JP2003270369A - Time correction method and time correction device for real time clock - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a highly precise real time clock at a low cost. <P>SOLUTION: This real time clock RTC 2 with an error correction function forms a real time clock. An AT quartz resonator 3 is provided with a flatter temperature characteristic in comparison with a tuning fork type quartz resonator 1 within a usage temperature range. A counter 4 counts a reference clock, which inputs only oscillations of the AT quartz resonator 3 in an initial time, as a source of time counting. A memory 5 stores a count value. In an actual use time, the counter 4 counts oscillations of the AT quartz resonator 3 as a source of time counting for the real time clock. A CPU 6 computes a correction value minimizing an error of the real time clock from a differential between the count value of the actual use time and the count value of the memory 5 to set it to the real time clock RTC 2. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a time correction method and a time correction device for a real-time clock that keeps time in a device that requires a clock function.

[0002]

2. Description of the Related Art Conventionally, in a device that needs to be operated in close relation to time, a real-time clock (hereinafter abbreviated as RTC) circuit is used for clocking. R
There are the following methods for time compensation of the TC circuit. (A) The load capacitance of the crystal unit, which is the source of the RTC time count, is trimmed one by one, and the error of the clock is adjusted within the allowable range. (B) The RTC time is periodically corrected based on the frequency of the commercial power source. (C) Receives time signal such as FM broadcast, and R
Correct the time of TC. (D) A high-precision crystal oscillator is mounted separately from the RTC, and the time of the RTC is regularly corrected based on the frequency of the crystal oscillator. (E) Using the RTC with error correction function, a correction value is calculated based on a reference clock from the outside at the time of factory shipment, and time compensation is automatically performed thereafter.

[0003]

However, the conventional technique (1) has a problem that man-hours and cost of the manufacturing process increase because it is necessary to manually adjust each one. In the conventional technique (2), a circuit for detecting the frequency of the commercial power source is required, so that there is a problem that the circuit scale increases and the cost increases. In addition, there is a problem that it cannot be used in areas where the power supply frequency is not stable. In the conventional technology (3), FM is used for each device.
There was a problem that a tuner was needed. The conventional technique (4) requires a high-priced high-precision crystal oscillator, which causes a problem of high cost.

On the other hand, in the prior art (5), time compensation can be realized at low cost. However, a tuning-fork type low-frequency crystal unit that can be used in combination with an RTC with an error correction function is generally used in the operating temperature range of electronic devices (for example, 0 to 40 ° C).
Since the internal temperature deviation is large, there is a problem that the corrected RTC error is limited to about 2 seconds per day, and highly accurate time compensation cannot be realized.

The present invention has been made to solve the above problems, and an object of the present invention is to provide a time correction method and time correction device for a real time clock capable of realizing a highly accurate real time clock at low cost. To do.

[0006]

According to the present invention, a correction value is set for a real-time clock circuit that generates a real-time clock with a period of counting a predetermined number of oscillations of a first vibration source as a unit time, and this correction is performed. In a real-time clock time correction method for correcting the real-time clock by increasing or decreasing the predetermined number according to a value, oscillation of a second vibration source having a flatter temperature characteristic than that of the first vibration source within an operating temperature range. The frequency is measured by using a reference clock that is input only at the initial stage as a timing source, and the measured oscillation frequency is stored, and the oscillation frequency of the second vibration source is measured by the real-time clock. Of the oscillation frequency measured in the actual measurement procedure and the oscillation frequency stored in the initial measurement procedure. Know calculates the correction value the smallest error of the real time clock is the correction value which was set to run a setup procedure for setting the real time clock circuit. Further, according to the time correction method of the real-time clock of the present invention, the oscillation of the second vibration source having a flatter temperature characteristic than that of the first vibration source is measured within the operating temperature range by measuring the reference clock input only at the initial time. Of the second vibration source, and counts the oscillation of the second vibration source during the predetermined measurement time by counting for the predetermined measurement time and storing the measured count value. During use measurement procedure, calculating the correction value that minimizes the error of the real-time clock from the difference between the count value measured in the actual use measurement procedure and the count value stored in the initial measurement procedure,
A setting procedure for setting the correction value in the real-time clock circuit is executed. Further, one configuration example of the time correction method of the real-time clock of the present invention is one in which the measurement time of the actual-use measurement procedure is made larger than the measurement resolution.

