JPS6216570B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an oscillation circuit that can oscillate at two oscillation frequencies simultaneously.

It has been known that two vibration modes occur simultaneously in one vibrator. For example, JP-A-55-73120 discloses that a tuning fork type crystal resonator generates a bending vibration mode and a torsional vibration mode simultaneously, and the bending vibration and torsional vibration are mutually coupled to improve the temperature frequency characteristics of the bending vibration mode. The configuration to be improved is memorized. In addition to the above citations,
It is known that GT resonators improve their temperature-frequency characteristics by utilizing two vibration modes in the vertical and horizontal directions within a flat crystal plate, and the same applies to thickness-shear resonators. There are some that use principles. When one oscillation circuit loop is connected to such a vibrator, it oscillates at the frequency of a vibration mode with low impedance. However, in this case, the vibration mode of the oscillation frequency is vibrating including the displacement component of the other vibration mode, which affects the output frequency, so the frequency of each of the two vibration modes is adjusted to obtain the best temperature-frequency characteristics. and it is necessary to obtain the desired output frequency. This requires an oscillation circuit that extracts and outputs both of the two vibration modes occurring in one vibrator.

A conventional oscillation circuit for the above purpose is shown in FIG. 1 as an example. The names and functions of each part in the diagram are as follows:
It is as follows. Reference numerals 11 and 13 have one input terminal of an exclusive OR circuit set to a logic level of 0, and each serves as an amplifier for two oscillation circuit loops.
The resistors 14 and 15 and the capacitor 17 are 13,
It serves as a phase shifter that delays the phase of the 11 output waves, improves the phase conditions of each oscillation, and stabilizes the oscillation frequency. 16 is a resonator such as a crystal resonator or a ceramic resonator, which must have a Q value of at least 10,000 or more. The capacitor 18 and the coil 20', and the capacitor 19 and the coil 21' are:
Tuned to each oscillation frequency, two oscillation frequencies obtained on the output side of the 16 oscillators (0 in the figure)
The mixed wave of fB and fC is filtered to separate each frequency, and also serves to invert the phase and feed it back to the amplifiers 11 and 13. Reference numeral 12 denotes an amplifier composed of an exclusive OR circuit, which serves as a buffer amplifier for fC waves. Numeral 10 is a circuit similar to 12, which mixes the fB and fC waves, and further a reduction pass filter 22' outputs the difference frequency Δf between fB and fC. 10, 11, 12, and 13 are circuits all composed of MOS transistors (field effect transistors). The problem with the conventional oscillation circuit shown in Figure 1 is that when two oscillation frequencies fB and fC are close to each other, for example, the frequency difference △f between them and the ratio △f/fC of either fB or fC is 10 ^- In case of ³ , 18
, 20', 19, and 21' require a Q value of 10 ³ or more, making it difficult to separate both frequencies with a normal L and C tuning circuit with a Q value of about 10 ² . Become. In particular, in the above-mentioned tuning fork type crystal resonator, fC was 200 KHz, while fB was 190 KHz, making it extremely difficult to separate frequencies using the circuit shown in Figure 1. Furthermore, if the frequency difference between fB and fC fluctuates significantly compared to the half-width frequency of the tuning circuit, the values of the coil and capacitor must be changed each time, which is inconvenient. It was hot.

The purpose of the present invention is to significantly improve these drawbacks and to:
Easily convert two frequencies contained in one oscillator,
Moreover, it is an object of the present invention to provide an oscillation circuit that can be separated reliably.

A diagram of the principle of the oscillation circuit according to the present invention is shown in FIG. 2a. The names and functions of each part in the diagram are as follows. Amplifiers 20 and 21 perform either inverting or non-inverting amplification of the oscillation circuit loops fC and Δf, respectively. 22 is an amplitude modulator which serves to amplitude modulate the carrier wave fC at a modulation frequency Δf. 23 is a vibrator having two adjacent resonant frequencies. 24 is a capacitor that serves as a load capacitance for the vibrator, and 23 and 24
A feedback circuit 30 is formed. 25 is a detector that detects the envelope of the amplitude modulated wave after passing through the vibrator. 2
6 and 27 are buffer amplifiers of fC and Δf, respectively. 28 and 29 are output terminals.

