RO127778B1 - Method and system for measuring time intervals - Google Patents

Abstract

The invention relates to a method and system for precisely measuring the time intervals in a wide range of values, in case of using microsystems based on micro-controllers, in particular. The claimed method uses a microsystem provided with a counter which registers the clock periods on a time interval from the moment when the initiation of an event is effectively sensed to the moment when the termination of the event is effectively detected, in conjunction with an alternative method which ensures the precise measurement of the small amounts of time on the duration of the interval fractions to be measured where the counter is inactive, and also on the duration of time where the counter is active, although the event is terminated, the microsystem switching rapidly between the two methods, the total measured time being obtained as a sum of the time determined by the counter and the times determined for the situation when the counter is inactive on the duration of the external impulse where from there is subtracted the time determined for the situation when the counter is active beyond the moment when the external event is terminated. The claimed system comprises a pulse shaper (5) producing an input binarized signal for controlling a circuit implementing a method of measuring the small amounts of time with high precision, comprising a current generator (8), a sampling-memorizing circuit (9) and a circuit (6) which implements the time-to-voltage conversion and also some circuits (11) ensuring the measurement by the method of counting by a programmable system provided with a controllable counter.

Description

The invention relates to a method and a system for measuring time intervals in a wide range of values, which is adapted to the implementation on common microsystems.

Several methods for measuring time intervals are known and electronic systems that implement such methods. These methods are usually divided into the following categories:

The method based on the counting of the impulses from a precision oscillator (clock) within the measuring time interval; this method is also called the counting method. The advantage of the direct counting method is that it offers the possibility of measuring long time intervals, but the accuracy of the measurement is directly proportional to the clock frequency. Some methods for improving the accuracy of measuring time intervals using the counting method are presented in the papers: “MeasurementTechniques”, Volume 23, no. 7, published in October 2007, Ya. Mints, V. N. Chinkov, Article "Accuracy of Time Interval Measurement by Puise Counting", pp. 603-606, "MeasurementTechniques", Volume 36, no. 12, published in 1993, O. N. Muzyehenko, the article Time and Frequency Measurements Improvingthe Accuracy of Time Interval Measurements by Puise Counting ”, pp. 1358-1362.

The time-voltage conversion (CTT) method is based on the charge of a capacitor during the event (measuring time interval), the measurement of the voltage at which that capacitor is charged and the conversion of voltage into time, according to the inverse function of the charge function. Charging can occur through a resistance of usually high value, according to the well-known exponential charging law, or with the help of a constant current generator, in which case the charge is linear in time. Implementations of this method are described in patents US 7460441 (B2), published 2.12.2008 and US 7330803 (B2), published on 12.02.2008 and patent applications WO 9713198 (A1), published on 10.04.1997 , US 20080169826 (A1), published on 17.07.2008 and US 3983481, published on 28.09.1976. The time-voltage conversion method consists of the conversion of time into amplitude (voltage), followed by analog-to-digital conversion. The principle of measurement of this method consists in charging a precision capacitor at a known current throughout the external event that can be delimited by a digital signal generated, for example, by a monostable circuit. At the end of the interval, the potential corresponding to the charge accumulated in the capacitor is measured using an analog-digital converter (CAD). Because the voltage on the capacitor is proportional to its charge time, the value read by the analog-digital converter allows the duration of the event to be determined. Another approach to this method of implementing the CTT method is using a fixed voltage source that achieves capacitor charging through a resistor during the measurement event. In this case, the time can be extracted from the analog-to-digital conversion result by reversing the exponential charge law of the capacitor.

A combined variant of the above methods is called the method of extension (expansion) of time intervals and consists of charging a capacitor during the event, followed by unloading the capacitor on a much higher value resistor, or using a value current generator much less than when loading, the download time up to a preset value being measured by the counting method. The method of expanding time intervals was introduced in the paper Application Note AN1016, MicrochipTechnology Inc., K. Blake, S. Bible, "Detecting Small CapacitiveSensors Usingthe MCP6291 and PIC16F690 Devices", published in 2005.

The Vernier method in various embodiments is known from patent applications WO 03081266 (A1), published on 02.10.2003, US 2003006750, published on 09.01.2003, CA 2379071 (A1), published on 26.09.2002, US 4908784, published on 13.03.1990. The method uses two oscillators with different frequency and determines, on the one hand, by counting

RO 127778 Β1 direct, the number of whole periods of one of the oscillators in the given interval and the number of 1 periods of the second, the difference between these numbers determining the fraction of interval of tact period left unmeasured on the basis of the first oscillator. The disadvantage of method 3 is the relatively limited cost and accuracy of measuring the interval fraction.

