1996 IEEE Int. Freq. Contr. [107] T. Lynch and W. Riley, “Tactical Rubidium Frequency Standard
Symp., Honolulu, HI, USA, June 1996, pp. 1011-1016. (TRFS),” RADC-TR-87-166, Final Technical Report, vol. I and vol. 2,
[75] K. Suzuki, et al., “Small-Sized Rubidium Oscillator,” in Proc. 1998 Oct. 1987.
IEEE Int. Freq. Contr. Symp., Pasadena, CA, USA, May 1998, pp. 73- [108] A. J. Curtis and R. D. McKain, “A Quantitative Method of Tailoring
79. Input Spectra for Random Vibration Screens,” J. Env. Sci., Sept. - Oct.
[76] J. Ho, et al., “New Rubidium Frequency Standard Design for Tele- 1987.
communications Applications,” in Proc. 1998 IEEE Int. Freq. Contr. [109] D. Watts, et al., “Specifying Commercial Atomic Frequency Standards
Symp., Pasadena, CA, USA, May 1998, pp. 80-83. Using Statistics,” in Proc. 1996 IEEE Int. Freq. Contr. Symp., Honolulu,
[77] Y. Koyama, et al., “An Ultra-Miniature Rubidium Frequency Standard,” HI, USA, June 1996, pp. 1023-1031.
in Proc. 2000 IEEE Int. Freq. Contr. Symp., Kansas City, MO, USA, [110] W. Weidemann, “Application Critical Parameters for Rubidium Stand-
June 2000, pp. 694-699. ards,” in Proc. 1998 IEEE Int. Freq. Contr. Symp., Pasadena, CA, USA,
[78] W. J. Riley, “A Modern Militarized Rubidium Frequency Standard,” May 1998, pp. 84-87.
Microwave Journal, May 2002.
[79] Data Sheet, “Quartzlock A10 Rubidium Time and Frequency Stand- Rb Lamps
ards,” Quartzlock, Ltd, 2002. [111] W. E. Bell, A. L. Bloom, and J. Lynch, “Alkali Metal Vapor Spectral
[80] W. J. Riley, “Recent Trends in Rubidium Frequency Standard Technol- Lamps,” Rev. Sci. Instrum., vol. 32, no. 6, pp. 688-692, June 1961.
ogy,” Tutorial at the 34th Precise Time and Time Interval (PTTI) Sys- [112] R. G. Brewer, “High Intensity Low Noise Rubidium Light Source,” Rev.
tems and Applications Meeting, Reston, VA, USA, Dec. 2002, Sci. Instrum., vol. 32, no. 12, pp. 1356-1358, Dec. 1961.
[81] J. Willison, “Rubidium Oscillator Design,” Stanford Research Systems, [113] T. English, E. Jechart, and T. M. Kwon, “Elimination of the Light Shift
Inc., Presentation at the National Synchronization Workshop, May 2001. in Rubidium Gas Cell Frequency Standards Using Pulsed Optical Pump-
ing,” in Proc. 10th Annu. Precise Time and Time Interval (PTTI) Appli-
RFS Environmental Sensitivity cations and Planning Meeting, Greenbelt, MD, USA, Nov. 1978, pp.
[82] W. J. Riley, “The Physics of the Environmental Sensitivity of Rubidium 147-168.
Gas Cell Atomic Frequency Standards,” IEEE Trans. Ultrason., [114] N. Oura, et al., “Composite-Type 87Rb Optical Pumping Light Source
Ferroelectr. Freq. Control, vol. 39, no. 2, pp. 232-240, Mar. 1992. for the Rubidium Frequency Standard,” in Proc. 13th Annu. Precise
[83] M. Huang, C. M. Klimcak, and J. C. Camparo, "Vapor-Cell Clock Fre- Time and Time Interval (PTTI) Applications and Planning Meeting,
quency and Environmental Pressure: Resonance-Cell Volume Changes,” Washington, DC, USA, Dec. 1981, pp. 803-815.
in Proc. 2010 IEEE Int. Freq. Contr. Symp, Newport Beach, CA, USA, [115] C. H. Volk, et al., “Lifetime and Reliability of Rubidium Discharge
June 2010., pp. 208-211. Lamps for Use in Atomic Frequency Standards,” in Proc. 38th Annu.
Freq. Contr. Symp., Philadelphia, PA, pp. 387-400, May 1984.
RFS Testing [116] R. A. Cook and R. P. Frueholtz, “An Improved Rubidium Consumption
[84] T. J. Lynch, W. J. Riley, and J. R. Vaccaro, “The Testing of Rubidium Model for Discharge Lamps Used in Rubidium Frequency Standards,”
Frequency Standards,” in Proc. 43rd Annu. Symp. Freq. Contr., Denver, in Proc. 42th Annu. Freq. Contr. Symp., Baltimore, MD, June 1988, pp.
CO, USA, May 1989, pp. 257-262. 525-531.
[85] Standard General Requirements for Electronic Equipment, MIL-STD- [117] J. Comparo and R. Mackay, “A Mechanism of Rubidium Clock Degra-
454. dation: Ring-Mode to Red-Mode Transition in RF-Discharge Lamps,” in
[86] Environmental Test Methods, MIL-STD-810. Proc. 2007 IEEE Int. Freq. Contr. Symp., Geneva, Switzerland, May-
[87] Electromagnetic Interference Characteristics, Requirements For 1)
2)
June 2007, pp. 45-48.
Equipment, MIL-STD-461.
[88] Electronic Equipment, Airborne, General Specification For, MIL-E- Lamp Exciters
5400. [118] R. Shaw, “Spontaneously Generated Ion Acoustic Waves in a Weakly
[89] Testing, Environmental, Airborne Electronic and Associated Equipment, Ionized Plasma,” M.S. thesis, Cornell Univ., Ithaca, NY, USA, 1964.
MIL-T-5422.
[90] Electronic Equipment, Naval Ship and Shore, General Specification,
MIL-E-16400.
