Digital calibration circuit for digital-to-analog converter based on random stepping of dynamic elements
By using a digital-to-analog converter digital calibration circuit with random stepping of dynamic components in Sigma Delta ADC, the in-band noise problem caused by DAC mismatch error is solved, the signal-to-noise ratio and algorithm efficiency are improved, and the occurrence of in-band tones is avoided. .
Patent Information
- Application Number
- CN202210753018.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-06-29
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2042-06-29
AI Technical Summary
The mismatch error of DAC in Sigma Delta ADC results in an increase in in-band noise, reducing the number of significant bits of the ADC. The existing DWA algorithms reduce efficiency when the input signal frequency is low or the amplitude is hourly, introducing tone.
Using a digital-to-analog converter digital calibration circuit based on random stepping of dynamic components, the stepping signal and component selection code are generated through adder, ADWA control unit, transcoder and ADWA logic unit to avoid periodic selection of components and reduce in-band noise.
The signal-to-noise ratio of the modulator is improved, the power of the in-band tone is avoided, the algorithm efficiency is improved, and the calibration capability of DAC mismatch error is enhanced.
Smart Images

Figure CN115001496B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the technical field of digital calibration, and in particular relates to a digital calibration circuit of a digital-to-analog converter based on random stepping of dynamic elements. Background Art
[0002] In recent years, in the fields of mobile communications, audio, biomedicine, etc., the requirements for data accuracy are getting higher and higher. As a bridge between the analog domain and the digital domain, ADC needs to be able to process weaker analog signals, so a higher precision ADC is needed. Sigma Delta ADC is the mainstream architecture of high-precision ADC at present. Sigma Delta ADC uses oversampling (OSR) and noise shaping technology to exchange a lower-bit quantizer for a higher effective number of bits. Increasing the oversampling rate, modulator order (L) and quantizer bit number can increase the ADC effective number of bits. However, the oversampling rate is limited by clock frequency and power consumption. Too high an order will introduce loop stability problems. Therefore, if you want to further increase the effective number of bits, you need to have a built-in multi-bit quantizer. When using a multi-bit quantizer, the feedback loop of the ADC must use a multi-bit DAC. The multi-bit DAC uses capacitors or current elements. Ideally, the value of each element is equal. Due to the error of the manufacturing process, there will be mismatch in the elements. When the input of an ordinary DAC is the same, the output is generated by the same element, and the mismatch error of the components used is the mismatch of the output level.
[0003] Sigma Delta ADC is mainly composed of loop filter, quantizer, DAC feedback loop to form a closed loop, in which the signal to be processed and the output signal of the feedback loop are input into the loop filter, the output of the loop filter is quantized by the quantizer, and the output of the quantizer is also the output of the ADC, which is fed back to the loop filter through the DAC feedback loop. Among them, the transfer function of the noise introduced by the DAC is:
[0004]
[0005] Among them, H(z) is the overall transfer function of the loop filter, U(z) is the ADC input signal, Y(z) is the quantizer output signal, D(z) is the DAC mismatch error, and Q(z) is the quantizer quantization error. The absolute value of the transfer function of D(z) is 1, which is equivalent to affecting the accuracy of the input signal and generating a large amount of noise in the band, thereby reducing the effective number of bits of the entire ADC. Therefore, the DAC mismatch error needs to be calibrated.
