The present invention relates to a wristwatch comprising an atomic oscillator. It also relates to a method of transmitting a time reference signal for a wristwatch by an atomic oscillator.
The quest for precision is one of the driving forces for technical innovation in watchmaking. This precision is in great part determined by the performance of an oscillator, the oscillation frequency of which generates a time signal that determines the timebase exploited by the mechanism of a wristwatch for finally indicating the time on a display.
A first solution in the prior art consists of a mechanical oscillator, based on a flywheel, called a balance wheel, coupled to a spiral spring. The stability of a mechanical oscillator is of the order of one second per day, despite the efforts of innovation based on the choice of particular materials, as is described for example in the documents EP 0 886 195 and EP 1 422 436.
A second solution in the prior art consists of a quartz oscillator, which can achieve a precision of one second per month, or even one second per year using more complicated temperature-compensated devices in order to avoid any drift caused by temperature variations, as is described in document WO 2008/125646.
Finally, a third solution, which is relatively theoretical as it is tricky to carry out in practice, is envisioned in the documents EP 1 852 756 and EP 1 906 271 using an atomic oscillator, based on the known effect of coherent population trapping (CPT), which makes it possible to measure a light intensity transmitted through a mixture of atoms, such as cesium or rubidium atoms. In theory, this solution makes it possible to obtain an oscillator which is more precise than that of the first two solutions. However, these documents do not provide information about the specific construction of an atomic oscillator within a wristwatch. For example, the atomic oscillator is used intermittently without any explanation as to the specific stable implementation of such a principle. Nor is it specified how to achieve both consumption and volume compatible with implementation in a wristwatch.
Thus, the aim of the invention is to provide a wristwatch oscillator that makes it possible to achieve great precision, while respecting the severe constraints of a very restricted volume and low available power within a wristwatch.
For this purpose, the invention is based on a wristwatch that relies on a system for detecting the beat frequencies obtained by the Raman effect in order to obtain a time reference of great precision.
Aspects of the invention are more particularly defined by the claims.
These objects, features and advantages of the present invention will be explained in detail in the following description of particular embodiments given by way of nonlimiting example in relation to the following figures.
The solution adopted is based on the use of an atomic oscillator based on the Raman effect, which relies on the irradiation of reference atoms at an optical resonance frequency which induces the emission of photons with an optical frequency shifted from the hyperfine frequency of the reference atoms. By combining the two resulting signals it is possible to obtain a detectable beat, the frequency of the signal of which serves as the timebase for the wristwatch.
Incidentally, part of the output signal 14 is optionally, but advantageously, used to modulate the laser injection current by microwave injection into the laser 1, this part of this signal being represented by the arrow 7. This arrangement makes it possible to achieve a signal-to-noise ratio at the output 14 which is of better quality and easier to exploit. This principle is equivalent to amplitude modulation of the laser.
It should be pointed out that the cell 3 has been positioned within a magnetic field B, thereby making it possible to lift the degeneracy of the Zeeman substates of the atoms. As a variant, the cell could be placed in a zero magnetic field, making it possible to superpose the energy levels, to obtain a high signal and a simplified oscillator.
These various components make it possible to operate the laser 1 that acts on the optical device 10 of the oscillator, a simplified representation of which has been shown with reference to
It should be noted that CPT-type atomic clocks all use a complex architecture and include a VCO (voltage-controlled oscillator), for correcting the local oscillator, and an electronic device for controlling the oscillator, representing in total a high power consumption. The atomic oscillator of the Raman type described above has the advantage of much greater simplicity for a greatly reduced power consumption.