Further, the real time clock time correction device of the present invention generates the real time clock with a period of counting a predetermined number of oscillations of the first vibration source as a unit time and at the same time according to the set correction value. A real-time clock circuit with an error correction function that corrects the real-time clock by increasing or decreasing the number, a second vibration source having a flatter temperature characteristic than the first vibration source within an operating temperature range, and the second vibration. An initial measurement means for measuring the oscillation frequency of the source using a reference clock input only at the initial time as a timing source, a storage means for storing the oscillation frequency measured by the initial measurement means, and the second vibration. An actual use measuring means for measuring the oscillation frequency of the source as the real time clock source, and an oscillation frequency measured by the actual use measuring means. And a setting means for calculating the correction value that minimizes the error of the real-time clock from the difference between the stored oscillation frequency and setting the correction value in the real-time clock circuit with an error correction function. . In addition, the time correction device for the real-time clock of the present invention generates the real-time clock with a period of counting a predetermined number of oscillations of the first vibration source as a unit time, and increases or decreases the predetermined number according to a set correction value. A real-time clock circuit with an error correction function for correcting the real-time clock, a second vibration source having a flatter temperature characteristic than the first vibration source within an operating temperature range, and an oscillation of the second vibration source. The initial time measuring means for counting for a predetermined measuring time by using a reference clock input only at the initial time as a time source, a storage means for storing the count value measured by the initial time measuring means, and the second means.
Of the vibration source of the actual use measuring means for counting during the predetermined measurement time by using the real time clock as a time source, and the count value measured by the actual use measuring means and the stored count value The correction value that minimizes the error of the real-time clock is calculated from the difference, and the correction value is set in the real-time clock circuit with an error correction function. Also,
The time correction device of the real-time clock of the present invention is such that the measurement time of the actual-use measuring means is made larger than the measurement resolution.

[0008]

BEST MODE FOR CARRYING OUT THE INVENTION [First Embodiment] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an electronic device according to a first embodiment of the present invention. The electronic device according to the present embodiment is provided with an RTC 2 with an error correction function that uses the tuning fork type crystal resonator 1 that is the first vibration source as a clock source, and the oscillation clock of the AT crystal resonator 3 that is the second vibration source. A counter 4 which serves as an initial measurement means for counting and an actual use measurement means, and a non-volatile memory 5 such as a flash ROM which serves as a storage means that can be written / read at any time.
It has a CPU 6 as a setting means for calculating and setting a correction value of the RTC 2 with an error correction function.

As such an electronic device, there is, for example, a controller for controlling / managing facilities in a building.
The clock precision of the RTC 2 with the error correction function depends on the precision and load capacitance of the tuning fork type crystal unit 1 to be connected (capacitance of the oscillation circuit incorporated in the RTC and stray capacitance of the printed board). In the present embodiment, the adjustment of the variable capacitance and the like is not performed, and the error is calculated by comparing the count value of the reference clock with the count value of the output clock of the RTC 2 with the error correction function,
By setting the correction value that minimizes the error in the RTC2 with the error correction function, the room temperature (temperature condition when setting the correction value at the initial stage. The temperature condition for shipping adjustment is specified from the assumed site environment) ± 3 ppm (ideal value after correction ±
The accuracy of 1.5 ppm is added to the accuracy of the reference clock and the error of automatic measurement).

Hereinafter, the RT in the electronic equipment as shown in FIG.
The time correction method of C will be described. First, at the time of factory shipment (initial time), a pre-calibrated reference clock CLKref is input to an electronic device, and the oscillation clock of the AT crystal unit 3 is countered for a predetermined measurement time measured using this reference clock CLKref. Count at 4. And CPU6
Stores the count value of the measurement result of the counter 4 in the memory 5 (initial measurement procedure).

Next, during the measurement time measured by using the reference clock CLKref, the clock output of the RTC 2 with error correction function is counted by a counter (not shown) in the CPU 6, and the count value of the measurement result and the known adjustment target count value are counted. R with error correction function is obtained by taking the difference with (count value when error of RTC2 with error correction function is 0).
A difference corresponding to the error of TC2 is obtained. And CP
U6 is the R with error correction function based on the difference between the count values.
The correction value of TC2 is calculated, and the calculated correction value is set in the RTC2 with an error correction function. This completes the factory default settings.

