Automating PCIe 6.0 Tx Equalizer Calibration Using a Multivariate Approach

Abstract

A receiver test and measurement system includes a signal generator that provides a stressed signal. The stressed signal must be calibrated according to several parameter specifications. Some parameters, including a set of two or more equalization parameters, can be calibrated on a test and measurement instrument that captures the stressed signal and performs the calibration process. A multivariable model is created that establishes the relationship between the parameters set on the signal generator and the measurements of those parameters on a test and measurement instrument, such as an oscilloscope. The multivariable model calculates coefficients that allow calculation of the settings that need to be set on the signal generator for any combination of two or more equalization parameters. This allows receiver testing to be performed in a calibrated manner for any combination of required equalization parameters.

Description

[Technical Field]

[0001] The present disclosure relates generally to testing electronic circuits, and more particularly to automatic calibration of stressed eye signals for receiver (Rx) testing of a PCIe bus of a device under test (DUT). [Background technology]

[0002] PCIe (Peripheral Component Interconnect Express) is a high-speed serial computer expansion bus standard widely used in electronic devices such as personal computer motherboards, graphics cards, and solid-state drives (SSDs). PCIe 6.0 is the latest generation of the PCIe standard and includes data transfer rates of up to 64 Giga transfers per second (GT / s). Devices under test (DUTs) containing PCIe 6.0 buses must be tested to ensure that bus components, such as transmitters (Tx) and receivers (Rx) included in the DUT, are operating properly.

[0003] Testing the receiver Rx of the PCIe 6.0 bus involves stressed eye calibration testing, which involves calibrating a signal generator, such as a pulse pattern generator (PPG) in a bit error rate tester (BERT), to generate a signal with a stressed eye that is supplied to the receiver Rx of the DUT. This signal with a stressed eye is referred to as the "stressed signal" in this description.

[0004] The stress signal generated by the signal generator must be calibrated to provide a worst-case signal with parameters (such as eye height and eye width) that fall within the tolerances specified in the PCIe 6.0 standard. This stress signal is then fed to the receiver Rx of the DUT to verify that the receiver makes the correct decision on this signal even when receiving this worst-case stress signal.

[0005] Accurate calibration of the stress signal for receiver Rx testing is required as part of silicon validation and compliance testing performed on the physical layer of the DUT's PCIe 6.0 bus. Calibration of the signal generator is performed by coupling a test measurement device (typically an oscilloscope) to the signal generator and calibrating it to provide a stress signal that replicates the worst-case PCIe 6.0 transmitter Tx as defined in the PCIe 6.0 standard. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] US Patent Application Publication No. 2015 / 0103874 [Non-patent literature]

[0007] [Non-Patent Document 1] "DPO70000SX ATI Performance Oscilloscope" introduction site, also introducing the DPS70004SX model, Tektronix, [online], [searched October 16, 2025], Internet<https: / / www.tek.com / ja / products / oscilloscopes / dpo70000sx> [Non-patent document 2] "MP1900A Signal Quality Analyzer-R" introduction site, Anritsu Corporation, [online], [Retrieved October 16, 2025], Internet<https: / / www.anritsu.com / ja-jp / test-measurement / products / mp1900a> [Non-patent document 3] "Arbitrary waveform generator" introduction site, Tektronix, [online], [searched October 16, 2025], Internet<https: / / www.tek.com / ja / products / arbitrary-waveform-generators> Summary of the Invention [Problem to be solved by the invention]

[0008] One of the operating characteristics that must be calibrated for a signal generator that provides a stress signal is equalization (EQ), which includes pre-cursor (i.e., preshoot) and post-cursor (i.e., de-emphasis) parameters. Previous generations of PCIe used only one pre-shoot parameter and one de-emphasis parameter, and these two parameters were largely independent of each other and could be calibrated independently.

[0009] In contrast, the PCIe 6.0 standard includes first and second preshoot parameters, along with a de-emphasis parameter, to provide the equalization necessary to accommodate the increased data transfer rates in this standard. The required signal generator setting for this second preshoot parameter is also a function of the measurements of the first and second preshoot parameters and the de-emphasis parameter. This significantly complicates the process of calibrating the signal generator to provide the required stressed eye signal.