FIG. 2b is an equivalent circuit of the vibrator used in FIG. 2a. The resonant circuits indicated by LB, CB, and RB in the figure are the equivalent circuits of the vibration mode corresponding to the frequency of fB, and the resonant circuits indicated by LC, CC, and RC are the equivalent circuits of the vibration mode corresponding to the frequency of fC. Represents a circuit. C0 is the electrostatic capacitance between the electrodes.
The oscillation operation in the oscillation circuit shown in FIG. 2a is performed as follows. The voltage waveform of the main part of the oscillation circuit is
As shown in the figure. The two resonant frequencies of the vibrator are fB and fC
The equivalent impedance when each oscillates as
Let ZB and ZC, and assume that ZB>ZC holds. Under this condition, first, the signal voltage at the oscillation frequency of fC starts to rise. In this case, the output of the vibrator 23 is a sine wave with a frequency fC, and the frequency signal fC that has been saturated and amplified by the amplifier 20 is sent to the amplitude modulator 2.
2 to the vibrator 23. Taking the tuning fork type crystal resonator mentioned above as an example, fC
starts oscillating at a frequency of approximately 200KHz, which is the frequency of higher-order bending vibration. However, the tuning fork type crystal resonator formed by the cut angle disclosed in the above-mentioned Japanese Patent Application Laid-Open No. 55-73120 has a special mode in which a torsional vibration mode is simultaneously generated in the fC bending vibration mode. The frequency-temperature characteristics of fC are improved by combining the two modes. Such a tuning fork type crystal resonator can oscillate with torsional vibration at the resonant frequency fB (approximately 190 KHz), but when only the oscillation circuit loop at the frequency fC is connected, it will not oscillate independently at the frequency fB. , so fC
has no effect on the sine wave. However, if you connect an oscillation loop with frequency fB, the frequency fB
starts torsional vibration, and torsional vibration exists independently of the vibration of fC. In the present invention
Detector 2 that detects the frequency difference △f signal between fC and fB
5, an amplifier 21, and an amplitude modulation circuit 22 to form an oscillation loop of frequency fB. The output of a tuning fork type crystal resonator that oscillates at frequency fC is basically a sine wave, but due to noise etc., a minute Δf component is superimposed on fC. Then, it is detected by the detector 25 and the amplifier 2
1 and input to the amplitude modulation circuit 22, fC
A signal obtained by modulating the frequency signal with Δf is input to the tuning fork type crystal resonator. As will be explained in detail later, there is an fC−△f, or fB component in the modulation signal, which causes the tuning fork crystal resonator to adjust the frequency.
Oscillation also starts due to fB. The output waveforms of the main parts in this state are shown in FIG. 5. FIG. . 3 is the output signal of the amplitude modulation circuit 22; 4
is the oscillator 2 when excited by the input signal 3
This is the output signal of No. 3. When the carrier wave of fC is amplitude-modulated at the modulation frequency △f, the amplitude modulated wave on the input terminal side of the vibrator (1 point in the figure) is observed with a spectrum analyzer as three frequency components: fC, fC±△f, and others. It is composed of harmonic components of. After this passes through the vibrator 23, the frequency component of the voltage appearing on the output side 2 in the figure consists of three components, fC and fC±Δf, due to the filter action of the vibrator. fC＋
The components △f, fC - △f are obtained on the output side because the operating point of the resonator is normally used in a region where the relationship between the excitation voltage and the displacement of the resonator is a straight line, so the voltage proportional to the input voltage is obtained on the output side. is obtained on the output side. Note that the fC component on the input terminal side will fluctuate intermittently, but since the period during which fC is stopped by △f is very short, the vibrator will continue to vibrate due to Q, and there will be almost no vibration. There is no attenuation.

