JP2016158199A - Digital temperature compensated oscillator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a digital temperature compensated oscillator, capable of precise temperature compensation over a wide temperature range.SOLUTION FOR THE PROBLEM: A digital temperature compensated oscillator includes a voltage controlled oscillator 1, a temperature-sensitive oscillator 2, a temperature digital data generation part 3, a temperature compensation circuit 4, a ΔΣ modulator 5, a passive multi-stage LPF 6, and a D/A converter 7. The temperature compensation circuit 4 preliminarily stores residual error correction data, which is a difference between correction data for intended temperature compensation and polynomial approximated correction data approximated with polynomial approximation, in a memory by extracting the residual error correction data by sampling at a discrete temperature within an operating temperature range, calculates residual error correction data corresponding to temperature digital data from a temperature digital data generation part 3 with a digital filter, based on residual error correction data corresponding to four anteroposterior sampling points of the temperature digital data, and obtains temperature compensation digital data by adding the calculated data to polynomial approximation correction data corresponding to the temperature digital data.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a digital temperature-compensated oscillator that compensates for frequency fluctuations due to temperature changes based on digitized temperature information.

  In recent years, as mobile environments such as high-function mobile terminals, such as smartphones, become small cells such as femtocells, picocells, and microcells, demands for the frequency stability of these base station side clocks have become stricter. In general, a temperature compensated oscillator is used as a clock for an application that requires high frequency stability.

  Various temperature-compensated oscillators have been known, but as a digital temperature-compensated oscillator, a temperature sensor with a large variation in oscillation frequency due to temperature is used, and external temperature change information detected by the temperature sensor is digitally controlled. A digital temperature code is generated by a circuit, and a digital temperature compensation code generated by a digital temperature compensation circuit is converted into an analog signal by a D / A converter based on the temperature digital code to generate a temperature compensation voltage (frequency control voltage). A configuration is known in which the temperature compensation voltage is input to a voltage controlled oscillator to control the oscillation frequency of the voltage controlled oscillator. And the temperature compensation digital data for compensating for the change of the oscillation frequency generated by the voltage controlled oscillator with respect to the environmental temperature is correction data for canceling the change of the oscillation frequency due to the temperature drift of the voltage controlled oscillator. Conventionally, it is generally obtained by a polynomial approximation formula.

  However, the correction data based on this polynomial approximation equation does not match the original correction data to be temperature compensated, that is, the actually measured correction data, and has a difference that cannot be corrected (hereinafter referred to as a residual). For example, as shown in FIG. 13, in the case where the ambient temperature is in the range of −40 to 100 ° C., the polynomial approximation correction data represented by a solid line approximated by a 10th order polynomial is used to correct the residual difference with respect to the correction data obtained by actual measurement. In addition, it is necessary to add a residual correction amount represented by a dotted line which is a difference between the actually measured correction data and the polynomial approximation correction data. In FIG. 13, the display range for the polynomial approximation correction is ± 25,000 ppb, and the display range for the residual correction is one-tenth ± 2,500 ppb (this value is a frequency of 25 ° C. Normalized value).

  In order to obtain the residual correction amount, correction data is obtained by actual measurement every certain temperature interval, and when the environmental temperature does not correspond to the correction data obtained by actual measurement, the correction is performed based on the correction data obtained by actual measurement. In general, the residual correction data is obtained. Conventionally, as a measure for calculating the residual correction data by calculation, a linear function is used as an approximate expression, spline interpolation using a higher-order polynomial of 2nd order or higher is used, or a least square method is used. It has been proposed to use the approximate curve utilized (Patent Document 1).

JP 2013-98865 A

  However, according to this conventional proposal, when a linear function is used as an approximate expression, there is an inconvenience that residual correction data cannot be calculated with high accuracy unless the measurement interval is reduced. In addition, when using spline interpolation using higher-order polynomials of the second or higher order, the coefficients used for the calculation are different in each temperature section, so a large number of non-volatile memories for storing the coefficients are required. There is an inconvenience. For example, in the case of spline interpolation using a quadratic polynomial, if the number of temperature intervals is M, the number of words in the nonvolatile memory is 3M, and in the case of spline interpolation using a cubic polynomial, 4M is required. When an approximate curve using the least square method is used, there is an inconvenience that residual correction data cannot be calculated with high accuracy.