[0010] The current industry standard method for calibrating signal generators uses a "brute force" approach (i.e., manually adjusting the values ​​of preshoot and de-emphasis parameters) to provide the stress signals required to test the receiver Rx in a DUT according to the PCIe 6.0 standard, which can take up to 18 hours or more. Therefore, an improved method and system for calibrating signal generators to provide the stress signals required to perform testing of the receiver Rx in a DUT according to the PCIe 6.0 standard is needed. [Means for solving the problem]

[0011] Embodiments of the disclosed technology use a multivariate model to determine the relationship between the settings and measurements of the equalization parameters used under the PCIe 6.0 standard, namely, first preshoot (PS1, preshoot 1), second preshoot (PS2, preshoot 2), and de-emphasis (DE), in calibrating a stressed signal for receiver Rx testing. In some embodiments, the multivariate model, process, or algorithm can reduce the time required to calibrate a signal generator to approximately 15 minutes, compared to the typical 18 hours or more required using a traditional iterative calibration process for stressed eye calibration in accordance with the PCIe 6.0 standard. The PCIe 6.0 standard defines 127 preset combinations, i.e., "presets," of the allowable values ​​of the equalization parameters PS1, PS2, and DE used in calibrating a signal generator. In some embodiments, the multivariate model utilizes two different models as part of the signal generator calibration process: 1) A linear two-variable model of the PS1 and DE equalization parameters for the preset combinations, with the equalization parameter PS2=0. 2) A linear three-variable model for preset combinations with equalization parameter PS2 ≠ 0.

[0012] A two-variable model is calculated using several different combinations of the parameters PS1 and DE (PS2=0), and a three-variable model is calculated using different combinations of the parameters PS1, PS2, and DE. For each of these models, an oscilloscope is coupled to the signal generator and captures the stress signal provided by the signal generator for each of these different combinations of the equalization parameters PS1, PS2, and DE. The calibration process then utilizes a data set formed by these captured stress signal samples to calculate dependent parameters that estimate the relationship or dependency between each set of the signal generator's equalization parameters PS1, PS2, and DE and actual or measured equalization parameters generated from the stress signal samples captured by the oscilloscope. Once each of these models is calculated, the signal generator can be calibrated by setting the equalization parameters PS1, PS2, and DE in the signal generator based on the estimated dependent variables of the corresponding model.

[0013] In this description, the preset combinations of equalization parameters PS1, PS2, and DE configured or set in the signal generator may be referred to as "set" equalization parameters and may be denoted as PS1(set), PS2(set), and DE(set). Conversely, the equalization parameters PS1, PS2, and DE measured by one or more test and measurement instruments may be referred to as "measured" equalization parameters and may be denoted as PS1(measure), PS2(measure), and DE(measure). [Brief explanation of the drawings]

[0014] [Figure 1] FIG. 1 is a block diagram of a test and measurement system including a master test and measurement instrument that performs a calibration operation to calibrate a signal generator that provides a stress signal in accordance with the PCIe 6.0 standard, in accordance with some embodiments of the present disclosure. [Figure 2]FIG. 2 is a block diagram of a test and measurement system for providing a stress signal and performing receiver testing of a DUT after calibration of the signal generator of FIG. 1 according to some embodiments of the present disclosure. [Figure 3] FIG. 3 is a flowchart of an example multivariable calibration process performed by the master test and measurement instrument of FIG. 1 according to some embodiments of the present disclosure. [Figure 4] FIG. 4 is a graph illustrating expected, measured, and error values ​​for the first preshoot parameter calculated by the multivariate calibration process of FIG. 3 in accordance with some embodiments of the present disclosure. [Figure 5] FIG. 5 is a graph illustrating expected, measured, and error values ​​for the second preshoot parameter calculated by the multivariate calibration process of FIG. 3 in accordance with some embodiments of the present disclosure. [Figure 6] FIG. 6 is a graph illustrating expected, measured, and error values ​​of de-emphasis parameters calculated by the multivariate calibration process of FIG. 3 in accordance with some embodiments of the present disclosure. [Figure 7] FIG. 7 is a graph illustrating expected, measured, and error values ​​for the first preshoot parameter calculated by the multivariate calibration process of FIG. 3 when the second preshoot value is zero, in accordance with some embodiments of the present disclosure. [Figure 8] FIG. 8 is a graph illustrating expected, measured, and error values ​​of the de-emphasis parameters calculated by the multivariate calibration process of FIG. 3 when the second preshoot value is zero, in accordance with some embodiments of the present disclosure. DETAILED DESCRIPTION OF THE INVENTION

[0015] Those skilled in the art will appreciate that any block diagrams herein represent conceptual views of illustrative systems embodying the principles of the present disclosure, and that flowcharts, flow diagrams, and the like represent various processes that may be substantially embodied on a computer-readable medium and executed by a computer or processor.