Next, the time correction after shipment from the factory will be described. Since the tuning fork crystal unit 1 has temperature characteristics as shown in FIG. 2, if the ambient temperature fluctuates during the operation of the electronic device, the oscillation frequency of the tuning fork crystal unit 1 fluctuates, and the RTC 2 with the error correction function is provided. There is an error in the time of. On the other hand, the AT crystal unit 3 has a temperature of 0 to 50 ° C. as shown in FIG.
Since the temperature characteristic at is flat, the fluctuation of the oscillation frequency is sufficiently small even if the ambient temperature changes. Utilizing such a relationship, the error of the RTC 2 with an error correction function due to the temperature deviation of the AT crystal unit 3 is obtained, and the correction value is automatically reset.

The counter 4 is an RTC 2 with an error correction function.
The oscillation clock of the AT crystal oscillator 3 is counted during a predetermined measurement time measured by using (measurement procedure during actual use). The CPU 6 calculates the difference between the count value stored in the memory 5 at the time of factory shipment and the count value of the measurement result of the counter 4. This difference is RTC2 with error correction function
Corresponds to the error of. Based on the obtained difference, the CPU 6 calculates a correction value that minimizes the error of the RTC 2 with the error correction function, and resets the correction value of the RTC 2 with the error correction function.

According to the above procedure, an inexpensive (that is, low precision) crystal oscillator is used, and an RT with an error correction function is used.
The error of C2 per day can be set to less than 1 second under the worst condition. If one AT crystal unit 3 also serves as the crystal unit required for the operation clock of the CPU 6 and the crystal unit of the counter 4, it is possible to suppress the cost increase.

[Second Embodiment] Next, the second embodiment of the present invention will be described.
The embodiment will be described. FIG. 3 is a block diagram showing the configuration of the electronic device according to the second embodiment of the present invention. The present embodiment more specifically describes the first embodiment. The electronic device according to the present embodiment is the first
Tuning fork type crystal oscillator 11 which is a vibration source, an RTC 12 having an error correction function, an AT crystal oscillator 13 which is a second vibration source, and a storage means which can be written / read at any time, such as a flash ROM. Non-volatile memory 15
And a CPU 16 as a setting means for calculating and setting a correction value of the RTC 12 with an error correction function, and a reference clock CL
It has a digital input circuit 17 for inputting Kref to the CPU 16.

The oscillation frequency of the tuning fork type crystal unit 11 for driving the RTC 12 with an error correction function is 32.768 kHz.
Is. The RTC 12 with an error correction function counts up for 1 second for each 32768 pulses of the clock generated by the tuning fork crystal unit 11. This RTC 12 with error correction function is connected to the CPU 16 via a bus,
From U16, time reading / writing and correction value setting are performed.

The oscillation frequency of the AT crystal oscillator 13 for operating the CPU 16 is 4 MHz. Below, this A
The oscillation of the T crystal oscillator 13 is called original oscillation. Inside the CPU 16, a system clock that is an integral multiple of the original oscillation (3 in this case)
2 MHz) is generated.

The CPU 16 is a processor core 101.
, Memory controller 102, 32-bit timers 103-1 and 103-2, general-purpose timer 104, interface 105, internal bus interface unit 106, and external bus interface unit 107.
And have.

32-bit timers 103-1 and 103-2
A clock signal obtained by dividing the system clock (the division ratio is variable) is used as a clock source for driving the.

A general-purpose timer 104 different from the timers 103-1 and 103-2 is an RTC built in the CPU 16,
Clock signal PITRTC obtained by dividing the system clock
It is driven by a clock having a frequency obtained by dividing LK (variable division ratio) by 9600. The 32-bit timer 103-1 and the general-purpose timer 104 correspond to the counter 4 of the first embodiment.

The reason for using the external RTC 12 instead of the general-purpose timer 104 with a built-in CPU is that
This is because it is necessary to use the RTC with an error correction function in order to realize within 1 second, and because only the RTC 12 is backed up by a battery so that the RTC operates even when the power is cut off due to a power failure or the like.