[0006] Currently, calibration is generally performed through additional digital circuits, such as the DWA algorithm, which has excellent performance and is easy to integrate. The traditional DWA algorithm selects components in a rotating manner so that each component is selected close to the number of times within a period of time, thereby producing a first-order shaping effect on the noise caused by DAC mismatch and reducing the in-band noise caused by mismatch. However, when the input signal frequency is low or the amplitude is small, the quantizer output may be periodic, resulting in the periodic selection of fixed components, which will reduce the efficiency of the DWA algorithm and introduce tone in the band. Summary of the invention
[0007] In order to solve the above problems existing in the prior art, the present invention provides a digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements. The technical problem to be solved by the present invention is achieved by the following technical solutions:
[0008] A digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements, comprising an adder A, an ADWA control unit, an adder B, a first register, a transcoder, and an ADWA logic unit; wherein,
[0009] The adder A is connected to the output end of the quantizer and is used to generate an end pointer of this cycle according to the output code of the quantizer;
[0010] The ADWA control unit is connected to the output end of the quantizer, and is used to generate a step signal S, an enable signal enable, and a starting element position signal SPR in an element selection cycle according to the output code of the quantizer;
[0011] The adder B is connected to the adder A and the ADWA control unit, and is used to output a start pointer according to the end pointer of the current cycle and the step signal S;
[0012] The first register is connected to the adder B and the adder A, and is used to store the start pointer output by the adder B in this cycle, and output it to the adder A in the next cycle to serve as the start pointer of the next cycle;
[0013] The transcoder is connected to the adder A, the adder B and the first register, and is used to generate an initial component selection code according to the end pointer and the start pointer;
[0014] The ADWA logic unit is connected to the ADWA control unit and the transcoder, and is used to shift the initial component selection code according to the enable signal enable and the starting component position signal SPR in the component selection cycle to obtain a final component selection code.
[0015] In one embodiment of the present invention, the adder A is a modulo 16 adder, which is specifically used to add the output code of the quantizer of this period to the start pointer of this period to obtain the end pointer of this period, and each time the accumulation reaches 16, the carry end is set to logic "1" to control the working mode of the transcoder, and then it is reset to zero and starts counting again.
[0016] In one embodiment of the present invention, the adder B is a modulo 16 adder, which is specifically used to add the end pointer of the current cycle and the carry of the adder A according to the step signal S, and send them to the register to determine the position of the first element selected in the next cycle, and return to zero and restart counting each time the accumulation reaches 16.
[0017] In one embodiment of the present invention, the ADWA control unit includes a first modulo 16 adder, a second register, a NOT logic gate, a first AND logic gate, a second AND logic gate, a second modulo 16 adder and a third register; wherein,
[0018] The input end of the first modulo 16 adder is connected to the quantizer output end of the DAC to be calibrated and the output end of the second register; the output end of the first modulo 16 adder is respectively connected to the input end of the second register, the input end of the first AND logic gate and the input end of the second AND logic gate;
[0019] The input end of the NOT logic gate is connected to the quantizer output end of the DAC to be calibrated, and the output end of the NOT logic gate is connected to the other input end of the first AND logic gate;
[0020] The output end of the first AND logic gate is connected to the other input end of the second AND logic gate and the input end of the second modulo-16 adder, and at the same time serves as the first output end of the ADWA control unit to output a carry signal;
[0021] The output end of the second AND logic gate serves as the second output end of the ADWA control unit to output an enable signal;
[0022] An output end of the second modulo-16 adder is connected to an input end of the third register;
[0023] The output end of the third register is connected to the other input end of the second modulo-16 adder, and simultaneously serves as the third output end of the ADWA control unit to output the starting element position signal of this cycle.
[0024] In one embodiment of the present invention, the transcoder includes two binary to thermometer code transcoders and a plurality of logic gate units; wherein,
[0025] The two binary to thermometer code transcoders are used to transcode the end pointer and the start pointer respectively to obtain corresponding thermometer codes;
[0026] Each logic gate unit includes an XOR logic gate, a NOT logic gate and a 2-to-1 selector, which are respectively used to output selection codes of corresponding components.
[0027] Beneficial effects of the present invention:
[0028] The digital calibration circuit of the digital-to-analog converter based on random stepping of dynamic elements provided by the present invention not only has a first-order shaping effect on the in-band noise, thereby improving the signal-to-noise ratio of the modulator; it also selects the starting element through the stepping signal, so that when the input signal is periodic, the periodic selection of elements can be avoided, thereby moving the power of the in-band tone to the outside of the band, reducing the in-band tone without sacrificing the signal-to-noise ratio, thereby improving the efficiency of the algorithm.