  An object of the present invention is to provide a digital temperature-compensated oscillator that eliminates these disadvantages and enables highly accurate temperature compensation over a wide temperature range.

  In order to achieve the above object, a digital temperature compensated oscillator according to claim 1 of the present invention generates an oscillation frequency (first oscillation frequency) whose frequency changes in response to a change in environmental temperature, and the first voltage is controlled by a control voltage. A voltage-controlled oscillator that generates an oscillation signal that can control the oscillation frequency of the oscillation circuit, and an oscillation frequency that has a larger rate of change than the first oscillation frequency of the voltage-controlled oscillator with respect to a change in the environmental temperature (second And a digital information obtained by comparing the first oscillation frequency generated by the voltage-controlled oscillator and the second oscillation frequency generated by the temperature-sensitive oscillator. A temperature digital data generation unit that generates an environmental temperature as temperature digital data calculated based on the temperature digital data. A temperature compensation circuit for generating temperature compensation digital data for compensating for a change of the first oscillation frequency generated by the voltage controlled oscillator with respect to the environmental temperature, and a passive multistage low-pass filter (hereinafter referred to as LPF) comprising a ΔΣ modulator and passive elements. D / A converter configured to convert the temperature compensation digital data into the control voltage represented by an analog voltage, and the temperature compensation circuit includes a correction data and a polynomial approximation that should originally be temperature compensated. The difference from the polynomial approximation correction data approximated in step (3) is used as residual correction data, sampled at discrete temperatures within the operating temperature range of the digital temperature compensated oscillator, and the sampled residual correction data is extracted from the temperature compensation circuit. Temperature digital data stored in memory and obtained from the temperature digital data generator Data, and residual correction data corresponding to four or more sampling points before and after the temperature digital data, the residual correction data in the temperature digital data is calculated by a digital filter, and the result is calculated as the temperature digital data. It is added to the polynomial approximation correction data corresponding to the data to obtain temperature compensation digital data.

  Similarly, in order to achieve the above object, a digital temperature compensated oscillator according to claim 2 of the present invention uses a polyphase FIR filter as the digital filter in the configuration of the invention of claim 1, and also uses polyphase FIR filter interpolation (hereinafter referred to as the following). The residual correction data is calculated by combining linear interpolation with FIR interpolation.

  Similarly, in order to achieve the above object, a digital temperature compensated oscillator according to claim 3 of the present invention, in the configuration of claim 2, the operating environment temperature range is divided into a high temperature side and a low temperature side, and a sampling point for each part. Are sequentially set from the maximum temperature value and the minimum temperature value toward the opposite side, and are calculated by different FIR filter interpolation formulas for each part.

  According to the digital temperature-compensated oscillator according to claim 1 of the present invention, it is possible to deal with the residual difference that cannot be dealt with by the temperature correction data by polynomial approximation with high accuracy, and thus temperature compensation in a wide temperature range is possible. Has the effect of becoming.

  According to the digital temperature compensated oscillator of claim 2 of the present invention, in addition to the effect of the invention of claim 1, the use of the polyphase FIR filter has the effect that the residual difference can be efficiently calculated. .

  According to the digital temperature compensated oscillator of claim 3 of the present invention, in addition to the effect of the invention of claim 2, one sampling point is added to each of the maximum temperature side and the minimum temperature side of the operating environment temperature range. This is sufficient, and it is not necessary to add a plurality of sampling points, so that an effect of being efficient can be obtained.

1 is a block diagram showing the overall configuration of an embodiment of a digital temperature compensated oscillator according to the present invention. The block diagram of a temperature compensation circuit similarly. The graph which similarly shows the example of a sampling of residual correction data. Similarly, the data No. of the residual correction data sampled. The schematic explanatory drawing of the look-up table which shows the correspondence with temperature. The graph which shows typically the relationship between the temperature digital data similarly output from the temperature digital production | generation part, and residual correction data. The graph which shows typically the residual correction data by FIR interpolation and linear interpolation similarly. The table | surface which shows an example of the coefficient for FIR interpolation calculation similarly. The graph which similarly shows the coefficient of FIG. The block diagram which similarly shows an example of a FIR interpolation residual correction arithmetic circuit. The graph which shows typically an example which requires two extra sampling points on the high temperature side. The graph which shows typically an example in which one extra sampling point is enough on the high temperature side. The graph which shows typically an example in which one extra sampling point is enough on the low temperature side. Similarly, a graph showing the relationship between the polynomial approximation correction and the residual correction.