Hereinafter, the RTC in such electronic equipment
The time correction method will be described. First, at the time of factory shipment,
Input the reference clock CLKref calibrated within 1ppm to the electronic equipment. The reference clock CLKref is input to the CPU 16 via the digital input circuit 17,
16 is sent to the internal bus via the external bus interface unit and the internal bus interface unit. Using this reference clock CLKref, the number of counts of the original oscillation (correctly, a clock generated by dividing the original oscillation) per unit time is measured.

Here, a counter having a cycle is used so that the error due to the resolution of the counter does not pose a problem even in the measurement time of about 10 minutes. Therefore, the 32-bit timer 103-1 is used as the counter which serves as the initial measurement means, and the clock CLKx obtained by dividing the system clock by 16 is used as the clock source to be measured. If the frequency of this clock CLKx is f _CPM , the system clock is 32 MHz, so the frequency f _CPM is 2 MHz.

The count value obtained by counting the clock CLKx by the 32-bit timer 103-1 during the measurement time M seconds measured using the reference clock CLKref is Cx.
At this time, the count value Cx12 obtained by converting the count value Cx12 for M seconds into a value for 12 minutes is obtained by the following equation. Cx12 = Cx × 60 (sec) × 12 (min) / M = 720 × Cx / M (count) (1) The processor core 101 includes a 32-bit timer 103-
The count value Cx12 calculated from the measurement result Cx of 1 is stored in the memory 15.

The errors included in the above initial measurement procedure are as follows. (A) Circuit delay from clock CLKx input to interrupt generation. Up to 125 μsec. (B) S / W processing time from interrupt generation to 32-bit timer reading. To read during interrupt processing (A)
It is very small compared to and can be ignored. (C) Resolution of the 32-bit timer 103-1. When the frequency f _{CPM of the} clock CLKx is 2MHz, 1 / 2MH
z = 500 nsec. From the above (a), (b), and (c), the measurement error is set to 200 μsec with a margin. If the measurement time is 12 minutes, the error is 200 (μsec) / 720 (sec) =
It becomes 0.278 ppm.

Next, using the reference clock CLKref,
The correction value of the RTC 12 with the error correction function is calculated and set. Here, a 32-bit timer 103 is used as a counter.
-2, RTC12 with error correction function for measurement target
The clock output of is counted by the 32-bit timer 103-2 through the interface 105, and the difference between the count value after measurement and the count value (known) of the adjustment target is calculated (measurement time is adjusted in advance according to the required clock accuracy). Further, the adjustment target value is set in consideration of the temperature characteristics of the tuning fork type crystal unit 11 so that the day difference is minimized within the operating temperature range of the product).

The processor core 101 calculates the correction value of the RTC 12 with the error correction function from the difference of the count values, sets the calculated correction value in the clock error correction register of the RTC 12 with the error correction function through the interface 105, and stores it in the memory. -Stored in the memory 15 through the controller 102.

After that, as long as the RTC 12 with the error correction function is backed up by the battery, the time is automatically corrected as described later. When the battery backup is interrupted, the correction value stored in the memory 15 is reset in the RTC 12 with the error correction function.

Next, a method of calculating the correction value of the RTC 12 with the error correction function will be specifically described. Here, the adjustment target oscillation frequency ft of the RTC 12 with the error correction function is set to 32.
The measurement time is 10 minutes at 768 kHz. Therefore, 10
The adjustment target count value Ct when measured for a minute is 327
68 x 60 (sec) x 10 (min) = 196608
It becomes 00 count.

If the actual oscillation frequency to be adjusted, that is, the oscillation frequency of the tuning fork type crystal oscillator 11 is fx, then fx
When> ft (when the clock is advanced), the correction value α is defined to be obtained by the following equation. α = (fx−ft + 0.1) / (fx × 3.051 × 10 ⁻⁶ ) ≈ (fx−ft) × 10 + 1 (2)

When the count value obtained by counting the clock output of the RTC 12 with the error correction function by the 32-bit timer 103-2 is Cy for 10 minutes measured using the reference clock CLKref, the equation (2) is rewritten as the following equation. be able to. α = (Cy−Ct) / 600 × 10 + 1 (3)