[0016] FIG. 1 is a block diagram of a test and measurement system 100 including master and extension test and measurement instruments 102A and 102B. The master and extension test and measurement instruments 102A and 102B capture samples of a stressed signal SE provided by a signal generator 104 and perform a calibration process. That is, the master and extension test and measurement instruments 102A and 102B perform a process to calibrate the signal generator using a calibration section 106 and provide the stressed signal SE in accordance with a specified specification or standard for testing a device under test (DUT). In the embodiment of FIG. 1, the master test and measurement instrument 102A operates the calibration section 106 to calculate a multivariate model that defines the relationship or dependency between the equalization parameters PS1(SET), PS2(SET), and DE(SET) set in the signal generator 104 and the actual or measured equalization parameters PS1(MEAS), PS2(MEAS), and DE(MEAS) measured by the test and measurement instruments 102A and 102B, based on a data set defined by the stressed signal SE samples captured by the test and measurement instruments 102A and 102B. Testing of the DUT may then be performed using the calibrated signal generator 104. In some embodiments, the calibrated signal generator 104 may be utilized to perform testing of the receiver Rx of the PCIe 6.0 bus of the DUT. A more detailed description of the receiver Rx testing, the calibration unit 106, and the multivariate model utilized by the calibration unit are described in more detail below with reference to FIGS. 2 and 3.

[0017] The test and measurement system 100 embodiment of FIG. 1 and other embodiments of the present disclosure described below are described in relation to the PCIe 6.0 standard. However, embodiments of the proposed multivariate processing and calibration process are not limited to the PCIe 6.0 standard. In further embodiments, the multivariate model and calibration process may be applied to many other high-speed input / output (IO) data transfer technologies, such as the Ethernet standard and the Universal Serial Bus (USB) standard (e.g., USB4 version 2.0), as well as any similar data transfer technology or standard developed in the future.

[0018] The stress signal SE provided by the signal generator 104 is a differential signal provided on a lane of the PCIe 6.0 bus of the DUT under test after calibration of the signal generator 104, as described in more detail with reference to FIG. 2. A lane of the PCIe 6.0 bus is a communication path for transmitting and receiving data between electronic devices coupled to the bus. In PCIe 6.0, each lane has two differential signal pairs: one pair for transmitting data and the other pair for receiving data. Thus, each lane has four conductive traces. In the test and measurement system 100, the stress signal SE generated by the signal generator 104 is a differential signal pair. One signal of the differential signal pair is provided to the master test and measurement instrument 102A via an appropriate communication link 108A, and the other signal is provided to the extension test and measurement instrument via an appropriate communication link 108B. Each communication link 108A, 108B will typically include a suitable interconnect structure, such as a cable, for interconnecting the signal generator 104 and the test and measurement instruments 102A, 102B. However, in some embodiments, the communication links 108A and 108B may include other suitable interconnect fixtures or connectors as required for a particular application.

[0019] In the test and measurement system 100, the master test and measurement instrument 102A and the extension test and measurement instrument 102B form a multi-stack test subsystem. The test and measurement instruments 102A and 102B are synchronized to capture a pair of differential signals that form the stress signal SE from the signal generator 104. In some embodiments, each of the test and measurement instruments 102A and 102B is an oscilloscope. Designating each test and measurement instrument 102A and 102B as either master or extension configures each test and measurement instrument 102A and 102B to function accordingly, and the master test and measurement instrument 102A is configured to send control signals to the extension test and measurement instrument 102B over the communication link 110 to synchronize the test and measurement instruments to simultaneously capture or acquire the differential signals that form the stress signal SE.

[0020] In some embodiments, each of the test and measurement instruments 102A and 102B is a Tektronix oscilloscope (e.g., a DPS77004SX oscilloscope), and the communication link 110 is a Tektronix UltraSync multi-unit time synchronization bus. The master test and measurement instrument 102A sends control signals to the extension test and measurement instrument 102B over the communication link 110 to synchronize the acquisition of the stress signal SE by the two test and measurement instruments. In some embodiments, the master test and measurement instrument 104A provides a sample clock signal over the communication link 110 used by the test and measurement instruments 102A and 102B to synchronize the simultaneous acquisition of the pair of differential signals that form the stress signal SE. The master test and measurement instrument 102A may also provide a trigger signal over the communication link 110 to control the start and stop of acquisition cycles for the test and measurement instruments 102A and 102B. For example, the master test and measurement instrument 102A may provide a trigger signal to initiate a capture cycle of the stressed signal SE for a given preset combination of equalization parameters EQ on the signal generator 104. The master test and measurement instrument 102A may control the transfer of data from the extended test and measurement instrument 102B to the master test and measurement instrument via the communication link 110. This data transferred via the communication link 110 corresponds to one acquired sample of the differential signal of the stressed signal SE captured by the extended test and measurement instrument 102B during each capture cycle for each preset combination of equalization parameters EQ set on the signal generator 104.

