JP4780098B2 - Microcomputer - Google Patents

Description

  The present invention relates to a microcomputer including a CPU configured to be able to set a low power consumption mode in which an operation is temporarily stopped while maintaining an internal state.

In a microcomputer, when the CPU does not need to perform processing, the microcomputer shifts to a low power consumption mode that reduces power consumption by stopping the oscillation operation of the clock signal while maintaining the internal state. Is configured to be possible (for example, Patent Document 1).
Further, the microcomputer controls the external device connected to the terminal by changing the level of the external signal output terminal. In this case, the output level of the terminal and the period during which the terminal is active are set by the program.

Especially in microcomputers that use a battery as the power supply for operation, when the CPU does not need to perform any processing, power consumption is reduced by stopping the oscillation of the clock signal while maintaining its internal state. Some of them are configured to be able to shift to a low power consumption mode. In that case, as for the internal processing of the microcomputer and the control of the external device, for example, as shown in FIG. 37, the low power consumption mode and the normal mode are alternately shifted so as to be periodically executed. It is possible.
Alternatively, if it is necessary to perform processing irregularly in response to some event that occurs outside the microcomputer, the event can be used as a factor to cancel the low power consumption mode, so that the normal mode can be entered when the event occurs. It is also possible to execute processing.
JP-A-11-305888

  However, in the conventional microcomputer as disclosed in Patent Document 1, if external noise is applied during the transition to the low power consumption mode, the output level of the external output terminal may change. For example, if the microcomputer is configured to control driving the motor, the level of the external output terminal is controlled so as to stop the rotation of the motor when shifting to the low power consumption mode. In this state, if the output level of the external output terminal changes due to the influence of external noise, the motor will rotate.

For example, when the program is executed in the normal mode, the level of the external output terminal is set in the program (for example, H), the level is held during the output period, and when the output period ends, the level is Set the level back (for example, L). And in order to grasp | ascertain the said output period, a program uses the timer which operate | moves based on the operation clock (machine clock) etc. of CPU.
Further, the program can execute other processes in parallel during the output period, but such a process that can be executed in parallel does not always exist. In that case, as a result, the CPU is activated only to hold the output level of the terminal. In addition, since the operation clock continues to oscillate during the period, there is a problem that current consumption increases.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a microcomputer capable of reducing the influence of external noise as much as possible when the low power consumption mode is set. An object of the present invention is to provide a microcomputer capable of setting the level of an external output terminal even in a state of shifting to a power consumption mode.

According to the microcomputer of the first aspect , the signal level changing means changes the level of the external signal output terminal within a period in which the CPU is set in the low power consumption mode. Therefore, since the CPU does not need to continue the normal operation mode only for changing the level of the external signal output terminal, the power consumption can be reduced.

According to the microcomputer of the first aspect , the signal level changing means compares the count value of the level change timer with the value of the register in which the level change timing is set, and the two match each other. When the trigger signal is output, the level of the external signal output terminal is changed. Therefore, the level of the output terminal can be changed at the timing set in the register.

According to the microcomputer of the second aspect , since the register is configured such that an arbitrary data value can be set by the CPU, the level change timing can be arbitrarily set by the user program.

According to the microcomputer of the third aspect, the level of the external signal output terminal can be reset by the CPU when the low power consumption mode is canceled, so that the time point when the CPU performs necessary processing. The output terminal level can be reset immediately.

According to the microcomputer of claim 4, wherein the interrupt generation means, after the signal level change means has changed the level of the external signal output terminal, the low power consumption mode is canceled by a cancellation factor, releasing timer To generate an interrupt request to the CPU at the timing when the low power consumption mode is scheduled to be released.

That is, the level of the external signal output terminal changed by the signal level changing means may be required to maintain the level to a minimum (for example, for the external device to which the external signal output terminal is connected). Depending on control requirements). In that case, since the maintenance period of the level is determined according to the relationship with the period in which the low power consumption mode is set, when the low power consumption mode is released earlier than planned due to the occurrence of the release factor The CPU cannot know the original signal level maintaining period.
Therefore, if the interrupt generation means generates an interrupt request to the CPU at the above timing, the CPU can recognize the sustain period. After recognizing the interrupt request, the level of the output terminal is set as necessary. Can be reset.

According to the microcomputer of the fifth aspect , the level change timer can be configured in a simpler manner by sharing the level change timer with the release timer.
According to the microcomputer of the sixth aspect of the present invention , the release timer can be more easily used in common with a watchdog timer for monitoring the operation of the CPU or a free-run timer that counts independently of the CPU operation clock. Can be configured.

(First embodiment)
A first embodiment when the present invention is applied to a microcomputer will be described below with reference to FIGS. FIG. 1 is a functional block diagram showing an electrical configuration of a microcomputer 1. The microcomputer 1 is configured around a CPU 2, and a main clock signal MCLK (for example, a frequency of 4 MHz, a first clock signal) oscillated and output by a main clock oscillation unit (oscillation circuit) 3 is supplied to the CPU 2. It has come to be supplied as. Although not specifically shown, the main clock oscillator 3 is configured using a crystal oscillator.

In addition, the microcomputer 1 includes a sub clock oscillation unit (CR oscillation circuit) 4 configured by a CR oscillation circuit. A sub clock signal SCLK (for example, frequency 25 kHz, second clock signal) oscillated and output by the sub clock oscillating unit 4 is supplied to the watch dog timer 5 as a count clock signal.
As is well known, the watchdog timer 5 has a function of monitoring the runaway of the CPU 2, and overflows and outputs a reset signal to the CPU 2 when the periodic clear operation is not executed by the CPU 2. Further, the watchdog timer 5 in this embodiment is configured to output a reset signal even when the count value of the timer 5 does not exceed the site of the painter when the clear operation image is performed.

  The sub clock correction unit (correction circuit) 6 is configured to correct the oscillation frequency of the sub clock signal SCLK using the main clock signal MCLK. On the other hand, the main clock monitoring unit (monitoring circuit) 7 is configured to monitor the main clock signal MCLK using the sub clock signal SCLK. When an abnormality in the main clock signal MCLK is detected, a reset signal is sent to the CPU 2. It is designed to output.

  FIG. 2 shows a detailed configuration centering on the sub-clock correction unit 6. The sub clock correction unit 6 includes an edge detection circuit 8, a pulse counter 9, a count number setting register 10, a comparison adjustment unit 11, and a resistance adjustment circuit 12. The sub clock oscillating unit 4 includes a ladder resistor 13 as a variable resistor, a capacitor 14, an inverter 15, and a buffer 16 whose resistance value is variable by a control signal from the resistance adjusting circuit 6.

  The edge detection circuit 8 of the sub clock correction unit 6 detects the edge of the oscillation output from the sub clock oscillation unit 4 and generates a detection signal. The pulse counter 9 is a digital counter that counts output pulses of the main clock signal MCLK. The count number setting register 10 is a register that stores an appropriate count number of MCLK pulses that should correspond to one cycle of the sub clock signal SCLK. That is, the count number set in the count number setting register 10 is set to a numerical value for adjusting the oscillation period of the sub clock signal SCLK according to the time during which the MCLK pulse is generated by the count number.

When receiving the edge detection signal from the edge detection circuit 8, the comparison / adjustment means 11 compares the pulse count by the pulse counter 9 with the count stored in the register 4, and based on the comparison result of both counts, An adjustment signal for adjusting the oscillation period of the sub clock signal SCLK is output.
The resistance adjustment circuit 12 is a circuit that generates a control signal based on the adjustment signal from the comparison adjustment unit 11 and adjusts the resistance value of the ladder resistor 13 of the sub-clock oscillator 4 based on the control signal.

  The operation of the subclock correction unit 6 is disclosed in detail in Japanese Patent Application Laid-Open No. 2001-111389, and an outline thereof will be described below. When one cycle of the sub clock signal SCLK elapses and the edge of the clock pulse is detected by the edge detection circuit 8, the detection signal is output to the comparison adjustment unit 11. Then, the comparison / adjustment unit 11 reads the pulse count number accumulated from the previous detection signal accumulated by the pulse counter 9 and the count number set in the count number setting register 10 and compares them. .

  Here, the count number set in the register 10 is an appropriate integrated value of pulses that should correspond to one cycle of the sub clock signal SCLK. Therefore, the comparison / adjustment unit 11 determines how much the one period of the subclock signal SCLK is larger or smaller than an appropriate set value based on how many MCLK pulses are accumulated in one period of the subclock signal SCLK. can do. The comparison adjustment unit 11 generates an adjustment signal for adjusting the oscillation period of the sub clock signal SCLK based on the determination result, and the resistance adjustment circuit 12 generates and outputs a control signal based on the adjustment signal. The resistance value of the ladder resistor 13 is adjusted. FIG. 3 is a flowchart showing the above correction processing.

FIG. 4 shows a detailed configuration of the main clock monitoring unit 7. The main clock monitoring unit 7 is configured around an M counter 17 and an S counter 18. The M counter 17 is a counter that counts the pulses of the main clock signal MCLK supplied via the 1 / N frequency divider 19, and the S counter 18 is a sub clock signal that is supplied via the 1 / N frequency divider 20. It is a counter that counts SCLK pulses. The S counter 18 is reset when the MSB of the count data in the M counter 17 rises, and outputs a reset signal to the CPU 2 when it overflows without being reset.
The frequency division values of the 1 / N frequency dividers 19 and 20 can be set by a program of the CPU 2. Further, the frequency dividers 19 and 20 are configured to output the clock signal as it is without frequency division.

  That is, in the main clock monitoring unit 7, if the frequency of the main clock signal MCLK is appropriate, the MSB of the count data in the M counter 17 is set before the S counter 18 that counts the output pulses of the sub clock signal SCLK overflows. Thus, the S counter 18 is set to be reset. When the frequency of the main clock signal MCLK decreases, the CPU 2 is reset because the S counter 18 overflows before the MSB of the M counter 17 rises.

Further, the overflow cycle of the S counter 18 in the main clock monitoring unit 7 is configured to be shorter than the overflow cycle of the watchdog timer 5.
The main clock oscillating unit 3, the sub clock oscillating unit 4, the sub clock correcting unit 6, and the main clock monitoring unit 7 constitute a clock control circuit device 21. The microcomputer 1 includes other peripheral circuits such as a ROM, a RAM, an I / O, and an A / D converter, although not specifically shown.

Then, by combining the functions of the sub-clock correction unit 6, the main clock monitoring unit 7 and the watchdog timer 5 of the clock control circuit device 21, the following monitoring mode is taken comprehensively.
(1) MCLK: Normal, SCLK: Frequency reduction If the correction is made within the range of the subclock correction unit 6 and the correction range is exceeded, the watchdog timer 5 resets the CPU 2.
(2) MCLK: Normal, SCLK: Increase in frequency When the correction is made within the range of the sub clock correction unit 6 and exceeds the correction range, the main clock monitoring unit 7 resets the CPU 2.
(3) MCLK: Frequency reduction, SCLK: Normal Since the SCLK becomes relatively faster, the main clock monitoring unit 7 resets the CPU 2.
(4) MCLK: frequency increase, SCLK: normal Since the SCLK becomes relatively slow, the watchdog timer 5 resets the CPU 2.