 31

GPS Rb Clocks [140] B. Ghosal and G. M. Saxena, “Design Verification Model of Rubidium
Frequency Standard for Space,” J. Modern Phys., vol, 5, no. 3, pp. 128-
[119] D. E. Ringer, J. Gandy, and E. Jechart, “Spaceborne Rubidium Frequen- 135, Jan. 2014.
cy Standard for NAVSTAR GPS,” in Proc. 7th Annu. Precise Time and
Time Interval (PTTI) Applications and Planning Meeting, Greenbelt, MIL (Tactical) RFS
MD, USA, Dec. 1975, pp. 671-696.
[120] “Design for a Rubidium Frequency Standard for the GPS Program,” [141] TRC-170(V) troposcatter radio, Forecast International, July 1999.
General Radio Company, Bolton, MA, USA, Final Report, Rockwell [142] Regency Net HF radio, Forecast International, July 1997.
Contract No. M5N3DGS-649024D, unpublished report, Feb., 1976. [143] Have Quick frequency hopping radios, Wikipedia.
[121] W. J. Riley, “A Rubidium Clock for GPS,” in Proc. 13th Annu. Precise [144] SCOTT (TSC-124) MILSTAR terminal, Forecast International, Nov.
Time and Time Interval (PTTI) Applications and Planning Meeting, 1997.
Washington, DC, USA, Dec. 1981, pp. 609-630. [145] M. E. Frerking and D. E. Johnson, “Rubidium Frequency and Time
[122] T. J. Lynch and W. J. Riley, “Test Results for Prototype GPS Rubidium Standard for Military Environment,” in Proc. 26th Annu. Symp. Freq.
Clocks,” in Proc. 15th Annu. Precise Time and Time Interval (PTTI) Contr., Atlantic City, NJ, USA, June 1972, pp. 216-222.
Applications and Planning Meeting, Washington, DC, USA, Dec. 1983, [146] “Phase 2A Bench Model Development, Tactical Rubidium Frequency
pp. 269-280. Standard,” EG&G Frequency Products, Salem, MA, USA, Final Tech-
[123] S. Goldberg, T. J. Lynch, and W. J. Riley, “Further Test Results for nical Report, Rome Air Development Center, Contract No. F19628-82-
Prototype GPS Rubidium Clocks,” in Proc. 17th Annu. Precise Time C-0055, Dec. 1982.
and Time Interval (PTTI) Applications and Planning Meeting, Washing- [147] W. J. Riley, “A Rubidium Clock for Seek-Talk,” in Proc. 14th Annu.
ton, DC, USA, Dec. 1985, pp. 145-155. Precise Time and Time Interval (PTTI) Applications and Planning Meet-
[124] F. Danzy and W. J. Riley, “Stability Test Results for GPS Rubidium ing, Greenbelt, MD, USA, Dec. 1982, pp. 141-154.
Clocks,” in Proc. 19th Annu. Precise Time and Time Interval (PTTI) [148] “Phase 2B Prototype Redesign, Tactical Rubidium Frequency Standard,”
Applications and Planning Meeting, Redondo Beach, CA, USA, Dec. EG&G Frequency Products, Salem, MA, USA, Final Technical Report,
1987, pp. 267-274. Rome Air Development Center, Contract No. F19628-82-C-0055, Jan.
[125] W. J. Riley, “Rubidium Atomic Frequency Standards for GPS Block 1983.
IIR,” in Proc. 22nd Annu. Precise Time and Time Interval (PTTI) Appli- [149] T. M. Kwon, B. C. Grover, and H. E. Williams, “Rubidium Frequency
cations and Planning Meeting, Vienna, VA, USA, Dec. 1990, pp. 221- Standard Study,” RADC-TR-83-230, Final Technical Report, Rome Air
230. Development Center, Oct. 1983.
[126] W. J. Riley, “Rubidium Atomic Frequency Standards for GPS Block [150] T. M Kwon, et al., “A Miniature Tactical Rb Frequency Standard,” in
IIR, Proc. 6th Eur. Freq. and Time Forum, Noordwijk, Netherlands, pp. Proc. 16th Annu. Precise Time and Time Interval (PTTI) Applications
231-238, Mar. 1992. and Planning Meeting, Greenbelt, MD, USA, Nov. 1984, pp. 143-155.
[127] W. J. Riley, “Rubidium Atomic Frequency Standards for GPS Block [151] H. Fruehauf, W. Weidemann, C. Colson, S. Stein, T. English, H.
IIR,” in Proc. 5th Technical Meeting of the Satellite Division of the In- Vorwerk, and R. Paul, “Militarized Rubidium Oscillator,” RADC-TR-85-
stitute of Navigation., Albuquerque, NM, pp. 537-545, Sept. 1992. 23, Final Technical Report, Rome Air Development Center, Apr. 1985.
[128] W. J. Riley, “Early Stability Test Results for GPS Block IIR Rubidium [152] “Tactical Rubidium Frequency Standard (TRFS),” EG&G Frequency
Clocks,” Proc. 5th Symp. Freq. Stds. & Metrol., Woods Hole, MA, Products, Salem, MA, USA, Final Technical Report, Hanscom Air Force
USA, Oct. 1995, pp. 403-405. Base Contract No. F19628-83-0175, Dec. 1986.
[129] W. J. Riley, “Early In-Orbit Performance of GPS Block IIR Rubidium [153] T. M Kwon, et al., Final Technical Report, Rome Air Development
Clocks,” in Proc. 29th Annu. Precise Time and Time Interval Systems Center, RADC-TR-87-140, Sep. 1987.
and Applications Meeting, Long Beach, CA, USA, Nov. 1997, pp. 213- [154] T.J. Lynch and W.J. Riley, “Tactical Rubidium Frequency Standard,”,
220. EG&G Frequency Products, Salem, MA, USA, Final Technical Report,
[130] R. Beard, et al., “GPS Block IIR Rubidium Atomic Frequency Standard Rome Air Development Center, RADC-TR-87-166, Oct. 1987.
Life Test,” in Proc. 30th Annu. Precise Time and Time Interval (PTTI) [155] T. J. Lynch and J. R. Vaccaro, “Ultrafast Warm-up Rubidium Refer-
Systems and Applications Meeting, Reston, VA, USA, Dec. 1998, pp. ence,” in Proc. 1996 IEEE Int. Freq. Contr. Symp., Honolulu, HI, June
145-160. 1996, pp. 993-1001.