[0021] Each of the communication links 108A and 108B is coupled to a channel or test port 112A and 112B of a corresponding test and measurement instrument 102A and 102B, respectively. The embodiment of the system 100 of FIG. 1 includes two test and measurement instruments 102A and 102B because, if these instruments are oscilloscopes, there is a limited number of high-bandwidth test ports 112A and 112B available on each oscilloscope. As those skilled in the art will appreciate, high-bandwidth channels or test ports on oscilloscopes and other test and measurement instruments are expensive, and as a result, the number of such high-bandwidth test ports available on each oscilloscope is limited. In PCIe 6.0, each of the test ports 112A and 112B must have sufficient bandwidth to capture stressed signals at data rates up to 40 GT / s. A typical oscilloscope capable of capturing signals with a 40 GT / s transfer rate has only one test port 112A and 112B capable of capturing such high-bandwidth signals. As a result, testing the receiver Rx of each PCIe 6.0 lane requires a pair of oscilloscopes, each of which receives one of the differential pair signals of the stress signal SE. If the oscilloscopes or other test and measurement devices 102A and 102B have multiple channels with high enough bandwidth to acquire signals at data rates up to 40 GT / s, only one such oscilloscope or other test and measurement device is required. Thus, in such an embodiment, the test and measurement system 100 would include only a single test and measurement device 102 with a pair of test ports 112, each of which would be coupled to receive a respective one of the differential signals forming the stress signal SE.

[0022] Each of the test and measurement instruments 102A and 102B includes a processor 114A and 114B coupled to a memory 116A and 116B. Each memory stores appropriate software or firmware that the processor 114A and 114B executes to control the overall operation of the instrument 102A and 102B. The memories 116A and 116B also store data sets corresponding to acquired samples of the stress signal SE. The memory 114A of the master test and measurement instrument 102A also stores appropriate software or firmware that is executed by the processor 114A to implement the calibration section 106. The signal generator 104 also includes a processor 118 coupled to a memory 120 that stores appropriate software or firmware that, when executed by the processor, controls the overall operation of the signal generator. For example, in an embodiment of system 100, signal generator 104 may be a bit error rate tester (BERT) (e.g., an Anritsu MP1900A BERT). If signal generator 104 is a BERT, processor 118 executes instructions stored in memory 120 to control a pulse pattern generator (PPG) of the BERT to generate stress signal SE. Signal generator 104 is not limited to being a BERT, and in further embodiments of system 100, the signal generator may be another type of signal generator, such as an arbitrary waveform generator (AWG).

[0023] In an embodiment of the disclosed technology, the calibration section 106 may be implemented via one or more processors 114A on the master test and measurement instrument 102A, but in other embodiments, the one or more processors may be on the extended test and measurement instrument 102B, on the pattern signal generator 104, on one or more external computing devices, including cloud-based processors, or distributed among one or more of the master test and measurement instrument 102A, the extended test and measurement instrument 102B, the signal generator 104, and the one or more external computing devices.

[0024] FIG. 2 is a block diagram of a test and measurement system 200 for providing a stress signal SE to test a receiver Rx of a DUT 202 after calibrating a signal generator 204 in accordance with some embodiments of the present disclosure. The signal generator 204 corresponds to the signal generator 104 of FIG. 1 after calibration to generate the stress signal SE required by the PCIe 6.0 standard for testing the receiver Rx of the DUT 202. The signal generator 204 is coupled to a receiver Rx 208 within the DUT 202 via a communication link 206. The DUT 202 may be a variety of electronic circuits, and while only one receiver Rx is shown in FIG. 2 for simplicity, a DUT would typically have multiple receivers Rx 208, one for each lane of the DUT's PCIe 6.0 bus. The receiver Rx 208 is part of the components of the PCIe 6.0 bus included in the DUT 202 and is tested to verify proper operation when receiving the stress signal SE from the signal generator 204. As described above, the stress signal SE emulates a worst-case signal from a PCIe 6.0 transmitter Tx based on the specifications of the PCIe 6.0 standard.

