Multi-channel signal link offset voltage consistency calibration method and system

Abstract

The invention discloses a multichannel signal link offset voltage consistency calibration method and system, and the method comprises the steps: S1, inputting the same standard input data to each channel during calibration, and obtaining an ADC sampling output result of each channel; s2, calculating a standard value according to an ADC sampling output result of each channel; s3, calculating a difference value between an ADC sampling output result of each channel and a standard value; s4, calculating an input compensation value of each channel by using the difference value of each channel; and S5, during normal work, compensating the input of each channel by using the input compensation value of each channel. The standard value is calculated as the compensation reference of each channel, so that the overall drift is avoided.

Description

Technical Field

[0001] This invention belongs to the field of optical chip technology, and in particular relates to a calibration method and system for the consistency of offset voltage in multi-channel signal links. Background Technology

[0002] Optical computing systems utilize optical devices to achieve parallel data processing, offering advantages such as high speed and low power consumption, and are widely used in fields such as artificial intelligence and signal processing. To improve computational efficiency, these systems typically employ multiple parallel transmit circuits (TX circuits) based on digital-to-analog converters (DACs) to achieve high-speed data processing through multi-channel parallel transmission. These TX circuits generally include a DAC and a driver circuit, and the consistency of their output voltage directly affects the signal transmission accuracy and computational reliability of the optical computing system.

[0003] Due to non-ideal characteristics such as inconsistent circuit component parameters, temperature drift, and process variations, offset voltage mismatch exists between different TX channels. Specifically, when the same DAC data is input, the driver output voltages of each channel differ, leading to data inconsistency in subsequent signal processing and ultimately reducing the computational accuracy of the optical computing system. For example, in high-precision signal processing scenarios, offset voltage mismatch can cause computational errors exceeding 10%, severely impacting system performance.

[0004] Chinese patent application CN202411940225.5 disclosed by the applicant discloses a calibration method for multiple channels of a photonic computing chip. The method involves inputting the same calibration optical signal to each channel of the computing module to obtain the calibration output value of each channel; selecting one channel as a reference channel and using its calibration output value as the reference calibration output value; and using the remaining channels as channels to be calibrated. The input optical signal to each channel to be calibrated is adjusted according to an external control electrical signal, so that when the calibration optical signal is input to the light source module, the calibration output value of each channel to be calibrated in the computing module is the reference calibration output value. Errors originally caused by manufacturing process errors and fiber coupling are eliminated by adjusting the input optical signal, resulting in a consistent final output structure.

[0005] In this application, one channel is selected as the reference channel, and the remaining channels are adjusted based on the reference channel. In actual use, it was found that if the selected reference channel itself has a large offset voltage or hardware failure, it will cause overall calibration deviation, resulting in overall output drift. Summary of the Invention

[0006] The purpose of this invention is to provide a calibration method and system for the consistency of offset voltage in multi-channel signal links, which partially solves or alleviates the above-mentioned deficiencies in the prior art. By calculating standard values ​​as compensation references for each channel, overall drift is avoided.

[0007] To solve the aforementioned technical problems, the present invention specifically adopts the following technical solution: A first aspect of the present invention provides a calibration method for the offset voltage consistency of a multi-channel signal link, wherein the multi-channel signal link includes a plurality of parallel channels and an ADC that samples the output of each channel; each channel is provided with a cascaded DAC and a driver; accordingly, the calibration method includes the following steps: During calibration, S1 inputs the same standard input data to each channel and obtains the ADC sampling output result of each channel; S2 calculates the standard value based on the ADC sampling output of each channel; S3 calculates the difference between the ADC sampling output of each channel and the standard value; S4 calculates the input compensation value for each channel using the difference between the channels; When S5 is working normally, it uses the input compensation values ​​of each channel to compensate for the input of each channel.

[0008] Furthermore, methods for calculating standard values ​​include: S21 sorts the ADC sampling output results of all channels and then groups the sorted ADC sampling output results into groups in order. S22 selects one of the ADC sampling outputs from each group as the reference value for that group; S23 utilizes the formula: , Calculate the deviation between the sampling output of all channels of the ADC and the reference value; where S n The deviation of the reference value n of the sampling output results of all channels ADC, where m is the number of channels, i is the sequence number, and x is the deviation of the reference value n. i Let x be the ADC sampling output of the i-th channel. n The baseline value is n; S24 uses the reference value with the smallest deviation as the standard value.