[131] A. Wu and B. Feess, “Development of GPS Space Clocks for GPS III
and Beyond,” in Proc. 32nd Annu. Precise Time and Time Interval RbXO
(PTTI) Systems and Applications Meeting, Reston VA, USA, Nov. 2000,
[156] J. R. Vig and V. J. Rosati, “The Rubidium-Crystal Oscillator Hybrid
pp. 389-400.
Development Program,” in Proc. 16th Annu. Precise Time and Time In-
[132] R. F. Dupuis, T. J. Lynch, and J. R. Vaccaro, “Rubidium Frequency
terval (PTTI) Applications and Planning Meeting, Greenbelt, MD, USA,
Standard for the GPS IIF Program and Modifications for the RAFSMOD
Nov. 1984, pp. 157-165.
Program,” in Proc. 2008 IEEE Int. Freq. Contr. Symp., Honolulu, HI,
[157] “Rubidium-Crystal Oscillator (RbXO) Development Program,” EG&G
USA, May 2008, pp. 655-660.
Electronic Components, Salem, MA, USA, Research and Development
Space RFS (Non-GPS) Technical Report, US Army Laboratory Command, Contract No.
SLCET-TR-84-0410-F, unpublished report, Apr. 1986.
[133] “Atomic Time and Frequency Reference System,” Varian Associates, [158] W. J. Riley and J. R. Vaccaro, “A Rubidium-Crystal Oscillator
Palo Alto, CA, USA, Final Report, Part I, Contract No. NAS 9-5425, (RbXO),” in Proc. 40th Annu. Freq. Contr. Symp., Philadelphia, PA,
Jan. 1966 to Nov. 1967. USA, May 1986, pp. 452-464.
[134] D. Cree, “Development of a Spacecraft Atomic Timing System,” NASA [159] W. J. Riley and J. R. Vaccaro, “A Rubidium-Crystal Oscillator
Manned Spacecraft Center, Houston, TX, USA, NASA Internal Note (RbXO),” IEEE Trans. Ultrason., Ferroelectr. Freq. Control, vol.
MSC-67-EE-12, Rev. A, May 1968. UFFC-34, no. 6, pp. 612-618, Nov. 1987.
[135] R. W. Frank, et al., “Design Concepts for an Aerospace Rubidium
Standard,” General Radio Company, Bolton, MA, USA, unpublished re- Rb Signal Parameters
port, Oct. 1969.
[160] C. Audoin, V. Candelier, and N. Dimarcq, “A Limit to the Frequency
[136] “Rubidium Spacecraft Atomic Timing System,” General Radio Compa-
Stability of Passive Frequency Standards Due to an Intermodulation Ef-
ny, W. Concord, MA, USA, Final Report, NASA Contract NAS 9-5425,
fect,” IEEE Trans. Instru. Meas., vol. 40, no. 2, pp. 121-125, Apr. 1991.
Part II, Dec., 1969.
[161] J. Q. Deng, et al., “Frequency Stability of Cell-Based Passive Frequency
[137] M. Bird, et al, “Rubidium Ultra-Stable Oscillators at Titan: The Huygens
Standards: Reducing the Effects of Local-Oscillator PM Noise,” in Proc.
Doppler Wind Experiment,” NASA Technical Reports Server, Aug.
1998 IEEE Int. Freq. Contr. Symp., Pasadena, CA, USA, May 1998, pp.
1997.
95-98.
[138] Y. Lv, et al,, “Characteristics of BeiDou-3 Experimental Satellite
[162] J. Vanier and L. Bernier, “On the Signal-to-Noise Ratio and Short-Term
Clocks,” Remote Sensing, Nov. 2018.
Stability of Passive Rubidium Frequency Standards,” IEEE Trans.
[139] L. A. Mallette, J. White, and P. Rochat, “Space Qualified Frequency
Instru. Meas., vol. IM-30, no. 4, pp. 277-282, Dec. 1981.
Sources (Clocks) for Current and Future GNSS Applications,” presented
[163] J. Vanier, “Light Shifts in Passive Rubidium Frequency Standards Using
at Position Location and Navigation Symp. (PLANS), 2010.
Hyperfine Filtering,” Laboratorire D’Electronique Quantique,
 32

Departement de Genie Electrique, Universite Laval, unpublished report, trical and Instrumentation Engineering Department, Thapar University,
Apr. 1981. New Delhi, July 2009.

Kenschaft Model Phase Noise

[164] R. P. Kenschaft, “Response of Quantum-Mechanical Resonances to [183] J. Vanier, M. Tetu, and L. Bernier, “Transfer of Frequency Stability
Modulated Radio Frequency Signals,” Ph. D. Thesis, University of from an Atomic Frequency Reference to a Quartz Crystal Oscillator,” in
Pennsylvania, 1969. Proc. 32nd Annu. Symp. Freq. Contr., Atlantic City, NJ, USA, June
[165] K. D. Lyon and W. J. Riley, “Application of the Kenschaft Model to the 1978, pp. 520-526.
Analysis of the Rubidium Gas Cell Frequency Standard,” in Proc. 1996 [184] F. Walls, “Stability of Frequency Locked Loops,” Appendix in Time
IEEE Int. Freq. Contr. Symp., Honolulu, HI, USA, June 1996, pp. 1032- Domain Frequency Stability Calculated from the Frequency Domain
1040. Description, NIST Publication NISTIR 89-3916, Sep. 1989.
[166] R. P. Kenschaft, “Mixed Modulation Signal and Noise Analysis and [185] J. Vaccaro, “Derivation of RAFS Open Loop VCXO Phase Noise Speci-
Simulation of the Rubidium Gas Cell Atomic Frequency Standard,” in fication,” PerkinElmer, Salem, MA, USA, unpublished report, Apr.
Proc. 1996 IEEE Int. Freq. Contr. Symp., Honolulu, HI, USA, June 2009.
1996, pp. 1041-1050.
Acceleration, Vibration and G-Sensitivity
C-Field [186] “The Effects of Acceleration on Precision Frequency Sources,” Elec-
tronics Technology and Devices Laboratory, Ft. Monmouth, NJ, USA,
[167] “Advantages of Fixed Minimum C-Field Operation of the EG&G GPS Research and Development Technical Report, US Army Laboratory
Rubidium Frequency Standard,” EG&G Electronic Components, Salem Command, Contract No. SLCET-TR-84-0410-F. July 1992.