The method of delay lines found in patent applications US 2010231437 5 • (Al), published on 16.09.2010, US 2003197498 (A1), published on 23.10.2003, US 2002131035 (Al), published on 19.09.2002, uses Delay lines with delays 7 known, but different, through which the impulse representing the event propagates in parallel.

By this method, the coincidences of the output pulse are determined, on the various 9 parallel delay lines, thus indirectly resulting in the duration of the event.

The interpolative method, also called the Nutt method, combines a method of measuring 11 short time intervals with a method of measuring long time intervals. The method is described in patent application GB 1487780 (A), published on 05.10.1977. 13

These methods have several disadvantages, the most important of which are listed below. 15

A disadvantage of the direct counting method is that high frequency tact generators (watches) are required to obtain good accuracy, the condition being 17 unattainable or difficult to achieve and expensive for very short time intervals, below 10 ns, and when the method is implemented with the usual microsystems, of 19 microcontrollers, the accuracy of less than 1 microsecond is difficult to obtain.

Other disadvantages of the direct counting method arise when it is implemented using microcontrollers or other microsystems that use interrupt processes or use loop monitoring to respond to the detection of the beginning and 23 end of the event. A microcontroller requires a certain amount of time to begin processing a interrupt routine when the external request occurs, while the latency of the interrupts is called 25. Due to this latency affecting the startup of the counter, the counting is always significantly delayed in the case of measurement systems based on 27 microcontrollers, in the case of the direct counting method. Similarly, there is a delay in the moment of stopping the counting, which occurs after the end of the event. For example, using 29 a microcontroller driven by a 40 MHz quartz. for which the execution time of an instruction with the duration of a machine cycle requires 4 touches, the corresponding maximum error 31 is 500 ns. These delays cannot be completely eliminated by calculation because the latency of the interrupts has a random variability due to the asynchrony between the internal event 33 and the machine cycles of the processor. Therefore, due to the uncertainty of the latency of the interrupts, variable delays occur on any processor, with a variation of one machine cycle or 35 more which is a cause of error. The error summates the variation of a machine cycle of the response time of the microcontroller to the external interruption for the beginning of the event 37 and the variation of a machine cycle upon noticing the end of the event. Therefore, the best accuracy that can be obtained using this method on a 39 microcontroller with uncertainty of the latency of one machine cycle interruption is two machine cycles. This corresponds, for example, to a microcontroller with a 41 40 MHz tact generator and a 4-touch machine cycle, at a 200 ns error.

The errors for determining the start and end times of the event are also 43 higher than those described above if the microcontroller accepts several types of interruptions, hierarchical, and the interrupts used for reporting the event do not have the highest priority.

EN 127778 Β1 in the case of the counting method implemented on a microcontroller, by increasing the frequency of tact, the machine cycle time decreases, which leads to the reduction of errors due to interruptions. This way of increasing the accuracy has the disadvantage that it requires faster microcontrollers, which determines the increase of the consumed power of the measurement systems.

The time-voltage conversion method (CTT) has disadvantages related to the range of the measured durations. The method is not applicable when the range of durations to be measured is large (several orders of magnitude) and the accuracy required is also high, because the short durations require capacitors of small values with fast loading, and the long durations require capacitors of large values, with slow loading. The time-voltage conversion method also requires high precision current generators with fast circuit switching, or precision capacitors for charging the capacitor.

The disadvantage of the CTT method is that it is not possible to measure very long and very short durations with the same precision using a single value of capacity. Thus, a reduced value capacity will allow the measurement accuracy to be increased at the expense of reducing, as a result of reaching the maximum capacitor charge value, the maximum range that can be measured. On the other hand, increasing the value of the capacity determines the increase of the maximum limit of the measurable interval, but this fact determines the reduction of the measurement accuracy.

The delay line method has disadvantages similar to the CTT method, which is not applicable when the range of lengths to be measured is large. In addition, the method has a relatively high cost.

The technical problem solved by the present invention is the precise measurement of the time intervals that can occupy wide ranges of values, of several orders of magnitude, with error independent of the length of the time interval and the compensation of the specific errors induced by the functioning of the microsystems.

The solution of the problem consists in the fact that an iterative type method is used, adapted according to the invention to the use on microsystems, so that the errors produced by the variability of the latency times of the microsystem and the asynchronism between the event and the tact of a time measuring counter are simultaneous. minimized. For this purpose, according to the invention, the method is implemented on a microsystem, for example on a microcontroller, equipped with an analog-digital converter (CAD) or with a delay line, as well as with an internal counter. The lengths of time in which the numerator is not activated, due to its asynchronous operation against the measured impulse, or due to the start and stop time of the processes necessary to activate and deactivate the numerator through a process of interruption or monitoring in the loop are determined by a method for measuring the accuracy of the short times, for example by the time-voltage conversion method, so as to eliminate the errors introduced by the interruption or loop monitoring processes, and the time durations of the event in which the counter is activated are measured by the counting method, the duration of the event being calculated based on the individual times thus measured.