In the loop of the frequency signal fC including the amplifier 20, the output signal 4 is saturated and amplified by the amplifier 20, thereby removing the fC±Δf component and obtaining a carrier wave fC with a constant amplitude. On the other hand, in the loop of the frequency signal △f including the amplifier 21, the output signal 4 is detected by a detector.
25 performs envelope detection to remove the fC component, and if necessary, a low-pass filter (described later) is installed to ensure that the △f component is obtained, and the amplifier 21 saturates and amplifies the modulated wave with a constant amplitude. Δf is formed. The amplifier 21 of the oscillation circuit loop of Δf is designed to have constant amplification and phase transfer characteristics in the frequency band that Δf can take. In this case, the vibrator 23 and the detector 2
5. The gain of the closed loop voltage transfer characteristic passing through the amplitude modulator 22 and amplifier 21 is △f=|fC−fB
| (absolute value), so the oscillation frequency of the oscillation circuit of Δf is locked to this frequency. The oscillation circuit according to the present invention is characterized in that the difference frequency between two resonant frequencies of the vibrator can be directly obtained.

Specific examples of the principle circuit of FIG. 2 are shown in the third and fourth sections.
As shown in the figure. The major differences between the oscillation circuits of FIG. 3 and FIG. 4 are as follows. The oscillation circuit shown in FIG. 3 is capable of sustaining oscillation even when the Q value of the vibrator is 10,000 or less. On the other hand, in Fig. 4, the Q value of the oscillator is 1.
Stable oscillation is possible when the number of oscillations exceeds 1,000,000.

The configuration of each oscillation circuit will be explained in detail below.

First, explanation will be given starting from FIG. In the figure, the names of each part and their functions are as follows. 300
is a modulator consisting of a C-MOS, NOR circuit, and a resistor 307 is a feedback resistor 30 for setting the operating point.
1,303,304,305 are 2 input C-
This is an amplifier that connects two gates of MOS and NOR circuits. Of course, an inverter can also perform the same function. Resistors 308, 309, 310, 3 connecting the gate and drain of these NOR circuits
11 sets the operating point of these amplifiers to the power supply voltage VDD.
This is the feedback resistor used to set it to 1/2 of the Optimize the operating range of the amplifier by setting the operating point of the amplifier to 1/2VDD. The capacitor 320 serves to reduce the high frequency gain of the amplifier 304 and eliminate the carrier fC component contained in the Δf signal after detection. That is, the capacitor 320 constitutes a low-pass filter that cuts out high frequency components based on the time constant determined between the capacitor 320 and the resistor 310, and the amplifier 304 is
No components are allowed to pass through, and only the Δf component can be processed. The two amplifiers 304 and 305 can be replaced with one non-inverting amplifier. 302 and 306 are inverters for buffer amplification of fC and Δf signals. Further, 322 is a vibrator, 321 is a diode, 312 is a load resistor that determines the input potential of the diode 321, 313,
317 is a resistor and a capacitor for smoothing a low frequency signal Δf generated by envelope detection of a modulated wave by a diode. The time constant created by these resistor R and capacitor C is selected so that approximately 1/fC<RC<1/Δf. 314 is a load capacitance of the vibrator 322, which is used to adjust the oscillation frequency. 315, 31
Coupling capacitors 6, 318, and 319 are for cutting the DC component of the signal. The capacitance values of these capacitors must be sufficiently large. The voltage waveforms of the main parts in the figure are observed as shown in FIG. Here, in order to establish a circuit correspondence with FIG. 2, the circuit blocks used in FIG. 2 are enclosed in a frame in FIG. 3 and given the same numbers as in FIG. 2. That is, 20 is a saturation amplifier for the fC component, 21 is a saturation amplifier for the Δf component, 30 is a feedback circuit, 22 is an amplitude modulator, 25 is a wave detector, and 26 and 27 are buffer amplifiers. Note that the NOR circuit 303 is an amplifier that linearly amplifies the modulated sine wave 4 for envelope detection, and the other amplifiers 301, 304, 30
5 is a saturation amplifier.