[0009] Furthermore, the median value of each group of ADC sampling output results is selected as the baseline value.

[0010] Furthermore, a baseline value is calculated by iteratively weighting and summing the ADC sampling output results for each group, specifically including: S221 utilization formula: , Calculate the ADC sampling data weights for each channel within the group; where ω i Let x be the weight of the i-th channel. iLet σ be the ADC sampling output of the i-th channel. In the first iteration, σ is the standard deviation of the ADC sampling output of the channels in the group, and in subsequent iterations, it is the standard deviation after the previous adjustment. μ is the mean of the ADC sampling output of the channels in the group in the first iteration, and in subsequent iterations, it is the center value. S222 normalizes the weights so that the sum of the weights of the ADC sampling output results of the channels within the group is 1. S223 calculates the weighted mean based on the normalized weights and uses this weighted mean as the new center value; S224 calculates the proportion of outliers within a group and dynamically adjusts the standard deviation σ based on the proportion of outliers; S225 Repeat steps S221-S224 until the rate of change of the weighted mean is less than the preset convergence threshold. S226 uses the weighted mean as the benchmark for this group.

[0011] Furthermore, using the formula: σ′=σ×(1+γ×outlier_ratio), Adjust the standard deviation; where σ' is the current adjusted standard deviation, σ is the standard deviation of the ADC sampling output of the channel within the group in the first iteration, and the standard deviation of the previous adjustment in subsequent iterations; γ is the adjustment coefficient, and outlier_ratio is the proportion of outliers.

[0012] Furthermore, the same standard input data is input to each channel several times, and the average value of the multiple ADC sampling output results is used as the ADC sampling output result of that channel.

[0013] Furthermore, using the formula: dac_dos_n=adc_dos_n*(vref_adc / vref_dac)*2 (M-N) , Calculate the input compensation value; where dac_dos_n is the input compensation value of the nth channel, adc_dos_n is the difference between the sampling output result of the nth channel ADC and the standard value, vref_adc is the reference voltage of the ADC, vref_dac is the reference voltage of the DAC, M is the accuracy of the DAC, and N is the accuracy of the ADC.

[0014] Furthermore, during normal operation, the input of each channel is compensated by subtracting the input compensation value of that channel from the input value of that channel.

[0015] Furthermore, a switch is provided on each channel behind the driver.

[0016] The present invention also provides a calibration system for the offset voltage consistency of a multi-channel signal link, comprising: The ADC sampling output acquisition module is used to input the same standard input data into each channel during calibration and acquire the ADC sampling output result of each channel. The standard value calculation module is used to calculate the standard value based on the ADC sampling output results of each channel; The difference calculation module is used to calculate the difference between the ADC sampling output result of each channel and the standard value; The input compensation value calculation module is used to calculate the input compensation value of each channel using the difference between each channel. The compensation module is used to compensate the input of each channel using the input compensation value of each channel during normal operation.

[0017] Beneficial effects: This invention sorts and groups the ADC sampling results, and combines an iterative weighted summation algorithm to dynamically adjust the standard deviation and weights. This effectively eliminates the interference of outliers, makes the weighted mean closer to the true signal distribution, and avoids the distortion of the benchmark value by extreme values.

[0018] By inputting the same standard data multiple times into each channel and taking the average, the influence of random noise in the ADC sampling process can be effectively reduced, making the original data closer to the true value and providing a reliable basis for subsequent benchmark calculation and compensation.

[0019] Grouping channels sequentially before calculating the baseline value can avoid the impact of local deviations in the overall data on the global baseline. This is especially suitable for scenarios where some channels in a multi-channel system have systematic deviations, thus improving the representativeness of the baseline value.

[0020] By iteratively calculating the weighted mean until the rate of change is less than a threshold, the benchmark value gradually converges to the optimal solution. This process can adapt to the noise levels and deviation distributions of different channels, and compared to a simple average value, it reflects the true characteristics of the channel group better, thus providing a more accurate standard reference for subsequent compensation.