  First, the overall configuration of a digital temperature compensated oscillator according to the present invention will be described with reference to FIG. 1 of the accompanying drawings. The digital temperature compensated oscillator is a D / A converter 7 comprising a voltage controlled oscillator 1, a temperature sensitive oscillator 2, a temperature digital data generator 3, a temperature compensation circuit 4, a ΔΣ modulator 5, and a passive 4-stage LPF 6. It has.

  The voltage controlled oscillator circuit 12 of the voltage controlled oscillator 1 has a known configuration and has an external vibrator 11. The voltage-controlled oscillator 1 generates an oscillation frequency (first oscillation frequency) whose frequency changes with a change in environmental temperature, and the first oscillation frequency by a control voltage output from the passive 4-stage LPF 6. An oscillation signal that can be controlled is generated. The generated oscillation signal having the first oscillation frequency is output to the buffer 8 and is input to the first counter 31 of the temperature digital data generation unit 3 via the frequency dividing circuit 9, while from the buffer 8. It is output to the outside via the buffer 10.

  The temperature-sensitive oscillator 2 is a ring oscillator formed by connecting an odd number of inverters in a ring shape, and the rate of change is greater than the first oscillation frequency generated by the voltage-controlled oscillator 1 with respect to changes in environmental temperature. An oscillation frequency (second oscillation frequency) is generated. The generated oscillation signal having the second oscillation frequency is output to the second counter 32 of the temperature digital data generation unit 3.

  The temperature digital data generation unit 3 includes the counters 31 and 32 and a control circuit / temperature conversion circuit 33. The counters 31 and 32 are controlled by a control signal of the control circuit. The counter 31 counts the first oscillation frequency generated by the voltage-controlled oscillator 1, and the counter 32 counts the second oscillation frequency generated by the temperature-sensitive oscillator 2. Each oscillation frequency is compared by the control circuit / temperature conversion circuit 33, the environmental temperature is calculated as digital data from the oscillation frequency of the temperature-sensitive oscillator 2, and temperature digital data is generated. The temperature digital data is output from the control circuit / temperature conversion circuit 33 to the arithmetic circuit 41 of the temperature compensation circuit 4.

  The temperature compensation circuit 4 includes an arithmetic circuit 41 and a memory 42, and generates temperature compensation digital data for compensating for the change of the first oscillation frequency generated by the voltage controlled oscillator 1 with respect to the environmental temperature. This temperature compensated digital data is digital data that is a correction for canceling a change in the oscillation frequency due to temperature drift of the voltage controlled oscillator 1, and this polynomial approximate correction data is obtained by a polynomial approximate expression obtained by a polynomial approximate expression. Then, residual correction data obtained by FIR interpolation and linear interpolation is added to the residual difference that remains uncorrected and is obtained.

  As shown in FIG. 2, an arithmetic circuit 41 to which temperature digital data of the temperature compensation circuit 4 is input includes a polynomial correction arithmetic circuit 41a for calculating polynomial approximation correction data, and cascade-connected FIR interpolation residuals for obtaining residual correction data. The difference correction calculation circuit 41b and the linear interpolation residual correction calculation circuit 41c are used. The polynomial approximation correction data obtained by the circuit 41a and the residual correction data obtained by the circuits 41b and 41c are added by an adder 41d. And output as temperature compensation digital data. On the other hand, the memory 42 includes a polynomial coefficient memory 42a that stores coefficients necessary for calculating polynomial approximation expressions, a residual data memory 42b that stores residual data necessary for FIR interpolation, and FIR interpolation. And an FIR coefficient memory 42c in which necessary coefficients are stored.

