EP3748438B1 - Measurement of the precision of a timepiece comprising an electromechanical transducer with continuous rotation in its device for analogue display of the time - Google Patents

Description

Technical area

The invention relates to the field of measuring the precision of a timepiece comprising a timepiece movement which incorporates a continuously rotating electromechanical transducer, which is either arranged in the kinematic chain connecting an energy source to an analog display of the hour, or in kinematic connection with such a kinematic chain. In particular, the invention relates to the measurement of the rate of such a horological movement, respectively of such a watch, and it also relates to the measurement of the precision of a quartz oscillator forming an internal electronic time base which allows to regulate the rotational speed of the rotor of the electromechanical transducer.

By step, this text means the daily temporal drift of the time displayed by the timepiece. The precision of the crystal oscillator can also be given in the form of a daily time drift. A daily time drift is measured relative to a very precise external time base which makes it possible to measure time intervals with very high precision.

According to two main embodiments of the invention, the electromechanical transducer is formed respectively by a small generator in connection with the kinematic chain connecting a barrel, forming a source of mechanical energy, to an analog time display and by a continuously rotating motor which is powered by an electrical energy source and which drives, via a kinematic chain, an analogue time display.

Technology background

The electromechanical transducers considered in the context of the invention are generally reversible, so that they can either produce electrical energy from a source of mechanical energy while allowing regulation of the speed of rotation of the rotor by braking this rotor in a controlled manner, or producing mechanical energy, more particularly motor torque, from an electrical power supply. In the latter case, driving electrical pulses can be supplied to the stator so as to ensure either a certain force torque, or a certain speed of rotation, in particular a nominal speed of rotation in a watch movement. Such transducers are also sometimes called 'electromagnetic transducers', since the rotor-stator coupling is of the electromagnetic type. Indeed, in motor mode, to go from an electric current to a mechanical driving force of a time display mechanism, provision is made for such an electric current to flow in at least one coil in such a way to generate a magnetic field which is coupled to permanent magnets carried by the rotor. In generator mode, to pass from a mechanical driving force of the rotor of the generator to an electric current, which can supply an electronic circuit for regulating the average rotational speed of the rotor, a force torque drives the rotor in rotation whose magnets then induce an electric current in the stator coil.

As regards constructions of watchmaking generators and possible operations of such generators, reference may be made in particular to the documents EP 0 679 968 , EP 0 822 470 , EP 0 935 177 , EP 1 099 990 , and WO 00/63749 . Concerning constructions of watchmaking motors with continuous rotation and possible operations of such motors with continuous rotation, one can refer in particular to the documents FR 2.076.493 , CH 714 041 and EP 0 887 913 . The document " Analyzer Twin" of May 31, 2011 from Witschi Electronic Ltd (http://www.witschi.com/assets/files/sheets/leaflet_analyzer-Twin.pdf, XP055636846 ) discloses a device and a method for measuring the average frequency of a signal from the quartz oscillator of a watch. The measuring device comprises a detection coil capable of detecting a variation of a magnetic field coming from the timepiece. Indeed, a variation of the magnetic field generates an induced voltage in the detection coil.

For conventional watches of the electromechanical type, that is to say watches comprising an electronic quartz movement associated with a stepper motor, it is known to be able to precisely measure the rate of such watches once they are nested and ready to use, without having to open a bottom or a battery door. To do this, there are measuring devices arranged to perform precise time measurements between the steps of the motor, using a magnetic sensor capable of precisely detecting a certain instant relative to each of the electrical pulses supplied to the stepping motors to train him. The electrical pulses generate magnetic pulses in the stator of the motor to rotate its rotor which is equipped with at least one permanent magnet. The magnetic pulses partially propagate out of the stator and they can be detected by a magnetic sensor outside the watch. Such measuring devices can accurately determine the rate of the electromechanical watch given that the driving pulses are generated at regular time intervals, in particular every second, these time intervals being determined by the internal electronic time base, that is that is to say by the quartz oscillator which is inhibited in a known manner to adjust the average frequency of this time base.

Unlike conventional watches of the electromechanical type which include a stepping motor, timepieces comprising an electromechanical transducer with continuous rotation in their movement, as explained previously, do not present a perfectly periodic event which is detectable from the exterior of the timepiece by a measuring device of the type described previously. Indeed, despite a regulation provided to control the average speed of rotation of the electromechanical transducer with continuous rotation so that the time displayed is correct on average and that there is no long-term time drift, the speed of instantaneous rotation varying around the nominal speed of rotation. Thus, in the particular case of a watch with generator undergoing a braking pulse in each half wave of the induced voltage signal generated in the coils of this generator, if the durations between these braking pulses are measured with appropriate means and that, as for the electromechanical watch with a motor step-by-step, an average of these measurements to obtain an average speed, a very long measurement period is then required, for example one day, to obtain the rate of the timepiece with sufficient precision, whereas for the above-mentioned electromechanical watch, it only takes two minutes, for example, to obtain the march with similar precision. The same problem arises in the particular case of a watch equipped with a continuous rotation motor which would receive a driving pulse at each period of the aforementioned induced voltage signal. Then, in the case where the braking pulses or the driving pulses are not provided regularly in each halfwave or each period of the induced voltage signal, the measurement becomes even more problematic. It is therefore understood that there is a real need to find a method for measuring the rate of a finished watch whose time display mechanism is in kinematic connection with a continuously rotating electromechanical transducer. 'Finished watch' means a watch whose watch case is closed with the movement mounted inside.

Summary of the invention

The object of the present invention is to provide a method for measuring the rate of a timepiece whose time display mechanism comprises a kinematic chain between a motor device and the time display which incorporates a continuously rotating electromechanical transducer or which is in kinematic connection with such a continuously rotating electromechanical transducer, taking into account that the rotational speed of its rotor is generally variable even if it is regulated to be on average equal to a rated rotational speed.

To this end, the invention relates to a method for measuring the average frequency of a digital signal which is derived from a periodic reference signal generated by an oscillator forming an electronic time base of a timepiece. The timepiece comprises a movement incorporating a mechanism formed by a kinematic chain which is arranged between a drive device of the movement and an analog time display device, this kinematic chain comprising or being kinematically connected to an electromechanical transducer with continuous rotation, the average speed of rotation of which is regulated by a regulation device, associated with the electronic time base, as a function of a nominal speed of rotation. In the case of a continuous rotation motor, it is understood that it forms the aforementioned motor device. The regulation device is arranged to supply the electromechanical transducer successively with regulation pulses to regulate its average rotational speed, these regulation pulses respectively defining the same events which are synchronized on the rising edges or on the falling edges of said digital signal and which are detectable, by a measuring device without galvanic contact with the movement, at respective instants of detection having the same temporal phase shift with said same events.

1. A) Mesure, sans contact galvanique avec le mouvement, d'une pluralité d'intervalles de temps successifs intervenant chacun entre deux instants de détection qui sont détectés pour deux impulsions de régulation respectives parmi les impulsions de régulation;
2. B) Détermination, pour chaque intervalle de temps de la pluralité d'intervalles de temps, d'un nombre entier correspondant qui est égal à l'arrondi, à l'entier le plus proche, du résultat de la division de cet intervalle de temps par la période moyenne théorique;
3. C) Sommation des nombres entiers déterminés à l'étape B) pour la pluralité d'intervalles de temps, pour obtenir ainsi un nombre total de périodes dudit signal digital;
4. D) Sommation des intervalles de temps mesurés de la pluralité d'intervalles de temps, pour obtenir ainsi une durée totale de mesure correspondant au nombre total de périodes;
5. E) Calcul de la fréquence moyenne dudit signal digital en divisant le nombre total de périodes par la durée totale de mesure.