[0021] The median value of each set of data can also be selected as the benchmark value, which enhances the robustness to abnormal data, especially when there are a few channels with large deviations, and can ensure the stability of the benchmark value.

[0022] During normal operation, the offset voltage of each channel can be corrected in real time by subtracting the input compensation value, ensuring that the signal link remains consistent during operation and avoiding the decrease in accuracy caused by device drift after traditional offline calibration.

[0023] The compensation value is calculated based on the difference between each channel and the standard value. The offset voltage of each channel can be corrected in a targeted manner, realizing one-by-one calibration, and ultimately significantly improving the consistency of multi-channel output.

[0024] In addition, switches are installed at the rear of each channel driver to isolate abnormal channels or selectively enable the calibration process during calibration, improving the flexibility and controllability of system operation, while also facilitating maintenance and troubleshooting. Attached Figure Description

[0025] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. The elements or parts in the drawings are not necessarily drawn to scale. Obviously, the drawings described below are some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative effort.

[0026] Figure 1 This is a schematic diagram of the multi-channel signal link in this invention; Figure 2 This is a flowchart of Example 1; Figure 3 This is a structural diagram of Example 2. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0028] In this document, suffixes such as "module," "part," or "unit" used to denote elements are used only for the purpose of illustrative purposes and have no specific meaning in themselves. Therefore, "module," "part," or "unit" may be used interchangeably.

[0029] In this document, the terms "upper," "lower," "inner," "outer," "front," "rear," "one end," and "the other end," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are used only for the convenience of describing the present invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the present invention. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0030] In this document, unless otherwise explicitly specified and limited, the terms "installed," "equipped with," "connected," etc., should be interpreted broadly. For example, "connection" can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection, a direct connection, or an indirect connection through an intermediate medium; it can be a connection within two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0031] In this document, "and / or" includes any and all combinations of one or more of the listed related items.

[0032] In this article, "multiple" means two or more, that is, it includes two, three, four, five, etc.

[0033] Example 1: like Figure 1 As shown, this embodiment provides a calibration method for the offset voltage consistency of a multi-channel signal link. The multi-channel signal link includes several parallel channels and an ADC that samples the output of each channel; each channel is equipped with a cascaded DAC and a driver. A switch is located after the driver on each channel. The entire link contains m parallel channels (1-m), each channel processing signals independently, achieving parallel data transmission and processing.

[0034] The DAC (Digital-to-Analog Converter) is used to receive digital input signals Dac_din_1-Dac_din_m and convert the digital quantity into an analog voltage.

[0035] The driver is used to amplify / condition the analog voltage output by the DAC, enhancing its driving capability, and outputting Vout_1-Vout_m.

[0036] The switch is used for calibration / operating mode switching. In calibration mode, any / all switches are turned on, thereby selecting the ADC for sampling for calibration. In operating mode, all switches are turned off and the output is sent to the load.

[0037] All channels share a high-precision ADC, and the Vout of each channel is sampled in a time-division manner to output a digital signal Adc_dout. Specifically, the channel connected to the ADC can be switched by controlling the on and off states of a switch.

[0038] Based on the aforementioned multi-channel signal link, this embodiment describes a calibration method for the consistency of multi-channel offset voltage in the link as follows: Figure 2 As shown, it specifically includes: During calibration, S1 inputs the same standard input data to each channel and obtains the ADC sampling output result of each channel.

[0039] The purpose of this step is to collect the ADC output result Adc_dout of each channel by inputting the same standard data to all channels, so as to provide basic data for subsequent calculation of standard value and compensation value.

[0040] To improve accuracy and avoid the impact of single data offsets, this embodiment inputs the same standard input data several times to each channel, and uses the average value of the multiple ADC sampling output results as the ADC sampling output result of that channel, thereby suppressing random noise and improving the signal-to-noise ratio.

[0041] S2 calculates the standard value based on the ADC sampling output of each channel.

[0042] The purpose of this step is to calculate, through appropriate strategies, a standard value that represents the ideal output of the system from the multi-channel ADC sampling results, and use it as a reference benchmark for offset voltage calibration.