The operation of such a circuit will be explained with reference to FIG. 3. When the crystal oscillator 322 starts to oscillate at two frequencies, fC and fB, the output terminal 4 outputs fC and fC±△f as shown in FIG. A combined output signal is generated. This output signal is the NOR circuit 30
1 and a feedback resistor 308, the signal is saturated and inverted amplified to generate the output signal shown in FIG. On the other hand, the output signal of 4 is the NOR circuit 30
3 and a feedback resistor 309, the diode 3
21, resistors 312 and 313, and a capacitor 317. This detector is 4
Detect the lower envelope of the output signal of the frequency △f
A signal 6 of (difference frequency between frequency fC and frequency fB) is output. Although not shown in the output signal 6 shown in FIG. 5, the output signal of this detector contains a slight frequency component of fC. This 6 output signal is saturated and inverted amplified by a saturated amplifier formed by a NOR circuit 304 and a feedback resistor 310. At this time, the fC component is cut off by the low-pass filter action of the capacitor 320 mentioned above. The output of this NOR circuit 304 is inverted again by a saturation amplifier formed by a NOR circuit 305 and a feedback resistor 311 to obtain output signal 2 shown in FIG. 1 frequency fC signal and 2 frequency △f signal are connected to NOR circuit 3
00 and the feedback resistor 307, and the output signal 3 of FIG.
occurs. As mentioned above, the output signal of 3 is the frequency
This is a signal that outputs a signal of fC intermittently, but since it is a signal with fC and fC±△f as frequency components, when such a signal is input to the crystal oscillator 322, the frequency fC and the frequency fB It continues to oscillate. At this time, it must be noted that 3 and 4 are in phase with respect to not only the Δf modulation signal but also the carrier wave. For this reason, two NOR circuits are used in the fC oscillation loop. At this time, the series circuit of the vibrator and the capacitor is equivalently resonating with a pure resistance, and especially when there is no capacitor, the phase angle between the voltage VQ applied to the vibrator and the flowing current IQ is becomes almost 0. At this time, the power p input to the vibrator is p=
VQ・IQ COSφ≒VQ IQ (φ is the angle formed by VQ and IQ), p is large, and oscillation is possible even with a low Q value oscillator. 6 in FIG. 5 is a waveform after amplification of 4 and detection along the lower envelope of the waveform.

Next, FIG. 4 shows another specific embodiment. In the figure, 401, 402, 404 are inverters configured by connecting two inputs of a C-MOS NOR circuit, and resistors 406, 404, respectively, are similar to FIG.
By connecting 07,410, it serves as an amplifier. Resistors 406, 407, and 410 are feedback resistors for setting the operating point of the amplifier to 1/2 VDD. 403 NOR circuit and resistors 408 and 409
constitutes a modulator. Resistors 408 and 409 are each
This is a feedback resistor for setting the operating point of the NOR circuit 403 to 1/2VDD. 400 and 405 are each
It serves as a buffer amplifier for fC and △f. 4
21 diodes, 412, 413 resistors, 41
The detector consists of 8 capacitors. 41
Coupling capacitors 4,416, 417, and 419 cut direct current. 425 trimmer capacitor,
The capacitor 415, the resistor 411, and the vibrator 422 all constitute a feedback circuit of the fC oscillation circuit loop. Here, in order to establish a circuit correspondence with FIG. 2, the circuit blocks used in FIG. 2 are enclosed in FIG. 4 and given the same numbers as in FIG. 2. That is 20
is a saturation amplifier for fC component, 21 is a saturation amplifier for Δf component, 30 is a feedback circuit, 22 is an amplitude modulator, and 25 is a saturation amplifier for fC component.
is a detector, and 26 and 27 are buffer amplifiers.
Note that the NOR circuit 402 is a linear amplifier, and the NOR circuit 402 is a linear amplifier.
Circuits 401 and 404 are saturated amplifiers. The capacitor 420 constitutes a low-pass filter for removing the fC component present in the detected Δf component signal, and its operation is similar to that of the capacitor 320 in FIG.
It is similar to Crystal oscillator 322 and capacitor 4
It is well known that the oscillator formed by 13,425 has a potential phase difference of 180° between the input and the output. Therefore, the number of inverting amplifiers in the fC oscillation circuit loop needs to be an odd number. FIG. 5 shows the voltage waveforms of the main parts in the figure.