The measurement method includes the following steps:

1. A) Measurement, without galvanic contact with the movement, of a plurality of successive time intervals each occurring between two instants of detection which are detected for two respective regulation pulses among the regulation pulses;
2. B) Determination, for each time interval of the plurality of time intervals, of a corresponding integer which is equal to the rounding, to the nearest integer, of the result of the division of this time interval by the theoretical average period;
3. C) Summing the integers determined in step B) for the plurality of time intervals, thereby obtaining a total number of periods of said digital signal;
4. D) Summing the measured time intervals of the plurality of time intervals, thereby obtaining a total measurement duration corresponding to the total number of periods;
5. E) Calculation of the average frequency of said digital signal by dividing the total number of periods by the total measurement duration.

For a timepiece having a quartz oscillator forming its internal electronic time base, it will be noted that this quartz oscillator is normally manufactured so that its own daily error is positive, that is to say that its natural frequency is slightly higher than a theoretical reference frequency, without however exceeding a maximum daily error, for example fifteen seconds per day.

According to a main mode of implementation of the measurement method, the digital signal is an inhibited digital signal which has periods of variable durations according to an inhibition of a certain number of periods of the periodic reference signal during successive inhibition cycles. Conventionally, the movement is arranged so that the average frequency of the inhibited digital signal determines an advance of the indicator members of the analog time display device.

According to a preferred variant of the main mode of implementation, the inhibition is carried out according to a method which distributes the inhibition of a certain number of periods of the periodic reference signal during each cycle of inhibition. Furthermore, the plurality of successive time intervals is provided so that the increase in the duration of any time interval among this plurality, resulting from the inhibition of one or more period(s) of the periodic signal reference during this time interval, i.e. at most equal to half of the theoretical average period of the inhibited digital signal.

Next, the accuracy of the analog time display device is determined by calculating a relative error given by the result of dividing the difference between the average frequency of the inhibited digital signal, obtained in step E) above, and the theoretical average frequency, for this inhibited digital signal, by this theoretical average frequency.

Finally, the rate of the timepiece is obtained by multiplying the aforementioned relative error by the number of seconds in a day.

The measurement method of the invention applies to a timepiece whose electromechanical transducer is either a generator or a continuous rotation motor.

Brief description of figures

• La Figure 1 montre partiellement une pièce d'horlogerie comprenant dans son mouvement une génératrice électromécanique à rotation continue et pour laquelle la méthode de mesure selon l'invention peut s'appliquer,
• La Figure 2 est une vue en coupe partielle du mouvement de la Figure 1, avec en plus divers éléments de ce mouvement représentés schématiquement,
• La Figure 3 montre schématiquement un mode de réalisation d'un circuit électronique formant le mouvement de la Figure 1,
• La Figure 4 est une vue schématique en perspective d'un dispositif de mesure permettant de mettre en œuvre la méthode de mesure selon l'invention,
• Les Figures 5A et 5B montrent un signal de tension aux deux bornes du stator de la génératrice du mouvement de la Figure 1 et la détection d'impulsions de champ magnétique reçus par le dispositif de mesure de la Figure 4 pour respectivement deux modes de régulation de la vitesse de rotation du rotor de la génératrice,
• La Figure 6 montre partiellement, de manière agrandie, le signal de tension représenté aux Figures 5A et 5B ainsi que divers signaux digitaux intervenant dans le circuit électronique du mouvement pour rythmer l'avance des organes de l'affichage de l'heure et pour permettre la régulation de la vitesse de rotation du rotor du transducteur électromécanique, et
• La Figure 7 est un tableau donnant un exemple d'un certain nombre d'intervalles de temps, mesurés au cours d'une période de mesure légèrement supérieure à un cycle d'inhibition, et divers nombres dérivés de ces intervalles de temps dans le cadre de la méthode de mesure selon l'invention.

The invention will be described below in detail using the accompanying drawings, given by way of non-limiting examples, in which:

• The Figure 1 partially shows a timepiece comprising in its movement a continuously rotating electromechanical generator and for which the measurement method according to the invention can be applied,
• The Figure 2 is a partial sectional view of the movement of the Figure 1 , with in addition various elements of this movement represented schematically,
• The Figure 3 schematically shows an embodiment of an electronic circuit forming the movement of the Figure 1 ,
• The Figure 4 is a schematic perspective view of a measuring device for implementing the measuring method according to the invention,
• The Figures 5A and 5B show a voltage signal at the two terminals of the stator of the motion generator of the Figure 1 and detecting magnetic field pulses received by the device for measuring the Figure 4 for respectively two modes of regulation of the speed of rotation of the rotor of the generator,
• The Figure 6 partially shows, in an enlarged manner, the voltage signal shown in Figures 5A and 5B as well as various digital signals intervening in the electronic circuit of the movement to punctuate the advance of the time display components and to allow regulation of the rotational speed of the rotor of the electromechanical transducer, and
• The Picture 7 is a table giving an example of a number of time intervals, measured over a measurement period slightly longer than an inhibition cycle, and various numbers derived from these time intervals as part of the method measurement according to the invention.

Detailed description of the invention

A method of implementing the measurement method according to the invention applied to a timepiece 2 comprising in its movement 4 an electromechanical generator 6 with continuous rotation (hereinafter 'the generator'), which has a kinematic connection 9 with a kinematic chain 8 which is arranged between a barrel 10, defining a source of mechanical energy and forming a motor device, and a time display 12. The kinematic chain 8 comprises , in the variant shown, a mobile 8A and a gear train 8B, shown schematically, engaged with the time display device 12 comprising the hands 14A, 14B, 14C.

In general, the generator 6 is formed by a rotor provided with permanent magnets and a stator comprising at least one coil through which passes a variable magnetic flux which is generated by the magnets of the rotor when the latter is rotating. In the variant shown, the stator 16 comprises a support 20 carrying three coils 22A, 22B and 22C arranged regularly around the axis of rotation 19 of the rotor and connected to an electronic circuit 24. The rotor 18 comprises a central shaft 32 carrying two flanges 28A, 28B, preferably made of ferromagnetic material, on each of which are arranged regularly, around the axis of rotation, six permanent magnets 30A and 30B having alternating polarities. In other words, two adjacent magnets 30A and 30B of the same flange have reversed polarities, whereas two magnets 30A just like two magnets 30B, carried respectively by the two flanges and aligned in the direction of the axis of rotation 19, have the same polarity. The shaft 32 of the rotor carries a pinion 34 meshing with the wheel of the mobile 8A. Thus, in the variant shown, the kinematic link 9 is formed by the meshing of the pinion 34 with the wheel of the mobile 8A.

Movement 4 further comprises a plate 36 and a bridge 38 in which are respectively arranged two bearings 40A and 40B each provided with an anti-shock device and in which the rotor 18 is pivoted.