The iterative hybrid metering mode can be implemented using low-cost microcontrollers, which include a sample-memory circuit and an internal or external analog-digital converter, as well as using ASIC or FPGA systems that include delay lines.

The method and the non-limiting embodiments of the present invention can be applied, respectively, using existing digital circuits, including currently marketed microcontroller-based microsystems, as well as other existing circuits, such as analog-digital converters, controlled current generators, and power lines. delay. The particular constructive solution can be determined using current computer-aided design methods.

RO 127778 Β1

The method of measurement and the systems of measurement of the time intervals, according to 1 invention, have the following advantages:

i) Allow achievements on common microsystems, including embedded systems (called 3 and embedded), at a low cost.

ii) Provides high accuracy. 5 iii) Ensures measurement over a wide range of durations.

The following is an example of embodiment of the invention in relation to the explanatory figures 7, which represents:

FIG. 1, an exemplary diagram of time subintervals when measuring a 9 time interval by the direct counting method, in case the counter is activated and deactivated directly by the measuring pulse or via a control circuit 11 which can be made with a microsystem like a microcontroller;

FIG. 2, an example of the operating time diagram of a measuring system 13 with microcontroller used in the implementation of the measurement method, in a non-limiting embodiment, according to the invention, if a simple instruction lasts a machine cycle of 4 tact periods, and the latency time of an interruption is 3 or 4 cycles machine and for the case of the combined use of time-voltage conversion and the direct counting method 17, as well as when the duration of the event is measured using maximum priority interruptions; 19

FIG. 3, the measurement method by the hybrid method according to the invention, for the case of the combined use of the time-voltage conversion and the direct counting, when for the 21 charging capacity (4) a constant current generator (8) is used;

FIG. 4, the measurement method by the hybrid method, according to the invention, for the case 23 of the combined use of the time-voltage conversion and the direct counting conversion, when the capacity loading (4) is realized by a resistor (7) from a voltage generator (1), 25

FIG. 5, the block diagram of the time measurement system, if an impulse trainer is used at the input;

- Fig. 6, a non-limiting example of a pulse formation circuit from an analog input signal, as well as the connection circuit to a microsystem equipped with a 29 sampling-memory circuit consisting of the resistor (7), the switch (2) and the capacitance (4) as well as with analog-to-digital converter (6) .31

The measurement method according to the invention is described in connection with FIG. 1 and 2, in which the notations are the following: Event - a binary impulse that appears at the entrance of the system of measurement at an indefinite time, some having an unknown duration, or an analogical process that can be characterized by such binary impulse; t _E - duration of the event 35; F _osc - the _tactile frequency of the microsystem used to measure the duration of the event; execution line pC instructions - conventional representation of a succession 37 instructions executed by the microcontroller; t _N - time measured by the counting method using the internal numerator; time counter - the signal that increases the counter 39 on the positive front; t _LH - the duration of the time interval measured by the alternative short-term measurement method, such as the time-voltage conversion method, from the beginning of the event 41; t _HL - the duration of the time interval measured by the alternative measurement method, from the end of the event; t _lR1 - the time elapsed since starting the interrupt routine 43 and the moment of starting the counter; t, _R2 - the time elapsed since the start of the interrupt routine treatment and the moment of stopping the counter. 45

The measurement method according to the invention combines in a manner that takes into account the errors that are introduced by the operation of the current microsystems, such as those based on 47 common microcontrollers, including errors due to latencies and their variability, a method of

EN 127778 Β1 measurement of large time intervals, namely the direct counting method, with an alternative measurement method for smaller, more precise time intervals. Namely, the measurement method according to the invention combines the application of a short time measurement method, for example the CTT method, during the intervals t _LH and t _HL of FIG. 1, with the application of the method of counting on long intervals that take place between the intervals t _LH and t _HL , long intervals that could not be determined with a specific method of measuring for short times. By using the method of measuring the short times, the errors given by the asynchronism of the event with the numerator and the variability of the interruption latency are eliminated. The application of the short time measurement method is facilitated by the fact that for their duration t _LH and t _HL their maximum value is known.