On the other hand, when fx <ft (when the clock is delayed), the correction value α is set to be obtained by the following equation. α = (fx−ft) / (fx × 3.051 × 10−6) ≈ (fx−ft) × 10 = (Cy−Ct) / 600 × 10 (4)

The processor core 101 converts the thus calculated correction value α into a 7-bit signed binary number and sets it in the clock error correction register of the RTC 12 with an error correction function. As described above, the RTC 12 with the error correction function
Normally counts up for 1 second for each 32768 pulses of the clock generated by the tuning fork crystal unit 11, but the clock error correction circuit (not shown) of the RTC 12 with error correction function has a second digit of 00 seconds, 20 seconds. , 40 seconds, the count value of 1 second is changed according to the value of the clock error correction register.

The correction value α in binary notation is set to F6, F5, F4,
If F3, F2, F1, and F0, the most significant bit F6 is 0
Then, ((F5, F4, F3, F2, F1, F0) -1) x
The count value increases by 2 and the clock is delayed. For example,
(F6, F5, F4, F3, F2, F1, F0) = (0, 0,
0, 1, 0:00), the count value is 32768+
(8-1) × 2 = 32782. The clock error correction circuit normally counts up for 1 second for each 32768 pulses of the clock, but once for 20 seconds, it counts up for 32782 pulses for 1 second.

When the most significant bit F6 is 1, ((bar F
5, bar F4, bar F3, bar F2, bar F1, bar F0) +
1) The count value decreases by 2 and the clock advances. For example, (F6, F5, F4, F3, F2, F1, F0) = (1,
1,1,1,1,10), the count value is 3276
8 + (− 2) × 2 = 32764. The clock error correction circuit counts up for 1 second with 32764 pulses once every 20 seconds.

When Cy = Ct, the processor core 101 sets any one of 0, +1, -64, and -63 in the clock error correction register of the RTC 12 with an error correction function. At this time, no correction is made.

The allowable value range of the correction value α is -62≤α≤6.
3 is set in advance. Accordingly, the allowable value of the difference (Cy−Ct) between the adjustment target count value Ct and the measurement result count value Cy can be expressed as follows.

The maximum correction value α is when fx> ft (0 <α, when the clock is advanced). Formula (3)
The following equation is obtained from (Cy−Ct) = (α−1) × Tm / 10 (5) The maximum value of the difference (Cy−Ct) is (Cy−Ct) max, and the maximum value of the correction value α is αmax. The following equation is obtained from (5). (Cy-Ct) max = (αmax-1) × Tm / 10 = (63-1) × Tm / 10 = 62 × Tm / 10 (6)

The minimum correction value α is fx <
This is the case of ft (α <0, clock is behind).
The following equation is obtained from the equation (4). (Cy−Ct) = α × Tm / 10 (7) If the minimum value of the difference (Cy−Ct) is (Cy−Ct) min and the minimum value of the correction value α is αmin, then from equation (7) The following equation is obtained. (Cy−Ct) min = αmin × Tm / 10 = −62 × Tm / 10 (8)

From the equations (6) and (8), the difference (Cy-C
The allowable range of t) is as follows. −62 × Tm / 10 ≦ (Cy−Ct) ≦ 62 × Tm / 10 (9)

With respect to the correct pulse output from the reference clock CLKref, the following error occurs due to the processing time from input of the output clock of the RTC 12 with error correction function to the CPU 16 → interrupt generation → 32-bit timer reading. (D) C of the output clock of the RTC 12 with error correction function
Delay of circuit from input to PU16 to generation of interrupt. The delay in the input photocoupler and RC filter is considered to be constant for each pulse and can be ignored. Since the clock of the state change detection circuit (16 kHz-50 / + 25%) of the digital input is used, the delay varies from 0 to 125 μs. As a result, the maximum measurement error is 125 μsec.

(E) S / W processing time from interrupt generation to 32-bit timer reading. Since it is read in the interrupt processing, it is much smaller than (d) and can be ignored. (F) 32-bit timer resolution. 32.768 kHz
In the case of, it is 30.5 μsec. Due to the above (d), (e), and (f), the measurement error is 300 μsec with a margin. If the measurement time is 10 minutes, the error is 300 (μsec) / 600 (sec) =
It becomes 0.5 ppm, which is within the allowable range as compared with the correction resolution of 3 ppm.