The operation of such a circuit will be explained with reference to FIG.
When the crystal oscillator 422 starts oscillating the two frequencies fC and fB, an output signal containing fC and fC±△f as shown in FIG. 5 4' is generated at the output terminal 4'. ing. This output signal is
A linear amplifier formed by a NOR circuit 402 and a feedback resistor 407 performs linear inversion amplification.
The output signal of such a NOR circuit is saturated and inverted amplified by a saturation amplifier formed by a NOR circuit 401 and a feedback resistor 406, and the output signal of FIG.
This occurs at the output of the NOR circuit 401. On the other hand, the output signal of the NOR circuit 402 which linearly inverts and amplifies the output signal of
13, is input to a detector formed by a capacitor 418. This detector detects the envelope of the output signal of the NOR circuit 402 at a position corresponding to the upper envelope of the output signal of the output terminal 4', and outputs the signal at the output terminal 5 as shown in FIG.
f signal 5 is generated. The output signal 5 shown in FIG. 5 contains a slight frequency component of fC. The output signal of this 5 is connected to the NOR circuit 404 and the feedback resistor 4.
The signal is inverted and amplified by the saturation amplifier formed by 10, and the signal 2 in FIG. 5 is outputted to the output terminal 2. At this time, the fC component is cut off by the low-pass filter action of the capacitor 420 mentioned above. 1
The frequency fC signal of 2 and the frequency △f signal of 2 are NOR
It is input to the amplitude modulator formed by the circuit 403 and the feedback resistors 408 and 409, and the fifth
Generates output signal 3 shown in the figure. Similar to FIG. 3, the vibrator 422 continues to oscillate due to this output signal 3. The amplitude modulator shown in FIG. 4 has a feedback resistor connected to each of its two inputs, but in FIG. 2 only one of the feedback resistors is connected. This is because the operating point becomes more stable and the oscillation circuit has better characteristics when the feedback resistor is provided, but since the output signal 1 is a saturated amplified signal, the operation can be performed even without the feedback resistor. The difference from FIG. 3 is that 3 and 4 are in phase with respect to Δf, but are out of phase with respect to fC. As a result, the power p input to the vibrator is p=VQ・IQ
cosΦ is the angle between the vibrator applied voltage VQ and the current IQ.
Close to 90°, p is small. Therefore, it is suitable for an oscillation circuit of a high Q value resonator.

Next, FIG. 6 is obtained by observing the frequency component of voltage 1 in FIGS. 3 and 4 with a spectrum analyzer. , the three observed frequency components
The relative energy intensity (spectrum) of fC, fC±△f is that the carrier wave fC is the largest, and the wave measurement component fC−
Comparing Δf and fC+Δf, fC−Δf is several dB larger. If it is an original amplitude modulated wave, its voltage waveform is e=E(1+k cos pt) cos wt, which is decomposed into e=A cos wt+kA/2{cos(w+p)t +cos(W-p)t } is. The wave measurement spectrum has the same value at (1/2 kA) ² . The energy level L ₂ of fC−△f in Fig. 5 is
The reason why the energy of fC + △f is larger than L ₁ (L ₂ > L ₁ ) is that one of the two resonance frequencies fC and fB of the used vibrator matches the carrier wave, and fB = fC - △f. , fC-Δf component increases.