  According to the present embodiment configured as described above, the sub clock correction unit 6 of the clock control circuit device 21 corrects the oscillation frequency of the sub clock signal SCLK based on the main clock signal MCLK, and the main clock monitoring unit 7. Monitors the oscillation state of the main clock signal MCLK based on the sub clock signal SCLK. Therefore, since the correction and monitoring are performed by the two main and sub clock signals, the reliability of the oscillation operation can be improved comprehensively.

Further, according to the present embodiment, the cycle of monitoring the main clock signal MCLK by the main clock monitoring unit 7 is set to be shorter than the cycle of the watchdog timer 5. Therefore, the main clock monitoring unit 7 can detect the occurrence of an abnormality that cannot be detected in the monitoring period of the watchdog timer 5. That is, when the main clock signal MCLK becomes late, the main clock monitoring unit 7 can reset the clock before the reset by the watchdog timer 5 is performed.
Then, by correcting the sub clock signal SCLK, the CPU 2 can reset the watch dog timer 5 as designed.

<The following is for reference>
A. An oscillation circuit for oscillating the first clock signal using an oscillator;
A CR oscillation circuit configured to oscillate a second clock signal having a frequency lower than that of the first clock signal and to adjust the oscillation frequency;
A correction circuit for correcting an oscillation frequency of the second clock signal based on the first clock signal;
A clock control circuit device comprising: a monitoring circuit that monitors an oscillation state of the first clock signal based on the second clock signal.
According to the clock control circuit device of A, the correction circuit corrects the oscillation frequency of the second clock signal based on the first clock signal, and the monitoring circuit oscillates the first clock signal based on the second clock signal. Monitor status. That is, since correction and monitoring are performed with the first and second clock signals, the reliability of the oscillation operation can be improved.

B. A clock control circuit device according to A above;
A CPU to which an operation clock signal generated based on the first clock signal is supplied;
A watchdog timer for monitoring the operation of the CPU based on the second clock signal,
The microcomputer is configured such that a period of monitoring the first clock signal by the monitoring circuit is shorter than the period of the watchdog timer.
According to the microcomputer B, the cycle for monitoring the first clock signal by the monitoring circuit is set to be shorter than the watchdog timer cycle. Therefore, the monitoring circuit can detect the occurrence of an abnormality that cannot be detected in the monitoring period of the watchdog timer.

(Second embodiment)
FIGS. 5 to 8 show a second embodiment of the present invention. The same parts as those in the first embodiment are designated by the same reference numerals, and the description thereof is omitted. Only different parts will be described below. The configuration of the second embodiment is basically the same as that of the first embodiment, and the adjustment operation by the comparison adjustment means 11 is slightly different.
That is, in the flowchart shown in FIG. 5, the processing step S11 of “set maximum value in CV” is inserted between steps S3 and S4. Steps S7 and S8 are replaced with steps S12 and S13. Here, “CV” is adjustment signal data output from the comparison adjustment unit 11.
That is, the execution of the flowchart shown in FIG. 5 is started (start) when the CPU 2 is reset or when the low power consumption mode such as the sleep mode is released. In that case, the initial value of the correction data is set to the maximum value (max) in step S11.

  FIG. 6 shows the processing contents in step S12. First, the comparison / adjustment unit 11 determines whether to perform high-speed correction (step A0). This setting is performed by the user, and is set, for example, in the user program or with a dip switch provided on the board. If it is not set to perform high-speed correction (“NO”), the comparison adjustment unit 11 sets the adjustment signal data CV to the minimum unit (LSB) of the correction data (step A4), and then proceeds to step A3. . In this case, since the correction is performed in the minimum unit, the low-speed correction is the same as in the first embodiment. In step A3, the adjustment signal data CV is output to the resistance adjustment circuit 12, thereby increasing the resistance value of the ladder resistor 13 in accordance with the data CV.

On the other hand, if it is set in step A0 to perform high-speed correction (“YES”), the comparison / adjustment unit 11 determines whether or not the adjustment signal data CV is the smallest unit at that time (step). A1). If the data CV is equal to the minimum unit (“YES”), the process proceeds to step A3.
In step A1, if the data CV is not equal to the minimum unit (“NO”), the comparison / adjustment unit 11 halves the adjustment signal data CV (step A2) and proceeds to step A3. In the process in step S13 shown in FIG. 7, the process in step A3 ′ is obtained by changing “increase” in step A3 to “decrease”.

Here, FIG. 8 shows a conceptual explanatory diagram of a state in which the high-speed correction operation is performed (that is, when step A3 is executed for step S12). That is, the first correction is performed based on the maximum value max of the adjustment range. For example, when the period data of the sub clock signal SCLK is larger than the set value of the register 10 (frequency is low), the maximum value is set in the direction in which the period becomes shorter. Shake at max.
In the next correction, since the cycle is shortened, it is shaken at a half of the maximum value max in the direction in which the cycle becomes longer. Further, in the next correction, since the cycle is still short, it is shaken at max / 4 in a direction in which the cycle becomes longer. Then, since the target value has been exceeded this time, the next correction is performed with max / 8 in a direction in which the cycle becomes longer.

As described above, correction is started at the maximum value max, and correction is performed so that the absolute value gradually decreases to 1/2, 1/4, and 1/8. That is, in terms of the data search method, so-called binary search is the same concept. That is, when the CPU 2 is released from reset or started up such as when the low power consumption mode is shifted to the normal mode, the oscillation state of the sub-clock oscillation unit 4 becomes considerably unstable, and the oscillation frequency is determined from the set frequency. It is estimated that it is in a greatly deviated state. Therefore, the adjustment operation in that period is performed by the binary search method, so that the periodic data of the sub clock signal SCLK is corrected so as to be equal to the set value of the register 10 faster (to converge to the target frequency).
When the adjustment signal data CV is repeatedly halved, the data reaches the minimum value (LSB) (step A1, “YES”). Thereafter, the correction is continued with the minimum value.

  In step A0, it is possible to select whether to perform high-speed correction for the following reason. That is, when the CPU 2 is configured so as to be able to execute the low power consumption mode, the state in which the oscillation frequency of the sub-clock oscillator 4 is far away from the set frequency when the CPU 2 shifts to the low power consumption mode during high-speed correction. There is a risk of migration. Therefore, by selecting the adjustment method according to the setting of the transition period to the low power consumption mode, the adjustment according to the design intention of the user can be performed.

As described above, according to the second embodiment, the comparison / adjustment unit 11 of the sub-clock correction unit 6 adjusts the oscillation frequency of the sub-clock oscillation unit 4 immediately after the CPU 2 is started. The absolute value was decreased by ½ every adjustment cycle. Therefore, the frequency of the sub clock signal SCLK can be converged faster than the set frequency.
Then, when the absolute value of the adjustment signal level reaches the minimum value (LSB), the comparison / adjustment unit 11 continues the adjustment with the minimum value, so that the adjustment is made when the oscillation frequency approaches the set frequency. Can be performed precisely. Further, the comparison adjustment unit 11 can select whether or not to perform high-speed correction in the adjustment operation immediately after the CPU 2 is activated, and thus can be adjusted according to the user's intention.

<The following is for reference>
C. An oscillation circuit for oscillating the first clock signal using an oscillator;
A CPU to which an operation clock signal generated based on the first clock signal is supplied;
A CR oscillation circuit configured to oscillate a second clock signal having a frequency lower than that of the first clock signal and to adjust the oscillation frequency;
A correction circuit for correcting an oscillation frequency of the second clock signal based on the first clock signal;
The correction circuit includes an edge detection circuit that detects an edge of the second clock signal, a pulse counter that performs a counting operation based on the first clock signal, and the pulse that should correspond to one cycle of the second clock signal. When a count number setting register for storing the count number of the counter and an edge detection signal is received from the edge detection circuit, the count number of the pulse counter is compared with the count number stored in the count number setting register. Comparison adjusting means for generating an adjustment signal for adjusting the oscillation frequency of the CR oscillation circuit based on the difference between the two, and an adjustment circuit for adjusting the oscillation frequency of the CR oscillation circuit based on the adjustment signal. Configured,
The microcomputer is configured to gradually decrease the absolute value of the adjustment signal level every adjustment cycle in the adjustment operation immediately after the CPU is activated.

  According to the microcomputer C, when the oscillation frequency of the CR oscillation circuit is adjusted immediately after the CPU is activated, the comparison adjustment unit gradually decreases the absolute value of the adjustment signal level every adjustment cycle. In other words, when the CPU is released from reset or started up, such as when shifting from the low power consumption mode to the normal mode, the oscillation state of the CR oscillation circuit becomes considerably unstable, and the oscillation frequency deviates significantly from the set frequency. Estimated to be in the state. Accordingly, the adjustment operation during that period can be converged more quickly with the set frequency if the adjustment signal level is adjusted so that the absolute value of the adjustment signal level is gradually decreased from a large value to a small value every adjustment cycle.

D. The microcomputer according to the above C, wherein, in the adjustment operation, the comparison adjustment means starts from the maximum value of the adjustment range and decreases the absolute value by ½ every adjustment period.
According to the microcomputer D, in the adjustment operation, the comparison adjustment means starts from the maximum value of the adjustment range and decreases the absolute value by ½ every adjustment cycle. That is, by using the same method as the so-called binary search in the data search, the oscillation frequency of the CR oscillation circuit can be converged faster by the set frequency.

E. The microcomputer according to the above C or D, wherein when the absolute value of the adjustment signal level reaches a minimum value, the comparison / adjustment means continues adjustment with the minimum value thereafter.
According to the microcomputer E described above, the comparison adjustment means continues the adjustment with the minimum value after the absolute value of the adjustment signal level reaches the minimum value, so that the oscillation frequency approaches the set frequency. If it becomes, it will become possible to adjust precisely.

F. The comparison adjustment means is configured to be able to select whether or not to perform an adjustment operation for gradually decreasing the absolute value of the adjustment signal level every adjustment cycle in the adjustment operation immediately after the CPU is started. The microcomputer according to any one of the above C to E.
According to the microcomputer F, the comparison and adjustment means can select whether or not to perform an adjustment operation for gradually decreasing the absolute value of the adjustment signal level every adjustment cycle in the adjustment operation immediately after the CPU is started. Is done. That is, when the CPU is configured to be able to execute the low power consumption mode, if the CPU shifts to the low power consumption mode during the adjustment operation for gradually decreasing every adjustment cycle, the oscillation frequency of the CR oscillation circuit becomes the set frequency. There is a risk of shifting in a state of being far away. Accordingly, if the adjustment method can be selected according to the setting of the transition period to the low power consumption mode, the adjustment can be performed in accordance with the user's intention.

C '. An oscillation circuit that oscillates a first clock signal using an oscillator, a CPU to which an operation clock signal generated based on the first clock signal is supplied, and a second having a frequency lower than that of the first clock signal A CR oscillation circuit configured to oscillate a clock signal and its oscillation frequency is adjustable, and a correction circuit that corrects the oscillation frequency of the second clock signal based on the first clock signal. An oscillation frequency adjusting method for adjusting an oscillation frequency of the second clock signal for a microcomputer,
The correction circuit detects an edge of the second clock signal, performs a counting operation based on the first clock signal, holds a count number corresponding to one cycle of the second clock signal, and detects the edge When the signal is received, the count number indicated by the count operation is compared with the held count number, and an adjustment signal for adjusting the oscillation frequency of the CR oscillation circuit based on the difference between the two is generated, Based on the adjustment signal, the oscillation frequency of the CR oscillation circuit is adjusted,
In the adjustment operation immediately after the CPU is started, the absolute value of the adjustment signal level is gradually decreased every adjustment cycle.