[0025] The signal generator 204 includes a processor 218 and a memory 220, corresponding to the processor 118 and the memory 120 of FIG. 1 . The memory 220 also stores software or firmware for executing routines for performing testing of the receiver Rx 208. During operation, the signal generator 204 establishes communication with the DUT 202 via a communication link 206 and transmits a test pattern bit stream encoded into a stressed signal SE for capture by the receiver Rx 208. The receiver Rx 208 captures the stressed signal SE, digitizes the signal, and decodes the digitized signal into a corresponding captured bit stream. The captured bit stream of the stressed signal SE at the DUT 202 should ideally correspond to the test pattern bit stream encoded into the stressed signal SE from the signal generator 204. To verify that the receiver Rx 208 is functioning properly in capturing and decoding the stressed signal SE, the captured bit stream generated by the receiver Rx 208 is transmitted to the signal generator 204 via a communication link 210. The signal generator 204 then compares the captured bit stream received from the DUT 202 via the communication link 210 with the test pattern bit stream encoded in the stress signal SE provided by the signal generator. Ideally, these two bit streams should be identical. Through this comparison, the signal generator 204 verifies whether the receiver Rx 208 is operating properly by ensuring that the receiver Rx's captured bit pattern meets a specified bit error rate (BER). While the communication link 210 is illustrated as being separate from the communication link 206, the communication link 210 may be part of the same lane of the PCIe 6.0 bus represented by the communication link 206. In this situation, the DUT 202's captured bit pattern is sent back to the signal generator 204 for verification via the same lane (i.e., the communication link 206) over which the stress signal SE is communicated.

[0026] 3 is a flowchart of a multivariable calibration process or procedure 300 performed by the master test and measurement instrument 102A of FIG. 1 in accordance with some embodiments of the present disclosure. The multivariable calibration process 300 corresponds to an embodiment of the calibration process or procedure performed by the calibration unit 106 of FIG. 1. In some embodiments of the present disclosure, the multivariable calibration process 300 utilizes a linear model of the PS1, PS2, and DE equalization parameters. This linear model assumes that the “set” preshoot equalization parameters PS1(set), PS2(set) and the de-emphasis equalization parameter DE(set) are linear functions or linearly dependent on the “actual” or “measured” equalization parameters PS1(measure), PS2(measure), and DE(measure). The multivariable linear model can be expressed as follows:

number

[0027] In this multivariate linear model, the following matrices need to be estimated:

number

[0028] The determination or calculation of the dependent variable coefficients α, β, and ω and the variable σ for Equation 3 will not be described in further detail and will be discussed with reference to the multivariable calibration process 300 of FIG. 3 and the test and measurement system 100 of FIG. 1. Process 300 begins with step 302, in which preset combinations of equalization parameters PS1, PS2, and DE are set in the signal generator 104. As described above, the PCIe 6.0 standard defines 127 preset combinations (i.e., "presets") that are combinations of allowable values ​​of the equalization parameters PS1, PS2, and DE used in signal generator calibration. Therefore, in step 302, the equalization parameters PS1(setting), PS2(setting), and DE(setting) of the signal generator 104 are set to one of these preset combinations. The signal generator 104 then provides a stressed signal SE generated using these equalization parameters PS1(setting), PS2(setting), and DE(setting) to the test and measurement instruments 102A and 102B.

[0029] Process 300 proceeds from step 302 to step 304, where test and measurement instruments 102A and 102B capture stress signal SE provided by signal generator 104, and from samples of these captured signals, master test and measurement instrument 102A calculates or determines equalization parameters PS1(mt), PS2(mt), and DE(mt). Each of equalization parameters PS1, PS2, and DE is defined in terms of electrical parameters of stress signal SE and is typically expressed in decibels (dB). The calculations performed by master test and measurement instrument 102A to determine the equalization parameters PS1(mt), PS2(mt), and DE(mt) for the captured stress signal SE are understood by those skilled in the art and will not be described in detail herein for the sake of brevity.

[0030] After step 304 is completed, process 300 proceeds to step 306, where it is determined whether the set PS1(set), PS2(set), and DE(set) equalization parameters on signal generator 104 and the corresponding PS1(measured), PS2(measured), and DE(measured) equalization parameters calculated in step 304 are N combinations. These N combinations of "set" equalization parameters on the signal generator and the corresponding "measured" equalization parameters calculated by calibration section 106 on master test and measurement instrument 102A define a calibration data set. This calibration data set is used to calculate the coefficients α, β, ω, and σ of the dependent variables of the multivariable model shown in Equation 1. If the determination in step 306 is negative, then N combinations of set and measured equalization parameters have not yet been created and calculated in steps 302 and 304. In this situation, if the determination at step 306 is negative, process 300 returns to step 302, where a new combination of PS1(SET), PS2(SET), and DE(SET) equalization parameters is provided to signal generator 104 for generation of a new stressed signal SE by the signal generator. Test and measurement instruments 102A and 102B then capture the new stressed signal SE provided by signal generator 104 at step 304, and master test and measurement instrument 102A calculates the corresponding PS1(MEAS), PS2(MEAS), and DE(MEAS) equalization parameters for the new stressed signal.