MA, USA, unpublished report, July 1982. [187] R. L. Filler, "The Acceleration Sensitivity of Quartz Crystal Oscillators:
[168] A. Stern, et al., “A Novel Compact Rubidium Frequency Standard with a A Review,” IEEE Trans. Ultrason., Ferroelectr. Freq. Control, vol. 35,
Low Sensitivity to Magnetic and Vibrational Disturbances,” in Proc. no. 3, pp. 297-305, May 1988.
42nd Freq. Contr. Symp., Baltimore, MD, USA, June 1988, pp. 519-524.
Buffer Gas
Magnetic Shielding
[188] P. Bender and E. C. Beaty, “Optical Detecting of Cesium Hyperfine
[169] W. G. Wadey, “Magnetic Shielding with Multiple Cylindrical Shells,” Transition,” in Proc. 12th Annu. Symp. Freq. Contr., Asbury Park, NJ,
Review of Scientific Instruments, vol. 27, no. 11, pp. 910-916, Nov. USA, May 1958, pp.593-605.
1956. [189] G. Missout and J. Vanier, “Pressure and Temperature Coefficients of the
[170] A. J. Mager, “Magnetic Shields,” IEEE Trans. Magn., vol. MAG-6, no. More Commonly Used Bufer Gases in Rubidium Vapor Frequency
1, pp. 67-75, Mar. 1970. Standards,” IEEE Trans. Instru. Meas., vol. 24, no. 2, pp. 180-184, June
[171] S. M. Freake and T. L. Thorp, “Shielding of Low Magnetic Fields with 1975.
Multiple Cylindrical Shells,” Review of Scientific Instruments, vol. 42, [190] J. Vanier, et al., “On the Pressure and Temperature Coefficients of Buff-
no. 10, pp. 1411-1413, Oct. 1971. er Gases Used in Optically Pumped Passive Rubidium Frequency Stand-
[172] S. A. Wolf, D. U. Gubser, and J. E.Cox, “Shielding of Longitudinal ards,” Laboratoire D’Electronique Quantique, Universite Laval, un-
Magnetic Fields with Thin, Closely-Spaced, Concentric Cylindrical published report, July 1981.
Shells with Applications to Atomic Clocks,” in Proc. 10th Annu. Pre- [191] C. W. Beer and R. A. Bernheim, “Hyperfine Pressure Shift of Cs-133
cise Time and Time Interval (PTTI) Applications and Planning Meeting, Atoms in Noble and Molecular Buffer Gases,” Phys. Rev. A 13, 1052-
Greenbelt, MD, USA, Nov. 1978, pp. 131-146. 1057, Mar. 1976.
[173] S. A. Wolf, D. U. Gubser, and J. E.Cox, “Shielding of Longitudinal [192] F. Strumia, et al, “Optimization of the Buffer Gas Mixture for Optically
Magnetic Fields with Thin, Closely-Spaced, Concentric Cylinders of Pumped Cs Frequency Standards,” in Proc. 30th Annu. Freq. Contr.
High Permeability Material,” Review of Scientific Instruments, vol. 50, Symp., June 1976, pp. 468-472.
no. 6, pp. 751-756, June 1979. [193] J. Vanier, et al, “On the Light Shift and Buffer Gas Shift in Passive
Rubidium Frequency Standard,” in 36th FCS 1982, pp.348-354.
Microwave Cavities [194] P. Chantry, et al, “Towards a Miniature Laser-Pumped Cesium Cell
Frequency Standard,” in Proc. 46th IEEE Freq. Contr. Symp., Hersey,
[174] J. Viennet, C. Audoin, and M. Desainfuscien, “Cavity Pulling in Passive PA, USA, May 1992, pp.114-122.
Frequency Standards,” IEEE Trans. Instru. Meas., vol. IM-21, no. 3, pp. [195] W. Stern and R. Novick, “Further Results on the Rubidium-87 Maser
204-209, Aug. 1972. Frequency Standard,” in Proc. 26th Freq. Contr. Symp., Atlantic City,
[175] H. E. Williams, T. M. Kwon, and T. McClelland, “Compact Rectangular NJ, USA, June 1972, pp. 223-224.
Cavity for Rubidium Vapor Cell Frequency Standards,” in Proc. 37th [196] E. Beaty, P. Bender, and A. Chi, “Hyperfine Transitions in Rubidium 87
Annu. Freq. Contr. Symp., Philadelphia, PA, USA, June 1983, pp. 12-17. Vapor,” in Proc. 13th Freq. Contr. Symp., Asbury Park, NJ, USA, May
[176] T. Sphicopoulos, “Cavites Compacte Pour Etalons De Frequence 1959, pp. 668-675.
Atomiques,” [Compact Cavities for Atomic Frequency Resonator] (in [197] M. Arditi and T. Carver, “Pressure, Light and Temperature Shifts in
French), PhD. thesis, Ecole Polythehnique Federale De Lausanne, Jan. Optical Detection of 0-0 Hyperfine Resonance of Alkali Metals”. Phys.
1986. Rev. 124, no. 3, Nov. 1961, pp. 800-809.
[177] G. H. Mei and J. T. Liu, “A Miniaturized Microwave Resonator for [198] F. Gong, Y. Jau, and W. Happer, "Nonlinear Pressure Shifts of Alkali-
Rubidium Frequency Standards,” in Proc. 1999 IEEE Int. Freq. Contr. Metal Atoms in Inert Gases,” Phys. Rev. Lett., vol. 100, 233002, June
Symp., Besançon, France, Apr. 1999, p. 601. 2008.
[178] J. Deng, “Subminiature Microwave Cavity for Atomic Frequency Stand- [199] H. G. Robinson and C. E. Johnson, “A New Heart for Rb Frequency
ards,” in Proc. 2001 IEEE Int. Freq. Contr. Symp., Seattle, WA, USA, Standards?: The Evacuated, Wall-Coated Sealed Cell,” IEEE Trans.
June 2001, pp. 85-88. Iustrum. Meas., vol. 32, no. 1, p. 198, Mar. 1983.