D '. The method for adjusting the oscillation frequency of the clock signal C ′ described above, wherein the adjustment operation starts from the maximum value of the adjustment range and decreases the absolute value by ½ every adjustment cycle.
E '. When the absolute value of the adjustment signal level reaches a minimum value, the adjustment is continued with the minimum value thereafter, and the oscillation frequency adjustment method for the C ′ or D ′ clock signal is characterized.
F '. In the adjustment operation immediately after the CPU is activated, it is possible to select whether or not to perform an adjustment operation for gradually decreasing the absolute value of the adjustment signal level every adjustment cycle. Signal oscillation frequency adjustment method.

(Third embodiment)
9 to 13 show a third embodiment of the present invention. FIG. 9 is a functional block diagram showing the configuration of the microcomputer 31. The microcomputer 31 is configured around the CPU 32. The CPU 32 is supplied with a clock signal generated and output from the clock generation circuit 33 as an operation clock signal. The clock signal is also supplied to other functional blocks 34 such as peripheral circuits.

The low power consumption control unit (terminal control means) 35 performs control related to a low power consumption mode such as a so-called sleep / stop mode in accordance with a control command given by the CPU 32. Here, the sleep mode stops the operation of the clock generation circuit 33 to stop the supply of the operation clock signal to the CPU 32, and the CPU 32 temporarily performs the processing while holding the data value of the internal register. It is a mode to stop. The sleep mode is periodically canceled (waked up) at set time intervals.
On the other hand, the stop mode is the same as the stop mode in that the CPU 32 stops processing, but is not released periodically and is released only when some event to be processed by the CPU 32 occurs. It is.

FIG. 10 shows a configuration of a functional part in which the low power consumption control unit 35 controls the sleep mode. That is, the low power consumption control unit 35 compares the count value of the restart timer 36 constituted by, for example, a CR oscillation circuit and the set data value of the internal SLEEP time setting register 37 by the coincidence comparator 38, and both are compared. When they match, a high level match signal is output to one input terminal of the AND gate 39. A SLEEP signal is given to the other input terminal of the AND gate 39.
Therefore, when the coincidence comparator 38 outputs a coincidence signal during a period when the high-level SLEEP signal is applied, the AND gate 39 outputs a high-level SLEEP cancellation signal. At this time, the restart timer 36 is reset.

On the other hand, FIG. 11 shows a configuration of a functional part in which the low power consumption control unit 35 controls the I / O terminal unit 40 when the stop mode is set. That is, a STOP signal and a port control signal are applied to the negative logic input terminal and the positive logic input terminal of the AND gate 41 of the low power consumption control unit 35, respectively. The AND gate 41 outputs a high impedance control signal to the I / O terminal unit 40.
The I / O terminal unit 40 has a configuration in which buffers 43A and 43B are bidirectionally connected to an input / output terminal (external signal terminal) 42. The output buffer 43A is controlled by a high impedance control signal output by the AND gate 41. Further, the I / O terminal unit 40 is configured such that input / output switching control (port control) and output level control are performed by the I / O control unit 44. The port control signal is output by the I / O control unit 44.

  Next, the operation of the third embodiment will be described with reference to FIGS. FIG. 12 is a flowchart illustrating processing when the CPU 32 shifts to the low power consumption mode. The CPU 32 determines whether or not the low power consumption mode is “sleep” (step B1). If it is “sleep” (“YES”), the sleep mode transition period data is set in the SLEEP time setting register 37 (step B2). Then, when a SLEEP command is output to the low power consumption control unit 35 (step B3), the process is terminated. In this case, the components shown in FIG. 9 of the low power consumption control unit 35 become active, the measurement of the transition period of the sleep mode is started, and the sleep mode is canceled when the period ends.

  On the other hand, when the low power consumption mode is “stop” (step B1, “NO”), the CPU 32 outputs a STOP command to the low power consumption control unit 35 (step B4). Then, the process ends. In this case, the components shown in FIG. 11 of the low power consumption control unit 35 act. That is, the STOP signal becomes high level, and the output level of the AND gate 41 becomes low, so that the output terminal of the output buffer 43A is in a high impedance state.

  FIG. 13 is a flowchart showing processing when the CPU 32 is activated (waked up) after the low power consumption mode is canceled. The CPU 32 determines whether or not the release is the sleep mode (step C1). If the sleep mode is the sleep mode (“YES”), the direction (input / output) setting value of each terminal at that time, and the output terminal The signal level setting is read from a ROM (not shown) or the like held as data (step C2). Then, by resetting the read setting value and level setting in the I / O control unit 44, the output signal level at each input / output terminal 42 is reset (step C3).

  Then, it is determined whether another event to be processed has occurred at that time (step C4). If it has occurred ("YES"), the event is processed (step C6). Transition processing to the low power consumption mode shown in FIG. 12 is performed (step C5). On the other hand, if an event to be processed has not occurred in step C4 ("NO"), the process proceeds to step C5 as it is.

  As described above, according to the third embodiment, the CPU 32 of the microcomputer 31 sleeps after resetting the signal output level of the input / output terminal 42 every time the sleep mode is periodically canceled by the restart timer 36. Change to mode. Therefore, even if the CPU 32 changes the signal output level of the input / output terminal 42 due to the influence of external noise during the transition to the sleep mode, the output level is regenerated every time the sleep mode is periodically released. Since it is set, the reliability of the microcomputer 31 can be improved.

  Further, the low power consumption control unit 35 of the microcomputer 31 sets the input / output terminal 42 to the high impedance state when the CPU 32 shifts to the low power consumption mode. Therefore, even if the CPU 32 is affected by external noise during the transition to the stop mode as described above, the input / output terminal 42 is in a high-impedance state and is therefore connected to the microcomputer 31 by the terminal 42. The external device is not driven, and the reliability can be improved.

(Fourth embodiment)
14 to 17 show a fourth embodiment of the present invention. FIG. 14 is a block diagram showing functional parts related to clock control of the microcomputer 51. The microcomputer 51 is configured around the CPU 52. The CPU 52 is selectively supplied with an operation clock (machine clock) signal via a selector 53.
The selector 53 is supplied with an original oscillation signal from an external oscillator 54 externally attached to the microcomputer 51 and a multiplied clock signal output by a DPLL (Digital Phase Locked Loop) circuit 55. The selector 53 is switched by a clock control unit (switch notification control unit) 56.

  The reference clock signal of the DPLL circuit (frequency multiplication circuit) 55 is also supplied via the selector 57, and the selector 57 outputs the original oscillation signal from the external oscillator 54 and the CR oscillation circuit 58. CR clock signal to be transmitted. The selector 57 is also switched by the clock control unit 56. For example, the original oscillation signal has a frequency of 4 MHz, and the CR clock signal has a frequency of 25 kHz. Note that the clock control unit 56 is configured to stop the oscillation operation of the external oscillator 54 when the CR clock signal is selected as the clock source of the DPLL circuit 55.

  The CR clock signal is also supplied to the CR counter 59 as a counting clock. The coincidence comparison unit 60 is provided with count data of the CR counter 59 and setting data of the oscillation stabilization time setting register 61. The setting data is initially set by hardware when the power is turned on, but can also be set by a user program executed by the CPU 52. The coincidence comparison unit 60 outputs a coincidence signal to the clock control unit 56 when the two data values are compared and matched.

  That is, if data corresponding to the oscillation stabilization time of the external oscillator 54 is set in the setting register 61, the lapse of the oscillation stabilization time can be determined by the timing at which the coincidence signal is output. The clock control unit 56 oscillates the DPLL circuit 55 by referring to the counter value of the state counter (sequence control means) 62 built in the DPLL circuit 55 and used to control the internal oscillation operation sequence. It is also possible to know the passage of the stabilization time.

  Here, the state counter 62 is the same as the control circuit 20 in the fifth embodiment, for example, and when the output clock of the external oscillator 54 is used as a reference clock signal, its count is one control cycle. It is supposed to be. Therefore, at the start of the multiplication operation, the multiplied clock signal as the first calculation result is output when the one control cycle has elapsed, and this period becomes the oscillation stabilization period.

  Then, the clock controller 56 outputs a clock switching completion interrupt signal to the interrupt controller 63 of the CPU 52 at those detection timings. The interrupt controller 63 performs enable control and mask control including other interrupt factors, and outputs an appropriate interrupt signal to the CPU 52.

  Next, the operation of the fourth embodiment will be described with reference to FIGS. FIG. 15 shows the transition state of the clock signal when the power is turned on or when the CPU 52 is activated from the sleep / stop mode. That is, when the mode transits to the sleep / stop mode, the clock signal oscillation operation is stopped except for the CR oscillation circuit 58. First, the external oscillator 54 starts an oscillation operation when the power is turned on. Since the oscillation operation, that is, the oscillation frequency is unstable at the beginning of oscillation, the CPU 52 prohibits the operation of the DPLL circuit 55.

When the counter value of the CR counter 59 matches the set data of the oscillation stabilization time setting register 61 and the match comparison unit 60 outputs a match signal to the clock control unit 56, the clock control unit 56 receives the interrupt signal Is output to the CPU 52 via the interrupt controller 63. Then, the CPU 52 recognizes that the oscillation operation of the external oscillator 54 is stabilized by the interrupt signal, and causes the DPLL circuit 55 to start the frequency multiplication operation.
The clock control unit 56 then refers to the counter value of the state counter 62 of the DPLL circuit 55, waits until the operation of the DPLL circuit 55 becomes stable, and outputs an interrupt signal when the operation becomes stable. To do. Therefore, the CPU 52 can recognize that the operation of the DPLL circuit 55 is stabilized by the interruption.

FIG. 16 is a diagram when the machine clock is switched by the user program in a state where the CPU 52 performs a normal operation. In this case, there are three types of clock sources: an external oscillator 54, a DPLL circuit 55 (the external oscillator 54 is a source), and a DPLL circuit 55 (a CR oscillation circuit 58 is a source). Switching is performed. In addition, since the DPLL circuit 55 can multiply or divide the frequency of the clock source by the CPU 52 setting the multiplication or frequency division data, the frequency switching is also performed in the DPLL circuit 55 alone. These patterns are shown below.
A: External oscillator → DPLL circuit (external source)
B: External oscillator → DPLL circuit (CR source)
C: DPLL circuit (external source) → DPLL circuit (CR source)
D: DPLL circuit (CR source) → external oscillator E: DPLL circuit (CR source) → DPLL circuit (external source)
X: Frequency change in DPLL circuit

FIG. 17 is a flowchart showing processing when the CPU 52 switches the machine clock signal based on the diagram shown in FIG. First, the CPU 52 selects a clock signal switching pattern from A to E and X (step D1). If the pattern is D or E (step D2, “YES”), the oscillation stabilization time data of the external oscillator 54 is set in the setting register 61 (step D3), and the clock is switched according to the pattern (step D4). ).
The clock switching is performed by switching the selectors 53 and 57 by the CPU 52 writing data in a setting register inside the clock control unit 56. If the switching pattern is other than D or E in Step D2 ("NO"), the process proceeds to Step D4 as it is.