In such an oscillator using the Raman effect, an incident laser beam at a first frequency interacts with an atomic vapor, thus stimulating, by light-atom interaction, the emission of a second beam having a second frequency through the Raman effect. As was mentioned, the beat between the first and second frequencies produces a third frequency, namely the beat frequency, which is used as timebase. In the case in which the vapor comprises for example rubidium 85 atoms and the laser is of the VCSEL (vertical-cavity surface-emitting semiconductor laser) type emitting a light beam at a wavelength lying in the region of 780 nm or 794 nm, the beat frequency is about 3 GHz with a bandwidth around one hundred kHz or so. This beat frequency is generally of very low level and has a very low spectral content. Detecting such a beat frequency output by the oscillator, to be used in a wristwatch, is a tricky technical problem, in particular for limiting the power consumption.
To solve this technical problem, a system for detecting a narrow-band signal (iPD) of high frequency (ωC), having a low current consumption, is proposed. The system comprises a generator, for delivering the signal (iPD) in the form of a current, and a parallel resonant circuit for varying the output impedance of the generator as a function of the frequency of the generated signal and for converting the current into a voltage. The system also includes an amplification stage for further increasing the gain, whilst minimally impairing the noise of the system, in order to be able to detect a signal of very low amplitude. The generator is the aforementioned photodetector 4 stimulated by electromagnetic radiation.
According to one embodiment of the detection system, shown in
The signal iPD to be detected appears in the form of a current at a node N that connects the inductor L1 to the photodiode PD. This node N is electrically coupled to the input of the amplifier MAMP and the amplified signal appears at the output of the amplifier MAMP. The node N thus configured is therefore associated with a parasitic capacitor CIN. This parasitic capacitor CIN together with the inductor L1 forms the parallel resonant circuit. The inductance of the inductor is determined so that its inductive reactance at the frequency of the signal to be detected is equal to the capacity reactance of the parasitic capacitor CIN. In other words, ωC L1=1/(ωC CIN). These conditions result in a low-pass filter with a quality factor Q and a mid-height width of 1/Q. With an inductor L1 integrated into the circuit, an equality factor Q of about 10 is obtained, whereas with an inductor L1 external to the circuit a quality factor Q of about 50 is obtained. The equivalent parallel resistance Rp is equal to ωL Q. Thanks to a high quality factor Q, it is possible to achieve a high gain without the consumption that would normally be associated therewith. Without the present invention, a broadband transimpedance amplifier with a bandwidth of 10 GHz would be used instead of that proposed. Typically, this kind of amplifier consumes about one watt, whereas the amplifier proposed above consumes less than two milliwatts.
Since the node N has a very high impedance, it is sufficient to use a simple MOS-type amplifier with a common low-noise source to further increase the gain, by minimizing the noise of the system, so as to enable a signal of very low amplitude to be detected. In one embodiment, the amplifier has a resistive load on the output. In another embodiment, profiting from the fact that the signal to be detected has a very narrow spectral content, which may be a single unmodulated frequency, the load at the output of the amplifier is provided by a second inductor L2, the inductance of which is chosen to maximize the gain for a signal at the predetermined frequency ωC.
The input of the amplifier may be coupled in AC mode to the node N, that is to say with a coupling capacitor CC, and the input of the amplifier may therefore be biased by a voltage source Vb through a resistor Rb so that the input of the amplifier is at an optimum voltage.
In the production of a circuit according to the present invention, it may happen that the capacitance of the parasitic capacitor CIN or the inductance of the inductor L1 varies from one batch to another or from one component to another. This would have the effect of shifting the resonance frequency of the resonant circuit to outside the frequency band suitable for detecting a signal at the predetermined frequency. For this reason, it is proposed to adjust the capacitance of the capacitor associated with the node N. This may be accomplished in various ways, for example by using a trimming capacitor or by using several capacitors that may be connected to or disconnected from the node N, for example by the targeted deposition of metal during fabrication. It may also be accomplished by a laser-trimming system in which the node N is connected to a capacitor, the capacitance of which is adjusted by laser ablation at the moment of testing the system.