[0029] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS
[0030] Figure 1 This is the schematic diagram of the traditional DWA algorithm;
[0031] Figure 2 is a schematic diagram of an ADWA algorithm provided by an embodiment of the present invention;
[0032] Figure 3 It is a schematic diagram of the structure of a digital calibration circuit of a digital-to-analog converter based on random stepping of dynamic elements provided by an embodiment of the present invention;
[0033] Figure 4 is a circuit structure diagram of an ADWA control unit provided by an embodiment of the present invention;
[0034] Figure 5 is a circuit structure diagram of a transcoder provided by an embodiment of the present invention;
[0035] Figure 6 The figure is a working principle diagram of a digital calibration circuit of a digital-to-analog converter based on random stepping of dynamic elements provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0036] The present invention is further described in detail below with reference to specific embodiments, but the embodiments of the present invention are not limited thereto.
[0037] Embodiment 1
[0038] See also Figure 1 , Figure 1This is the principle diagram of the traditional DWA algorithm. The traditional DWA will use the next element of the last element used to generate the output of this cycle as the first element that can be selected in the next cycle; Figure 1 As shown, the first cycle input is 2, using component U 0 To U 1 , the second cycle input is 4, then use component U 2 To U 5 ; and use 16 elements as a selection cycle, after each element is selected once, start again from the first element U 0 Select from; for example, input 10 in the 4th cycle, select U 14 To U 15 and U 0 To U 10 For the same input, the errors introduced by the selected components are different. At a certain moment, the total error introduced is 0. The mismatch error of the DWA algorithm at time k can be expressed as:
[0039] y mis (k)=INL(ptr(k))-INL(ptr(k-1))
[0040]
[0041] Among them, C i is the actual size of each component, C m is the mean value of the components, and IM is the accumulation of component mismatch errors. mis The z-domain expression for (k) can be written as
[0042] Y mis (z) = (1-z -1 )IM(PTR(z))
[0043] It can be seen that the use of the traditional DWA algorithm can perform first-order shaping on the noise generated by the in-band mismatch.
[0044] This embodiment proposes an improved data weighted average (ADWA) calibration algorithm based on the traditional DWA algorithm. The algorithm further considers the component selection cycle to select the components used in this cycle. Each component is selected once as a component selection cycle. When a cycle ends, the starting component of the component selection cycle steps by 1, and the next selected component starts from the starting component of the component selection cycle. In particular, if the quantizer outputs full amplitude, if the starting component of the selection cycle steps, a certain component will be selected twice, so it is stipulated that the starting component of the component selection cycle will not step at this time. Figure 2 As shown, Figure 2: is a schematic diagram of the ADWA algorithm provided by an embodiment of the present invention. The horizontal direction is the component list, the vertical direction is the input of each cycle, and the filled ones represent the selected components. The traditional DWA algorithm selects components in a loop without stepping. The ADWA algorithm will step the starting point of the component selection loop by 1 after the component selection loop. According to the legend, the number of component selection loops is distinguished. The slash is the first selection loop, the dark gray is the second selection loop, and the square is the third selection loop. When the input is 13, the first selection loop ends and meets the stepping condition, so the third selected component is selected from U 1 When the input is 16, in order to prevent a component from being selected twice at the same time, the starting point of the selection cycle is still U 1 .
[0045] In view of the above ADWA algorithm, this embodiment provides a digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements. Figure 3 , Figure 3 1 is a schematic diagram of a digital calibration circuit structure of a digital-to-analog converter based on random stepping of dynamic elements provided by an embodiment of the present invention, which includes:
[0046] Adder A1, ADWA control unit 2, adder B3, first register 4, transcoder 5, ADWA logic unit 6, wherein,
[0047] The adder A1 is connected to the output end of the quantizer and is used to generate the end pointer of this cycle according to the output code of the quantizer;
[0048] The ADWA control unit 2 is connected to the output end of the quantizer and is used to generate a step signal S, an enable signal enable and a starting element position signal SPR in an element selection cycle according to the output code of the quantizer;
[0049] The adder B3 is connected to the adder A1 and the ADWA control unit 2, and is used to output a start pointer according to the end pointer of this cycle and the step signal S;
[0050] The first register 4 is connected to the adder B3 and the adder A1, and is used to store the start pointer output by the adder B in this cycle, and output it to the adder A1 in the next cycle as the start pointer of the next cycle;
[0051] The transcoder 5 is connected to the adder A1, the adder B3 and the first register 4, and is used to generate an initial component selection code according to the end pointer and the start pointer;
[0052] The ADWA logic unit 6 is connected to the ADWA control unit 2 and the transcoder 5, and is used to shift the initial component selection code according to the enable signal enable and the starting component position signal SPR in the component selection cycle to obtain the final component selection code.