  Subsequently, the CPU 52 determines whether or not the switching pattern is D or E as in step D2 (step D5), and if it is the pattern D or E (“YES”), waits until an external oscillation interrupt occurs (step S5). Step D6). When an interrupt occurs (“YES”), the process proceeds to step D7. In step D5, if the pattern is other than patterns D and E, the process proceeds to step D7.

  In step D7, the CPU 52 determines whether or not it is the switching pattern D. If it is not the pattern D ("NO"), the CPU 52 waits for the occurrence of a PLL interrupt (step D8, "YES") and ends the process. If the pattern is D in step D7 ("YES"), the process is terminated. That is, in the case of the pattern E, the CPU 52 recognizes the clock stabilization by waiting for the generation of both the external oscillation interrupt and the PLL interrupt. In the case of the pattern D, only the external oscillation interrupt is recognized, and in the other patterns, the clock stabilization is recognized by waiting for the occurrence of only the PLL interrupt.

As described above, according to the fourth embodiment, the clock control unit 56 of the microcomputer 51 monitors the oscillation operation of the DPLL circuit 55 by the state counter 62, and when the frequency of the multiplied clock signal is switched, the switching is performed. An interrupt request is generated to the CPU 52 at a timing when the accompanying oscillation operation is stabilized.
Therefore, the CPU 52 does not need to set a redundant waiting time, and the processing efficiency can be improved. The clock control unit 56 can clearly recognize that the oscillation operation of the DPLL circuit 55 is stable by referring to the sequence counter 62.

Further, since the operation clock signal of the CPU 52 can be selected from a plurality of clock signals by the selectors 53 and 57, the control range can be expanded by selecting more various clock signals.
The clock control unit 56 also monitors the oscillation operation of the external oscillator 54 and generates an interrupt request to the CPU 52 even when the oscillation operation associated with the operation switching is stabilized. Therefore, the CPU 52 generates the interrupt request. Thus, it becomes possible to know an appropriate timing for starting the operation of the DPLL circuit 55.

<The following is for reference>
G. A frequency multiplication circuit that multiplies and outputs the frequency of the reference clock signal;
A CPU capable of supplying a multiplied clock signal output by the frequency multiplication circuit as an operation clock signal;
A switching notification control unit that monitors the oscillation operation of the frequency multiplication circuit and, when the frequency of the multiplied clock signal is switched, generates an interrupt request to the CPU at a timing when the oscillation operation associated with the switching is stabilized; A microcomputer comprising:
According to the microcomputer G, the switching notification control unit monitors the oscillation operation of the frequency multiplying circuit, and when the frequency of the multiplied clock signal is switched, the switching notification control unit notifies the CPU at a timing when the oscillation operation accompanying the switching is stable. An interrupt request is generated. Therefore, when the frequency of the multiplied clock signal is switched, the CPU can know that the oscillation operation is stabilized by an interrupt, so that it is not necessary to set a redundant standby time, and the processing efficiency can be improved. it can.

H. The frequency multiplication circuit includes sequence control means for outputting a control signal for controlling an oscillation operation sequence based on the reference clock signal,
The G microcomputer according to claim 1, wherein the switching notification control unit monitors an oscillation operation of the frequency multiplication circuit based on the control signal.
According to the microcomputer of H, the switching notification control unit monitors the oscillation operation of the frequency multiplication circuit based on the control signal output by the sequence control means that controls the oscillation operation sequence of the frequency multiplication circuit. It can be clearly recognized by the control signal that the oscillation operation is stable.

I. There are one or more clock signals that can be supplied as the operation clock signal in addition to the multiplied clock signal.
The G or H microcomputer described above, wherein the operation clock signal can be selected from the plurality of clock signals.
According to the microcomputer I, the operation clock signal of the CPU is configured to be selectable from a plurality of clock signals including the multiplied clock signal. Can be obtained.

J. et al. The switching notification control unit also monitors the oscillation operation of an external oscillator that outputs a clock signal that can be supplied as the operation clock signal, and issues an interrupt request to the CPU even when the oscillation operation is stabilized due to the operation switching. A microcomputer according to the above I, characterized in that:
According to the microcomputer J, the switching notification control unit also monitors the oscillation operation of the external oscillator, and generates an interrupt request to the CPU even when the oscillation operation associated with the operation switching is stabilized. Therefore, when the clock signal output from the external oscillator becomes the reference clock signal of the frequency multiplication circuit, the CPU can know the appropriate timing for starting the operation of the frequency multiplication circuit by the generation of the interrupt request. It becomes.

(5th Example)
18 to 23 show a fifth embodiment when the present invention is applied to a vehicle ECU (Electronic Control Unit). The basic portion of the configuration described below is disclosed in Japanese Patent Application Laid-Open No. 2000-357947, and will be briefly described below. FIG. 18 is a functional block diagram showing an electrical configuration of the ECU (microcomputer) 71. The ECU 71 is configured as a semiconductor integrated circuit (IC), and has a clock synchronization circuit such as a CPU 72, a memory 73, and a gate array 74 as its internal circuit. The ECU 71 is externally attached to a crystal oscillator 75, and an oscillation circuit (reference clock oscillation circuit) 76 applies a bias to the crystal oscillator 75 and outputs a reference clock signal PREF having a frequency of 16 KHz. .

The reference clock signal PREF is supplied to a frequency multiplying circuit 77. The frequency multiplying circuit 77 generates a frequency 8 MHz multiplied clock signal POUT obtained by multiplying the reference clock signal PREF by 512, and the CPU 72, the memory 73, and the gate array 74. Output to the clock input terminal. The frequency multiplication circuit 77 is configured as a so-called DPLL (Digital Phase Locked Loop) circuit, and the multiplication number is set according to the value of the multiplication number setting data DV given from the CPU 72.
Further, the frequency multiplication circuit 77 outputs a mode control signal PA (stop control) for switching the operation mode of the ECU 71 between the low power consumption mode and the standby mode by a low power consumption control circuit (low power consumption control means) 78. It has come to be.

The key detection switch 79 is a switch that detects whether or not a car key is inserted into a key cylinder (none of which is shown). If the key detection switch 79 does not output a key detection signal to the low power consumption control circuit 78, the low power consumption control circuit 78 sets the mode control signal PA to a low level to cause the ECU 71 to enter the low power consumption mode (sleep mode). / Stop).
When the key detection switch 79 outputs a key detection signal, the low power consumption control circuit 78 sets the mode control signal PA to a high level and switches the ECU 71 from the low power consumption mode to the standby mode. The frequency multiplication circuit 77 and the low power consumption control circuit 78 constitute a clock control circuit (oscillation circuit device) 71a.

FIG. 19 is a functional block diagram showing a schematic configuration of the frequency multiplying circuit 77 (for further details, see Japanese Patent Laid-Open No. 8-265111). The control circuit (sequence control means) 80 is supplied with a reference clock signal PREF output from the oscillation circuit 76. The control circuit 80 has a built-in sequence counter composed of three flip-flops (not shown).
The sequence counter counts the number of input pulses of the reference clock signal PREF, and eight cycles of the reference clock signal PREF are set as one sequence control cycle. Various control timing signals are synchronized with the reference clock signal PREF by various control timing signals DCO (Digital Controlled). (Oscillator, digitally controlled oscillator) 81 and counter / data latch circuit 82.

  The DCO 81 includes a ring oscillator (multiphase clock signal output means) 83 inside. As shown in FIG. 21, the ring oscillator 83 includes a NAND gate 84 and a plurality of INV (inverter) gates 85 as a logic inversion circuit. Each of the INV gates 84 is connected in a ring shape to the input terminal of the next stage, and one input terminal of the NAND gate 85 is connected to the output terminal of the final stage INV gate (not shown). The other input terminal is given a stop control signal from the outside. Multiphase clock signals R1 to R16 are output from the output terminals of the INV gate 84, respectively.

  Referring to FIG. 19 again, the counter / data latch circuit 82 is supplied with control timing signals UCE and CLR output from the control circuit 80. These control timing signals UCE and CLR have a pulse width corresponding to one cycle of the reference clock signal PREF, and are signals output in the third and seventh cycles of the sequence control cycle in the control circuit 80, respectively.

The counter / data latch circuit 82 is supplied with the clock signal R13 output from the ring oscillator 83 as RCK, and performs a counting operation by an internal up counter (16 bits) by the clock signal RCK. The counter / data latch circuit 82 counts the time corresponding to one cycle of the reference clock signal PREF by causing the counter to perform an up-count operation while the control timing signal UCE is being output.
The count data is latched when the control timing signal DLS output from the control circuit 80 in the fifth cycle of the sequence control cycle is latched at the timing of the latch signal DLC provided via the DCO 81, and is latched when the control timing signal CLR is output. Data is cleared.

The counter / data latch circuit 82 right-shifts the counted 16-bit data DT16 to DT1 according to the multiplication number setting data DV given by the CPU 72, and latches 12 bits of the shifted data. The latched 12-bit data is output to the DCO 81 as CD12 to CD1. The multiplied clock signal POUT ′ output from the DCO 81 is divided by two through a frequency dividing circuit 86 for adjusting the duty ratio, and is output as the multiplied clock signal POUT.
The mode control signal PA is also supplied to the control circuit 80 and, for example, as an operation start signal PSTB to the control circuit 80 via a delay circuit 87 that gives a delay time of about one cycle of the reference clock signal PREF. Is also given.

  FIG. 20 is a functional block diagram showing a detailed configuration of the DCO 81. Of the latch data CD12 to CD1 supplied from the counter / data latch circuit 82, the upper eight bits CD12 to CD5 are loaded as count data of the down counter 88 at a predetermined timing. The down counter 88 down-counts the count data loaded by the clock signal R13 output from the ring oscillator 83.

  Further, among the latch data CD12 to CD1, the lower four bits CD4 to CD1 are supplied to the data input terminal D of the register 90 through the adder 89. The register 90 outputs the output data of the adder 89 as 5-bit data D5 to D1 according to the timing signal output from the timing control unit 91, and the lower 3 bit data D4 to D1 among them are While being supplied to the pulse selector 92, it is input to the adder 89 as an added value. The data D5 output from the register 90 corresponds to a carry signal generated in response to the addition in the adder 89, and is supplied to the timing control unit 91.

The pulse selector 92 is supplied with the multiphase clock signals R16 to R1 output from the ring oscillator 83, and among the multiphase clock signals R16 to R1, the values of the data D4 to D1 output from the register 90 are provided. One corresponding to (number corresponding to decimal value + 1) is selected and output to the timing control unit 91 from one of the output terminals P1 (R8 to R1) and P2 (R16 to R9). It has become. The timing controller 91 is provided with a clock signal R5 output from the ring oscillator 83.
The down counter 88 counts down the loaded count data. When the count value becomes “2”, the output signal CN2 is set to high level. When the count value becomes “1”, the output signal CN1 is set to high level. And output to the timing control unit 91.

  The outline of the above operation is as follows. That is, every eight cycles of the reference clock signal PREF, the count data DT16 to DT1 corresponding to one cycle of the reference clock signal PREF is counted, and the 12-bit data CD12 to CD1 shifted to the right according to the multiplication data Dv is obtained. It is given to DCO81. When the upper eight bits CD12 to CD5 are down-counted, any one of the multiphase clock signals R16 to R1 corresponding to the value (+1) of the lower data D4 to D1 given from the register 90 is selected and multiplied. It is output as the clock signal POUT '.