In the case exemplified in FIG. 2, the alternative method is considered to be the time-voltage conversion method, the notations being consistent with the use of this method. The two measurement methods are called successively, so that in the time intervals t _LH and ¢, where the direct counting method is erroneously inactive or erroneously active, the alternative, precision method is called and performs the time measurement. In this way, three time values are determined, namely the T ₁ value obtained by direct counting, the T ₂ value obtained by the alternative method, T ₂ corresponding to the measured value for t _LH according to fig. 2, as well as the value T ₃ obtained by the alternative method, T ₃ corresponding to the measured value for t _HL according to fig. 2. The values T _1: T ₂ and T ₃ are used to determine the time measured for the respective event according to the relation T _m = + T ₂ - T _{3 The} operations of storing the values

Ț, T ₂ and T ₃ and the calculation for T _m are performed by the microsystem. The value of T is stored during the event after starting the numerator, the rest of the operations being performed after the event cessation, in the breaks between events whose duration is measured. The programming details corresponding to the operation described above are known to the engineers in the field.

The time diagram corresponding to the operation of the measurement system according to this non-limiting embodiment is shown in FIG. 2. A machine cycle specific to the microcontroller considered in this non-limiting example lasts 4 clock ticks. I niţierea preceding the start time conversion, respectively, stopping the counter, and the delays Îjri and t, each _{R 2} ^is compensated by the variation of the decrease in the latency time error · interruptions to the example of Fig. 2, latency which is between 3 and 4 machine cycles for the case exemplified in fig. 2, is eliminated by the alternative method of measurement, provided that the instructions and their execution order are as in fig. 2, to ensure the compensation of the respective durations. The instructions executed and their order must be as follows: initialization of the fast (alternative) method and simultaneously interrupt request activated by the external event; effective start of the interruption, in which the first instruction executed is the start of the numerator, and the second instruction stops the alternative method; reading the first value resulting from the alternative method, T1, as well as storing it; setting the option to accept the interrupt on the descending front of the event; the continuation of counting for an indefinite duration, until beyond the time of the occurrence of the end of the event, that is, until the interruption of notification of the end of the event becomes active and the instruction to stop the counter is executed.

The alternative method reduces to an negligible level the error of asynchrony between the counting tact of the numerator and the external event, because the respective phases are included in the time intervals measured by the alternative method.

In a non-limiting embodiment of the measurement system, according to the invention, the method of short-time conversion is used as the time-voltage conversion method, as in FIG. 3, 4 and 5.

In a particular, non-limiting embodiment, according to the invention, method 1 is used.

CTT based on a capacitor 4 charged through a resistor 7 by the pulse representing the event, according to FIG. 3. Another non-limiting embodiment, illustrated in FIG. 4, 3 uses a current generator 8 to charge the capacitor 4.

The capacitor 4, considered here to be integrated in the sampling-memory circuit 9 5 is charged by a digital input signal that is in state 1 logically throughout the event and in state 0 logically. This signal is connected both to the 7 interrupt pin of the microcontroller and to the analog input of the internal sampling circuit. When the interrupt treatment routine starts, generated 9 by the beginning of the event, considered here to be produced by the trainer. pulses 5, in converter 6 both the analog-to-digital conversion begins by switching off the capacitor 4 through 11 opening the switch 2 and counting the pulses of the tact generator. according to the embodiment of FIG. 3, it is made from a 13 voltage generator 1. At the end of the event, another interrupt generated by the decreasing front of the signal describing the event will take place to perform a new analog-digital conversion 15 and stop the counting. Switch 2 is open only during the analog-to-digital conversion period, and the rest is closed. 17

We present for a non-limiting embodiment the main elements of the implementation of the hybrid method on a microcontroller of the PIC18FXXX series. The measuring system 19 corresponds to the diagram in fig. 6, wherein the charging resistor of the measuring capacitor 4 is formed by the group of resistors in series Rj and R ₂ - in the variant illustrated 21 in fig. 6, the system accepts analog signals at the input, and the event consists in framing the input signal between two preset amplitude values. 2. 3

For this non-limiting example of realization, the pseudocode of critical instructions for the accuracy of the measurement that are performed in the interrupt treatment routine is; 25

11: start AD conversion;

I2: selects start or stop numerator depending on event front: positive front 27 - negative event start - end event;

I3: start counter for positive front, respectively, stop counter for negative front 29;

14: setting the interrupt acceptance on the negative / positive front; 31

I5: waiting loop completion AD conversion;

I6: read and save the value given by the AD converter. 33

At both fronts of the signal describing the event, the same instructions 11, I2 and I3 are executed which, for a non-limiting example of realization, require a single 35 cycle machine. The alternation of starting and stopping the counter, respectively, can be ensured by an auxiliary variable that memorizes the next value of the control register of the 37 counter. The modification of the content of the auxiliary variable according to the current state of the numerator (on or off) is carried out in this case in the non-critical program sequence 39 for the accuracy of the measurement.