To the Figure 3 , the electronic circuit 24 is connected to the terminals 44A and 44B of the coils of the stator 16. When the rotor 18 is rotated, a variable magnetic flux, generated by the magnets of the rotor, passes through the coils and generates in each of they an alternating induced voltage. Given that the coils are three in number, that the magnets carried by each flange are six in number with alternating polarities, and that these magnets and these coils are arranged regularly around the axis of rotation of the rotor, the three Voltages induced respectively in the three coils are substantially in phase. In a first variant, the three coils are arranged in series and the peak voltages substantially add up. It will be noted that in a second variant the three coils can be arranged in parallel. The three coils deliver together, when the rotor is driven in rotation, an alternating voltage U ₁ to the electronic circuit 24 which comprises a rectifier 46, which supplies a _{substantially continuous voltage U 1} * to a voltage regulator 48. The voltage regulator supplies a supply voltage U ₂ to the electronic circuit, in particular to a circuit 50 for regulating the average rotational speed of the rotor 18.

The regulation circuit 50 comprises a switch 52, formed by a transistor, which is controlled by a control unit 54. The switch 52 is arranged between the two terminals 44A and 44B of the stator 16, so that when this switch is closed , that is to say on, these two terminals are electrically connected and the voltage U ₁ is zero, the coils 22A - 22C of the stator then being short-circuited. When the switch is open, that is to say not conducting, the voltage U ₁ is proportional to the voltage induced in the three coils by the magnets of the rotating rotor. The average rotational speed of the generator 6 is regulated, as a function of a nominal rotational speed, by a regulating device formed by the regulating circuit 50. The regulating circuit is associated with an electronic time base 25 which is formed by: - a quartz oscillator 26 which generates a periodic reference signal S _PR , - a first frequency divider 60 which receives the periodic reference signal S _PR and which supplies a periodic digital signal S _DP whose frequency F _DP is equal to the natural frequency F _NR of the periodic reference signal S _PR divided by a given integer, for example two, and - a second frequency divider 62 which receives the signal S _DP and which supplies an inhibited digital signal S _DI to a logic unit 64, which processes this inhibited digital signal to generate a clock signal S _Ho . The inhibited digital signal S _DI is also supplied to the control unit 54. It will be noted that the first divider and the second divider generally form the first two stages of a division unit which also forms at least a first part of the logical unit 64.

In general, given that the manufacture of quartz oscillators does not make it possible to obtain a very precise natural frequency, it is intended to produce quartz oscillators having a natural frequency higher than a theoretical reference frequency F _RT , in a certain given frequency value range. In general, the reference frequency theoretical F _RT is equal to 32'768 Hz. In the variant described, the frequency divider 60 is a divider by two, so that the theoretical frequency FT _DP of the digital signal S _DP is equal to 16'384 Hz and the period corresponding theoretical PT _DP is worth 1/16'384 second. For example, the daily error of uninhibited crystal oscillators is predicted to be between one and twenty seconds.

The second frequency divider is associated with an inhibition unit 66 which, conventionally, inhibits a determined number of pulses in the digital signal S _DP to correct a predetermined error of the quartz oscillator 26 resulting from manufacturing tolerances and the fact that, as already indicated, the quartz crystals are produced in such a way as to present a too high natural frequency in a certain range of frequencies above a theoretical reference frequency F _RT . Then, for each quartz oscillator produced, its natural frequency F _NR is determined and a number of inhibitions per inhibition cycle is calculated, this number of inhibitions being introduced into the inhibition unit 66. In general, the inhibitions are distributed over each of the successive inhibition cycles. In a known variant, an inhibition cycle lasts 64 seconds and the determined number of inhibitions is divided by this number of seconds to obtain a unit number of inhibitions per second. This last number is a real number. At each second during an inhibition cycle, the unit inhibition number is added to a counter and the whole part of the result of the addition performed by this counter is inhibited, then only keeping in the counter the part remaining fractional. Let us take two simple examples: a) the determined inhibition number is 32 and the unitary inhibition number is therefore 0.5, so that the inhibition of a period of the periodic digital signal every two seconds is provided; b) the determined inhibition number is 96 and the unitary inhibition number is 1.5, so that one inhibition and two inhibitions are alternately provided during successive seconds of an inhibition cycle. It will be noted that, advantageously, when the unit inhibition number is greater than one, inhibitions carried out during the same second are not accumulated in the same period of the digital signal inhibited, but are separated by a certain unit time interval, for example substantially 125 ms (1/8 second).

It will be noted that the inhibition of periods of the reference signal generated by the quartz, to adjust the precision of an electronic watch and thus reduce its rate, is a technique well known by the person skilled in the art who knows various ways of implementing it. . The present invention is therefore not limited to a single possible implementation, but to several known variants insofar as certain conditions remain valid, as explained below.

To regulate the speed of the generator, the clock signal S _Ho determines a reference value for the frequency of the voltage induced in the coils, which corresponds to the frequency of the voltage signal U ₁ . This set point value is a function of the nominal speed of rotation of the generator and it is determined by the time base 25, so that it is vitiated by an error corresponding to that of the time base. A voltage comparator 58, one input of which is connected to one of the terminals 44A, 44B and the other input to a reference voltage 59, generates a signal F _UG which is supplied to a reversible counter 56 and to the control unit 54. More particularly, the signal F _UG is a digital signal whose period corresponds to the electric period of the generator, that is to say to the period of the voltage induced in its stator and therefore of the voltage U ₁ . This signal F _UG decrements the reversible counter 56 at each electrical period detected while the logic unit 64 increments this reversible counter at each period of the clock signal S _Ho . Thus, the reversible counter integrates, from an initial instant, a time drift of the generator and therefore of the analog time display relative to a setpoint advance determined by the setpoint value which is derived from the inhibited digital signal supplied by the internal time base 25. The state of the reversible counter is supplied to the control unit 54 which manages the average rotational speed of the generator according to a given method.

The regulation circuit 50 is arranged to supply the generator successively with regulation pulses to regulate its average speed of rotation so that it is as close as possible to a nominal speed of rotation provided for the rotor of the generator. The regulation pulses are formed here by braking pulses of the rotor of the generator which are each generated by a momentary short-circuit of the coil or coils forming the stator of this generator. The nominal speed of rotation is determined by the construction of the movement 4, in particular by the kinematic chain 8 and the kinematic link 9. In the variant described here, the nominal speed of rotation is equal to 64 / 9 = 7.1111 revolutions per second. For the generator described above, the nominal electrical frequency of the alternating voltage signal U ₁ is that of the voltage induced in its three coils. It is worth three times the nominal speed of rotation, i.e. 64 / 3 = 21.3333 Hz. Thus, the nominal electrical period is worth 46.875 ms and the nominal duration of an alternation of signal U ₁ is exactly equal to 23.4375 ms .

To the Figure 4 schematically shown is a measuring device 70 capable of implementing the measuring method according to the invention, using appropriate software, the content of which will become apparent on reading the detailed description of this measuring method. The measuring device 70 comprises a detection coil 72 capable of detecting a variation of a magnetic field coming from the timepiece 2. Indeed, a variation of the magnetic field generates an induced voltage in the detection coil. By way of example, the measurement device 70 can be materially a device called 'Analyzer Twin' from the company Witschi Electronic SA in Büren in Switzerland, in which specific software is implemented for the implementation of the measurement method according to the invention. Other similar measuring devices for electronic watches can also be used. Indeed, it is not useful that the measuring device can also be used for mechanical watches, as is the case with the 'Analyzer Twin' model.