  As shown in FIG. 21, an analog switch 93 is arranged in the pulse selector 92 corresponding to each of the multiphase clock signals R1 to R16. Then, any one of the plurality of analog switches 93 is encoded by the encoder 99 based on the control data D1 to D4 and selectively turned on, so that one of the multiphase clock signals R1 to R16 is switched to the timing control unit 91. To output.

FIG. 22 shows a part of the internal configuration of the counter / data latch circuit 82. That is, the control data CD12 to CD1 (only CD4 to CD1 are shown) are held in a double latch configuration by the frequency setting circuit section 94 and the frequency information holding circuit section (data holding means) 95. These circuit units 94 and 95 are constituted by flip-flops 96 and 97, respectively, and control clocks 1 and 2 are given to clock input terminals.
A reset signal is given to the reset terminal of the flip-flop 96 constituting the frequency setting circuit unit 94. On the other hand, a reset signal and a stop control signal in the low power consumption mode are given to the reset terminal of the flip-flop 96 constituting the frequency information holding circuit unit 95 via the OR gate 98.

  The frequency information holding circuit unit 95 is used for resynchronizing the control data 1 to 4 synchronized with the control clock 1 by the frequency setting circuit unit 94 with the control clock 2 during normal operation. The one having the same function is also present in Japanese Patent Laid-Open No. 2000-357947 (however, when shifting to the low power consumption mode, the data is reset and disappears. )

Next, the operation of the fifth embodiment will be described with reference to FIG. FIG. 23 is a timing chart showing a case where the frequency multiplying circuit 77 is activated and performs an oscillation operation from a state where the frequency multiplying circuit 77 is in the low power consumption mode. In the low power consumption mode (1), the oscillation operations of the ring oscillator 83 and the oscillation circuit 76 are stopped. Then, when the low power consumption mode is canceled in (2), they start an oscillation operation.
Further, during the low power consumption mode, the data of the frequency setting circuit unit 94 is reset, but the data of the frequency information holding circuit unit 95 is held. This retained data is data (oscillation control condition) used for the oscillation operation performed by the frequency multiplication circuit 77 before shifting to the low power consumption mode of (1), and the clock of the ring oscillator 83 is used. This is data obtained by measuring the period of the reference clock signal of the oscillation circuit 76 by the signal.

  Therefore, when the low power consumption mode is canceled in (2), the frequency multiplying circuit 77 is immediately based on the data held in the frequency information holding circuit unit 95 without measuring the reference clock signal period again. Oscillation operation can be started. When the control clock 1 falls in (3), the frequency setting circuit unit 94 outputs control data inputs 1 to 4 newly set by the CPU 72 after wakeup, and the frequency information holding circuit unit 95 The control data inputs 1 to 4 are latched at the rising edge of the control clock 2, and the newly set control data 1 to 4 are output at the falling edge of (4). In (4), F_E, which is an enable signal for the multiplied clock signal output (oscillation output), becomes active, and the clock signal POUT is output to the CPU 72.

  As described above, according to the fifth embodiment, when the oscillation operation of the frequency multiplying circuit 77 configured as a digitally controlled DPLL circuit is temporarily stopped by shifting to the low power consumption mode, it is set at that time. The control data CD12 to CD1 thus held are held by the frequency information holding circuit unit 95, and the oscillation operation is restarted based on the held control data CD12 to CD1. Therefore, unlike the conventional case, it is not necessary to reset the oscillation control condition again when the oscillation operation is resumed, so that the oscillation state at the time of resumption can be stabilized more quickly. And further reduction in power consumption can be achieved.

  Further, since the frequency multiplying circuit 77 includes the ring oscillator 83 in which a plurality of logic inverting circuits are connected in a ring shape, a configuration for obtaining a multi-phase clock signal necessary for the oscillation operation of the digital control system is obtained. This can be realized with a small number of gates. Further, since the low power consumption control circuit 78 stops the oscillation operation of the ring oscillator 83 when the operation of the clock synchronization circuit is stopped to shift to the low power consumption mode, the power consumption can be further reduced. . The frequency multiplying circuit 77 includes a DCO 81 that generates and outputs an n-multiplied clock signal obtained by multiplying the frequency of the reference clock signal PREF output by the oscillator circuit 76, so that high-precision oscillating operation can be performed by a digital control method. Can be done.

  Here, assuming that the control data is reset in the frequency multiplying circuit 77, the period of the reference clock signal PREF is measured based on the period of the multiphase clock signal accordingly, and the multiplying operation is performed based on the measured value. Will be redone. On the other hand, if the control data is held, it is not necessary to repeat the period measurement of the reference clock signal PREF, so that the oscillation operation can be stabilized at a very high speed.

  Further, assuming that the oscillation operation is temporarily temporarily stopped in the conventional analog PLL circuit 200, the voltage signal output from the phase comparator 201 is temporarily A / D converted and the digital data is latched. In the case where the oscillation operation is resumed by, for example, performing the above-described operation, the digital data held must be D / A converted and output. On the other hand, in the case of the digital control type frequency multiplication circuit 77 as in the fifth embodiment, since the control data is consistently in a digital format, the data can be held very easily. It is also advantageous compared to

<The following is for reference>
K. In a digitally controlled oscillation circuit device configured to be able to oscillate at a frequency determined according to a set oscillation control condition and to be able to temporarily stop the oscillation operation,
When the oscillation operation is temporarily stopped, the oscillation control condition set at that time is configured to be held, and the oscillation operation is configured to be restarted based on the held oscillation control condition. A featured oscillation circuit device.
According to the oscillation circuit device of K, when the oscillation operation of the digital control system is temporarily stopped, the oscillation control condition set at that time is held, and the oscillation operation is performed based on the held oscillation control condition. To resume. Therefore, unlike the conventional case, it is not necessary to reset the oscillation control condition again when the oscillation operation is resumed, so that the oscillation state at the time of resumption can be stabilized more quickly. And further reduction in power consumption can be achieved.

L. The oscillator circuit device for K described above, comprising a ring oscillator in which a plurality of logic inversion circuits are connected in a ring shape.
According to the oscillation circuit device of the above L, since it is configured to include a ring oscillator in which a plurality of logic inversion circuits are connected in a ring shape, a configuration for obtaining a multiphase clock signal necessary for an oscillation operation of a digital control system Can be realized with a smaller number of gates, and power consumption can be further reduced.
M.M. The oscillation circuit device for K or L described above, wherein the oscillation control condition can be changed before restarting the oscillation operation.
According to the M oscillation circuit device, the oscillation control condition can be changed before the oscillation operation is resumed, so that the oscillation frequency can be switched when the oscillation operation is resumed.

(Sixth embodiment)
FIG. 24 shows a sixth embodiment of the present invention. In the sixth embodiment, an example is shown in which the CPU 52 performs processing so as to change the frequency of the n-multiplied clock signal by the frequency multiplying circuit 77 at the same time when the low power consumption mode is released. That is, the data held in the frequency information holding circuit unit 95 is configured to be readable and writable by the CPU 72.
In the flowchart shown in FIG. 24, the CPU 72 first reads data held in the frequency information holding circuit unit 95 (step E1). Then, the multiplication rate at the next activation is determined according to the data, and the data corresponding to the frequency at the activation is written in the frequency information holding circuit unit 95 (step E2). Then, it shifts to the low power consumption mode (step E3).

  As described above, according to the sixth embodiment, since the control data CD12 to CD1 can be changed before the current oscillation operation is stopped, for example, the CPU 52 controls the oscillation operation and sets the oscillation control condition. In the case of the configuration to be set, if it is desired to change the oscillation frequency when the operation is resumed next time before the CPU 52 stops the oscillation operation, an oscillation control condition corresponding to the oscillation frequency can be set in advance. .

<The following is for reference>
N. The oscillation circuit device of M, wherein the oscillation control condition is configured to be changeable before the current oscillation operation is stopped.
According to the N oscillation circuit device, the oscillation control condition can be changed before stopping the current oscillation operation. For example, the CPU controls the oscillation operation and sets the oscillation control condition. In this case, if it is desired to change the oscillation frequency when the operation is resumed next time before the CPU stops the oscillation operation, an oscillation control condition corresponding to the oscillation frequency can be set in advance.

O. A multi-phase clock signal output means for generating and outputting a multi-phase clock signal having a predetermined phase difference by an oscillation operation by digital control, and the period of the reference clock signal output by the reference clock oscillation circuit And a frequency multiplication circuit for generating and outputting an n-multiplied clock signal obtained by multiplying the frequency of the reference clock signal by n with the phase difference of the multiphase clock signal as a resolution based on the measured value. The oscillation circuit device according to any one of the above K to L, which is configured.
According to the oscillation circuit device of O, the frequency multiplication circuit that includes the multi-phase clock signal output means and generates and outputs an n-multiplied clock signal obtained by multiplying the frequency of the reference clock signal output by the reference clock oscillation circuit by n. Therefore, a highly accurate oscillation operation can be performed by a digital control method.
Assuming that the oscillation control condition is reset in the configuration including the frequency multiplication circuit, the period of the reference clock signal is measured based on the period of the multiphase clock signal, and the measurement value is The multiplication operation will be redone. On the other hand, when the oscillation control condition is maintained, it is not necessary to repeat the period measurement of the reference clock signal, and the oscillation operation can be stabilized extremely quickly.

P. Oscillation of the multi-phase clock signal output means when the operation of the clock synchronization circuit that is operated by the n-multiplied clock signal output from the frequency multiplier circuit being applied to the clock input terminal is stopped to shift to the low power consumption mode Low power consumption control means for stopping operation;
The oscillation circuit device of O, further comprising data holding means for holding period measurement data of the reference clock signal as the oscillation control condition.
According to the oscillator circuit device of P, the low power consumption control means stops the oscillation operation of the multiphase clock signal output means when the operation of the clock synchronization circuit is stopped to shift to the low power consumption mode, and the data holding The means holds period measurement data of the reference clock signal. Therefore, when the low power consumption mode is canceled, the oscillation operation based on the period measurement data held by the data holding unit can be immediately performed.

(Seventh embodiment)
25 to 28 show a seventh embodiment of the present invention. FIG. 25 shows an example of the configuration of a single chip microcomputer (microcomputer) used in, for example, a vehicle body ECU (Electronic Control Unit). The microcomputer 101 includes a main oscillation circuit 103, a clock control circuit 104, a CR oscillation circuit 105, a port control circuit (signal level changing means) 106, and the like around the CPU 102.

  The main oscillation circuit 103 oscillates a crystal oscillator 107 connected to the outside of the microcomputer 101 and supplies an operation clock signal (machine clock or system clock) to the CPU 102. When the CPU 102 is instructed by the CPU 102 to start the sleep / stop mode via the address bus 108 and the data bus 109, the clock control circuit 104 outputs a clock stop signal to the main oscillation circuit 103. The output of the clock signal is stopped, and the CPU 102 (or the entire microcomputer 101 including the CPU 102) is set to the sleep / stop mode.

  Here, the sleep mode refers to a mode in which a timer built in the clock control circuit 104 automatically cancels once a predetermined time has elapsed, and the stop mode refers to an external mode once the transition is made. A mode that is not canceled until a wake-up factor (a wake-up signal, which will be described later) is generated. That is, these are modes in which the power consumption is reduced by lowering the operation state of the CPU 102 than in the normal state.