Here, generation of temperature compensation digital data will be described in detail. For example, the polynomial approximation correction data in the operating temperature range of the voltage controlled oscillator 1 in the range of −40 to 100 ° C. is a 10th order polynomial if the target value of error is about ± 200 ppb, TC ₁₀ = A ₀ + A ₁ (T−T ₀ ) + A ₂ (T−T ₀ ) ² + A ₃ (T−T ₀ ) ³ + A ₄ (T−T ₀ ) ⁴ + A ₅ (T−T ₀ ) ⁵ + A ₆ (T− T ₀ ) ⁶ + A ₇ (T−T ₀ ) ⁷ + A ₈ (T−T ₀ ) ⁸ + A ₉ (T−T ₀ ) ⁹ + A ₁₀ (T−T ₀ ) ¹⁰ (see FIG. 13) ). Here, TC ₁₀ is polynomial approximate correction data output from the polynomial correction arithmetic circuit 41a, A _{0 to} A ₁₀ are coefficients stored in advance in a polynomial coefficient memory, and (T−T ₀ ) is a control circuit / temperature conversion. This is temperature digital data output from the circuit 33.

  FIG. 3 shows an example in which residual correction data (see FIG. 13) is sampled and extracted at a temperature interval of 4 ° C. The sampled residual correction data is stored in advance in the residual data memory 42b. ing. FIG. 4 shows the data No. of the residual correction data at 4 ° C. intervals. The correspondence relationship with temperature is shown as a look-up table (hereinafter referred to as LUT). In the residual data memory 42b, the data No. Is stored as an address and residual correction data is stored as data. Is associated with temperature. In FIG. 4, i, n, and N are all integers, and the temperature interval ΔT is {T (N) −T (0)} / N. In this example, ΔT = 4 (° C.).

  When the temperature digital data T output from the control circuit / temperature conversion circuit 33 is a temperature corresponding to the address of the residual correction data, the residual correction data is stored in the residual data memory 42b. For example, if the temperature digital data T is between the temperature T (n) and the temperature T (n + 1) in the LUT of FIG. 4, it is necessary to obtain residual correction data by calculation. FIG. 5 schematically shows the relationship between the temperature digital data and the residual correction data shown in FIG. 4 sampled in advance in the range of the temperature digital data T (n−3) to T (n + 3). T) is the residual correction data to be obtained.

  In this embodiment, the residual correction data y (T) shown in FIG. 5 is converted into four temperature increments T (n−1), T (n), and T (n + 1) before and after the temperature digital data T is sandwiched. , T (n + 2) is calculated by 4-phase FIR interpolation, and the result is further calculated by linear interpolation, in other words, by linear function interpolation. More specifically, as shown in FIG. 6, FIR interpolation is performed to divide the portion between T (n) and T (n + 1) before and after the temperature digital data T in FIG. 5 into four to obtain three interpolation points. Here, the data No. corresponding to the interpolation point. Is n + k / 4, temperature digital data is T (n + k / 4), residual correction data is y (n + k / 4), and k is an integer satisfying 0 <k ≦ 4. Since k (4) is T (n + 1), three interpolation points T (n + 1/4), T (n + 1/2), and T (n + 3/4) may be obtained as interpolation points.

  Then, residual correction data y (n + 1/4), y (n + 1/2), y () of the three interpolation points T (n + 1/4), T (n + 1/2), and T (n + 3/4), respectively. n + 3/4), the residual correction data of the temperature digital data before and after the temperature digital data T is located. In the example of FIG. 6, the residual correction data of T (n + 1/2) and T (n + 3/4) is a straight line. Interpolation is performed to obtain residual correction data y (T) corresponding to the temperature digital data T. Therefore, in this example, the residual correction data y (T) can be obtained without obtaining the residual correction data corresponding to the interpolation point T (n + 1/4).

Specifically, as shown in FIG. 5, the temperature digital data T (n−1), T (n), T (4) before and after the temperature digital data T, that is, two low temperature sides and two high temperature sides are inserted. n + 1), T (n + 2) residual correction data y (n−1), y (n), y (n + 1), y (n + 2) and 4-phase polyphase FIR filter coefficients c (i, j) The FIR interpolation data is obtained by substituting into the following equation.
y (n + 1/4) = c (3,1) * y (n-1) + c (3,2) * y (n) + c (3,3) * y (n + 1) + c (3,4) * y (n + 2)
y (n + 1/2) = c (2,1) * y (n-1) + c (2,2) * y (n) + c (2,3) * y (n + 1) + c (2,4) * y (n + 2)
y (n + 3/4) = c (1,1) * y (n-1) + c (1,2) * y (n) + c (1,3) * y (n + 1) + c (1,4) * y (n + 2)
y (n + 1) = c (0,1) * y (n-1) + c (0,2) * y (n) + c (0,3) * y (n + 1) + c (0,4) * y (n + 2 )
The above arithmetic expressions correspond to the FIR filter 3 (phase 3), the FIR filter 2 (phase 2), the FIR filter 1 (phase 1), and the FIR filter 0 (phase 0) in order from the top. The coefficient c (i, j) is obtained in consideration of the filter characteristics of the cutoff frequency, the pass band, and the stop band in the temperature region of the residual correction data. In the 16-tap polyphase FIR filter, Examples used in the above arithmetic expressions are shown in FIGS.