[179] X. Huang, et al., “A Microwave Cavity with Low Temperature Coeffi- [200] S. Knappe and H. G. Robinson, “Double-Resonance Lineshapes in a
cient for Passive Rubidium Frequency Standards,” in Proc. 2001 IEEE Cell with Wall Coating and Buffer Gas,” New Journal of Physics, Vol-
Int. Freq. Contr. Symp., Seattle, WA, USA, June 2001, pp. 105-107. ume 12, June 2010.
[180] J. Hu, et. al., “A Subminiature Microwave Cavity for Rubidium Atomic
Frequency Standards,” in Proc. 2007 IEEE Int. Freq. Contr. Symp., Ge- Relativity
neva, Switzerland, May-June 2007, pp. 599-601. [201] C. O. Alley, “Introduction to Some Fundamental Concepts of General
[181] Y. Wang, et. al., “A Downsized Microwave Cavity for the Rubidium Relativity and to Their Required Use in Some Modern Timekeeping
Vapor Cell Frequency Standard,” in Proc. 2007 IEEE Int. Freq. Contr. Systems,” 1979 FCS Tutorial Transcript, pp. 687-724, 1982.
Symp., Geneva, Switzerland, May-June 2007, pp. 595-598. [202] N. Ashby, “A Tutorial on Relativistic Effects in the Global Positioning
[182] S. S. Raghuwanshi, “Development & Analysis of Microwave Cavity and System,” NIST Contract No. 40RANB9B8112, Feb. 1990 (Similar pa-
Magnetic Shielding for Rubidium Atomic Clock,” M. Eng Thesis, Elec- per).
 33

[203] P. Kline, “An Experimental Investigation of Relativistic Effects in [224] R. Lutwak, “Chip Scale Atomic Clock,” presented at DARPA MTO
GPS,” Navigation, vol. 48, no. 4, pp. 297-305, Winter 1988-1999. Technology Symp., San Jose CA, Mar. 2007.
[204] N. Ashby, “Relativity and the Global Positioning System,” Physics [225] R. Lutwak, et al., “The Miniature Atomic Clock – Pre-Production Re-
Today, pp. 41-47, May 2002. sults,” in Proc. Eur. Freq. and Time Forum and IEEE Int. Freq. Contr.
Symp., Geneva, Switzerland, May 2007, pp. 1327-1333 (and White Pa-
Baseplate Temperature Controllers per).
[226] R. Lutwak, “The Chip-Scale Atomic Clock – Recent Developments,” in
[205] PerkinElmer, Inc, “Improved Baseplate Temperature Controller (BTC) Proc. 2009 IEEE Int. Freq. Contr. Symp., Besançon, France, Apr. 2009,
for GPS III RAFS,” unpublished report, July 2009. pp. 573-577.
[227] Introduction to Symmetricom’s SA.45s Chip Scale Atomic Clock
Radiation Hardening (CSAC), Symmetricom.
[228] J. Kitching, “Chip-Scale Atomic Devices,” Appl. Phys. Rev., vol. 5,
[206] T. M. Flanagan and R. E. Leadon, “Radiation Effects in Crystal and 031302, 2018.
Atomic Frequency Standards,” in Proc. 7th Annu. Precise Time and
Time Interval (PTTI) Applications and Planning Meeting, Greenbelt, CPT Rubidium Clocks
MD, USA, Dec. 1975, pp. 125-135. [229] J. Deng, et al., “A Commercial CPT Rubidium Clock,” in Proc. 2008
[207] “Radiation Hardening of EG&G Rb Frequency Standard,” Final Report Eur. Freq. and Time Forum, Toulouse, France, Apr. 2008.
No. ED-7875, U.S. Naval Research Laboratory, June 1982 (un-
published). Aging
[208] T. English, “Radiation Hardness of Efratom M-100 Rubidium Frequen-
cy Standard,” in Proc. 14th PTTI Meeting, Greenbelt, MD, USA, Dec. [230] J. C. Camparo, “Does the Light Shift Drive Frequency Aging in the
1982, pp. 547-575. Rubidium Atomic Clock?,” IEEE Trans. Ultrason., Ferroelectr. Freq.
[209] T. C. English, G. L. Malley, and R. Korde, “Neutron Hardness of Photo- Control., vol. 52, no. 7, pp. 1075-1078, Aug. 2005.
diodes for Use in Passive Rubidium Frequency Standards,” in Proc. [231] W. J. Riley, “RAFS Drift and its Potential Reduction,” Hamilton Tech-
42nd Freq. Contr. Symp., Baltimore, MD, USA, June 1988, pp. 532-539. nical Services, unpublished report, June 2007
[232] J. C. Camparo, “Influence of the Atmosphere on a Rubidium Clock’s
Reliability Frequency Aging,” in Proc 39th Annu. Precise Time and Time Interval
(PTTI) Systems and Applications Meeting, Long Beach, CA, USA, pp.
[210] MIL-HDBK-217F, Notice 2, “Military Handbook for Reliability Predic- 317-322, Nov. 2007.
tion of Electronic Equipment,” Feb. 1995. [233] M. Epstein, et al., “Long-Term Clock Behavior of GPS IIR Satellites,”
[211] SR-332, Issue 2, “Reliability Prediction Procedure for Electronic in Proc 39th Precise Time and Time Interval (PTTI)Systems and Appli-
Equipment,” Telcordia, Sep. 2006. cations Meeting, Long Beach, CA, USA, pp. 59-78, Nov. 2007.

Standards Rubidium
[234] D.A. Steck. "Rubidium-87 D Line Data,” Los Alamos National Labora-
[212] MIL-O-55310B, “Military Specification, Oscillators, Crystal, General tory Technical report LA-UR-03-8638, Sep. 2001.
Specification For,” May 1988.
[213] IEEE Standard 1139-1988, “Standard Definitions of Physical Quantities Patents
for Fundamental Frequency and Time Metrology”. [235] R. H. Dicke and T. R. Carver, “Atomic or molecular oscillator circuit,”
[214] IEEE Standard 1139-1999, “IEEE Standard Definitions for Physical U.S. Patent 2 836 722, May 27, 1958.
Quantities for Fundamental Frequency and Time Metrology – Random [236] R. H. Dicke, “Gas Cells for microwave spectroscopy and frequency
Instabilities”. stabilization,” U.S. Patent 2 882 493, Apr. 14, 1959.