In general, the measurement method according to the invention provides for measuring, in particular for a timepiece 2 such as a wristwatch or for a movement 4 ready to be cased, the average frequency of a digital signal internal to the electronic circuit of the movement 4, this digital signal being derived from the periodic reference signal S _PR generated by the quartz oscillator 26 forming the electronic time base 25 of this movement 4. It is expected that the average rotational speed of the generator 6 is regulated by a regulation circuit, associated with the electronic time base, according to a nominal speed of rotation. The regulation device is arranged to be able to successively supply the generator with braking pulses by short-circuiting the terminals 44A and 44B of the coils of the stator 16 of the generator in order to regulate its average speed of rotation. The control unit 54 of the regulation device each generates braking pulses in the following manner: When it is planned to generate a braking pulse with a view to regulating the speed of rotation of the generator, in particular as a function of the state of the reversible counter 56 or possibly also other detected events, the control unit waits to detect in the digital signal F _UG coming from the comparator 58, depending on the variant, either a next rising edge, or a following edge among the edges risers and descenders; then it triggers the braking pulse directly or after a given delay, via the control signal Scom which it supplies to switch 52, by closing this switch at a time td _n , n = 1, 2, 3, . ... In a specific variation, as shown in Figure 6 , the control signal Scom changes from its logic state '0' (switch open) to its logic state '1' (switch closed and therefore on) on the first rising edge of the inhibited digital signal S _DI , received by the control unit to temporally manage the braking pulses, following the considered _edge of the signal F UG . In another specific variant, provision is made to start a braking pulse on the first detected edge, rising or falling, of signal S _DI following detection of the considered zero crossing of voltage signal U ₁ .

In the context of the invention, the regulation pulses respectively define the same events which are synchronized on the rising edges or on the falling edges of the inhibited digital signal S _DI and which are detectable, by a measuring device without galvanic contact with the movement and preferably by a magnetic field sensor 72, at corresponding detection instants. In a main mode of implementation of the measurement method of the invention described using the figures, this event is the end of each braking pulse. As shown at Figure 6 , the respective ends tf _n , n = 1, 2, 3, ..., of the braking pulses BP _n are synchronized and moreover in phase with the rising edges of the inhibited digital signal S _DI and also with the rising edges of the signal periodic digital S _DP . It will be noted that, by the generation of the signal S _DI , the rising edges of this signal S _DI are in phase with corresponding rising edges of the periodic digital signal S _DP . The braking pulses BP _n are identified in the figures either by corresponding control pulses of the control signal S _Com ( Figures 5A and 5B ), or by the extended (i.e. non-point) areas of the voltage U1 where the latter has a zero value ( Figure 6 ), resulting from the control pulses. The braking pulses BP _n have braking durations T _BPn .

In the variant represented, the signal S _DI has an average frequency FM _DI which is, over an inhibition cycle, slightly less than a quarter of the average frequency FM _DP of the periodic digital signal S _DP . The inhibited digital signal S _DI is derived from the signal S _DP with the application of the inhibition provided to correct the relative error of the crystal oscillator. To generate the inhibited digital signal S _DI , the periodic digital signal S _DP is divided twice by two in the divider 62 by applying the inhibition during the first of these two successive divisions by two. To explain how inhibition takes place, we have introduced Figure 6 a dummy inhibited signal S _FI having, outside the periods undergoing inhibition, the frequency of the signal S _DP . Without inhibition, the period P _DI of the signal S _DI is exactly four times the period P _DP of the signal S _DP . On the other hand, when an inhibition 'Inh' occurs during the first division by two of the signal S _DP , a period P _DP of this signal is inhibited, i.e. it is ignored and therefore not taken into account. account, so that the period P _DI * of the signal S _DI generated during this inhibition is greater than that of the period P _DI , since the period P _DI * has in fact a duration equal to five times the period P _DP . We therefore understand that P _DI * = 1.25·P _DI (+25%). The inhibited digital signal S _DI is therefore characterized by an average frequency FM _DI and an average period PM _DI . As the clock signal S _Ho is determined by the signal S _DI and this clock signal determines a reference value for the frequency of the voltage induced in the coils of the generator, there is provided for the signal S _DI a theoretical average frequency FMT _DI and a corresponding theoretical average period PMT _DI which are functions respectively of the nominal electrical frequency and of the nominal electrical period of the voltage U ₁ (which are equal to those of the induced voltage). Over an inhibition cycle, the frequency F _DP of the periodic digital signal S _DP can also vary slightly, so that over an inhibition cycle C _Inh and also over the total measurement duration T _Mes the signal S _DP has a frequency average FM _DP and a corresponding average period PM _DP. Then, to the period P _DP of the signal S _DP and to the mean period PM _{DP there} corresponds a same theoretical period PT _DP , also called theoretical mean period PT _DP , and a same _{corresponding theoretical frequency FT DP} , also called theoretical mean frequency. The theoretical frequency FT _DP is, by construction of the time base oscillator, lower than the average frequency FM _DP .

In the variant described in the figures, the theoretical frequency FT _DP =16'384 Hz and the theoretical period PT _DP =1/16'384 second. Then, the theoretical average frequency FMT _DI equals FT _DP /4, ie FMT _DI = 4'096 Hz, and the theoretical average period PMT _DI = 1 /4'096 second. Finally, it will be noted that the natural frequency F _NR of the periodic reference signal S _PR also presents, over an inhibition cycle or a total duration of measurement, an average natural frequency FM _NR which is twice the average frequency FM _DP of the signal S _DP . To these frequencies F _NR and FM _NR corresponds the theoretical reference frequency F _RT = 32,768 Hz, which is, by construction of the oscillator, lower than the natural frequency F _NR .

1. A) Mesure par le dispositif de mesure 70, qui comprend ou est associé à une base de temps externe très précise, d'une pluralité d'intervalles de temps successifs TI_n, n = 1, 2, 3, ..., N, intervenant chacun entre deux instants de détection correspondant respectivement à deux instants finaux tf_n-1 et tf_n de deux impulsions de freinage successives BP_n-1 et BP_n ;
2. B) Détermination, pour chaque intervalle de temps TI_n de la pluralité d'intervalles de temps TI_n, n = 1, 2, 3, ..., N, d'un nombre entier M_n(S_DP) qui est égal à l'arrondi, à l'entier le plus proche, du résultat NR_n(S_DP) de la division de cet intervalle de temps TI_n par la période théorique PT_DP du signal digital périodique S_DP, soit NR_n (S_DP) = TI_n/PT_DP = TI_n·FT_DP, ou/et d'un nombre entier M_n(S_Di) qui est égal à l'arrondi, à l'entier le plus proche, du résultat NR_n(S_DI) de la division de l'intervalle de temps TI_n par la période moyenne théorique PMT_DI du signal digital inhibé S_DI, soit NR_n(S_DI) = TI_n/PMT_DI = TI_n·FMT_DI ;
3. C) Sommation des nombres entiers M_n(S_DP), respectivement M_n(S_Di) déterminés à l'étape B) pour la pluralité d'intervalles de temps TI_n, n = 1, 2, 3, ..., N, pour obtenir ainsi un nombre total de périodes TNP (S_DP), respectivement TNP (S_DI) du signal digital périodique S_DP, respectivement du signal digital inhibé S_DI ;
4. D) Sommation des intervalles de temps TI_n de la pluralité d'intervalles de temps mesurés à l'étape A), pour obtenir ainsi une durée totale de mesure T_Mes correspondant au nombre total de périodes TNP (S_DP), respectivement TNP (S_DI) ;
5. E) Calcul de la fréquence moyenne FM_DP, respectivement FM_DI du signal S_DP ou/et du signal S_DI en divisant le nombre total de périodes TNP (S_DP), respectivement TNP (S_DI) par la durée totale de mesure T_Mes, soit FM_DP = TNP (S_DP)/T_Mes et FM_DI = TNP (S_DI)/T_Mes.