As described above, the RTC 12 with the error correction function is time-corrected at the time of factory shipment (initial time). Next, the time correction after shipment from the factory will be described. A counter using an original oscillation with a small temperature deviation is run for 24 hours. Since the count value of the 12-minute original oscillation measured using the reference clock CLKref is stored in the memory 15, the 24-hour original oscillation count value and the memory 15 measured using the output clock of the RTC 12 with error correction function are stored. RT with error correction function by comparing with the count value stored in
The error of C12 can be calculated.

When a 32-bit timer is used as the counter as in the case of factory shipment, the clock source cannot be slower than 2 MHz, so that the count value for one hour is, for example, 2000000 × 60 (sec) × 60 (m).
in) = 72000000, and the 32-bit counter overflows, so S / W processing is required for each carry of 32 bits. Therefore, here, the general-purpose timer 104, which can use a clock source having a large frequency division ratio from the original oscillation, is used as a counter (measurement means during actual use).

The clock signal PITRTCLK, which is the clock source of the general-purpose timer 104, is the original oscillation divided by 4, that is, the system clock divided by 32. Therefore, the frequency of the clock signal PITRTCLK is 1
MHz. The general-purpose timer 104 uses the clock signal PI
Since the clock obtained by dividing TRTCLK by 9600 is counted, the general-purpose timer 104 uses 1 / {1 (MHz) / 96
00} = 1 count up every 9.6 msec.

The error due to the resolution of the general-purpose timer 104 for 24 hours is 9.6 (msec) / 24 (H) =
It is 0.111 ppm and there is no problem. Compared to this error, the error due to the S / W processing from the interrupt of the RTC 12 with the error correction function to the reading of the general-purpose timer is sufficiently small and is ignored.

The clock signal PITRTCLK is set to 9 for 24 hours measured by the RTC 12 with the error correction function.
Assuming that the count value obtained by counting the clock divided by 600 by the general-purpose timer 104 is Cr, it is 32 during the same 24 hours.
The count value Cz obtained by counting the clock CLKx by the bit timer 103-1 is as follows. Cz = Cr ×
f _CPM × 9600 ・ ・ ・
(10)

The measurement time at the initial stage is 12 minutes (more precisely, M seconds is converted to 12 minutes), and the measurement time at the time of actual use is 24 hours. It is necessary to convert the count value so that Count value Cx1 stored in memory 15
A count value Cx120 obtained by converting 2 into a value of 24 hours is obtained by the following equation. Cx120 = Cx12 × 5 × 24 (H) (11) Since the count value Cx120 is a value measured by using the reference clock CLKref, it is a true value of the count value Cz.

Therefore, the RTC 12 with the error correction function
The count number e of the clock CLKx corresponding to the error of 24 hours can be obtained from the equations (10) and (11) by the following equation. e = Cx120-Cz = (Cx12 × 5 × 24) - (Cr × f CPM × 9600) = Cx12 × 2 3 × 3 × 5-Cr × f CPM × 2 7 × 3 × 5 2 ··· (12) In Expression (12), when the count number e is positive, the clock is advanced, and when the count number e is negative, the clock is delayed.

The correction value of the RTC 12 with the error correction function is set to ±
If you increase / decrease by 1, a pulse of 32768Hz will be ± every 20 seconds.
Increases or decreases by 2. Therefore, RTC1 with error correction function
RT with error correction function when the correction value of 2 is corrected by ± 1
The correction amount T _ADJ of C12 per 24 hours is as follows. T _ADJ = {2 (pulse) / (32768 (pulse) × 20 (sec))} × 60 (sec) × 60 (min) × 24 (H) = 3 3 × 5/2 9 (sec) ≒ 0. 264 (sec) ... (13)

During the time of the correction amount T _ADJ , the count value Ch when the clock CLKx is counted by the 32-bit timer 103-1 is given by the following equation. _{Ch = T ADJ × f CPM ×} 1000000 = 3 3 × 5/2 9 × f CPM × 1000000 = f CPM × 3 3 × 5 7/2 3 ··· (14)