[215] IEEE Standard 1193-2000, “IEEE Guide for Measurement of Environ- [237] R. H. Dicke, “Method and system employing photon absorption by a
mental Sensitivities of Standard Frequency Generators”. microwave resonant medium,” U.S. Patent 2 884 524, Apr. 28, 1959.
[216] MIL-STD-461E, “Requirements for the Control of Electromagnetic [238] R. H. Dicke, “Stable oscillator,” U.S. Patent 2 927 278, Mar. 1, 1960.
Interference Characteristics of Subsystems and Equipment”. [239] M. Arditi, “Atomic frequency standard,” U.S. Patent 2 951 992, Sept. 6,
1960.
Rubidium Masers
[240] M. Arditi, “Gas cell for frequency selective system,” U.S. Patent 2 955
[217] P. Davidovits and R. Novick, “The Optically Pumped Rubidium Maser,” 262, Oct. 4, 1960.
Proc. IEEE, vol. 54, no. 2, pp. 155-170, Feb. 1996. [241] A. L. Bloom, “Electrodeless discharge method and apparatus,” U.S.
Patent 2 975 330, Mar. 14, 1961.
Laser Pumped Atomic Clocks [242] P. C. Robison, “Tuning arrangement utilizing optical pumping,” U.S.
[218] G. Mileti, J. Deng, and F. L. Walls, “Laser-Pumped Rubidium Frequen- Patent 3 038 126, June 5, 1962.
cy Standards: New Analysis and Progress,” IEEE Journal of Quantum [243] H. G. Dehmelt, “Optical absorption monitoring of oriented or aligned
Electronics, vol. 14, no. 2, pp. 233-237, Feb. 1998. quantum systems,” U.S. Patent 3 071 721, Jan. 1, 1963.
[219] S. Micalizio, et al., “Pulsed Optically Pumped Rubidium Clock,” in 8th [244] W. E. Bell and A. L. Bloom, “Electrodeless discharge lamp apparatus,”
Symp. on Freq. Stds. and Metrol., Podsdam, Germany, Oct. 2015. U.S. Patent 3 109 960, Nov. 5, 1963.
[245] M. E. Packard, A. L. Bloom, and W. E. Bell, “Quantum resonance stabi-
Chip Scale Atomic Clocks lized frequency source,” U.S. Patent 3 129 389, Apr. 14, 1964.
[220] R. Lutwak, et al., “The Chip-Scale Atomic Clock – Coherent Population [246] H. G. Dehmelt, “Modulation of a light beam by absorbing quantum
Trapping vs. Conventional Interrogation,” in Proc. 34th Annu. Precise systems exhibiting a periodically varying alignment,” U.S. Patent 3 150
Time and Time Interval (PTTI) Systems and Applications Meeting, 313, Nov. 22, 1964.
Reston, VA, USA, Dec. 2002, pp. 539-550. [247] R. M. Whitehorn, “Atomic frequency standard,” U.S. Patent 3 159 797,
[221] R. Lutwak, et al., “The Chip-Scale Atomic Clock - Recent Development Dec. 1, 1964.
Progress,” in Proc. 35th Annu. Precise Time and Time Interval (PTTI) [248] R. H. Dicke, “Gas cell frequency stabilization,” U.S. Patent 3 165 705,
Systems and Applications Meeting, San Diego, CA, USA, Dec. 2003, pp. Jan. 12, 1965.
467-478. [249] W. E. Bell, “Electrodeless discharge lamp apparatus,” U.S. Patent 3 170
[222] R. Lutwak, et al., “The Chip-Scale Atomic Clock – Low-Power Physics 086, Feb. 16, 1965.
Package,” in Proc. 36th Annu. Precise Time and Time Interval (PTTI) [250] M. Arditi, “Atomic clock,” U.S. Patent 3 174 114, Mar. 16, 1965.
Systems and Applications Meeting, Dec. 2004, Washington, DC, USA, [251] W. E. Bell, “Quantum oscillators,” U.S. Patent 3 187 251, June 1, 1965.
pp. 339-354. [252] P. L. Bender and E. C. Beaty, “Alkali vapor frequency standard utilizing
[223] R. Lutwak, et al., “The MAC – a Miniature Atomic Clock,” in Proc. optical pumping,” U.S. Patent 3 192 472, June 29, 1965.
2005 IEEE Int. Freq. Contr. Symp., Vancouver, BC, Canada, Aug. 2005, [253] W. A. Marrison, “Electrodeless vapor discharge lamp with auxiliary
pp. 752-757. voltage triggering means,” U.S. Patent 3 196 312, July 20, 1965.
 34

[254] E. L. Garwin and L. O. Heflinger, “Magnetic Shield Arrangements,” [287] S. Goldberg, “Miniaturized atomic frequency standard having both filter
U.S. Patent 3 222 449, Dec. 7, 1965. cell and absorption cell in resonator cavity,” U.S. Patent 4 494 085, Jan.
[255] W. A. Marrison, “Electrodeless vapor discharge lamp with auxiliary 15, 1985.
radiation triggering means.” U.S. Patent 3 227 923, Jan. 4, 1966. [288] T. M. Kwon et al, “Cavity resonator for atomic frequency standard,”
[256] L. Malnar, “Resonance cells for optical pumping,” U.S. Patent 3 242 U.S. Patent 4 495 478, Jan. 22, 1985.
423, Mar. 22, 1966. [289] W. Weidemann, “Integrated microwave cavity resonator and magnetic
[257] W. E. Bell, “Atomic stabilized frequency source,” U.S. Patent 3 246 shield for an atomic frequency standard,” U.S. Patent 4 661 782, Apr.
254, Apr. 12, 1966. 28, 1987.
[258] D. J. Farmer, “Optically pumped combination gas cell and microwave [290] A. De Marchi and G. D. Rovera, “Method and a device for the frequency
resonating cavity,” U.S. Patent 3 248 666, Apr. 26, 1966. control of an atomic or molecular beam frequency standard,” U.S. Patent
[259] H. G. Dehmelt, “Optical pumping of hyperfine states by light at the 4 692 716, Sept. 8, 1987.