When the wakeup signal WKUP is output in response to the occurrence of the wakeup factor, the clock control circuit 104 stops outputting the clock stop signal and restarts the output of the clock signal by the main oscillation circuit 103. ing. Even in the sleep mode, the mode is canceled by the output of the wakeup signal WKUP.
For example, the oscillation frequency of the main oscillation circuit 103 is about 4 MHz, and the oscillation frequency of the CR oscillation circuit 105 is about 25 kHz. The CPU 102 operates at a clock frequency of 16 MHz by multiplying a clock with a frequency of 4 MHz by 4 with a built-in PLL oscillation circuit (not shown).

The port control circuit 106 is a circuit that controls the level of the output terminal (external signal output terminal) 110 of the microcomputer 101 in accordance with the setting performed by the CPU 102 on the internal register. The port control circuit 106 is provided with a SLEEP signal output by the clock control circuit 104 during the sleep mode.
Although not specifically shown, the microcomputer 101 includes other peripheral circuits such as an interrupt controller, a DMA controller, a watchdog timer, an A / D converter, a D / A converter, and a serial communication circuit.

  FIG. 26 is a functional block diagram showing the internal configuration of the port control circuit 106. The timer (level change timer) 111 performs a counting operation based on the CR clock signal output from the CR oscillation circuit 105, and the count data is given to the comparator (comparison circuit) 112. Note that the timer 111 is enabled to count when the SLEEP signal is active.

  The setting data of the output timing register 113 is also given to the comparator 112, and the comparator 112 outputs a coincidence signal to the level setting unit 115 via the coincidence holding unit 114 when the two data values coincide. The output timing register 113 is a register configured so that data can be written by the CPU 102. The coincidence holding unit 114 is a flip-flop that holds the active state of the coincidence signal output by the comparator 112.

  The level setting unit 115 outputs the level signal set by the level selection unit 116 to the output terminal 110 of the microcomputer 101 using a coincidence signal given via the coincidence holding unit 114 as a trigger. Similar to the output timing register 113, the level selection unit 116 is a register configured so that data can be written by the CPU 102. The output state of the coincidence holding unit 114 is configured to be reset (S / W reset) by the CPU 102 writing to a reset register provided separately.

The function of the port control circuit 106 is intended to operate as follows. That is, when the CPU 102 wakes up from the sleep mode, it is necessary to perform some processing on an external device such as a control target connected to the terminal of the microcomputer 101. At that time, when it is required to activate the function of the external device (for example, to turn on the operation power supply) as a pre-stage of processing, if all steps are taken after the CPU 102 wakes up, It takes a long time to complete the process.
Therefore, in the seventh embodiment, even when the CPU 102 shifts to the sleep mode, the level of the output terminal 110 is controlled by hardware in the stage before the release to activate the function of the external device in advance. With the goal.

  Next, the operation of the seventh embodiment will be described with reference to FIGS. FIG. 27 is a flowchart showing the contents of the sleep mode transition process by the CPU 102. First, the CPU 102 sets the level of the output terminal 110 to be changed during the sleep mode by writing to the level selection unit 116 (for example, L → H, step F1). Subsequently, by writing to the output timing register 113, it is set at which timing in the sleep mode the level of the output terminal 110 is changed (step F2).

  Then, the CPU 102 sets the duration of the sleep mode by writing to an internal register (not shown) of the clock control circuit 104 (step F3), and issues a sleep command to the clock control circuit 104 (step F4). . Then, the clock control circuit 104 receives it and outputs a clock stop signal to the main oscillation circuit 103 to stop the oscillation operation. In parallel, the clock control circuit 104 outputs a SLEEP signal to the port control circuit 106.

  FIG. 28 is a timing chart during the transition to the sleep mode. When the SLEEP signal becomes active (high) (e), the machine clock output by the main oscillation circuit 103 is stopped (d). Then, the timer 111 of the port control circuit 106 starts the count operation from the rise of the SLEEP signal (b). For example, setting data “10” is set in the output timing register 113 (a). When the count value of the timer 111 reaches “10”, the comparator 112 outputs a coincidence signal.

  At that time, the level setting unit 115 changes the signal level of the output terminal 110 from L (low) to H (high). Here, when the signal level of the output terminal 110 changes to high, the external device connected to the output terminal 110 becomes active. Thereafter, when the CPU 102 wakes up after the set sleep duration period elapses, the CPU 102 can immediately perform necessary processing on the external device. After executing the processing, the signal level of the output terminal 110 is returned to the original level by performing a S / W reset on the coincidence holding unit 114.

As described above, according to the seventh embodiment, the port control circuit 106 of the microcomputer 101 changes the level of the output terminal 110 during the period when the CPU 102 is set to the sleep mode. Accordingly, since the CPU 102 does not need to continue the normal operation mode only for changing the level of the output terminal 110, the power consumption can be reduced.
Then, the port control circuit 106 compares the count value of the timer 111 with the register value of the output timing register 113, and the comparator 112 changes the level of the output terminal 110 when the trigger signal is output because they match. Therefore, the level of the output terminal 110 can be changed at the timing set in the register 113. Since the register 113 is configured so that any data value can be set by the CPU 102, the level change timing can be arbitrarily set by the user program.

  In addition, since the output terminal level 110 can be reset by the CPU 102 when the sleep mode is canceled, the output terminal level 110 can be immediately reset when the CPU 102 performs necessary processing. it can.

(Eighth embodiment)
29 and 30 show an eighth embodiment of the present invention. The same parts as those in the seventh embodiment are designated by the same reference numerals and the description thereof is omitted, and only different parts will be described below. The port control circuit 117 in the eighth embodiment is obtained by adding a function to the configuration of the seventh embodiment. That is, a sleep time setting register 118 is added, and the setting data in the register 118 is compared with the count value of the timer 111 by the comparator 119.

  The comparator 119 outputs a coincidence signal to the interrupt generation unit 120 when the two coincide. However, when an external wake-up factor occurs before the coincidence signal is output by the comparator 119 and the sleep mode is canceled, the interrupt generation unit 120 receives the coincidence signal and causes the CPU 102 to generate an interrupt request. It is like that. That is, the timer 111, the sleep time setting register 118, the comparator 119, and the interrupt generation unit 120 constitute an interrupt generation unit 121.

Next, the operation of the eighth embodiment will be described with reference to FIG. First, the functional purpose of the interrupt generation means 121 will be described. The sleep mode is canceled by the occurrence of an external factor even before the sleep time elapses. However, after the port control circuit 106 changes the level of the output terminal 110 by the function of the seventh embodiment, the external factor When the release is performed by the occurrence, the following problem occurs.
That is, as described in the seventh embodiment, the purpose of the port control circuit 106 to change the level of the output terminal 110 is to enable the CPU 102 to immediately execute processing for the external device when the sleep mode is canceled. It is. Therefore, the time from when the level of the output terminal 110 changes until the sleep mode is canceled by the timer is normally set to the minimum time necessary to activate the external device. This is because the current consumption increases if redundancy is provided for this time.

  Under such circumstances, if the sleep mode is canceled before the sleep time elapses, the required time has not elapsed since the level of the output terminal 110 was changed when the CPU 102 was activated. May still be inactive. Then, when the CPU 102 is started, there is no means for knowing how long the original required time will be later.

Therefore, the eighth embodiment aims to notify the CPU 102 of the originally required time by interruption. That is, as shown in FIG. 30, the port control circuit 117 changes the level of the output terminal 110 in the same manner as in the seventh embodiment, and then (f) when the sleep mode is canceled early due to the occurrence of an external factor ( d, e), the interrupt generation unit 120 becomes active. Then, when the original sleep time has elapsed and the comparator 119 outputs a coincidence signal, an interrupt is generated for the CPU 102 at that time.
Therefore, the CPU 102 can recognize that the external device is activated by the interrupt signal, and after performing necessary processing, resets the level of the output terminal 110 by S / W.

  As described above, according to the eighth embodiment, the interrupt generation means 121 is not shown when the sleep mode is canceled due to the generation of the cancellation factor after the port control circuit 117 signal changes the level of the output terminal 110. An interrupt request is generated to the CPU 102 at the timing when the sleep mode is scheduled to be released by the release timer. Therefore, the CPU 102 can grasp the sustain period of the signal level changed for the output terminal 110, and can reset the level of the output terminal 110 as necessary after recognizing the interrupt request.

(Ninth embodiment)
31 to 34 show a ninth embodiment of the present invention. In FIG. 31 showing the configuration of the microcomputer 131, the CPU 132 has a 16-bit data bus, for example, and reads and executes programs and data stored in the ROM 133. The ROM 133 is composed of, for example, a flash ROM, and has a data bus width of 32 bits, which is twice that of the CPU 132.

A ROM controller (memory interface circuit device) 134 is interposed between the two, and the CPU 132 reads data from the ROM 133 via the ROM controller 134. The ROM controller 134 outputs an address ROM_A obtained by performing a predetermined operation on the read address A output by the CPU 132 and reads the data ROM_D from the ROM 133. The data is output as data D on the data bus on the CPU 132 side at an appropriate timing.
That is, in the microcomputer 101, since the reading speed of the ROM 133 constituted by the flash ROM is slower than the processing speed of the CPU 132, the data bus width on the ROM 133 side is doubled that of the CPU 132, and arbitration control between the two is performed by the ROM. The controller 134 is configured to perform.

  FIG. 32 is a block diagram showing an internal configuration of the ROM controller 134. The ROM controller 134 includes a control unit (data reading unit, reading control unit) 135, an address path unit 136, and a data path unit (reading control unit) 137. The control unit 135 includes an address path control unit 137, a data path control unit 138, an address hit detection unit 139, a ROM control unit 140, and a CPU control unit 141.

  The address path control unit 137 and the data path control unit 138 control the address path unit 136 and the data path unit 137, respectively. The address hit detection unit 139 detects a match (hit) between the address held by the address path unit 136 and the address output by the CPU 132. The CPU control unit 141 outputs a read control signal (acknowledge signal ACKN) to the CPU 132. Note that the read control signal ACKN is at a high level when the CPU 132 waits for ROM data read, and indicates that the acknowledge is active at a low level. The ROM control unit 140 controls these in an integrated manner.

The address path unit 136 outputs an address to the ROM 133 via a multiplexer (MUX) 142. A through path 143 directly connected to the address bus on the CPU 132 side is connected to one input side of the MUX 142, and a counter side path 144 is connected to the other input side.
The through path 143 is also connected to the input side of the buffer (latch) 145, and the output side of the buffer (address storage means) 145 is connected to the address hit detection unit 139. The buffer 145 stores an address output when the CPU 132 executes a data read cycle.

  Further, the bit output unit 146 is a bit (valid bit) indicating ROM read enable when a data read target address of the ROM 133 and data corresponding to the address are stored in the buffer 145 and a data buffer 154 described later. ) Is set, and the output side is also connected to the address hit detection unit 139. The bit output unit 146 is reset by hardware.

On the other hand, a MUX 147, a counter buffer (address holding unit) 148, and a counter (+4) 149 are arranged on the counter side path 144 side. A through path 143 is connected to one input side of the MUX 147, and an output side of the counter 149 is connected to the other input side. The output terminal of the MUX 147 is connected to the input side of the counter 149 and the counter side path 144 via the counter buffer 148.
Similarly to the through path 143 side, the counter side path 144 is also connected to the input side of the buffer 150, and the output side of the buffer 150 is connected to the address hit detection unit 139. The output side of the bit output unit 151 is also connected to the address hit detection unit 139.