[0043] Specifically, the steps for calculating the standard value in this embodiment include: S21 sorts the ADC sampling output results of all channels and then groups the sorted ADC sampling output results into equal groups according to the order.

[0044] The original channel's ADC sampling output result X=[x1,x2,…,x] for each sampling is... m [m is the number of channels, after sorting, the ordered array X] sorted =[x (1) ,x (2) ,…,x (m) ], satisfying x (1) ≤x (2) ≤⋯≤x (m) Divide the sorted array into k groups, each containing n = m / k data points. x1, x2, ..., x m These represent the ADC sampling output results corresponding to the 1st, 2nd, ..., mth channels, respectively (or, if each channel is sampled multiple times, then x1, x2, ..., xm). m (These are the average values ​​of the ADC sampling output results calculated after multiple samplings of the 1st, 2nd, ..., mth channels, respectively).

[0045] Of course, in cases where the data cannot be evenly distributed, such as when m=17 and k=4, the first three groups should contain four data points, and the fourth group should contain five data points, in order to ensure that the groups are evenly distributed as much as possible.

[0046] S22 selects one of the ADC sampling outputs from each group as the baseline value for that group.

[0047] The baseline value serves as a representative of each group for the calculation of subsequent standard values, and therefore should be as unaffected as possible by outliers. This embodiment provides two methods for selecting the baseline value: one is the median value, and the other is calculating the baseline value through iterative weighted summation.

[0048] The median method leverages the resistance to extreme values ​​in the middle position of the sorted data, excluding outliers. Specifically, it takes the middle value of the sorted data within a group; if the group size n is odd, the median value is x. (n / 2+1) If n is even, the median value is (x (n / 2) +x (n / 2+1) ) / 2.

[0049] The median method is efficient and computationally simple. However, it has low data utilization and is suitable for scenarios where outliers are clearly identified and hardware resources are limited. For example, when the electrical performance of devices in each channel is not significantly different, the output results of most channels are relatively balanced and concentrated in the middle region, approximating a normal distribution.

[0050] The specific steps for calculating the baseline value using iterative weighted summation include: S221 utilization formula: , Calculate the ADC sampling data weights for each channel within the group; where ω i Let x be the weight of the i-th channel. i Let σ be the ADC sampling output of the i-th channel. In the first iteration, σ is the standard deviation of the ADC sampling output of the channels in the group, and in subsequent iterations, it is the standard deviation after the previous adjustment. μ is the mean of the ADC sampling output of the channels in the group in the first iteration, and in subsequent iterations, it is the center value.

[0051] The weight decreases exponentially with the distance between the data point and the center value; the closer the distance, the higher the weight, which naturally suppresses outliers.

[0052] The weight varies with the distance d = |x from the center value of the data point i -μ∣ decays exponentially: When d=0, the weight is at its maximum. ; When d=σ, the weights decrease to ; When d=3σ, the weight is only It is close to 0.

[0053] In the above formula, the larger σ is, the flatter the function curve, making it less sensitive to distance and suitable for scenarios with many outliers; the smaller σ is, the steeper the curve, and slight changes in distance result in significant weight differences, making it suitable for scenarios with concentrated datasets. In this embodiment, σ is dynamically adjusted to suit the usage scenario.

[0054] In the initial iteration, μ = the group mean, which is the initial center of the data distribution, while in subsequent iterations it is the weighted mean of the previous iteration; in the initial iteration, σ = the group standard deviation, which characterizes the dispersion of the data, while in subsequent iterations it is the adjusted standard deviation of the previous iteration. The calculation of the mean and standard deviation is a prior art technique and will not be elaborated here.

[0055] S222 normalizes the weights so that the sum of the weights of the ADC sampling output results of the channels within the group is 1.

[0056] After normalization, the weights are equivalent to the probability density, making the weighted mean equivalent to the mathematical expectation. If not normalized, the sum of the weights may be greater than 1, causing the weighted mean to deviate from the true center.

[0057] S223 calculates the weighted mean based on the normalized weights and uses this weighted mean as the new center value.