Next, FIG. 7 shows a specific example different from the modulator used in FIGS. 3 and 4. In the C-MOS, NOR, or NAND modulators described above, the carrier wave is completely clamped to the logic level of 1 or 0, so a charging/discharging spike voltage appears on the output side of the oscillator, making the oscillator a mere capacitor. , undesirable. 7th
The circuit shown in the figure can improve this, with a transmission gate of 70, resistors of 72 and 73, and a
Consists of 1 C-MOS inverter. resistance 72
serves as a DC bias resistor and a carrier wave bypass resistor for controlling the degree of modulation. Also, resistance 7
3 is the power supply voltage at the gate of inverter 71
This is to set the operating point of the amplifier by applying a DC bias of 1/2 of VDD. Due to the presence of the resistor 73, it takes 70 seconds until the oscillation of the modulation signal △f rises.
The transmission gate is in a semi-conducting state, allowing the carrier wave fC to rise. Further, this circuit is compatible with the amplitude modulator made of the NOR circuit shown in FIGS. 3 and 4.

As described above, if the oscillation circuit according to the present invention is used, for example, two of the vibrators used in the oscillation circuit can be used.
The temperature characteristics of the two resonant frequencies are the carrier frequency fC
If the temperature change in one mode fB is extremely small, and the temperature change in frequency of one mode fB is extremely large, then
It is possible to use fC as a frequency standard and the other as a temperature detection signal, making it possible to apply it to temperature sensors. Furthermore, the above-mentioned circuit can be integrated into one chip.
It is also possible to convert it into an IC, creating a small, low-power temperature sensor.

As explained above, the oscillation circuit according to the present invention is
This has the effect of reliably separating the frequencies of two vibration modes existing at frequencies close to one vibrator with a simple circuit configuration.

[Brief explanation of the drawing]

FIG. 1 shows a conventional oscillation circuit that can oscillate at two oscillation frequencies simultaneously. Figure 2 a, b
1 is a principle diagram of an oscillation circuit that performs the same function as that in FIG. 1 according to the present invention. 3 and 4 are diagrams showing specific embodiments of the oscillation circuit according to the present invention. FIG. 5 is a diagram showing voltage waveforms of main parts in FIGS. 2a, 3, and 4. FIG. 6 is a diagram showing the frequency spectrum of the voltage waveforms 4 and 4' in FIGS. 3 and 4. FIG. FIG. 7 is a diagram showing another specific embodiment of an amplitude modulator that can be used in the amplitude modulator of the oscillation circuit according to the present invention. 16, 23, 322, 422... vibrator with two resonance frequencies, 22... amplitude modulator, 20,
21...Amplifier, 24...Detector.

Claims (1)

[Claims] 1. In an oscillation circuit that simultaneously outputs two frequency signals, one vibrator 23 vibrates in a first vibration mode with a first frequency fC and a second vibration mode with a second frequency fB, An amplitude modulator 22 that modulates at a third frequency Δf, feedback circuits 23 and 24 connected to the amplitude modulator and composed of the vibrator and a passive circuit element 24, and detecting the output wave of the feedback circuit to generate the third frequency Δf. a detector 25 that outputs a detected output signal with f; a first amplifier 20 that directly amplifies the output wave of the feedback circuit and outputs it as the first frequency to the amplitude modulator; and a first amplifier 20 that amplifies the output signal of the detector. a second amplifier 21 that outputs a modulation signal to the amplitude modulator, the first frequency and the second frequency are close to each other, and the third frequency Δf serving as the modulation signal is equal to the first frequency. An oscillation circuit characterized in that the frequency is a difference between fC and the second frequency fB, and one output signal of the oscillation circuit is the first frequency fC, and the other output signal is the third frequency Δf. .