According to one embodiment of the present invention, the resonant circuit comprises an electromechanical resonator of the BAW (bulk acoustic wave) type as illustrated in
Another technical problem encountered when implementing the oscillator using the Raman effect within a wristwatch is to achieve sufficient stability, while allowing precise operation over a satisfactory time period. This problem is solved by the operation described above in relation to
Feedback of the RF signal detected at the optical laser frequency, so as to control the emission frequency of the laser, is always recommended in the prior art for obtaining a stable high-precision atomic oscillator, in particular for atomic clocks of the CPT type. In the present case, it has been found that it is almost impossible to control the operation of the Raman oscillator repeatedly and reliably in closed-loop mode with respect to the optical frequency of the laser. Synchronous detection, for stabilizing the frequency of a laser, is not appropriate in the case of a Raman oscillator in closed-loop mode.
Surprisingly, it is possible to operate the Raman oscillator without optical frequency feedback control of the laser, that is to say with zero frequency feedback control, or in other words with no active control of the optical frequency of the laser, i.e. operation in open-loop mode with respect to the laser frequency.
Stability tests were carried out according to the above principle that demonstrated great stability. At a temperature of 87.5° C., the Raman oscillator will vary by one second every 160 years and operate in a stable manner for several days at least continuously.
The temperature of the cell, having an active length of 5 mm, was also lowered to below the melting point of rubidium (39.3° C.). Lowering the temperature from 90° C. to 35° C. corresponds to reducing the saturation vapor pressure by two orders of magnitude (˜10−4 torr to 10−6 torr). The stability depends on the temperature of the cell, but this remains acceptable up to a temperature of 35° C. This is because at a temperature of 40° C., the Raman oscillator still operates satisfactorily with a stability of one second every 16 years, something which is remarkable. At 35° C., the Raman signal is still present and sufficiently stable. This unexpected observation makes it possible to envision an atomic oscillator with no cell heating according to one embodiment, operating for example only when the temperature around the cell is high enough, for example around 35° C., preferably around 40° C. Thus, according to one embodiment, the atomic oscillator may operate at a temperature of 40° C. or below, or even 35° C. or below. It is also conceivable to reduce the operating temperature by using Cs instead of Rb in the cell, the melting point of cesium being even lower than that of rubidium (28.5° C. as opposed to 39.3° C.). Thus, a process for emitting a time signal within a wristwatch using an atomic oscillator may comprise a temperature feedback control, the operation thereof being maintained within the abovementioned temperature ranges, and/or a temperature-dependent correction of the time signal.
An additional technical problem is encountered when the oscillator is running. Specifically, the solution explained above shows how to obtain stable high-performance operation of the oscillator when it is in cruise mode on the basis of the devices described in relation to
To do so, it has been found that there is a reduced laser injection current range, i.e. a corresponding frequency range, close to the optical absorption peak of the gas in the cell, which makes it possible, when laser irradiation on the cell starts in open-loop mode, to switch thereafter to closed-loop mode as described above in order to make the oscillator resonate so as to achieve the optimum operating regime described above. Thus, by judiciously choosing the laser injection current upon priming the laser and then placing into closed-circuit mode with respect to the laser injection current as explained above, the oscillator naturally reaches its optimum operating regime. This phenomenon results in self-priming of the oscillator and enables it to be used intermittently.
This operating range is more precisely illustrated in
The above considerations make it possible to implement the priming process using an oscillator for a wristwatch based on the Raman effect, which forms part of the process of emitting a time signal by an atomic oscillator according to the invention.
A first phase consists of seeking the optimum laser injection current i, that is to say the range from i1 to i2. This first phase comprises the following steps:
To give an example, for rubidium and a VCSEL laser used for the experiments, the laser injection current must be chosen to be between 2.25760 mA and 2.25824 mA, with V1 being 15% of Vmax−Vmin above Vmin and V2 being at 67% of Vmax−Vmin above Vmin.
This first phase of the priming process may be carried out before each priming of the oscillator so as to obtain the greatest possible precision, thereby making it possible to modify the preceding values over time according to any drift of the device or of the measurement conditions. As a variant, this phase is carried out only once, in order to calibrate the device, and the data is stored so as to be used at each priming.