Using the Figure 4 , 5A , 6 and 7 , the measurement method according to the invention will be described in more detail for a first mode of regulation of the average rotational speed of the electromechanical transducer in which the regulation device is arranged to generate the regulation pulses so that, in normal operation , any two successive regulation pulses have between their respective starts td _n approximately the same positive whole number of half-waves of the induced voltage signal which is generated by the magnets of the rotor in the coil(s) of the stator when this rotor is in rotation. In this first mode of regulation, the regulation of the average speed of rotation of the rotor is obtained by a variation of the duration T _BPn of the regulation pulses. In the variant described here for a generator whose average speed of rotation is regulated by braking pulses, provision is made to generate a braking pulse at each alternation. The measurement method includes the following steps:

1. A) Measurement by the measuring device 70, which comprises or is associated with a very precise external time base, of a plurality of successive time intervals TI _n , n = 1, 2, 3, ..., N , each occurring between two instants of detection corresponding respectively to two final instants tf _n-1 and tf _n of two successive braking pulses BP _n-1 and BP _n ;
2. B) Determination, for each time interval TI _n of the plurality of time intervals TI _n , n = 1, 2, 3, ..., N, of an integer M _n (S _DP ) which is equal the rounding, to the nearest integer, of the result NR _n (S _DP ) of the division of this time interval TI _n by the theoretical period PT _DP of the periodic digital signal S _DP , i.e. NR _n (S _DP ) = TI _n /PT _DP = TI _n ·FT _DP , or/and an integer M _n (S _Di ) which is equal to the rounding, to the nearest integer, of the result NR _n (S _DI ) of the division of the time interval TI _n by the theoretical average period PMT _DI of the inhibited digital signal S _DI , i.e. NR _n (S _DI ) = TI _n /PMT _DI = TI _n · FMT _DI ;
3. C) Summation of the integers M _n (S _DP ), respectively M _n (S _Di ) determined in step B) for the plurality of time intervals TI _n , n = 1, 2, 3, ..., N, to thus obtain a total number of periods TNP (S _DP ), respectively TNP (S _DI ) of the periodic digital signal S _DP , respectively of the inhibited digital signal S _DI ;
4. D) Summation of the time intervals TI _n of the plurality of time intervals measured in step A), to thus obtain a total measurement duration T _Mes corresponding to the total number of periods TNP (S _DP ), respectively TNP ( _SDI );
5. E) Calculation of the average frequency FM _DP , respectively FM _DI of the signal S _DP or/and of the signal S _DI by dividing the total number of periods TNP (S _DP ), respectively TNP (S _DI ) by the total duration of measurement T _Mes , i.e. FM _DP = TNP (S _DP )/T _Mes and FM _DI = TNP (S _DI )/T _Mes .

In step A), the final instants are detected here by a magnetic sensor 72 of the measuring device which is arranged to be able to detect short pulses of induced voltage DE _n , n = 1, 2, 3, ..., intervening at the end of the braking pulses BP _n given the sudden drop in the current induced in the coils of the stator of the generator when the switch 52 is open (turned off) at the end of each braking pulse. To precisely detect the same specific instant of the induced voltage pulses DE _n , two parallel comparators are provided which detect, on the rising edge of these pulses, the instant when the induced voltage reaches a threshold voltage Us or -U _S respectively for the positive and negative pulses which follow one another alternately, given that the braking pulses are carried out at each alternation of the voltage U ₁ at the terminals of the stator 16 of the generator 6. It will be noted that the detection instants have the same small time shift with the respective ends of the corresponding braking pulses.

As indicated above, in the context of the invention, provision is made to measure either the average frequency FM _DI of the inhibited digital signal S _DI , so as to be able to finally determine the rate of the timepiece, or the frequency average FM _DP of the periodic digital signal S _DP so as to be able to determine the precision of the oscillator 26 (generally a quartz oscillator) supplying the periodic reference signal S _PR . Thus, in a first variant, the digital signal is the periodic digital signal S _DP whose average frequency FM _DP is equal to the average natural frequency FM _NR , over the total measurement duration T _Mes , of the periodic reference signal S _PR divided by a given whole number, for example by two. The accuracy of the oscillator is determined by calculating a relative error ER (S _DP ) given by the result of dividing the difference between the average frequency FM _DP of the signal S _DP , obtained in step E), and the frequency theoretical FT _DP of this signal S _DP by this theoretical frequency, ie ER (S _DP ) = (FM _DP - FT _DP )/FT _DP . It will be noted that the relative error of the periodic reference signal S _PR generated by oscillator 26 is identical, ie ER (S _PR )=ER (S _DP ). In a second variant, the digital signal is therefore the inhibited digital signal S _DI which has periods P _DI and P _DI * of variable durations depending on an inhibition of a certain number of periods of the periodic reference signal during successive inhibition cycles. The average frequency FM _DI of the inhibited digital signal determining an advance of the indicator members 14A to 14C of the analog time display device 12, the precision of the analog time display device is determined by calculating a relative error ER (S _DI ) given by the result of dividing the difference between the average frequency FM _DI of the inhibited digital signal S _DI , obtained in step E), and the theoretical average frequency FMT _DI of this signal S _DI by this frequency theoretical average, or ER (S _DI ) = (FM _DI - FMT _DI ) / FMT _DI . The rate of the timepiece is obtained by multiplying the relative error ER (S _DI ) by the number of seconds in a day, ie _{Rate = ER (S DI} )·86'400 [s/day].

Avec FT_DP = 16'384 Hz et FMT_DI = 4'096 Hz, on obtient : $$\mathrm{ER}\left({\mathrm{S}}_{\mathrm{PR}}\right)=\mathrm{ER}\left({\mathrm{S}}_{\mathrm{DP}}\right)=105\cdot {10}^{-6}=105\mathrm{ppm},\mathrm{et}\mathrm{ER}\left({\mathrm{S}}_{\mathrm{DI}}\right)=0.5525\mathrm{ppm}.$$

ER (S_PR) correspond ici environ à 9 s/jour alors que ER (S_DI) correspond à une Marche = 0.0477 [s/jour], et donc à une erreur annuelle d'environ 17.5 s pour une fréquence de référence moyenne annuelle qui correspondrait à la fréquence de référence moyenne FM_NR donnée par le double de FM_DP, soit FM_NR = 32'771.5 Hz.By way of example, taking the measurement results given in the table in the Picture 7 , we have a total measurement duration T _Mes = 64.007533 seconds, the total number of TNP periods (S _DP ) = 1,048,810 and the total number of TNP periods (S _DI ) = 262,175. We obtain : $${\mathrm{FM}}_{\mathrm{DP}}={\mathit{16}}^{\prime }385.7276,\mathrm{and}{\mathrm{FM}}_{\mathrm{DI}}=4^{\prime }096.002263.$$

With FT _DP = 16'384 Hz and FMT _DI = 4'096 Hz, we obtain: $$\mathrm{RE}\left({\mathrm{S}}_{\mathrm{PR}}\right)=\mathrm{RE}\left({\mathrm{S}}_{\mathrm{DP}}\right)=105\cdot {10}^{-6}=105\mathrm{ppm},\mathrm{and}\mathrm{RE}\left({\mathrm{S}}_{\mathrm{DI}}\right)=0.5525\mathrm{ppm}.$$

ER (S _PR ) corresponds here to approximately 9 s/day while ER (S _DI ) corresponds to a Walk = 0.0477 [s/day], and therefore to an annual error of approximately 17.5 s for an average annual reference frequency which would correspond to the average reference frequency FM _NR given by the double of FM _DP , _{i.e. FM NR} = 32'771.5 Hz.