[0053] Specifically, when the mismatch calibration circuit obtains the input of this cycle, adder A obtains the end pointer according to the start pointer generated in the previous cycle obtained from the register to determine the position of the last element selected in this cycle. The transcoder selects the element through the carry judgment end of adder A to obtain the corresponding element selection code. At the same time, the ADWA control unit determines whether the starting element of the element selection cycle of this cycle needs to be stepped according to the input, and then outputs the step signal S, and whether to transform the element selection code, outputs the enable signal enable, and records the starting element position SPR of the latest element selection cycle. Then the ADWA logic unit transforms the element selection code output by the transcoder through the enable end. If the enable end is valid, the element selection code is transformed, otherwise, the transcoder output is directly output. At the same time, after the ADWA control unit outputs the carry, adder B accumulates the carry value, and refreshes the start pointer through the register before the next quantizer output arrives, which is used for transcoding of the next cycle.
[0054] In this embodiment, adder A1 is a modulo 16 adder, which is specifically used to add the output code of the quantizer of this period to the start pointer of this period to obtain the end pointer of this period, and each time the accumulation reaches 16, the carry end is set to logic "1" to control the working mode of the transcoder, and then it is reset to zero and starts counting again.
[0055] Furthermore, adder B3 is also a modulo 16 adder, which is specifically used to add the end pointer of this cycle and the carry of adder A according to the step signal S, and send them to the register to determine the position of the first element selected in the next cycle, and return to zero and restart counting each time the accumulation reaches 16.
[0056] This embodiment uses the modulo 16 adder A and the modulo 16 adder B to implement the calibration of the hexadecimal input code.
[0057] For further information, see Figure 4 , Figure 4 : is a circuit structure diagram of an ADWA control unit provided by an embodiment of the present invention, which includes:
[0058] A first modulo 16 adder 21, a second register 22, a NOT logic gate 23, a first AND logic gate 24, a second AND logic gate 25, a second modulo 16 adder 26 and a third register 27; wherein,
[0059] The input end of the first modulo 16 adder 21 is connected to the quantizer output end of the DAC to be calibrated and the output end of the second register 22; the output end of the first modulo 16 adder 21 is respectively connected to the input end of the second register 22, the input end of the first AND logic gate 24 and the input end of the second AND logic gate 25;
[0060] An input end of the NOT logic gate 23 is connected to the quantizer output end of the DAC to be calibrated, and an output end of the NOT logic gate 23 is connected to another input end of the first AND logic gate 24;
[0061] The output end of the first AND logic gate 24 is connected to the other input end of the second AND logic gate 25 and the input end of the second modulo 16 adder 26, and at the same time serves as the first output end of the ADWA control unit 2 to output a carry signal;
[0062] The output terminal of the second AND logic gate 25 serves as the second output terminal of the ADWA control unit 2 to output an enable signal;
[0063] An output terminal of the second modulo-16 adder 26 is connected to an input terminal of the third register 27;
[0064] The output end of the third register 27 is connected to the other input end of the second modulo-16 adder 26 and simultaneously serves as the third output end of the ADWA control unit 2 to output the starting element position signal of this cycle.
[0065] Specifically, the working principle of the ADWA control unit is as follows:
[0066] Each component selection cycle means selecting 16 components, and each time 16 is added up, a selection cycle is completed. Therefore, by judging whether an accumulator records all the inputs so far as an integer multiple of 16, it can be judged whether the starting component of the component selection cycle needs to be stepped. The first modulo 16 adder completes this task. When the sum of the inputs is an integer multiple of 16, the starting component of the component selection cycle, that is, SPR, needs to step by 1, that is, the step signal S outputs a high level 1. In general, if the current cycle is fully added up to 16, it proves that a component selection cycle is completed, and the carry end of the first modulo 16 adder outputs a high level. At the same time, the second modulo 16 adder records the number of times the carry is high, and the position of the starting component of the current component selection cycle is obtained, that is, the SPR signal.