The priming process also implements the following steps for specifically priming the laser and the oscillator:
According to an advantageous method of implementation, this process includes a prior step of measuring the optical power of the laser, since the frequency of the oscillator may depend on the optical power interacting with the atoms. This operation may be carried out by measuring the optical power by means of a photodiode of the device and by comparing the photovoltage thus generated with a stable reference voltage source. By adjusting the laser injection current and the laser temperature, it is then possible to obtain the nominal optical power and nominal optical frequency of the oscillator. The process may include a step of adjusting the power of the laser.
According to another advantageous method of implementation, this process includes a prior step of setting the temperature of the gas cell and of the laser, since the operation of the oscillator depends on the temperature, as mentioned previously. There is a correlation between the frequency of the Raman oscillator in closed-loop form and the temperature of the cell. This property enables the frequency to be controlled during the phases of starting and stopping the oscillator, by a single temperature measurement.
Thus, depending on the embodiment chosen, the Raman oscillator includes a temperature feedback control loop. To do this, it includes a temperature sensor, which may be a photodiode, and a heater to increase the temperature if said photodiode is below a temperature setpoint.
The steps described above of the priming process are automatically controlled by the oscillator on the basis of the hardware and software means of the processing unit 23 mentioned above, especially under microprocessor control.
The above atomic oscillator is thus implemented within a wristwatch.
According to a first wristwatch embodiment, the Raman oscillator is used intermittently, to complement a conventional oscillator of the prior art, for example a quartz oscillator. In this embodiment, the atomic oscillator transmits a timebase, which sets the quartz oscillator, corrects it and enables the precision thereof to be greatly increased over time. This intermittent operation of the atomic oscillator has the advantage of controlled additional consumption compared with a conventional wristwatch. Since the priming of this oscillator is controlled by the process described above, the performance of this first implementation in a wristwatch is very high. The atomic oscillator priming period is chosen according to a compromise between power consumption and precision of the wristwatch: the more this oscillator is used, the more precise the clock becomes, but the higher the power consumption. When the additional oscillator of lower precision is corrected by the atomic oscillator, the latter is turned off.
According to a second wristwatch embodiment, the Raman oscillator is used by itself as a replacement for the usual conventional oscillator, as a single timebase, and therefore used for permanent operation. The highest precision is obtained in this embodiment, but at the expense of greater power consumption.
The atomic oscillator described above is also produced with a compact structure, facilitating the insertion thereof into a wristwatch.
These three embodiments differ by the means used to direct the beam onto the cell and the photodetectors and by the means used to prevent the beam reflected by the mirror from interfering with the laser source.
A more complete embodiment example corresponding to the second embodiment is illustrated in
In
According to a standard embodiment, the light 112 produced by the laser 102 is linearly polarized and attenuated by a neutral absorbent filter 104a. A different type of filter may be used in other embodiments. The presence of this filter is not necessary for the invention. A half-wave plate 104b may be used to modify the angle of the linear polarization of the laser source. In combination with the miniature splitter cube 101, the half-wave plate 104b acts as a variable attenuator. In other embodiments, the use of the half-wave plate 104b may be omitted and the ratio of the light intensities of the beams transmitted and reflected by the splitter cube 101 is adjusted by an appropriate orientation of the linear polarization axis of the light emitted by the laser relative to the splitter cube. A quarter-wave plate 105 is placed on the output side of the splitter cube against that face from which the laser beam deflected by the splitter 101 is output, i.e. at right angles to the beam incident on the splitter cube. The fast axis of the quarter-wave plate 105 is oriented in such a way that the incident linear polarization 113 is modified to a circular polarization 114 in a first rotation direction. In other embodiments, the quarter-wave plate 105 is oriented in such a way that the incident linear polarization 113 is modified to a circular polarization in a rotation direction the reverse of the first. The circularly polarized laser beam 114 passes through the gas cell 106 and reaches the mirror 107. The latter reflects the beam only partially and part of the beam passes through the mirror 107 to be directed onto the photodetector 109. According to a standard embodiment, the gas cell is made of glass-silicon-glass by MEMS (microelectromechanical system) techniques with an internal volume of typically 1 mm3 and filled with an absorbent medium of the alkali metal (rubidium or cesium) atomic vapor type and a buffer gas mixture. According to a standard embodiment, the gas cell is filled with natural rubidium and a nitrogen/argon mixture as buffer gas. In other embodiments, other types of cell may be filled with different buffer gases. According to one particular embodiment, a miniature cylindrical cell may be used. In another particular embodiment, the gas cell may be integrated into the PBSC 101. The cell 106 may be filled with other types of alkali metal vapor (rubidium 85, rubidium 87 or cesium 133 for example) and other types of buffer gas (Xe or Ne for example).