[0058] Specifically, using the formula: , Calculate the weighted mean; where μ new Let x be the new center value, i.e., the weighted mean. i This represents the ADC sampling output of the i-th channel. The weight is the normalized weight of the i-th channel.

[0059] In the first iteration, μ is the mean of the ADC sampling output of the channels within the group, which may be affected by outliers. Through weighting, the weight of outliers is reduced. For example, the weight of data 3σ away from the center is ≈0.01. The weighted mean shifts towards the normal data area, so that after each iteration, the new center value is closer to the true distribution center.

[0060] S224 calculates the proportion of outliers within a group and dynamically adjusts the standard deviation σ based on this proportion; specifically, it uses the formula: σ′=σ×(1+γ×outlier_ratio), Adjust the standard deviation; where σ' is the current adjusted standard deviation, σ is the standard deviation of the ADC sampling output of the channel within the group in the first iteration, and the standard deviation of the previous adjustment in subsequent iterations; γ is the adjustment coefficient, which is set manually; outlier_ratio is the proportion of outliers.

[0061] The outlier ratio is typically based on the current center value μ and standard deviation σ, considering data points that deviate from the center value by a certain range as outliers. For example, if data points satisfy |x i If -μ|>k×σ (where k is the threshold, such as 2 or 3), it is considered an outlier. The number of outliers is counted and divided by the total number of data points to obtain the outlier_ratio.

[0062] When the outlier_ratio is high, meaning there are many outliers within the group, the value of (1+γ×outlier_ratio) in the formula increases, leading to an increase in σ. Conversely, if there are few outliers, σ′ is close to the original σ. By amplifying the standard deviation, the sensitivity of outliers in the weight calculation is reduced, thereby weakening the impact of outliers on the benchmark value.

[0063] During the iteration process, if a calculation reveals a high proportion of outliers, it indicates that the current data distribution is affected by abnormal values. In this case, increasing the standard deviation makes the weight calculation more lenient in judging the degree of deviation, avoiding misclassifying normal data as outliers, while simultaneously reducing the weight of true outliers. Dynamically adjusting the standard deviation can prevent sudden changes in the standard deviation due to outliers in a single iteration, making the iteration process smoother. For example, when outliers are assigned low weights, the center values ​​in subsequent iterations will be closer to the true data distribution, thus gradually reducing the proportion of outliers and forming a virtuous cycle.

[0064] S225 Repeat steps S221-S224 until the rate of change of the weighted mean is less than the preset convergence threshold.

[0065] The rate of change of the weighted mean is the ratio of the absolute value of the difference between the current weighted mean and the previous weighted mean to the previous weighted mean. When it is less than a preset convergence threshold, such as 0.1%, the iteration stops.

[0066] S226 uses the weighted mean as the benchmark for this group.

[0067] Once the convergence condition is met, the weighted mean value after the current iteration is output as the benchmark value of the group, thereby eliminating outlier interference and reflecting the true level of the normal channel signal within the group.

[0068] In the iterative weighted summation method for calculating the baseline value, each iteration uses the weighted mean as the new center, pushing the baseline value towards the data-dense area; σ is adjusted in real time according to the proportion of outliers, so that the width of the weight function adapts to the current data dispersion. Compared with the center value method, it has upgrades in noise resistance mechanism, data utilization, and adaptability.

[0069] S23 utilizes the formula: , Calculate the deviation between the sampling output of all channels of the ADC and the reference value; where S n The deviation of the reference value n of the sampling output results of all channels ADC, where m is the number of channels, i is the sequence number, and x is the deviation of the reference value n. i Let x be the ADC sampling output of the i-th channel. n Let n be the baseline value.

[0070] After calculating the baseline value representing each group, it is necessary to select another baseline value from multiple baseline values ​​as the standard value for all channels in the entire link. In this embodiment, the strategy for selecting the standard value is to compare it with the standard deviation of the ADC sampling output results of all channels, and select the baseline value that minimizes the sum of the deviations of all channels as the system standard value. This conforms to the least squares estimation principle, minimizing the sum of the squared errors between the standard value and the data of all channels, and this standard value is the maximum likelihood estimate, which theoretically has the best accuracy.