[0067] When the accumulated input of this cycle is not an integer multiple of 16 and the SPR is stepped, transcoding is required and enable outputs a high level.
[0068] In particular, when the quantizer output is 16, in order to avoid an element being selected twice, the SPR does not need to be stepped, so the logic gates 23 and 24 are used to exclude the case where the quantizer output is 16.
[0069] In addition, if the sum of all input values so far is exactly an integer multiple of 16, the starting element of the element selection cycle, that is, SPR needs to be stepped, but the element selection code of the ADWA algorithm is the same as the initial element selection code, so no transcoding is required, enable is 0, and logic gate 25 completes this function.
[0070] For further information, see Figure 5 , Figure 5 is a circuit structure diagram of a transcoder provided by an embodiment of the present invention; it includes two binary to thermometer code transcoders and a plurality of logic gate units; wherein,
[0071] Two binary to thermometer code transcoders are used to transcode the end pointer and the start pointer respectively to obtain the corresponding thermometer code;
[0072] Each logic gate unit includes an XOR logic gate, a NOT logic gate and a 2-to-1 selector, which are respectively used to output selection codes of corresponding components.
[0073] Specifically, the working principle of the transcoder is as follows: convert the end and start pointers into corresponding thermometer codes, perform XOR operations on each bit, and when the carry end of adder A is 1, it means that U 16 Selected and taken from U 0 Start selecting, at this time the selector selects the inverse output of the XOR to obtain the unshifted component selection code.
[0074] In this embodiment, the ADWA logic unit 6 is mainly composed of a decoder, a shifter, and a capacitor array. The detailed circuit can refer to the prior art, and this embodiment will not be described in detail here. Among them, the decoder can output a unique hot code according to the value of SPR, and the position of 1 corresponds to the position of the element selecting the loop start element. When the shift enable signal is valid, the combinational logic detects the code element to be shifted and shifts it, and then sets the code element corresponding to SPR to zero to obtain the final element selection code element; the capacitor array selects the element to generate output according to the output code element of the ADWA logic unit.
[0075] See also Figure 6 , Figure 6 1 is a working principle diagram of a digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements provided by an embodiment of the present invention. Figure 3 and Figure 6 , the working principle of the digital calibration circuit of the digital-to-analog converter based on random stepping of dynamic elements provided in this embodiment is introduced in detail.
[0076] Specifically, when the rising edge of its clock CLK_Q arrives, the quantizer samples and outputs the thermometer code at the same time, and the DAC immediately selects the element to generate the output according to the thermometer code output by the comparator in this cycle. When CLK_Q is at a low level, the sampling capacitor is reset and the comparison result is latched; for the DAC using the ADWA algorithm, when the rising edge of CLK_Q arrives, the adder A obtains the position of the last element used in this cycle (end pointer) according to the DAC input and start of this cycle. The transcoder outputs the element selection code when it is not shifted according to the start and end pointers. The ADWA control unit determines the starting element step flag S of the element selection cycle according to the input accumulation, records and outputs the position SPR of the starting element of the element selection cycle, determines whether transcoding is required, and outputs the enable terminal. When the shift enable terminal is valid, the ADWA logic unit translates SPR into a one-hot code, shifts the input unshifted element selection code, obtains the final element selection code, and then selects the element to obtain the DAC output. When CLK_Q is low, the comparator does not perform comparison, but the output result is latched, the working output of adder B is refreshed start, and the adder of the ADWA control unit is refreshed SPR.
[0077] The digital calibration circuit of the digital-to-analog converter based on random stepping of dynamic elements provided by the present invention not only has a first-order shaping effect on the in-band noise, thereby improving the signal-to-noise ratio of the modulator; it also selects the starting element through the stepping signal, so that when the input signal is periodic, the periodic selection of elements can be avoided, thereby moving the power of the in-band tone to the outside of the band, reducing the in-band tone without sacrificing the signal-to-noise ratio, thereby improving the efficiency of the algorithm.