The backreflected (incident and Raman) light beams 115 pass through and interact a second time with the atomic medium (two-pass arrangement). The quarter-wave plate 105 converts these circularly polarized light beams into linearly polarized light beams 116. These (incident and Raman) light beams 119 are predominantly reflected and reach the first photodetector 108a, which records the beat frequency between the incident beam and the Raman beam. In a standard Raman embodiment, the first photodetector 108a is a high-speed semiconductor (silicon or gallium arsenide) photodetector which is positioned at the focus of the concave mirror 107. In other Raman embodiments, various types of high-speed photodetector may be used. The second photodetector 108b records the light 118 coming directly from the laser 102 and initially transmitted by the miniature splitter cube 101. In this way, the output power of the laser diode 102 may be measured and regulated. As an option, the photodetector 121 records the backreflected beam 117 transmitted by the splitter 101. The diaphragms 110 and 111 are used to prevent undesirable light from reaching the photodetectors if their dimensions are greater than those of the miniature splitter cube 101.
This use of the semitransparent mirror 107 makes it possible for the light that has interacted with the atoms of the cell to be detected by the photodetector 109. To prevent the beams backreflected by the mirror from interfering with the laser source 102, it is also advantageous to place a polarizer 103 in front of the laser source 102, and with a transmission axis parallel to the polarization of the beam emitted by the laser source 102.
As an option, the following elements may also be used:
It should be noted that, in these embodiments described in relation to
Number | Date | Country | Kind |
---|---|---|---|
11405232 | Mar 2011 | EP | regional |
Number | Name | Date | Kind |
---|---|---|---|
5881026 | Baur et al. | Mar 1999 | A |
6806784 | Hollberg et al. | Oct 2004 | B2 |
7077562 | Bourgeois et al. | Jul 2006 | B2 |
7391273 | Seki et al. | Jun 2008 | B2 |
7697377 | Agesawa et al. | Apr 2010 | B2 |
20020163394 | Hollberg et al. | Nov 2002 | A1 |
20070076776 | Lust et al. | Apr 2007 | A1 |
20090128820 | Nomura | May 2009 | A1 |
20090256638 | Rosenbluh et al. | Oct 2009 | A1 |
20100188081 | Lammegger | Jul 2010 | A1 |
Number | Date | Country |
---|---|---|
0886195 | Dec 1998 | EP |
1422436 | May 2004 | EP |
1852756 | Nov 2007 | EP |
1906271 | Apr 2008 | EP |
2008125646 | Oct 2008 | WO |
2011026252 | Mar 2011 | WO |
Entry |
---|
Author: Natasa Vukicevic, Alexander S. Zibrov, Leo Hollberg, Fred L. Walls, John Kitching and Hugh G. Robinson Title: Compact Diode-Laser Based Rubidium Frequency Reference Publisher: IEEE Traansactions on Ultrasonics, Ferroelectrics, and Frequency control vol. 47, No. 5, Sep. 2000. |
European Search Report (ESR) of EP 11 40 5232, mailing date Aug. 18, 2011. |
Number | Date | Country | |
---|---|---|---|
20120229222 A1 | Sep 2012 | US |