It will be noted that the time intervals TI _n follow one another without interruption. Thus, the total measurement duration T _Mes consists of a plurality of time intervals TI _n , n = 1, 2, 3, ..., N, which are contiguous, these time intervals being measured by the device very precise measurement. The total measurement duration T _Mes therefore corresponds to a period of time without interruption between an initial instant tfo and a terminal instant tf _N . This advantageous variant is optional for measuring the average frequency of the periodic digital signal S _DP , but it is preferable for the inhibited digital signal S _DI because the inhibitions generally do not occur at each time interval TI _n and these inhibitions are not not necessarily distributed in a perfectly homogeneous manner over time.

It will be noted that the total measurement duration T _Mes is provided to be very slightly greater than the duration of an inhibition cycle C _Inh which here is theoretically 64 seconds. In fact, the last time interval TI _N corresponds to the time interval, between two ends tf _N-1 and tf _N of braking pulses, during which the end of a time measurement of a cycle d inhibition C _Inh from the final instant tfo of an initial braking pulse BP ₀ , this instant tfo being selected as the start of the measurement. The temporal measurement of an inhibition cycle is also carried out by the measurement device which comprises or is associated with a very precise external time base, for example an atomic time base. In the variant shown, the total number N of contiguous time intervals measured is equal to 2731, ie N=2731. The nominal electrical frequency of the voltage signal U ₁ is equal to 64/3 Hz. The nominal electrical period is therefore equal to 46.8750 milliseconds. Thus, the nominal duration of an alternation of the voltage signal U ₁ is equal to 23.4375 ms. 2731 vibrations at this nominal duration gives a total duration slightly greater than 64 s, or 64.0078125 s. It will be noted that the nominal duration of an alternation corresponds exactly to 96 theoretical average periods PMT _DI = 1/4'096 s of signal S _DI and to 384 theoretical periods PT _DP = 1/16'384 s of signal S _DP .

The table of the Picture 7 gives the plurality of time intervals TI _n , n = 1, 2, 3, ..., N = 2731, obtained in step A) of the measurement method, as well as the real numbers NR _n (S _DP ) and NR _n (S _DI ) and the corresponding rounded integers M _n (S _DP ) and M _n (S _Di ) obtained in step B) of this measurement method. Given that the rotational speed of the generator varies, it is observed that the integers M _n (S _DP ) and M _n (S _Di ) are variable around the respective nominal integers 384 and 96. As a factor '4' is provided between the nominal integers 96 and 384, and given that the detected events DE _n are synchronous with the rising edges of the inhibited digital signal S _DI , the nominal integers M _n (S _DP ) are even numbers in the absence inhibition during corresponding time intervals TI _n and odd numbers when inhibition occurs during the corresponding time intervals (at most one inhibition per time interval is provided in the variant described here). Thus, one can easily determine on the table of the Picture 7 the time intervals during which inhibitions occur.

The total number of inhibitions in the variant described is equal to 110. This number is equal to the difference between the total number of periods TNP(S _DP ) = 1,048,810 and the total number of periods TNP(S _DI ) = 262'175 multiplied by the aforementioned factor '4'. Thanks to the roundings carried out in the measurement method according to the invention, it is possible to determine both the effective number of periods of the periodic digital signal S _DP , which is not inhibited, and the effective number of periods of the inhibited digital signal S _DI , which is derived from the signal S _DP with the application of the inhibition process to correct the error of this signal S _DP . The rounding carried out on the real numbers NR _n (S _DI ) to obtain the integers M _n (S _Di ) has the consequence that these integers M _n (S _Di ) are independent of whether an inhibition takes place or not during the corresponding time interval TI _n. Thus, thanks to the measurement method according to the invention, despite the fact that the electromechanical transducer has a variable rotational speed, the effective numbers of periods of the inhibited digital signal S _{DI are determined} during the time intervals TI _n which are function of the regulation pulses applied to the electromechanical transducer, these regulation pulses being able to intervene or not during each of these time intervals. Moreover, within the framework of the measurement method according to the invention, it is possible to determine the effective numbers of periods of the periodic digital signal S _DP , which is not inhibited, during the time intervals TI _n and thus determine, in addition to the precision of the internal oscillator, the number of inhibitions per inhibition cycle which has been provided for the timepiece in question and which is stored, at the time of the measurement, in a memory of the inhibition unit 66 or an internal memory accessible to this inhibition unit. It will be noted that this number of inhibitions can generally be replaced or corrected, in particular following a finding that the rate of the timepiece is not optimal or outside a specific range provided for the timepiece in question . The actual theoretical number NT _IC of inhibitions per inhibition cycle to be provided is easily calculated by multiplying the duration of an inhibition cycle C _Inh by the relative error ER(S _PR ) of the frequency of reference and dividing the result by the mean period PM _DP of the periodic digital signal S _DP on which the inhibitions are performed, i.e. NT _IC = C _Inh ER(S _DP )/PM _DP because ER(S _PR ) = ER(S _DP ). For the variant described, we obtain NTic=110.112.

In another variant, a braking pulse is provided at each period of the voltage U ₁ , so that only the positive induced voltage pulses DE _2n-1 or only the negative induced voltage pulses DE _2n appear (see Figure 5A ), depending on whether the braking pulses are applied during the rising edges or the falling edges of the voltage signal U ₁ , and they are detected using a single voltage comparator with the threshold voltage U _S , respectively - U _S . The theoretical average duration of the time intervals is then equal to 46.8750 ms.

To guarantee high precision of the measurement method according to the invention, three conditions set out below are advantageously to be observed.