[0071] S24 uses the reference value with the smallest deviation as the standard value.

[0072] Traditional fixed reference values, such as randomly selecting the output of one channel as the standard, may lead to large overall deviations due to anomalies in a single channel. Dynamically selecting the reference value with the smallest deviation ensures that the average deviation between the standard value and the data from all channels is minimized. The previous step suppresses outliers within the group through iterative weighting; this step further eliminates reference values ​​that are locally optimal but have large global deviations through global deviation comparison, forming a double guarantee against noise, thereby improving the accuracy of calibration to a certain extent and thus improving the consistency of the signal link offset voltage of each channel.

[0073] S3 calculates the difference between the ADC sampling output of each channel and the standard value.

[0074] The purpose of this step is to quantify the offset of each channel relative to the system's optimal reference, in units consistent with the ADC output.

[0075] S4 calculates the input compensation value for each channel using the difference between the channels.

[0076] Since the final calculated difference is the result of the ADC, it still needs to be converted into the input of the DAC, specifically using the formula: dac_dos_n=adc_dos_n*(vref_adc / vref_dac)*2 (M-N) , Calculate the input compensation value; where dac_dos_n is the input compensation value of the nth channel, adc_dos_n is the difference between the sampling output of the nth channel ADC and the standard value, vref_adc is the reference voltage of the ADC, vref_dac is the reference voltage of the DAC, M is the precision of the DAC (e.g., 16 bits), and N is the precision of the ADC (e.g., 12 bits).

[0077] The compensation value dac_dos_n essentially converts the ADC sampling deviation into a voltage adjustment amount that the DAC can output. vref_adc / vref_dac are used to unify the voltage reference of the ADC and DAC. (M-N) It is the precision conversion factor, which converts the digital scale of the ADC to the scale of the DAC.

[0078] When S5 is working normally, it uses the input compensation values ​​of each channel to compensate for the input of each channel.

[0079] This step is the final execution stage of the entire calibration process. Its core objective is to use the input compensation values ​​of each channel obtained from the preceding steps (S23-S24 to calculate the deviation, S3 to calculate the difference, and S4 to derive the compensation value) to correct the actual input signal in real time, eliminate hardware errors such as ADC offset and gain error, and thus improve sampling accuracy.

[0080] During normal operation, the input of each channel is compensated by subtracting its input compensation value from the input value of that channel. That is, the compensation value dac_dos_n is converted into an analog voltage by a DAC and then superimposed on the original input signal. For example: If adc_dos_n>0 (sampling value is too high), the compensation voltage is negative to cancel out the high part of the input signal; If adc_dos_n < 0 (sampling value is too low), the compensation voltage is positive, boosting the input signal to the standard value.

[0081] By compensating for inconsistencies between channels, the sampling results of different channels are unified under the standard value benchmark, thereby improving the overall accuracy of the system.

[0082] Example 2: like Figure 3 As shown, this embodiment provides a calibration system for the offset voltage consistency of a multi-channel signal link, including: The ADC sampling output acquisition module is used to input the same standard input data into each channel during calibration and acquire the ADC sampling output result of each channel. The standard value calculation module is used to calculate the standard value based on the ADC sampling output results of each channel; The difference calculation module is used to calculate the difference between the ADC sampling output result of each channel and the standard value; The input compensation value calculation module is used to calculate the input compensation value of each channel using the difference between each channel. The compensation module is used to compensate the input of each channel using the input compensation value of each channel during normal operation.

[0083] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0084] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a computer terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0085] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A calibration method for the consistency of offset voltage in a multi-channel signal link, characterized in that: The multi-channel signal link includes several parallel channels and an ADC that samples the output of each channel; each channel is equipped with a cascaded DAC and a driver; correspondingly, the calibration method includes the following steps: During calibration, S1 inputs the same standard input data to each channel and obtains the ADC sampling output result of each channel; S2 calculates the standard value based on the ADC sampling output of each channel; S3 calculates the difference between the ADC sampling output of each channel and the standard value; S4 calculates the input compensation value for each channel using the difference between the channels; When S5 is working normally, it uses the input compensation values ​​of each channel to compensate for the input of each channel.