Therefore, the RTC 12 with the error correction function
The correction amount X of the correction value of is expressed by the following equation from the equations (12) and (14). X = (RTC error count) / (correction value ± 1 count) = e / Ch = (Cx12 × 2 ³ × 3 × 5-Cr × f _CPM × 2 ⁸ × 3 × 5 ² ) ^{_{/ (f CPM × 3 3 ×}} 5 7/2 3) = (Cx12 × 2 3 -Cr × 2 8 × 5) / (f CPM × 3 2 × 5 6/2 3) = (Cx12-Cr × 2 5 × 5) × 2 ⁶ / (f _CPM × 3 ² × 5 ⁶ ) = (Cx12−160 × Cr) × 64 / (f _CPM × 140625) (15)

The processor core 101 calculates the correction amount X by the formula (15) from the count value Cx12 stored in the memory 15 and the measurement result Cr by the general-purpose timer 104, and rounds the correction amount X to the rounded value. The correction value is corrected and reset by adding it to the current correction value of the clock error correction register of the RTC 12 with the error correction function, and the corrected correction value is stored in the memory 15.

However, if the sign of the correction value before and after the change is different, or if the correction value becomes 0 after the correction, an additional ± 1 is added (when the correction amount to be added to the correction value is positive, an additional +1 is added). If the correction amount to be added is negative,
Add 1). As described above, R with error correction function
TC12 is automatically corrected.

In addition, in order to restore the factory settings,
The correction value at the time of factory shipment is left in the memory 15. Since the general-purpose timer 104 does not count during power-off, the counter of the general-purpose timer 104 is invalid until the first correction timing after power-on (correction is performed only 24 hours after power-on).

The error of this correction method can be calculated as follows. (G) Reference clock CLKre used at factory shipment
The error of f itself is 1.0 ppm. (H) 0.278 pp error included in factory settings
m. (I) 0.111 ppm error due to the resolution of the general-purpose timer 104 after factory shipment. (J) Error due to resolution of RTC correction value 1.526 pp
m.

(K) The worst value of 6.0 ppm, which is the difference between the factory shipment setting and the correction after the factory shipment due to the temperature deviation of 3.0 ppm at 10 to 55 ° C. of the original oscillation. (L) Aging change of crystal is 2.0 ppm. Above (g), (h), (i), (j), (k), (l)
The total maximum error is 10.915 ppm, which is within ± 1 second per day (error 11.574 ppm).

In the first and second embodiments, the AT
Instead of directly measuring the oscillation frequency of the crystal unit 13,
Although the clock obtained by dividing the original oscillation of the AT crystal oscillator 13 is counted, the frequency of the original oscillation of the AT crystal oscillator 13 or the frequency of the clock obtained by dividing the original oscillation is measured to determine the error of the RTC with the error correction function. May be requested.

[0059]

According to the present invention, within the operating temperature range, the first
The oscillation frequency of the second oscillation source having a flatter temperature characteristic than that of the oscillation source, the reference clock that is input only at the initial time is measured as a timing source, and the measured oscillation frequency is stored. The real-time measurement procedure for measuring the oscillation frequency of the vibration source of 2 as a real-time clock source and the difference between the oscillation frequency measured in the actual measurement procedure and the oscillation frequency stored in the initial measurement procedure By calculating the correction value that minimizes the clock error and performing the setting procedure to set the correction value in the real-time clock circuit, high-precision time correction can be realized and a high-precision real-time clock is generated. can do. As a result, if the real-time clock circuit and its time correction device are mounted on, for example, an equipment controller, a time lag when performing scheduled operation of equipment is reduced, and a highly reliable system can be provided.

Further, by counting the oscillation of the second vibration source instead of measuring the oscillation frequency of the second vibration source, similarly highly accurate time correction can be realized and a highly accurate real time clock is generated. can do.

Further, by making the measurement time of the actual use measurement procedure longer than the measurement resolution, the error due to the measurement resolution can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of an electronic device according to a first embodiment of the present invention.

FIG. 2 is a diagram showing temperature characteristics of a crystal unit.

FIG. 3 is a block diagram showing a configuration of an electronic device according to a second embodiment of the present invention.