  FIG. 9 calculates residual correction data y (n + 1/2) and y (n + 3/4) at interpolation points T (n + 1/2) and T (n + 3/4) before and after the above-described temperature digital data T is located. An example of a circuit configuration for doing this will be shown. Here, the FIR filter A (phase A) and the FIR filter B (phase B) indicate two FIR filters selected from the FIR filters 0 to 3 (phases 0 to 3). In this example, the FIR filter 1 (phase 1) corresponds to FIR filter 2 (phase 2). When the temperature digital data T is input, the interpolation temperature position calculation circuit 411 calculates a temperature position (position of the temperature T in FIG. 5) corresponding to the data T, and a residual data address calculation circuit 412 and an FIR coefficient address calculation circuit. It outputs to 413. The residual data address calculation circuit 412 receives a data No. for reading out residual correction data necessary for interpolation from the residual data memory 42b. The address corresponding to is calculated. Meanwhile, at the same time, the FIR coefficient address calculation circuit 413 calculates an address for reading the FIR filter coefficients of two phases before and after the temperature position.

  Based on the results of these address calculations, corresponding residual correction data y (n−1), y (n), y (n + 1), and y (n + 2) are called from the residual data memory 42b. Corresponding phase 1 and phase 2 FIR coefficients are called from the FIR coefficient memory 42c. Then, interpolation data y (n + 1/2) and y (n + 3/4) are respectively calculated in the two phases corresponding to these FIR coefficients based on the above-described arithmetic expressions, and each interpolation data y (n + 1/2), y is calculated. (n + 3/4) is added by the adder 416 and output to the linear interpolation residual correction calculation circuit 41c (see FIG. 2).

  The linear interpolation residual correction calculation circuit 41c shown in FIG. 2 corrects the residual corresponding to the temperature digital data T by a linear approximation expression based on the two interpolation data y (n + 1/2) and y (n + 3/4). Data y (T) is obtained, and this residual correction data y (T) is output to the adder 41d. The adder 41d adds the polynomial approximation correction data obtained by the polynomial correction calculation circuit 41a and the residual correction data y (T) to generate temperature compensation digital data.

  As shown in FIG. 1, the ΔΣ modulator 5 to which the temperature compensation digital data generated by the temperature compensation circuit 4 as described above is input includes a multiplier 51, adders 52a, 52b, 52c, a delay circuit 53a, 53b and 53c, multipliers 54a and 54b, a quantizer 55, and a PWM modulator 56. A dither signal is input to the adder 52b, and this dither signal is composed of lower-order bit data of the output of the counter 32 that has received the output of the temperature-sensitive oscillator 2. For example, if the output of the counter 32 that receives the output of the temperature-sensitive oscillator 2 is 18-bit data, it is preferable that the lower 4-bit data is the dither signal.

  The quantizer 55 is a multi-value quantizer with three or more values. For example, a quantized PDM signal quantized with four levels of “00”, “01”, “10”, and “11” is used. And output to the PWM modulator 56 and to the adder 52c. The PWM modulator 56 outputs a binary PWM signal with a multi-level pulse width of 3 levels or more. For example, if the quantizer 55 has 4 values (4 levels), “0” is similarly applied. ”,“ 1 ”,“ 2 ”,“ 3 ”of four levels of pulse width, converted into a binary PWM signal having a pulse width of a level corresponding to the level of the input PDM signal, and passive type Output to the 4-stage LPF 6. On the other hand, the adder 52c receives a signal input to the adder 52b as well as the output of the quantizer 55, and the adder 52c outputs a quantization error by the quantizer 55.