[0031] Process 300 continues executing steps 302-306 until step 306 returns a positive determination, indicating that execution of steps 302 and 304 has generated a calibration data set of N combinations of set and measured equalization parameters. This calibration data set includes N combinations of "set" equalization parameters provided to signal generator 104 and N corresponding "measured" equalization parameters calculated by calibration section 106 on master test and measurement instrument 102A. In one embodiment, N=16, so the calibration data set includes 16 combinations of set and measured equalization parameters. The 16 combinations of set and measured equalization parameters must have values ​​that cover the sampled region or range of expected values ​​for the PS1, PS2, and DE equalization parameters.

[0032] If the determination at step 306 is positive, process 300 proceeds to step 308, where calibration unit 106 uses an estimation process to calculate the coefficients α, β, ω, and σ of the dependent variables in Equation 1. In one embodiment, calibration unit 106 uses a minimum mean square error (MMSE) estimation process to calculate the coefficients α, β, ω, and σ of the dependent variables. To this end, at step 308, calibration unit 106 calculates the following matrix in Equation 1:

number

[0033] In one embodiment, the resulting equation for the Pulse Pattern Generator (PPG) of the BERT is given in Equation 2 as follows:

[0034]

number

[0035] The multivariate calibration process 300 utilizes the linear model of Equation 1, a more specific embodiment of which, as shown in Equation 2, was verified with various combinations of the desired PS1(setting), PS2(setting), and DE(setting) equalization parameters listed in the PCIe 6.0 standard for a data rate of 64 GT / s. Recall that, as discussed above, the PCIe 6.0 standard defines 127 preset combinations ("presets") that are allowable combinations of the PS1, PS2, and DE equalization parameter values ​​used in calibrating a signal generator. To perform this verification, the model of Equation 1 and Equation 2 were utilized to provide device settings for the signal generator 104. For each preset combination, the desired PS1(measured), PS2(measured), and DE(measured) values ​​were input into these model equations, and the corresponding PS1(setting), PS2(setting), and DE(setting) values ​​were calculated. These PS1(setting), PS2(setting), and DE(setting) setting values ​​were then used to configure the signal generator 104. These three equalization parameters PS1(measured), PS2(measured), and DE(measured) are then measured using oscilloscopes 102A and 102B. These measurements are then compared to expected or target values ​​to determine how well the model provided expected values ​​for the equalization parameters PS1(measured), PS2(measured), and DE(measured) measured after configuring signal generator 104 using the model.

[0036] 4-6 are graphs showing the results of validation tests performed on multivariable calibration process 300 using Equation 1 and Equation 2. The vertical axis of each graph shows expected and measured values ​​in decibels (dB), and the horizontal axis shows the 84 presets of the corresponding equalization parameters used in the validation. FIG. 4 is a graph showing the set PS1 (expected value), measured PS1 (measured value), and error values ​​for the PS1 equalization parameter in multivariable calibration process 300 of FIG. 3 according to some embodiments of the present disclosure. As shown in this graph, the error between the expected and measured values ​​of the PS1 equalization parameter is small. FIG. 5 similarly shows the same values ​​for equalization parameters PS2 (expected value) and PS2 (measured value) and the error between these two values. FIG. 6 similarly shows the same values ​​for equalization parameters DE (expected value) and DE (measured value) and the error between these two values.

[0037] In an embodiment of the disclosed technology, when the multivariable model of Equation 1 was used to calibrate the signal generator 104, the average error of the de-emphasis equalization parameter DE was 0.2 dB, the average error of the first preshoot equalization parameter PS1 was 0.11 dB, and the average error of the second preshoot equalization parameter PS1 was 0.15 dB. The PCIe 6.0 standard requires that the error between the expected and measured parameters be within + / - 0.5 dB for the signal generator 104 to pass the calibration standard or transmitter EQ test. The error achieved when using the multivariable model of Equation 1 was within the required transmitter Tx EQ test pass margin 99% of the time for the test cases used. The percentage of cases where the error was between 0.5 dB and 1 dB, and therefore did not meet the criteria for proper calibration according to the transmitter Tx EQ test defined in the PCIe 6.0 standard, was 1%.

[0038] In embodiments of the disclosed technology, the time required to calibrate the signal generator 104 is significantly reduced compared to current conventional approaches. Calibration of the signal generator 104 can be completed in approximately 15 minutes using the calibration process 300 of FIG. 3, which is 10 times faster than current conventional calibration processes, such as iteratively adjusting equalization parameters. The dependency of the equalization parameters, particularly the dependency of the PS2 parameter on the PS1 and DE parameters, makes such an iterative approach more difficult and time-consuming.