[Explanation of symbols]

1, 11 ... Tuning fork type crystal oscillator, 2, 12 ... RTC with error correction function, 3, 13 ... AT crystal oscillator, 4 ... Counter, 5, 15 ... Memory, 6, 16 ... CPU, 17 ... Digital input circuit , 101 ... Processor core, 102 ... Memory controller, 103-1, 103-2 ... 32-bit timer, 104 ... General-purpose timer, 105 ... Interface, 106 ... Internal bus interface unit, 1
07 ... External bus interface unit.

Claims (6)

[Claims] 1. A correction value is set for a real-time clock circuit that generates a real-time clock with a period of counting a predetermined number of oscillations of a first vibration source as a unit time, and the predetermined number is set according to the correction value. In a real-time clock time correction method for increasing and decreasing to correct the real-time clock, an oscillation frequency of a second vibration source having a flatter temperature characteristic than the first vibration source within an operating temperature range is input only at an initial stage. Initial measurement procedure of measuring a reference clock as a timing source and storing the measured oscillation frequency, and an actual use measurement procedure of measuring the oscillation frequency of the second vibration source as the real-time clock as a timing source. And the difference between the oscillation frequency measured in the actual use measurement procedure and the oscillation frequency stored in the initial measurement procedure, the real time A time correction method for a real-time clock, wherein the correction value that minimizes the clock error is calculated, and a setting procedure for setting the correction value in the real-time clock circuit is executed. 2. A correction value is set for a real-time clock circuit that generates a real-time clock with a period of counting a predetermined number of oscillations of the first vibration source as a unit time, and the predetermined number is set according to the correction value. In a time correction method of a real-time clock that increases or decreases to correct the real-time clock, the oscillation of a second vibration source having a flatter temperature characteristic than the first vibration source within an operating temperature range is input only at an initial stage. An initial measurement procedure for counting a reference clock as a time source during a predetermined measurement time and storing the measured count value, and an oscillation of the second vibration source for a predetermined measurement time with the real-time clock as a time source. Medium-count actual measurement procedure, count value measured in the actual measurement procedure, and count value stored in the initial measurement procedure And a setting procedure for setting the correction value in the real-time clock circuit so that the correction value that minimizes the error in the real-time clock is executed, and a time correction method for the real-time clock. 3. The time correction method for a real-time clock according to claim 1, wherein the measurement time of the actual-use measurement procedure is set larger than the resolution of measurement. 4. A real-time clock is generated with a period in which a predetermined number of oscillations of the first vibration source are counted as a unit time, and the real-time clock is corrected by increasing or decreasing the predetermined number according to a set correction value. A real-time clock circuit with an error correction function, a second vibration source having a flatter temperature characteristic than the first vibration source within an operating temperature range, and an oscillation frequency of the second vibration source are input only at an initial stage. Initial measurement means for measuring a reference clock as a time source, storage means for storing the oscillation frequency measured by the initial measurement means, and the oscillation frequency of the second vibration source for measuring the real time clock. The actual-use measuring means for measuring as the source of the rear-use and the difference between the oscillation frequency measured by the actual-use measuring means and the stored oscillation frequency, It calculates the correction value error is minimal time clock, the time correction apparatus of the real-time clock; and a setting means for setting the correction value to the error correcting function real-time clock circuit. 5. A real-time clock is generated with a period in which a predetermined number of oscillations of the first vibration source are counted as a unit time, and the real-time clock is corrected by increasing or decreasing the predetermined number according to a set correction value. A real-time clock circuit with an error correction function, a second vibration source having a flatter temperature characteristic than the first vibration source within an operating temperature range, and an oscillation of the second vibration source are input only at an initial stage. An initial measurement means for counting a reference clock as a time source during a predetermined measurement time, a storage means for storing a count value measured by the initial measurement means, and an oscillation of the second vibration source for the real time An actual-use measuring means that counts for a predetermined measuring time using a clock as a time source, and a count value measured by the actual-use measuring means and the stored value. The difference between the real-time clock and the error value of the real-time clock is calculated from the difference between the real-time clock circuit and the real-time clock circuit having the error correction function. Time correction device. 6. The time correction device for a real-time clock according to claim 4 or 5, characterized in that the measurement time of said actual-use measuring means is made larger than the resolution of measurement.