Zeeman frequency,” U.S. Patent 3 256 478, June 14, 1966. [291] W.J. Riley, Jr., “Power supply circuit for an alkali vapor spectral lamp,”
[260] H. G. Dehmelt, “Optical absorption monitoring of aligned alkali atoms,” U.S. Patent 4 721 890, Jan. 26, 1988.
U.S. Patent 3,267,360, Aug. 16, 1966. [292] A. Lepek and A. Stern, “Elimination of magnetic influence on atomic
[261] H. G. Dehmelt, “Apparatus for optical alignment and detection of atomic clocks,” U.S. Patent 4 953 148, Aug. 28, 1990.
energy states,” U.S. Patent 3 281 709, Oct. 25, 1966. [293] L. Lewis, “Saturated absorption double resonance system and appa-
[262] R. Novick and P. Davidovits, “Optically pumped atomic resonance ratus,” U.S. Patent 5 136 261, Aug. 4, 1992.
apparatus with improved optical pumping means,” U.S. Patent 3 304 [294] L. Lewis, “Compact optically pumped resonance system and apparatus,”
516, Feb. 14, 1967,. U.S. Patent 5 146 185, Sept. 8, 1992.
[263] G. T. Inouye, J. M. Andres, and G. L. Brown, “Gas cell frequency [295] M. Ohtsu, “Laser pumped atomic frequency standard with high frequen-
standard,” U.S. Patent 3 348 165, Oct. 17, 1967. cy stability,” U.S. Patent 5 148 437, Sept. 15, 1992.
[264] J. T. Arnold, “Adjustable frequency atomic frequency standard,” U.S. [296] P.J. Chantry, R. W. Weinert, S. H. Tallisa, B. R. McAvoy, and T. J.
Patent 3 363 193, Jan. 9, 1968. Smith, Jr., “Miniaturized atomic frequency standard,” U.S. Patent 5 192
[265] R. J. Rorden, “Automatic search sweep for atomic frequency standard,” 921, Mar. 9, 1993.
U.S. Patent 3 364 438, Jan. 16, 1968. [297] I. Liberman, “Gas cell for a miniaturized atomic frequency standard,”
[266] R. C. Rempel, B. E. Swartz, M. E. Packard, and R. J. Rorden, “Frequen- U.S. Patent 5 327 105, July 5, 1994.
cy stabilization apparatus,” U.S. Patent 3 382 452, May 7, 1968. [298] H. S. Schweda et al, “Atomic frequency standard,” U.S. Patent 5 387
[267] L. Malnar, B. Henri, “Atomic clocks with spin exchange collision,” U.S. 881, Feb. 7, 1995.
Patent 3 388 339, June 11, 1968. [299] I. Liberman, “Atomic time standard with piezoelectric stabilization of
[268] P. Davidovits, R. Novick and N. Knable, “Atomic resonance apparatus diode laser light source,” U.S. Patent 5 442 326, Aug. 15, 1995.
utilizing an improved buffer gas cell,” U.S. Patent 3 390 350, June 25, [300] T. C. English, F. Lee, R. E. White, and J. A. Carmichael, “Method and
1968. apparatus for reduction of light intensity decay in optical pumping de-
[269] I. Wieder, “Optically pumped maser and solid state light source for use vices,” U.S. Patent 5 457 430, Oct. 10, 1995.
therein,” U.S. Patent 3 403 349, Sept. 24, 1968. [301] J. D. Crockett, “Lamp oscillator for atomic frequency standards,” U.S.
[270] G. Broussaud and M. Leon, “Optical resonance cells,” U.S. Patent 3 418 Patent 5 489 821, Feb. 6, 1996.
565, Dec. 24, 1968,. [302] T. C. English, “Method and apparatus for reduction of atomic frequency
[271] G. R. Huggett, “Atomic resonance gas cell having an evacuated double standard phase noise,” U.S. Patent 5 491 451, Feb. 13, 1996.
end wall structure,” U.S. Patent 3 510 758, May 5, 1970. [303] T. C. English, “Evanescent field interrogation for atomic frequency
[272] W. Happer, Jr., “Off-resonant light as a probe of optically pumped alkali standards,” U.S. Patent 5 517 157, May 14, 1996.
vapors,” U.S. Patent 3 513 381, May 19, 1970. [304] H. H. Telle, “Frequency stabilized laser source,” U.S. Patent 5 553 087,
[273] H. G. Dehmelt, “Apparatus for optically monitoring the gyromagnetic Sept. 3, 1996.
resonance of quantum systems,” U.S. Patent 3 575 655, Apr. 20, 1971. [305] W. R. Verbanets, “Miniature atomic frequency standard controlled by a
[274] H. G. Dehmelt, “Apparatus for optically monitoring the gyromagnetic digital processor,” U.S. Patent 5 606 291, Feb. 25, 1997.
resonance of quantum systems,” U.S. Patent 3 584 292, June 8, 1971. [306] G. Skoczen, “Resonator package for atomic frequency standard,” U.S.
[275] E. Jechart, “Gas cell atomic frequency standard of compact design,” Patent 5 627 497, May 6, 1997.
U.S. Patent 3 798 565, Dec. 14, 1972. [307] J. D. Crockett and D. A. Watts, “Heater controller for atomic frequency
[276] F. C. Gabriel, “Electrodeless discharge lamp and power coupler there- standards,” U.S. Patent 5 656 189, Aug. 12, 1997.
fore,” U.S. Patent 3 873 884, 02/11/1974. [308] P. Thomann and A. Jornad, “Atomic frequency oscillator,” U.S. Patent 5
[277] E. Jechart, “Gas cell atomic frequency standard having selected alkali 656 974, Aug. 12, 1997.
vapor isotopic ratios,” U.S. Patent 3 903 481, Mar. 27, 1974. [309] J. C. Camparo and S. B. Delcamp, “Rubidium atomic clock with fluo-
[278] S. Murayama, M. Yamamoto, M. Ito, M. Yasuda, M. Watanabe, and K. rescence optical pumping and method using same,” U.S. Patent 5 657
Kayama, “High frequency discharge lamp for a spectral line source,” 340, Aug. 12. 1997.