[0039] Multivariate calibration process 300 according to embodiments of the disclosed technology is not limited to utilizing the linear model shown in Equation 1 for the dependencies or relationships between the equalization parameters PS1(SET), PS2(SET), and DE(SET), and PS1(MEASURED), PS2(MEASURED), and DE(MEASURED). In some embodiments, process 300 may utilize a higher-order model, such as a quadratic model, to estimate the dependencies between the set equalization parameters and the measured equalization parameters.

[0040] As mentioned above, in some embodiments, the multivariable calibration process 300 may utilize two different models as part of the calibration process for the signal generator 104. The embodiments of the multivariable calibration process 300 described above with reference to FIGS. 3-6 correspond to a linear three-variable model for a preset combination of equalization parameters PS1, PS2, and DE, where equalization parameter PS2≠0. In some embodiments, the multivariable calibration process 300 utilizes a preset combination of a linear two-variable model, with parameter PS2=0. In this situation, some embodiments of the process 300 utilize a two-variable multivariable model given by the following equation:

number

number

[0041] In one embodiment of the linear two-variable model, the resulting equation for the BERT's pulse pattern generator (PPG) is given in Equation 4 as follows:

number

[0042] Multivariable calibration process 300, utilizing the two-variable linear model of Equation 3 and Equation 4, was verified for various combinations of the expected PS1(setting) and DE(setting) equalization parameters (PS2(setting) = 0) listed in the PCIe 6.0 standard for a data transfer rate of 64 GT / s. FIGS. 7 and 8 are graphs showing the results of verification tests performed on multivariable calibration process 300 using Equation 3 and Equation 4. The vertical axis of each of these graphs represents the expected and measured values ​​in decibels (dB), and the horizontal axis represents the 42 presets of the corresponding equalization parameters used in the verification. While FIG. 7 is a graph showing the PS1(expected value), PS1(measured value), and error values ​​for the PS1 equalization parameter, FIG. 8 shows the same values ​​for the equalization parameters DE(expected value) and DE(measured value), as well as the error between these two values. As shown in FIGS. 7 and 8, the error between the expected and measured values ​​of the PS1 and DE equalization parameters is small, as shown in these graphs.

[0043] Aspects of the disclosed technology may operate on specially created hardware, firmware, digital signal processors, or specially programmed general-purpose computers, including processors that operate according to programmed instructions. The terms "controller" or "processor" herein contemplate microprocessors, microcomputers, ASICs, and dedicated hardware controllers, among others. Aspects of the disclosed technology may be implemented with computer-usable data and computer-executable instructions, such as one or more program modules, executed by one or more computers (including a monitoring module) or other devices. Generally, program modules include routines, programs, objects, components, data structures, etc., which, when executed by a processor in a computer or other device, perform particular tasks or implement particular abstract data types. Computer-executable instructions may be stored in computer-readable storage media, such as hard disks, optical disks, removable storage media, solid-state memory, RAM, etc. Those skilled in the art will appreciate that the functionality of the program modules may be combined or distributed as desired in various embodiments. Furthermore, such functionality may be embodied in whole or in part in firmware or hardware equivalents, such as integrated circuits, field programmable gate arrays (FPGAs), etc. Certain data structures may be used to more effectively implement one or more aspects of the disclosed technology, and such data structures are considered within the scope of the computer-executable instructions and computer-usable data described herein.

[0044] The disclosed aspects may, in some cases, be implemented in hardware, firmware, software, or any combination thereof. The disclosed aspects may also be implemented as instructions carried by or stored on one or more computer-readable media, which may be read and executed by one or more processors. Such instructions may be referred to as a computer program product. As used herein, computer-readable media refers to any medium that can be accessed by a computing device. By way of example, and not limitation, computer-readable media may include computer storage media and communication media. Example

[0045] The following examples are provided to aid in understanding the technology disclosed in this application. Embodiments of the technology may include one or more of the examples described below, and any combination thereof.

[0046] Example 1 is a method including the steps of: selecting expected values ​​for a set of two or more equalization parameters of a stress signal generated by a signal generator; setting values ​​of the set of two or more equalization parameters in the signal generator based on the expected values ​​and a multivariate model defining a relationship between the expected values ​​and measured values ​​of the two or more equalization parameters; providing the stress signal generated by the signal generator to at least one test and measurement instrument; measuring the set of two or more equalization parameters of the stress signal with the at least one test and measurement instrument; repeating the steps from selecting the expected values ​​to measuring the set of two or more equalization parameters N times to generate N combinations of measured sets of two or more equalization parameters from the N combinations of expected values ​​of the set of two or more equalization parameters; applying an estimation process to estimate coefficients of a dependent variable of the multivariate model defining a relationship between the expected set and the measured set of two or more equalization parameters; and generating setting values ​​of the set of two or more equalization parameters for calibrating the signal generator using the multivariate model including the estimated coefficients of the dependent variables.