  Similarly to the bit output unit 146, the bit output unit 151 enables ROM read when an instruction read target address of the ROM 133 and an instruction corresponding to the address are stored in the buffer 148 and an instruction buffer 153 described later. Is set so as to set a bit. When instruction reading by the CPU 132 is performed at consecutive addresses, the hit and miss hits are performed from the bit output units 151 and 153 and the address A on the CPU 132 side for each cycle between the ROM controller 153 and the ROM 133. Are detected alternately.

  The data path unit 137 includes a MUX 152, an instruction buffer 153, a data buffer 154, and a MUX 155. Of the 32-bit ROM_D, the upper 16 bits [31:16] and the lower 16 bits [15: 0] are connected to the input side of the MUX 152, respectively. The instruction buffer 153 stores lower 16-bit data when the CPU 132 executes an instruction read (instruction fetch) cycle for the ROM 133.

  The data buffer 154 stores 32-bit data when the CPU 132 executes a data read (data load) cycle with respect to the ROM 133. The MUX 155 selects either the upper side or the lower side of the MUX 152, the instruction buffer 153, and the data buffer 154 and outputs the selected data to the data bus on the CPU 132 side.

  FIG. 33 is a flowchart showing the control contents of the ROM controller 134 when the CPU 132 executes a read cycle for the ROM 133. FIG. 34 is a timing chart when the CPU 132 continuously executes instruction read cycles. Note that the ROM controller 134 is composed solely of hardware, and the flowchart in FIG. 33 shows the operation of the hardware.

The operation outline in FIG. 34 will be described. The CPU 132 reads an instruction in units of 16 bits (2 bytes) (a). In this case, the ROM controller 134 increments the address held in the buffer 148 by “4” in accordance with the data bus width of 32 bits of the ROM 133, and reads the instruction in units of 32 bits (4 bytes). (B).
The ROM controller 134 outputs the upper 16 bits of the 32-bit instruction simultaneously read out to the data bus on the CPU 132 side at the same time as reading, and temporarily stores the lower 16 bits in the buffer 153 (h). Then, when the CPU 132 executes the next read cycle, the instruction stored in the buffer 153 is output to the data bus (i).

  When the CPU address A matches the address held in the counter buffer 148, the control unit 135 outputs a count-up signal for incrementing the address by “4” in the next cycle (c). At the same time, an address stored in the counter buffer 148 is output to the buffer 150, and a load signal for loading the data read from the ROM 133 to the buffer 153 is output (e).

  In FIG. 33, the control unit 135 stands by until a read cycle for the ROM 133 is generated depending on the output state of the decode signal by an external decoder (not shown) (step G1). When the read cycle occurs (“YES”), it is determined whether or not the read cycle is “command read” (step G2).

A. <Instruction read: Address continuous>
When the control unit 135 refers to the access code or the like in the control signal output by the CPU 132 and determines that it is “command read” (step G2, “YES”), the address hits the command buffer 153 next. It is determined whether or not (step G3).
If the instruction corresponding to the address is stored in the instruction buffer 153, it is a hit (“YES”), so the control unit 135 uses the counter 149 to add “4” (via the MUX 147). The value of the buffer 148 is updated (step G4). Note that the control unit 135 determines whether the lower order of the instruction is output when the bit output unit 151 outputs “1” and the address stored in the buffer 150 matches the address A on the CPU 132 side. It is determined that 16 bits are hit.

  Then, the address is output to the address bus on the ROM 133 side via the MUX 142, and a 4-byte instruction is read (step G5). Subsequently, the control unit 135 causes the buffer 153 to output an instruction hitting the instruction buffer 153 of the data path unit 137 to the data bus on the CPU 132 side.

Reference is now made to FIG. The above case corresponds to cycle (3), for example. That is, when the CPU 132 outputs the address (A + 2), the instruction D2 corresponding to the address is stored in the instruction buffer 153 in the reading performed by the ROM controller 134 in the previous cycles (1) and (2). . Therefore, in this case, a buffer hit occurs, and the instruction D2 is output from the buffer 153.
At the same time, if there is a buffer hit, there is a high probability that instructions will be read continuously (A + 4) next time, so that instructions are read in advance using the address (A + 4) in preparation for the next instruction read.

  FIG. 33 will be referred to again. If no buffer hit occurs in step G3 (“NO”), the control unit 135 determines whether or not the address at that time matches the address currently being read (step G7). If they match ("YES"), the system waits until the setup time until the reading being performed at that time is completed (step G8, "YES"), and is read. The contents of the buffer 153 are updated with the lower 16 bits of the data (step G9). Further, the upper 16 bits of the read data are output as they are to the data bus on the CPU 132 side (step G10).

  The above case corresponds to, for example, the cycle (4) in FIG. That is, when the CPU 132 outputs the address (A + 4), the instruction D4 corresponding to the address is not stored in the buffer 153 and a buffer miss occurs, but the address (A + 4) is performed in cycles (3) and (4). Matches the read address (prefetch address). Therefore, the control unit 135 outputs the upper side D4 of the data read in cycle (4) to the data bus on the CPU 132 side, and stores the lower side D6 in the buffer 153.

B. <Instruction read: Branch>
In step G7, if the address at that time does not match the prefetch address (“NO”), this is a case where a branch instruction is executed in the program. At this time, the control unit 135 updates the prefetch address (counter buffer 148) (step G11), and updates the ROM address by switching the MUX 142 to the through path 143 side (step G12). And the process similar to step G8-G10 is performed (step G13-G15).

C. <Read data>
In step G2, the case where the control unit 135 determines “NO” and proceeds to step G16 is a case where the CPU 132 performs a data read cycle on the ROM 133. When the CPU 132 loads data, there are two cases of byte access and word (16 bit) access. Regardless of the case, the ROM controller 134 stores the data for 32 bits to which the read target data belongs to the ROM 133. Read more simultaneously. Further, the read address in that case is stored in the buffer 145.
The 32-bit data is always stored in the data buffer 154, and either the upper 16 bits or lower 16 bits to which the data to be read by the CPU 132 belongs are output to the data bus via the MUXs 152 and 155.

  That is, in step G16, the control unit 135 determines whether or not the read address has hit the data buffer 154 based on the address stored in the buffer 145 and the bit of the bit output unit 146. If there is a hit ("YES"), the 16-bit data on the hit side from the data buffer 154 is selected by the MUX 155 and output to the data bus on the CPU 132 side (step G17). If the read address does not hit the data buffer 154 at step G16 ("NO"), the same processing as steps G12 to G15 is performed on the data (steps G18 to G21).

  As described above, according to the ninth embodiment, the controller 135 of the ROM controller 134 holds the buffer 148 when the address output by the CPU 132 in the instruction read cycle matches the address held in the buffer 148. The address value is incremented and data is read from the ROM 133 in the next cycle. The instruction buffer 153 stores the data read by the control unit 135 as much as the bus width of the CPU 132 is exceeded, and the control unit 135 executes the instruction read cycle for the ROM 133 at the next consecutive address. The data held in the instruction buffer 153 is output to the data bus of the CPU 132.

  That is, even when the reading speed of the ROM 133 is slow, when the CPU 132 reads an instruction at successive addresses (incremental addresses according to the data bus width of the CPU 132), the instruction to be read from the next time is read this time. Are simultaneously read and stored in the instruction buffer 153, and the instruction is read from the instruction buffer 153 in the next reading. Therefore, it is possible to reduce the wait cycle conventionally required for the CPU 132 to read from the ROM 133 and improve the processing efficiency.

  When the CPU 132 executes a data read cycle for the ROM 133, the read data is stored in the data buffer 154, and the control unit 135 reads the read address when the CPU 132 executes the data read cycle next time. If the data buffer 154 is hit, the data stored in the buffer 154 is output to the data bus of the CPU 132. That is, since the data buffer 154 functions as a so-called data cache, the CPU can improve the data reading efficiency.

<The following is for reference>
Q. A memory interface circuit which is connected between the CPU and one or more ROMs having a data bus width wider than the CPU, and controls reading when the CPU attempts to read data from the ROM A device,
Instruction address holding means for holding an address value;
The address held in the address holding means is compared with the address output by the CPU to execute the instruction read cycle for the ROM, and if they match, the address value held in the address holding means Data read means for reading data from the ROM by increasing the ROM bus width exceeding the CPU bus width;
An instruction buffer for storing the data read by the data reading means for an amount that the ROM bus width exceeds the CPU bus width;
Read control means for controlling the data stored in the instruction buffer to be output to the data bus of the CPU when the CPU executes an instruction read cycle for the ROM at the next consecutive address. A memory interface circuit device.

  According to the memory interface circuit device of Q, the data reading means, when the address held in the address holding means matches the address output by the CPU in the instruction read cycle, the address value held in the address holding means is Read the data from the ROM in increments. Then, the instruction buffer buffers the data read by the data reading means as long as it exceeds the CPU bus width, and the read control means executes the instruction read cycle for the ROM at the next consecutive address. The data held in the instruction buffer is output to the CPU data bus.

  In other words, even when the ROM read speed is slow, if the CPU reads instructions at consecutive addresses, the instructions that the CPU is to read from the next time are read simultaneously in the current read and held in the instruction buffer. In the next reading, the instruction is read from the instruction buffer. Therefore, it is possible to reduce the wait cycle conventionally required for the CPU to read from the ROM, and to improve the processing efficiency.

R. When the CPU executes a data read cycle for the ROM, a data buffer for storing the data read from the ROM by the data read means as much as the bus width of the ROM exceeds the bus width of the CPU When,
Address storage means for storing the address,
When the CPU next executes the data read cycle for the ROM based on the address stored in the address storage means, the read control means stores data corresponding to the read address in the data buffer. The memory interface circuit device according to Q, wherein if it is determined that the data is stored, the data is controlled to be output to a data bus of the CPU.

According to the above R memory interface circuit device, when the CPU executes a data read cycle, the data read from the ROM is stored in the data buffer by the amount that the ROM bus width exceeds the CPU bus width. To do. At this time, the read address is stored in the address storage means. When the CPU executes the data read cycle next time and determines that the data corresponding to the read address is stored in the data buffer, the read control means outputs the data to the CPU data bus. .
That is, since the data buffer functions as a so-called data cache, when the data read address hits the data buffer, the CPU can read the data at high speed. Therefore, it is possible to improve the data reading efficiency.

(Tenth embodiment)
FIGS. 35 and 36 show a tenth embodiment of the present invention, and only the parts different from the ninth embodiment will be described. As shown in FIG. 35, in the tenth embodiment, the control unit 135A of the ROM controller 134A has a function of detecting that the CPU 132 is in a stalled state, and in this case, the current of the ROM 133 is cut off. It is configured to perform control.

In the timing chart shown in FIG. 36, the control unit 135A detects that the stall has occurred in the operation of the CPU 132, for example, at the timing of the cycle (5) when the address value (A + 4) is the same value for two consecutive cycles. .
After that, when the stall state of the CPU 132 continues for another two cycles, for example, the control unit 135A controls to cut the current consumption of the ROM 133. The term “stall” used here refers to a state in which access to the ROM 133 is not performed (for example, other resources such as RAM and I / O) in addition to stopping the arithmetic processing inside the CPU 132 in a general sense. In the case of access). In that case, it can be determined that the ROM 133 is not selected by the decode signal of the external decoder. For example, in step G1 in FIG. 33, it is possible to measure and determine the period during which “NO” is determined.