U.S. Patent 4 095 142, June 13, 1978. [310] I. Liberman and P. J. Chantry, “Miniature atomic frequency standard,”
[279] A. F. Podell and L. F. Mueller, “Compact microwave resonant cavity for U.S. Patent 5 670 914, Sept. 23, 1997.
use in atomic frequency standards,” U.S. Patent 4 349 798, Sept. 14, [311] K. Atsumi, Y. Nakajima, and Y. Koyama, “Rubidium atom oscillator
1982. with temperature stabilized frequency output,” U.S. Patent 5 712 597,
[280] G. Busca and L. Johnson, “Atomic frequency standard having micro- Jan. 27, 1998.
wave loop around absorption cell,” U.S. Patent 4 405 905, Sept. 20, [312] G. L.Skoczen, “Methods and apparatus for digital frequency generation
1983. in atomic frequency standards,” U.S. Patent 5 714 910, Feb. 3, 1998.
[281] W. R. Fowks, “Lamp housing assembly primarily for the lamp of a [313] J. D. Crockett and G. L. Skoczen, “Digital frequency generation in
rubidium frequency standard,” U.S. Patent 4 434 406, Feb. 28, 1984. atomic frequency standards using digital phase shifting,” U.S. Patent 5
[282] W. R. Fowks, “Programmable frequency synthesizer primarily for use in 721 514, Feb. 24, 1998.
an atomic clock,” U.S. Patent 4 446 446, May 1, 1984. [314] Y. Nakajima and Y. Koyama, “Rubidium atomic oscillator with laser
[283] W. R. Fowks, “Radio frequency source circuit primarily for igniting the diode wavelength adjustment,” U.S. Patent 5 751 193, May 12, 1998.
lamp of a rubidium frequency standard,” U.S. Patent 4 456 891, June 26, [315] P. J. Chantry and C. F. Petronio, “Apparatus and method for microwave
1984. field strength stabilization in cell type atomic clocks,” U.S. Patent 5 852
[284] W. R. Fowks, “Method for stabilizing the resonance frequency of a 386, Dec. 22, 1998.
rubidium frequency standard,” U.S. Patent 4 462 006, July 24, 1984. [316] A. Stern, B. Levi, and A. Saksonov, “Atomic frequency standard using
[285] W. J. Riley, Jr. “Methods and apparatus for rapid and accurate frequen- digital processing in its frequency lock loop,” U.S. Patent 6 130 583,
cy syntonization of an atomic clock,” U.S. Patent 4 476 445, Oct. 9, Oct. 10, 2000.
1984. [317] J. Deng, “Subminiature microwave cavity,” U.S. Patent 6 133 800, Oct.
[286] S. Goldberg, “Vapor discharge lamp assembly,” U.S. Patent 4 485 333, 17, 2000.
Nov. 27, 1984. [318] J. Deng, “Laser light quantum system,” U.S. Patent 6 172 570, Jan. 9,
2001.
 35

[319] R. H. Kern, “Metallic cell for optically activated atomic frequency

standards,” U.S. Patent 6 215 366, Apr. 10, 2001.
[320] G. Mei, D. Zhong, S. An, J. Liu, and X. Huang, “Miniaturized micro-
wave cavity for atomic frequency standard,” U.S. Patent 6 225 870, May
1, 2001.
[321] J. Vanier, “Atomic frequency standard based on coherent state prepara-
tion,” U.S. Patent 6 255 647, July 3, 2001.
[322] M. J. Delaney, K. N. Bonnette , and D. E. Janssen, “Atomic frequency
standard based upon coherent population trapping,” U.S. Patent 6 265
945, July 24, 2001.
[323] J. Vanier, “Atomic frequency standard”, U.S. Patent 6 320 472, Nov. 20,
2001.

Coherent Population Trapping

[324] G. Alzetta, et al, “An experimental method for the observation of RF

transitions and laser beat resonances in oriented Na vapor”, Il Nuovo
Cimento B, vol. 36, no. 1, pp. 5-20, Nov. 1976.
[325] G. Orriols, “Nonabsorption resonances by nonlinear coherent effects in a
three-level system”, Il Nuvo Cimento, vol. 53, no. 1, pp. 1-24, Sep. 1979.
[326] N. Cyr, N. Têtu and M. Breton, “All optical microwave frequency
standard: a proposal, IEEE Trans. Inst. Meas., vol. 42, no. 2, pp. 640-
649, Apr. 1993.
[327] F. Levi, A. Godone and J. Vanier, “On the use of a modulated laser for
hyperfine frequency excitation in passive atomic frequency standards”.
Proc. 11th European Frequency and Time Forum, Neuchatel, Switzer-
land, pp. 216-220, Mar. 1997.
[328] J. Vanier, A. Godone and F. Levi, “Coherent population trapping in
cesium: Dark lines and coherent microwave emission”, Phys. Rev. A,
vol. 58, no. 3, p. 2345, Sep. 1998.
[329] J. Vanier and C. Tomescu, The Quantum Physics of Atomic Frequency
Standards, Recent Developments, Boca Raton, London and New York:
CRC Press, 2016, pp. 127-135 and pp. 300-324.

BIOGRAPHY
William J. Riley, Jr. (S’61M’63SM’81
F’03LF’04) was born in Newark, NJ on Apr.
26, 1939. He graduated with a B.S.E.E. degree
from Cornell University, Ithaca, NY in 1962 and
earned an M.S.E.E. degree from Northeastern
University, Boston, MA in 1966.
His professional career began as a
Development Engineer at General Radio where
he pursued time and frequency technology
including rubidium frequency standards. In
1980 he became the Engineering Manager of the Rubidium Department at
EG&G where he directed the design of rubidium frequency standards,
including those used onboard the GPS satellites. His last position before
retirement was as Manager of Rubidium Technology at
FTS/Datum/Symmetricom. Mr. Riley is also the proprietor of Hamilton
Technical Services where he developed and sold the Stable32 program for
frequency stability analysis (now freely available from the IEEE UFFC)
and consults in the time & frequency field. He is the author of the
Rubidium Frequency Standard Primer, numerous papers and reports, and
holds several patents.
Mr. Riley served on several IEEE standards committees and the PTTI
Advisory Board. He received the 2000 IEEE International Frequency
Control Symposium Rabi Award and the 2011 Distinguished PTTI Service
Award.