  The delay circuits 53a, 53b, and 53c delay the quantization error by one cycle, two cycles, and three cycles. The output of the delay circuit 53a is multiplied by a predetermined coefficient by the multiplier 54a and then input to the adder 52a. The output of the delay circuit 53b is multiplied by a predetermined coefficient by a multiplier 54b and input to the adder 52a. On the other hand, the output of the delay circuit 53c is directly input to the adder 52a. In addition to these inputs, the output of the multiplier 51 is input to the adder 52a, and the inputs are added and output to the adders 52b and 52c.

  The passive four-stage LPF 6 is configured by four stages of LPFs composed of a resistance element and a capacitance element. For example, the total resistance value is 1 GΩ, the first-stage resistance element is 700 MΩ, and the other three resistance elements are each set to 100 MΩ. The capacitance value of each capacitive element is set, for example, such that the sum of the capacitance values is 100 pF, the capacitive element of the fourth stage as the final stage is 70 pF, and the other three capacitive elements are each set to 10 pF. Yes. In this way, if the first-stage resistance value and the final-stage capacitance value are set larger than other resistance values or capacitance values, the amount of attenuation in the low frequency region can be increased. The PWM signal is converted into an analog signal by passing through the passive four-stage LPF 6, and this analog signal is input to the voltage controlled oscillator circuit 12 of the voltage controlled oscillator 1 as a control voltage signal.

Next, the operation of the above-described digital temperature compensation oscillator will be described.
When the voltage controlled oscillator 1 generates and outputs an oscillation signal having a frequency (first oscillation frequency) corresponding to the environmental temperature, the oscillation signal is output from the buffer 8 to the external device via the buffer 10. Input from the buffer 8 to the frequency divider 9. The frequency divider 9 divides the input oscillation frequency and outputs it to the counter 31. On the other hand, the temperature-sensitive oscillator 2 generates an oscillation signal having a frequency (second oscillation frequency) having a change rate larger than the first frequency with respect to the environmental temperature, and outputs the oscillation signal to the counter 32.

  The oscillation frequencies counted by the counters 31 and 32 are compared by the control circuit / temperature conversion circuit 33. The control circuit / temperature conversion circuit 33 compares the digital information obtained by the comparison. The ambient temperature is generated as temperature digital data, and this temperature digital data is output to the arithmetic circuit 41.

  The arithmetic circuit 41 including the polynomial correction arithmetic circuit 41a, the FIR interpolation residual correction arithmetic circuit 41b, and the linear interpolation residual arithmetic circuit 41c is preliminarily configured from the input temperature digital data, as described above, in the polynomial coefficient memory 42a, the residual Calculations are made based on the respective arithmetic expressions, data and coefficient values stored in the memory 42 comprising the difference data memory 42b and the FIR coefficient memory to compensate for changes in the oscillation frequency of the voltage controlled oscillation circuit 12 with respect to the environmental temperature. Temperature compensation digital data is generated and output to the ΔΣ modulator 5.

  The temperature compensated digital data input to the ΔΣ modulator 5 is converted into a four-level (4-level) PDM signal by the quantizer 55, and further converted into a binary PWM signal by a PWM modulator 56 with a 4-level pulse width. Is done. The PWM signal is converted into an analog signal by a passive four-stage LPF 6 so that the temperature compensated digital data is converted into a control voltage represented by an analog voltage and input to the voltage controlled oscillation circuit 12 for voltage control. The oscillation frequency of the type oscillator 1 is controlled.

  In the above-described embodiment, since the use environment temperature range of −40 ° C. to 100 ° C. is sampled at intervals of 4 ° C., which is a divisor of 140, the lowest temperature value in the use environment temperature range is −40 ° C. Even if the sampling points are determined sequentially toward the low temperature side from the maximum temperature value of 100 ° C., both temperature values become sampling points. For this reason, when the temperature digital data T is between the lowest temperature value or the highest temperature value and the previous sampling point, it corresponds to the temperature data T by FIR interpolation using four residual correction values. In order to obtain the residual correction data to be obtained, it suffices if there is a sampling point that is one point ahead of the operating environment temperature range on either the low temperature side or the high temperature side. That is, in FIG. 5, if T (n + 1) is the maximum temperature value, T (n + 2) is sufficient on the high temperature side, and T (n + 3) or more is unnecessary. If T (n) is the lowest temperature value, T (n-1) is sufficient on the low temperature side, and T (n-2) or less is unnecessary.