The first condition imposes a maximum duration on the measured time intervals TI _n . The measurement of the plurality of successive time intervals TI _n in step A) is performed so that each is less than a maximum duration TI _Max which is equal to the theoretical average period for the digital signal considered divided by double the maximum relative error ER _Max for the natural frequency F _NR of the periodic reference signal S _PR relative to the theoretical reference frequency F _RT , i.e. TI _Max (S _DP ) = PT _DP /2 ER _Max (F _NR ) for measuring the average frequency FM _DP of the periodic digital signal S _DP , i.e. TI _Max (S _DI ) = PMT _DI /2 ER _Max (F _NR ) for measuring the average frequency FM _DI of the inhibited digital signal S _DI . As the measurement method is based on rounding to the nearest integer value, to obtain an integer number of periods M _n (S _DP ), respectively M _n (S _Di ) of the digital signal considered which corresponds for each time interval TI _n to the effective integer number of periods of the digital signal considered, it is necessary that each real number obtained NR _n (S _DP ), respectively NR _n (S _Di ) deviates by a maximum of half a period from the digital signal considered relative to the whole number M _n (S _DP ), respectively M _n (S _Di ). As PMT _DI = 4·PT _DP , it is understood that the condition is stricter for the measurement of the average frequency FM _DP of the signal S _DP and therefore of the precision of the oscillator of the internal time base. Also, for the _SDI signal, as there are inhibitions provided to correct the oscillator error and these inhibitions are usually distributed over the inhibition cycles, the first condition discussed here is not necessary to guarantee high measurement accuracy but it ensures high accuracy in all cases. As a numerical example, if we take twenty seconds/day as the maximum error for the oscillator, ER _Max (F _NR ) is approximately 230 ppm (0.00023), TI _Max (S _DP ) = 132.7 ms and TI _Max (S _DI ) = 530.8 ms. In the variant considered, the theoretical duration of an alternation of signal U ₁ is equal to 23.4375 ms, so that at least one braking pulse is required every five alternations to accurately measure the average frequency of the oscillator, respectively at least one braking pulse every twenty-two half-waves to accurately measure, in the absence of inhibition during at least one of the time intervals TI _n , the average frequency of the inhibited digital signal and therefore the operation of the timepiece.

The second condition relates to the maximum number of inhibitions which can occur during each time interval TI _n . In order to obtain an integer number of periods M _n (S _Di ) of the inhibited digital signal S _DI which corresponds, for each of the time intervals TI _n , to the effective integer number of periods of this inhibited digital signal, the plurality of successive time intervals is provided so that the increase in the duration of any time interval among this plurality, resulting from the inhibition of one or more period(s) of the periodic reference signal during this time interval, or at most equal to half of the theoretical mean period PMT _DI of the inhibited digital signal (it being understood that a number equal to one and a half integers is rounded off to this integer). In the variant described, these are periods of the periodic digital signal S _DP which are inhibited. As the ratio between the theoretical average period PMT _DI of the inhibited digital signal and the theoretical period PT _DP of the signal S _DP is four, i.e. PMT _DI = PT _DP /4, this second condition implies for this variant that there is at most two inhibitions per time interval TI _n . As the period P _DP of the signal S _DP is practically less than the theoretical period PT _DP , there is a certain margin by limiting the inhibitions per time interval measured to two inhibitions.

It will be noted that the second condition is advantageous in order to ensure high measurement accuracy in all cases, but it is not necessary in all cases. Indeed, in a mode of the inhibition method which distributes the inhibitions during an inhibition cycle according to a substantially uniform scheme, for example by distributing the number of inhibitions as well as possible in sub-periods of the cycles of inhibition and by avoiding carrying out in these sub-periods more than two pulses in a short period of time, it is possible to have more than two inhibitions per time interval if the time intervals TI _n are, in a variant, long enough. With a braking pulse every half-wave, as in the variant described previously, it is observed that the maximum number of inhibitions during each half-wave is indeed equal to two. On the table of Picture 7 , take the time interval TI ₂₃₃ where an inhibition already occurs, we have NR ₂₃₃ (S _DI ) = 94.240. If we added one more inhibition, we would obtain approximately NR(S _DI ) = 94.490 which rounds correctly to M (S _DI ) = 94. With three inhibitions, we would have an NR(S _DI ) greater than 94.50, which would generate an error in accounting for the effective number of periods of the inhibited digital signal. On the other hand, if the time interval TI _n had a sufficiently long duration for the error generated by the oscillator to be greater than a theoretical period PT _DP of the signal S _DP , then there could be three inhibitions during a such time interval and always have a correct rounding to the number of effective periods of the signal S _DI . According to the calculations and results given in connection with the first condition exposed previously, it can therefore be concluded that there could be three inhibitions during a time interval greater than 22 alternations of the voltage signal U1, i.e. at least 23 alternations between two braking pulses determining the time interval considered and preferably at least 24 alternations, ie 12 electrical periods. Thus, the person skilled in the art can understand that there is a certain link between the time intervals which are measured during the implementation of the measurement method of the invention and the inhibition method to be provided, and therefore that there is a certain relationship between the number of regulation pulses per unit time, when implementing the measurement method of the invention, and the mode of distribution of the inhibitions during the inhibition cycles .

The third condition for guaranteeing high measurement precision relates to the total measurement duration T _Mes for measuring the average frequency of the inhibited digital signal and the rate of the timepiece. As indicated, conventional inhibition methods plan to distribute the inhibitions during each cycle of inhibition. In a particular mode of implementation, the inhibitions, the maximum integer number of which per inhibition cycle is 255 or 511, are distributed per second. An inhibition cycle theoretically lasts 64 [s]. As already explained previously, in each sub-period of one second, an integer number of inhibitions is carried out, corresponding to the integer value of the total number of inhibitions provided divided by 64, and an additional inhibition is periodically added corresponding to the summation of fractional parts over seconds, each time this summation exceeds unity. In each sub-period of one second, it is planned to perform the inhibitions every TU = 125 ms starting at the start of the sub-period. Thus, if three pulses are provided in a given sub-period, the first occurs at time zero of this sub-period, the second after 125 ms and the third after 250 ms (= 2·TU). Then, there is no more inhibition in this sub-period, namely for a little less than 750 ms.

As it is not known at what instant of an inhibition cycle the first time interval TI ₁ of the measurement method begins, it is advantageously provided that the total measurement duration T _Mes encompasses as closely as possible entirely a cycle of inhibition to be sure that all the inhibitions planned for an inhibition cycle have occurred during the plurality of time intervals TI _n measured. However, as the time intervals are determined by the braking pulses which depend in particular on the variable rotational speed of the generator, it is practically not possible to obtain a total measurement duration T _Mes equal to exactly one cycle of 'inhibition. Consequently, in a preferred variant, provision is made to end the measurements of the time intervals at the first braking pulse following a period of time corresponding to an inhibition cycle. Thus, T _Mes = C _Inh + T _add . It will be noted that the probability of an inhibition pulse being counted in excess is high, or even more than one inhibition if the additional duration T _add were to exceed TU=125 ms. To avoid this, in a preferred variant, provision is made for the time intervals TI _{n to} be less than TU/2. In the variant considered, this means that at least one braking pulse is required at each electrical period of the voltage signal U ₁ . In addition, provision is made to start the first time interval TI ₁ at the end of the braking pulse which directly follows the detection of an inhibition. Thus, it is ensured that one inhibition is not counted too much relative to the total number of inhibition provided in an inhibition cycle. In the preferred variant set out here, provision is therefore made to perform measurements of time intervals between braking pulses and to perform the calculations set out in relation to the table of the Picture 7 before starting the measurement method for the plurality of time intervals TI _n determining the total measurement duration T _Mes .