  As described above, according to the tenth embodiment, the control unit 135A of the ROM controller 134A stops reading data from the ROM 133 when detecting that the CPU 132 is in a stalled state. When the stall state of the CPU 132 continues for a predetermined period, the current consumption in the ROM 133 is cut off. Therefore, useless power consumption can be reduced.

<The following is for reference>
S. The read control means is configured to be able to detect that a stall has occurred in the operation of the CPU,
The data reading means, when it is detected that the stall has occurred, and when the reading of data to the ROM is being executed, the reading process is continued and completed, Q, R, the memory interface circuit device of any one of U below.
According to the memory interface circuit device of S, the data reading means continues the reading process when data reading from the ROM is being executed even when it is detected that the CPU operation has stalled. To complete. That is, since reading is not immediately stopped when a stall occurs, reading efficiency can be improved.

T.A. The read control unit detects that a stall has occurred in the operation of the CPU, and controls the current consumption in the ROM to be cut off when the stall state continues for a predetermined period. The memory interface circuit device of any one of U below.
According to the memory interface circuit device of T, the read control means controls to cut off the current consumption in the ROM when the stall state of the CPU continues for a predetermined period, so that it is possible to reduce wasteful power consumption.

The present invention is not limited to the embodiments described above and shown in the drawings, and the following modifications or expansions are possible.
In the first embodiment, the 1 / N frequency dividers 19 and 20 may be provided as necessary.
In the second embodiment, the main clock monitoring unit 7 may be deleted.
In the second embodiment, steps A0 and A4 may be deleted.
In the second embodiment, the high-speed correction operation is not necessarily performed by the binary search method, but may be performed by starting from relatively large adjustment signal data and gradually decreasing the absolute value thereof.
In the fifth embodiment, the oscillation operation of the oscillation circuit device is not limited to the one accompanying the execution of the low power consumption mode. That is, the clock signal output from the oscillation circuit device is not limited to the one supplied as the operation clock signal for the CPU, and may be used for any purpose. Therefore, the present invention can be widely applied as long as it is necessary to temporarily stop the output of the clock signal according to the use and then restart the output.

In the seventh embodiment, the data set in the output timing register 113 may be configured to be fixed in hardware.
In the eighth embodiment, if the level change timer is shared with the release timer, the configuration can be simplified.
In the eighth embodiment, an interrupt generation timer may be provided independently of the level change timer.
If the microcomputer 101 is provided with a watchdog timer for monitoring the operation of the CPU 102, or a free-run timer that counts independently of the operation clock of the CPU 102, the level change timer is set to By using the timer in common, the configuration can be simplified. However, in this case, when used as a level change timer, the count value of the watchdog timer or the like at the start of the timer operation is read, and the sum of the count values corresponding to the output timing is set in the output timing register. . By using a common timer, software creation and management can be facilitated, and the possibility of bugs occurring can be reduced.

In the ninth embodiment, the bus width of the buffer built in the memory interface circuit device may be made wider than the data bus width of the ROM. If comprised in this way, the data read-out means will perform the read-out with respect to ROM several times, and it can pre-read about more data by storing data in the said buffer.
In the ninth or tenth embodiment, the ROM may be similarly implemented by arranging two 16-bit bus mats. In this case, the current consumption of the mat ROM that is not read by the ROM controller 134 may be cut off. Even if configured in this manner, useless power consumption can be reduced when the ROM is composed of a plurality of mats.

Further, in the tenth embodiment, when the CPU 132 shifts to a low power consumption mode such as sleep / stop, the current consumption of the ROM may be cut off. With such a configuration, useless power consumption can be reduced.
In the tenth embodiment, if the ROM controller 134A detects that the stall detection timing of the CPU 132 is, for example, cycle (4) and the address (A + 2) continues, the instruction data D2 being executed for the ROM 133 at that time , D4 can be continued if other conditions can be continued. With this configuration, the ROM controller 134A does not immediately stop reading from the ROM 133 even if it detects a stalled state, so that the reading efficiency can be improved.

<The following is for reference>
U. 32. The memory interface circuit device according to claim 30, wherein the buffer has a bus width wider than a data bus width of the ROM.
According to the memory interface circuit device of U, the bus width of the buffer is made wider than the data bus width of the ROM. Therefore, if the data reading means performs reading a plurality of times and stores data in the buffer, more data is stored. This data can be prefetched.

V. When the ROM is composed of a plurality of mats,
35. The memory interface circuit device according to claim 34, wherein the read control means controls so as to cut off a current consumption other than that of a ROM which is a read target by the data read means.
According to the memory interface circuit device of the above V, the read control means cuts off the current consumption other than the ROM that is read by the data read means, so that it is useless when the ROM is composed of a plurality of mats. Power consumption can be reduced.

W. The read control means controls the CPU to cut off all current consumption of the ROM when the CPU shifts to a low power consumption mode in which the operation is temporarily stopped while maintaining the internal state. The S or T memory interface circuit device.
According to the memory interface circuit device of W, the read control means cuts off all the current consumption of the ROM when the CPU shifts to the low power consumption mode, so that it is possible to reduce wasteful power consumption.

1 is a functional block diagram showing an electrical configuration of a microcomputer according to a first embodiment when the present invention is applied to a microcomputer. The figure which shows the detailed structure centering on a subclock correction | amendment part. The flowchart which shows the correction process by the sub clock correction part Shows the detailed configuration of the main clock monitoring unit FIG. 3 equivalent view showing a second embodiment of the present invention. The flowchart which shows the processing content in step S12 The flowchart which shows the processing content in step S13 Conceptual explanatory diagram of the state in which high-speed correction operation is performed Functional block diagram showing the configuration of a microcomputer according to a third embodiment of the present invention The figure which shows the structure of the functional part which a low power consumption control part controls about sleep mode The figure which shows the structure of the functional part in which a low power consumption control part controls an I / O terminal part when stop mode is set. A flowchart showing processing when the CPU shifts to the low power consumption mode. Flowchart showing processing when the low power consumption mode is canceled and the CPU starts (wakes up) The block diagram which shows the function part regarding the clock control of a microcomputer which is 4th Example of this invention. The figure which shows the transition state of the clock signal at the time of power-on and the wake-up when the CPU is activated from the sleep / stop mode Diagram of switching the machine clock by a user program while the CPU is performing normal operations A flowchart showing processing when the CPU performs switching of the machine clock signal based on the diagram shown in FIG. The functional block diagram which is 5th Example of this invention and shows the electrical structure of ECU Functional block diagram showing schematic configuration of frequency multiplier Functional block diagram showing detailed configuration of DCO Diagram showing the configuration of the ring oscillator Diagram showing a part of the internal configuration of the counter / data latch circuit Timing chart showing the case where the frequency multiplier circuit starts and oscillates from the low power consumption mode CPU flowchart showing a sixth embodiment of the present invention. The figure which is 7th Example of this invention and shows one structural example of a single chip microcomputer Functional block diagram showing the internal configuration of the port control circuit The flowchart which shows the contents of the sleep mode shift processing by CPU Timing chart during sleep mode transition FIG. 26 equivalent diagram showing an eighth embodiment of the present invention. 28 equivalent diagram The figure which is 9th Example of this invention and shows the structure of a microcomputer Block diagram showing the internal configuration of the ROM controller The flowchart which shows the control content of ROM controller in case CPU performs the read cycle with respect to ROM Timing chart when CPU executes instruction read cycle continuously FIG. 31 equivalent diagram showing the tenth embodiment of the present invention. Figure 33 equivalent 28 equivalent diagram showing the prior art

Explanation of symbols

  1 is a microcomputer, 2 is a CPU, 3 is a main clock oscillation unit (oscillation circuit), 4 is a subclock oscillation unit (CR oscillation circuit), 5 is a watchdog timer, 6 is a subclock correction unit (correction circuit), 7 Is a main clock monitoring unit (monitoring circuit), 8 is an edge detection circuit 8, 9 is a pulse counter, 10 count number setting register, 11 is a comparison adjustment means, 13 is a ladder resistor (variable resistor), 21 is a clock control circuit device , 31 is a microcomputer, 32 is a CPU, 35 is a low power consumption control unit (terminal control means), 42 is an input / output terminal (external signal terminal), 51 is a microcomputer, 52 is a CPU, 54 is an external oscillator, 55 Is a DPLL circuit (frequency multiplication circuit), 56 is a clock control unit (switch notification control unit), 62 is a state counter (sequence control means), 1 is an ECU (microcomputer), 71a is a clock control circuit (oscillation circuit device), 76 is an oscillation circuit, 77 is a frequency multiplication circuit, 78 is a low power consumption control circuit (low power consumption control means), and 80 is a control circuit ( Sequence control means), 81 a DCO, 83 a ring oscillator (multi-phase clock signal output means), 95 a frequency information holding circuit (data holding means), 101 a single chip microcomputer, 102 a CPU, 106 a port control Circuit (signal level changing means), 110 is an output terminal (external signal output terminal), 111 is a timer (level change timer), 112 is a comparator (comparison circuit), 113 is an output timing register, 117 is a port control circuit, 121 is an interrupt generating means, 131 is a microcomputer, 132 is a CPU, 133 is RO , 134 and 134A are ROM controllers (memory interface circuit devices), 135 and 135A are control sections (data reading means, reading control means), 137 is a data path section (reading control means), 145 is a buffer (address storage means), Reference numeral 148 denotes a counter buffer (address holding means), 153 denotes an instruction buffer, and 154 denotes a data buffer.

Claims (6)

A CPU configured to be able to set a low power consumption mode in which the operation is temporarily stopped while maintaining the internal state;
A release timer for periodically releasing the low power consumption mode;
A signal level changing means for changing the level of the external signal output terminal within a period in which the low power consumption mode is set ;
The signal level changing means is
A register for setting a timing for changing the level of the external signal output terminal;
A level change timer that starts a count operation from the time when the CPU enters the low power consumption mode;
Comparing the count value of the level change timer with the value of the register, and a comparator circuit that outputs a trigger signal when both match, and when the trigger signal is output, the level of the external signal output terminal A microcomputer characterized by being configured to change . It said register comprises a microcomputer according to claim 1, wherein the arbitrary data value is configured to be set by the CPU. Level of the external signal output terminal is changed by the signal level change means, wherein the low power consumption mode is released, according to claim 1, characterized in that it is configured to be capable of resetting by the CPU or 2. The microcomputer according to 2 . The low power consumption mode is set so as to be released even when a release factor other than the release timer occurs,
After the signal level changing means changes the level of the external signal output terminal, when the low power consumption mode is canceled due to the generation of the release factor, the low power consumption mode is scheduled to be released by the release timer. 4. The microcomputer according to claim 3 , further comprising interrupt generation means for generating an interrupt request to the CPU at timing. The level change timer includes a microcomputer according to any one of claims 1 to 4, characterized in that it is constituted by the common and the release timer. The release timer claim 1, characterized in that it is constituted by a common free-run timer count operation independently of the operation clock of the watchdog timer, or the CPU to monitor the operation of the CPU 5. The microcomputer according to any one of 4 to 4 .

Images