The method described can be used only when the duration of the measured pulse 41 covers the duration of the execution of all said operations, namely the conversion time, the duration of the interruption service and the duration of at least one counting pulse used to increase the counter 43.

In another non-limiting embodiment, according to the invention, described in connection with FIG. 4, a current generator 8 is used in the CTT method to charge capacitor 4 during the measurement. 47

RO 127778 Β1

An example of a block diagram corresponding to this embodiment is illustrated in FIG. 5, wherein the storage sampling circuit includes the capacitor 4 and the switch in FIG. 3 and 4, and the microsystem includes microcontroller 11 which, in turn, includes the numerator and converter AD 6.

In another non-limiting embodiment, according to the invention, the external event is represented by the passing of an analog type signal through a range of values. In this case, the input circuit of FIG. 6, a circuit consisting of a set of two comparators 12 which determines the thresholds of the predetermined analogue value range and a coincident logic circuit 13.

In a particular, non-limiting embodiment, illustrated in FIG. 6, the input circuit uses fast comparators 12 and a TTL logic gate type AN D 13 with rapid response, which performs the conversion into a digital pulse of the analog event determined by two amplitude values. Fig. 6 shows this external input circuit, of pulse type 5 and the internal circuit of the microcontroller from the input of the analog-digital converter 6, as this circuit appears for example for the case of the microcontroller PIC18F4480.

In all these embodiments, the maximum duration that the short time measurement method has to measure is well determined and is equal to the sum of the variable latency of the interruption latency, for example 4 machine cycles, the duration of a semi-period of the numerator increment tact. , the last duration being due to the asynchrony between this tact and event and the duration of the instructions executed until the numerator starts. If the incremental increment of the counter is the same as the increment of the microsystem, a duration of about 10 machine cycles is usually coverage for many of the current microsystems in applying the described method. These considerations allow the complete design of the circuits that implement the short time measurement method.

Claims (7)

Claims 1 1. Method of measuring the time intervals, according to the invention, characterized in that, for the purpose of determining the duration of events, it consists of counting the tact periods for the duration between the effective notification of the beginning of the event and the effective detection of the 5 end of the event, using a microsystem. endowed with a numerator, in conjunction with an alternative method, which ensures the accurate measurement of small times over 7 times the fractions of the measuring interval in which the numerator is inactive, as well as for the duration of time when the numerator is active, although its time event is concluded, 9 the microsystem rapidly switching between the two methods, the total time measured being obtained as the sum of the time determined by the numerator and the times determined for the situation of inactive numerator 11 during the external impulse, from which the time determined for the situation in which the numerator is active is subtracted beyond from the moment the external event ends. 13 2. Method of measuring the time intervals according to claim 1, characterized in that the alternative method of measurement determines the time interval between the beginning of the event 15 and the effective start of the count, a time interval that includes the time fractions due to the asynchronism of the event with the touch. the numerator and machine cycles of the 17 microsystem, as well as the variable latencies due either to interrupt processes or to a loop detection process for monitoring the external event, latencies that include 19 and the latencies of the interrupt processes involved. 3. The method of measuring the time intervals according to claim 1, characterized in that the alternative method of measurement determines the time interval between the end of the event and the effective stop of the counting, the time interval including the time fractions 23 due to the asynchronism of the event with the numerator touch. and with the machine cycles of the microsystem. as well as latencies due either to interrupt processes or to a 25-process loop detection process of the external event, latencies that also include the latencies of the interrupt processes involved. 27 4. Method of measuring the time intervals according to claim 1, characterized in that it uses as a method of measuring the short time intervals in which the counter is inactive during the event, as well as during the intervals in which the counter is active after completion. of the event, the method of time-voltage conversion by 31 loading during the event of a capacitor (4) through a precision resistor (7). 5. Method of measuring the time intervals according to claims 1 and 2.33, characterized in that it uses as a method of measuring the short time intervals the method of time-voltage conversion by charging during the event of a capacitor 35 (4) using a constant current generator (8). 6. Method of measuring the time intervals according to claims 1 and 2.37, characterized in that it uses the delay line method as a method for measuring the short time intervals. 7. Time interval measuring system, characterized in that it includes a pulse generator (5) that produces the binary input signal for the control of a 41 circuit that implements the method of measuring the small time, consisting of a current generator (8), a sampling-memory circuit (9) and a circuit that implements the 43 time-voltage conversion (6), as well as some circuits (11) that provide measurement by the counting method by a programmable system equipped with a controllable counter . 45