  However, when the operating environment temperature range of −40 ° C. to 100 ° C. is sampled at intervals of 3 ° C. or 8 ° C. which are not divisors of 140, −40 ° C., which is the lowest temperature in the operating environment temperature range, is set as the sampling point. Then, on the high temperature side, if it is 3 ° C. intervals, 98 ° C., and if it is 8 ° C., 96 ° C. is the highest temperature sampling point, and the highest temperature value 100 ° C. is not the sampling point (see FIG. 10). On the other hand, if the sampling temperature is 100 ° C, which is the highest temperature value in the operating environment temperature range, the lowest temperature sampling point is -38 ° C at 3 ° C intervals on the low temperature side, and -36 ° C at 8 ° C intervals. The minimum temperature of −40 ° C. is not a sampling point.

  For this reason, for example, as shown in FIG. 10, when the temperature digital data T is between the maximum temperature value in the operating environment temperature range that is not the sampling point and the sampling point T (n) immediately before, the four digital data T In order to obtain residual correction data corresponding to the temperature data T by FIR interpolation using the residual correction data, sampling points T (n + 1) and T (n + 2) up to two points after the maximum temperature value are required. . Therefore, it is necessary to set an extra sampling point and obtain the residual correction data from the actual measurement. In the above example, if the lowest temperature value is the sampling point, the range is -43 ° C to 104 ° C at 3 ° C intervals. In addition, since the range of −46 ° C. to 112 ° C. is an actual measurement range at 8 ° C. intervals, there is an inconvenience that the actual temperature range is unnecessarily widened.

  In order to eliminate this inconvenience, the lowest temperature value and the highest temperature value are used as sampling points, and the sampling points are determined from the sampling points toward the opposite side at a predetermined temperature interval, so that the temperature is almost within the operating environment temperature range. If the operating temperature range is -40 ° C to 100 ° C in the center, it is in a state of crossing each other around 25 ° C. For example, two sampling points on one side are provided on the other side from the center, and the sampling point on the other side is One is provided on one side from the center. On the high temperature side shown in FIG. 11, using the above-described arithmetic expression, interpolation data 0 equal to the residual correction data y (n + 1) is an expression corresponding to the FIR filter 0 (phase 0), and interpolation data 1 is the FIR filter 1 (Phase 1), the interpolation data 2 can be obtained by an expression corresponding to the FIR filter 2 (Phase 2), and the interpolation data 3 can be obtained by an expression corresponding to the FIR filter 3 (Phase 3).

On the other hand, on the low temperature side shown in FIG. 12, calculation is performed using the following arithmetic expression different from the above-described arithmetic expression.
y (n + 1) = c (0,4) * y (n-1) + c (0,3) * y (n) + c (0,2) * y (n + 1) + c (0,1) * y (n + 2 )
y (n + 1/4) = c (1,4) * y (n-1) + c (1,3) * y (n) + c (1,2) * y (n + 1) + c (1,1) * y (n + 2)
y (n + 1/2) = c (2,4) * y (n-1) + c (2,3) * y (n) + c (2,2) * y (n + 1) + c (2,1) * y (n + 2)
y (n + 3/4) = c (3,4) * y (n-1) + c (3,3) * y (n) + c (3,2) * y (n + 1) + c (3,1) * y (n + 2)
This calculation formula calculates FIR filter 0 (phase 0) for obtaining interpolation data 0 equal to residual correction data y (n), FIR filter 1 (phase 1) for obtaining interpolation data 1 and interpolation data 2 in order from the top. This corresponds to the FIR filter 2 (phase 2) and the FIR filter 3 (phase 3) for obtaining the interpolation data 3, respectively. The coefficient c (i, j) may be the one shown in FIGS.
For convenience of explanation, T (n + 1) is the highest temperature value in FIG. 11 and T (n) is the lowest temperature value in FIG.