To the Figure 5B are represented the control signal Scom, the voltage signal U ₁ and the voltage signal U _Det detected by the measuring device in an implementation of the measuring method according to the invention for a second speed regulation mode average rotation of the electromechanical transducer in which the regulation device is arranged to generate the regulation pulses BP _n in such a way that any two successive regulation pulses have between their respective starts td _n approximately a positive whole number of alternations of an induced voltage signal generated by the variable magnetic flux in the stator, formed by at least one coil, when the rotor of the electromechanical transducer is rotating. In the second regulation mode, the regulation pulses have, at least over a certain regulation period, substantially the same duration and the regulation of the average rotational speed of the rotor during this regulation period is obtained by a variation of the positive integer number of the aforementioned alternations between the regulation pulses. Otherwise, the measurement method remains similar to that explained previously for the first mode of regulation and the three conditions explained previously also apply. In the case of a timepiece provided with a generator, it is understood that it is preferable to implement the method of measurement when the barrel which drives this generator is wound up, so that the torque of force is relatively high and it is then necessary to perform sufficient braking pulses to regulate the speed of rotation of the generator.

Finally, all the teaching given in the present description of the invention in relation to a timepiece provided with a generator also applies, by analogy, to a timepiece provided with a continuous rotation motor. and an electric power supply for supplying this motor with driving electric pulses. In such an embodiment, the electromechanical transducer is thus a continuous rotation motor forming the motor device of the watch movement. This motor is formed by a rotor provided with permanent magnets and a stator comprising at least one coil through which passes a variable magnetic flux which is generated by the magnets of the rotor when the latter is rotating. In this case, the regulation impulses are motor impulses which are each generated by a momentary power supply of said at least one coil of the stator. To do this, the switch 52 of the regulation circuit is then arranged between an electric terminal of the stator and a terminal of the electric power supply capable of delivering a certain supply current to the coil.

Claims (12)

1. Method for measuring the mean frequency of a digital signal (S_DP, S_DI) which is derived from a reference periodic signal (S_PR) generated by an oscillator (26) forming an electronic time base (25) of a timepiece (2), this timepiece comprising a movement (4) incorporating a mechanism formed by a kinematic chain (8) which is arranged between a motor device (10) of the movement and an analogue time display device (12), this kinematic chain comprising or being kinematically linked to a continuous rotation electromagnetic transducer (6) whose mean rotational speed is regulated by a regulation device (50), associated with the electronic time base, according to a nominal rotational speed, this regulation device being arranged to successively supply the electromagnetic transducer with regulation impulses (BP_n) to regulate the mean rotational speed thereof, these regulation impulses defining respectively same events (tf_n) which are synchronised on the rising edges or on the falling edges of said digital signal and which are detectable, by a measurement device (70) without galvanic contact with the movement, at respective detection times having the same time phase-shift with said same events;
   the measurement method comprising the following steps:

   A) Measurement, without galvanic contact with the movement, of a plurality of successive time intervals (TI_n) each occurring between two detection times which are detected for two respective regulation impulses among said regulation impulses;

   B) Determination, for each time interval of the plurality of time intervals, of a corresponding whole number (M_n(S_DP), M_n(S_DI)) which is equal to the rounded result (NR_n(S_DP), NR_n(S_DI)), to the nearest integer, of the division of this time interval by the theoretical mean period (PT_DP, PMT_DI) given by said digital signal;

   C) Summation of the whole numbers determined in step B) for the plurality of time intervals, to thus obtain a total number of periods of said digital signal;

   D) Summation of the measured time intervals of the plurality of time intervals, to thus obtain a total measurement duration (T_Mes) corresponding to said total number of periods;

   E) Calculation of the mean frequency of said digital signal by dividing the total number of periods by said total measurement duration.
2. Measurement method according to claim 1, characterised in that the measurement of a plurality of successive time intervals in step A) is performed such that each is less than a maximum duration which is equal to the theoretical mean period for said digital signal divided by double the maximum relative error for the natural frequency (F_NR) of the reference periodic signal relative to a theoretical reference frequency (F_RT).
3. Measurement method according to claim 1 or 2, characterised in that said digital signal is a periodic digital signal (S_DP) wherein the mean frequency is equal to the mean natural frequency, over said total measurement duration, of the reference periodic signal divided by a given whole number.
4. Measurement method according to claim 3, characterised in that the precision of said oscillator is determined by calculating a relative error given by the result of the division of the difference between said mean frequency of the periodic digital signal obtained in step E) and a theoretical mean frequency, equal to the inverse of said theoretical mean period (PT_DP), by this theoretical mean frequency.
5. Measurement method according to claim 1 or 2, characterised in that said digital signal is an inhibited digital signal (S_DI) which has periods (P_DI, P_DI*) of variable durations according to an inhibition of a certain number of periods of the reference periodic signal during successive inhibition cycles; and in that the mean frequency of the inhibited digital signal determines a gain of the indicator organs of the analogue time display device.
6. Measurement method according to claim 5, characterised in that the precision of the analogue time display device is determined by calculating a relative error given by the result of the division of the difference between the mean frequency of the inhibited digital signal, obtained in step E), and a theoretical mean frequency, equal to the inverse of said theoretical mean period (PMT_DI), by this theoretical mean frequency.
7. Measurement method according to claim 6, characterised in that the rate of the timepiece is obtained by multiplying said relative error by the number of seconds in one day.
8. Measurement method according to any one of claims 5 to 7, characterised in that said inhibition is performed according to a process which distributes the inhibition of the certain number of periods of the reference periodic signal using each inhibition cycle; and in that the plurality of successive time intervals is envisaged such that the increase of the duration of any time interval among this plurality, resulting from the inhibition of one or more period(s) of the reference periodic signal during this time interval, is at most equal to half of one / said theoretical mean period of the inhibited digital signal.
9. Measurement method according to any one of the preceding claims, characterised in that said electromechanical transducer is a generator (6) formed by a rotor (18) equipped with permanent magnets and a stator (16) comprising at least one coil (22A,22B,22C) through which a variable magnetic flux, which is generated by the magnets of the rotor when the latter is rotating, passes; and in that said regulation impulses are braking impulses of the rotor each generated by a momentary short-circuit of said at least one coil.
10. Measurement method according to any one of claims 1 to 8, characterised in that said electromechanical transducer is a continuous rotation motor formed by a rotor equipped with permanent magnets and a stator comprising at least one coil through which a variable magnetic flux, which is generated by the magnets of the rotor when the latter is rotating, passes, the continuous rotation motor forming said motor device; and in that said regulation impulses are motor electrical impulses which are each generated by a momentary electrical power supply of said at least one coil.
11. Measurement method according to claim 9 or 10, characterised in that said regulation device is arranged to generate regulation impulses in such a way that, in normal operation, any two successive regulation impulses have between the respective starts (td_n) thereof the same positive whole number of alternations of an induced voltage signal generated by said variable magnetic flux in said at least one coil when the rotor is rotating; and in that the regulation of the mean rotational speed of the rotor is obtained by a variation of the duration of the regulation impulses.
12. Measurement method according to claim 9 or 10, characterised in that said regulation device is arranged to generate regulation impulses in such a way that, in normal operation, any two successive regulation impulses have between the respective starts (td_n) thereof the same positive whole number of alternations of an induced voltage signal generated by said variable magnetic flux in said at least one coil when the rotor is rotating; in that the regulation impulses have, at least over a certain regulation period, substantially the same duration; and in that the regulation of the mean rotational speed of the rotor during said regulation period is obtained by a variation of said positive whole number.

Images