[0078] The above contents are further detailed descriptions of the present invention in combination with specific preferred embodiments, and it cannot be determined that the specific implementation of the present invention is limited to these descriptions. For ordinary technicians in the technical field to which the present invention belongs, several simple deductions or substitutions can be made without departing from the concept of the present invention, which should be regarded as falling within the protection scope of the present invention.
Claims
1. A digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements, It is characterized in that It includes an adder A (1), an ADWA control unit (2), an adder B (3), a first register (4), a transcoder (5), and an ADWA logic unit (6); wherein, The adder A(1) is connected to the output end of the quantizer and is used to generate an end pointer of this cycle according to the output code of the quantizer; The ADWA control unit (2) is connected to the output end of the quantizer and is used to generate a step signal S, an enable signal enable and a starting element position signal SPR in an element selection cycle according to an output code of the quantizer; The adder B (3) is connected to the adder A (1) and the ADWA control unit (2), and is used to output a start pointer according to the end pointer of the current cycle and the step signal S; The first register (4) is connected to the adder B (3) and the adder A (1), and is used to store the start pointer output by the adder B in this cycle, and output it to the adder A (1) in the next cycle to serve as the start pointer of the next cycle; The transcoder (5) is connected to the adder A (1), the adder B (3) and the first register (4), and is used to generate an initial component selection code according to the end pointer and the start pointer; The ADWA logic unit (6) is connected to the ADWA control unit (2) and the transcoder (5), and is used to shift the initial component selection code according to the enable signal enable and the starting component position signal SPR in the component selection cycle to obtain a final component selection code.
2. The digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements according to claim 1, It is characterized in that The adder A(1) is a modulo 16 adder, which is specifically used to add the output code of the quantizer of this cycle to the start pointer of this cycle to obtain the end pointer of this cycle, and each time the accumulation reaches 16, the carry end is set to logic "1" to control the working mode of the transcoder, and then reset to zero and restart counting.
3. The digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements according to claim 1, It is characterized in that The adder B (3) is a modulo 16 adder, which is specifically used to add the end pointer of the current cycle and the carry of the adder A (1) according to the step signal S, and send them to the register for determining the position of the first element selected in the next cycle, and return to zero and restart counting every time the accumulation reaches 16.
4. The digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements according to claim 1, It is characterized in that The ADWA control unit (2) comprises a first modulo 16 adder (21), a second register (22), a NOT logic gate (23), a first AND logic gate (24), a second AND logic gate (25), a second modulo 16 adder (26) and a third register (27); wherein, The input end of the first modulo 16 adder (21) is connected to the quantizer output end of the DAC to be calibrated and the output end of the second register (22); the output end of the first modulo 16 adder (21) is respectively connected to the input end of the second register (22), the input end of the first AND logic gate (24) and the input end of the second AND logic gate (25); The input end of the NOT logic gate (23) is connected to the quantizer output end of the DAC to be calibrated, and the output end of the NOT logic gate (23) is connected to the other input end of the first AND logic gate (24); The output end of the first AND logic gate (24) is connected to the other input end of the second AND logic gate (25) and the input end of the second modulo-16 adder (26), and serves as the first output end of the ADWA control unit (2) to output a carry signal; The output end of the second AND logic gate (25) serves as a second output end of the ADWA control unit (2) to output an enable signal; The output end of the second modulo-16 adder (26) is connected to the input end of the third register (27); The output end of the third register (27) is connected to the other input end of the second modulo-16 adder (26), and simultaneously serves as the third output end of the ADWA control unit (2) to output the starting element position signal of this cycle.
5. The digital calibration circuit for a digital-to-analog converter based on random stepping of dynamic elements according to claim 1, It is characterized in that The transcoder (5) comprises two binary to thermometer code transcoders and a plurality of logic gate units; wherein, The two binary to thermometer code transcoders are used to transcode the end pointer and the start pointer respectively to obtain corresponding thermometer codes; Each logic gate unit includes an XOR logic gate, a NOT logic gate and a 2-to-1 selector, which are respectively used to output selection codes of corresponding components.