  As described above, in the above-described embodiment, the highest temperature value and the lowest temperature value are both sampling points, the high temperature side arithmetic expression and the low temperature side arithmetic expression reverse the order of the coefficients of each phase, and By using different formulas in which the order corresponding to the phases 1 to 3 is reversed, the temperature digital data T is the highest temperature value in the operating environment temperature range, or the lowest temperature value and the sampling point T (n) immediately before that. Even when there is a gap, when the residual correction data corresponding to the temperature data T is obtained by FIR interpolation using the four residual correction data before and after, the extra sampling points outside the operating environment temperature range are also on the high temperature side. One is sufficient on the low temperature side.

Note that the present invention is not limited to the above-described embodiment. For example, the number of sampling points of residual correction data used for calculation in a digital filter is not limited to four before and after the temperature digital data is sandwiched. Further, the polynomial for obtaining the polynomial approximation correction data is not limited to the 10th order, and, for example, a 7th to 9th order polynomial has no problem in terms of accuracy.
Furthermore, the coefficients used in the FIR interpolation calculation formula are not limited to those described above. Furthermore, when using different arithmetic expressions on the high temperature side and the low temperature side, the position that bisects the operating environment temperature range is not limited to the central portion.

1 Voltage Control Oscillator 2 Temperature Sensitive Oscillator 3 Temperature Digital Data Generation Unit 4 Temperature Compensation Circuit 5 ΔΣ Modulator 6 Passive 4-Stage LPF
7 D / A Converter 11 Vibrator 12 Voltage Control Oscillation Circuit 31, 32 Counter 33 Control Circuit / Temperature Conversion Circuit 41 Operation Circuit 41a Polynomial Correction Operation Circuit 41b FIR Interpolation Residual Correction Operation Circuit 41c Linear Interpolation Residual Correction Operation Circuit 41d adder 42 memory 42a polynomial coefficient memory 42b residual data memory 42c FIR coefficient memory 51 multipliers 52a, 52b, 52c adders 53a, 53b, 53c delay circuits 54a, 54b multiplier 55 quantizer 56 PWM modulation 411 Interpolation temperature position calculation circuit 412 Residual data address calculation circuit 413 FIR coefficient address calculation circuit 414 FIR filter A calculation circuit 415 FIR filter B calculation circuit 416 Adder

Claims (3)

A voltage controlled oscillator that generates an oscillation frequency (first oscillation frequency) whose frequency changes in response to a change in environmental temperature and generates an oscillation signal capable of controlling the first oscillation frequency by a control voltage;
A temperature-sensitive oscillator that generates an oscillation frequency (second oscillation frequency) having a rate of change larger than the first oscillation frequency of the voltage-controlled oscillator with respect to a change in the environmental temperature;
The comparison is performed by calculating based on digital information obtained by comparing the first oscillation frequency generated by the voltage-controlled oscillator and the second oscillation frequency generated by the temperature-sensitive oscillator. A temperature digital data generator for generating the ambient temperature as temperature digital data,
A temperature compensation circuit for generating temperature compensation digital data for compensating for a change in environmental temperature of a first oscillation frequency generated by the voltage controlled oscillator based on the temperature digital data;
A D / A converter configured by a ΔΣ modulator and a passive multi-stage low-pass filter composed of passive elements and converting the temperature-compensated digital data into the control voltage represented by an analog voltage;
The temperature compensation circuit uses the difference between the correction data that should be compensated for temperature and the polynomial approximation correction data approximated by polynomial approximation as residual correction data, and is sampled at discrete temperatures within the operating temperature range of the digital temperature compensated oscillator. Based on the temperature digital data obtained from the temperature digital data generation unit and residual correction data corresponding to four or more sampling points before and after sandwiching the temperature digital data, extracted and stored in the memory in advance, Digital temperature compensation characterized in that residual correction data in the temperature digital data is calculated by a digital filter and the result is added to the polynomial approximation correction data corresponding to the temperature digital data to obtain temperature compensated digital data Oscillator.   2. The digital temperature compensation oscillator according to claim 1, wherein a polyphase FIR filter is used as a digital filter in the temperature compensation circuit, and residual correction data is calculated by combining linear interpolation with polyphase FIR filter interpolation. The operating environment temperature range is divided into high and low temperatures, and sampling points are set for each part in order from the highest temperature value and the lowest temperature value to the opposite side, and calculated with different FIR filter interpolation formulas for each part. 3. The digital temperature compensated oscillator according to claim 2, wherein:

Images