2. The calibration method for the consistency of offset voltage in a multi-channel signal link according to claim 1, characterized in that... Methods for calculating standard values ​​include: S21 sorts the ADC sampling output results of all channels and then groups the sorted ADC sampling output results into groups in order. S22 selects one of the ADC sampling outputs from each group as the reference value for that group; S23 utilization formula: ， Calculate the deviation between the sampling output of all channels of the ADC and the reference value; where S n The deviation between the sampling output of all channels of the ADC and the reference value, where m is the number of channels, i is the index, and x is the value of the reference value. i Let x be the ADC sampling output of the i-th channel. n As the baseline value; S24 uses the reference value with the smallest deviation as the standard value.

3. The calibration method for the consistency of offset voltage in a multi-channel signal link according to claim 2, characterized in that: The median value of each ADC sampling output is selected as the baseline value.

4. The calibration method for the consistency of offset voltage in a multi-channel signal link according to claim 2, characterized in that... For each group of ADC sampling outputs, an iterative weighted summation is performed to calculate the baseline value, specifically including: S221 utilization formula: ， Calculate the ADC sampling data weights for each channel within the group; where ω i Let x be the weight of the i-th channel. i Let σ be the ADC sampling output of the i-th channel. In the first iteration, σ is the standard deviation of the ADC sampling output of the channels in the group, and in subsequent iterations, it is the standard deviation after the previous adjustment. μ is the mean of the ADC sampling output of the channels in the group in the first iteration, and in subsequent iterations, it is the center value. S222 normalizes the weights so that the sum of the weights of the ADC sampling output results of the channels within the group is 1. S223 calculates the weighted mean based on the normalized weights and uses this weighted mean as the new center value; S224 calculates the proportion of outliers within a group and adjusts the standard deviation σ based on the proportion of outliers; S225 Repeat steps S221-S224 until the rate of change of the weighted mean is less than the preset convergence threshold. S226 uses the weighted mean as the benchmark for this group.

5. The calibration method for the consistency of offset voltage in a multi-channel signal link according to claim 4, characterized in that... Using the formula: σ′=σ×(1+γ×outlier_ratio), Adjust the standard deviation; where σ' is the current adjusted standard deviation, σ is the standard deviation of the ADC sampling output of the channel within the group in the first iteration, and the standard deviation of the previous adjustment in subsequent iterations; γ is the adjustment coefficient, and outlier_ratio is the proportion of outliers.

6. The calibration method for the offset voltage consistency of a multi-channel signal link according to claim 1, characterized in that: The same standard input data is input to each channel several times, and the average value of the multiple ADC sampling output results is used as the ADC sampling output result of that channel.

7. The calibration method for the consistency of offset voltage in a multi-channel signal link according to claim 1, characterized in that... Using the formula: dac_dos_n=adc_dos_n*(vref_adc / vref_dac)*2 (M-N) , Calculate the input compensation value; where dac_dos_n is the input compensation value of the nth channel, adc_dos_n is the difference between the sampling output result of the nth channel ADC and the standard value, vref_adc is the reference voltage of the ADC, vref_dac is the reference voltage of the DAC, M is the accuracy of the DAC, and N is the accuracy of the ADC.

8. The calibration method for the consistency of offset voltage in a multi-channel signal link according to claim 1, characterized in that: During normal operation, the input of each channel is compensated by subtracting the input compensation value of that channel from the input value of that channel.

9. The calibration method for the consistency of offset voltage in a multi-channel signal link according to claim 1, characterized in that: Each channel has a switch located behind the driver.

10. A calibration system for the consistency of offset voltage in a multi-channel signal link, characterized in that... include: The ADC sampling output acquisition module is used to input the same standard input data into each channel during calibration and acquire the ADC sampling output result of each channel. The standard value calculation module is used to calculate the standard value based on the ADC sampling output results of each channel; The difference calculation module is used to calculate the difference between the ADC sampling output result of each channel and the standard value; The input compensation value calculation module is used to calculate the input compensation value of each channel using the difference between each channel. The compensation module is used to compensate the input of each channel using the input compensation value of each channel during normal operation.

Images