[0047] In a second embodiment, the method according to the first embodiment is characterized in that the set of two or more equalization parameters includes preshoot 1 (PS1), preshoot 2 (PS2), and de-emphasis (DE) equalization parameters.

[0048] A third embodiment is the method according to the first embodiment, further comprising: testing a device under test (DUT) using the calibrated signal generator.

[0049] Example 4 is the method according to example 1, in which the process of testing the DUT using the calibrated signal generator includes a process of performing a PCIe 6.0 receiver (Rx) test of the DUT.

[0050] Example 5 is the method according to example 1, in which the multivariate model is a linear model.

[0051] Example 6 is the method according to example 5, in which the set of two or more equalization parameters is preshoot 1 (PS1), preshoot 2 (PS2), and de-emphasis (DE) equalization parameters, and the multivariable model is given by a three-variable model defined by the following equation:

number

[0052] Example 7 is the method according to example 5, wherein the set of two or more equalization parameters is a preshoot 1 (PS1) and a de-emphasis (DE) equalization parameter, and the multivariable model is given by a two-variable model defined by the following equation:

number

[0053] An eighth embodiment is the method according to the first embodiment, in which the estimation process is a minimum mean square error (MMSE) calculation.

[0054] Example 9 is the method according to example 1, in which the signal generating device includes a bit error rate tester (BERT).

[0055] Example 10 is the method according to example 1, in which the at least one test and measurement instrument includes at least one oscilloscope.

[0056] Example 11 is the method according to example 10, wherein the at least one oscilloscope includes a master oscilloscope and an expansion oscilloscope.

[0057] Example 12 is a method according to Example 11, wherein a lane of the bus has a first differential signal coupled to the master oscilloscope and the second differential signal coupled to the extension oscilloscope.

[0058] Example 13 is a test and measurement system including: a signal generator configured to generate a stress signal having characteristics defined by parameters including a set of two or more equalization parameters; and at least one test and measurement device coupled to the signal generator to receive the stress signal and configured to capture the stress signal, the test and measurement device having a processor, the processor configured to: selecting expected values ​​for a set of two or more equalization parameters of the stress signal generated by the signal generator; setting values ​​of the set of two or more equalization parameters in the signal generator based on a multivariate model that defines a relationship between the expected values ​​and the measured values ​​of the two or more equalization parameters and the expected values; providing the stress signal from the signal generator to the at least one test and measurement instrument; measuring a set of two or more equalization parameters of the captured stress signal; applying an estimation process to estimate coefficients of dependent variables in a multivariate model that defines a relationship between a set of expected values ​​and a set of measurements of two or more equating process parameters; generating values ​​of the set of two or more equalization parameters for calibrating the signal generator using a multivariate model including estimated coefficients of dependent variables; The calibration process is configured to perform the following.

[0059] Example 14 is the test and measurement system according to Example 12, in which the set of two or more equalization parameters includes a preshoot 1 (PS1) equalization parameter, a preshoot 2 (PS2) equalization parameter, and a de-emphasis (DE) equalization parameter.

[0060] Example 15 is the test and measurement system according to Example 12, further comprising: a process of testing a device under test (DUT) using the calibrated signal generator.

[0061] Example 16 is a test and measurement system according to Example 14, in which the test of the DUT is a test of a PCIe 6.0 receiver (Rx) of the DUT.

[0062] Example 17 is the test measurement system according to example 13, in which the multivariate model is a linear model.

[0063] Example 18 is the test and measurement system according to Example 17, wherein the set of two or more equalization parameters is preshoot 1 (PS1), preshoot 2 (PS2), and de-emphasis (DE) equalization parameters, and the multivariable model is given by a three-variable model defined by the following equation:

number

[0064] Example 19 is the test and measurement system according to example 17, wherein the set of two or more equalization parameters is a preshoot 1 (PS1) and a de-emphasis (DE) equalization parameter, and the multivariable model is given by a two-variable model defined by the following equation:

number

[0065] Example 20 is the test and measurement system according to Example 13, in which the estimation process is a minimum mean square error (MMSE) calculation.

[0066] Example 21 is a method including a process of selecting expected values ​​of preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE) equalization parameters of a stress signal generated by a signal generator; a process of setting values ​​of the preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE) equalization parameters in the signal generator based on the expected values ​​and a multivariate model that defines a relationship between the expected values ​​and measured values ​​of the PS1, PS2 and DE equalization parameters; a process of receiving, from the signal generator, a stress signal that is generated based on the preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE) equalization parameters set in the signal generator; and a process of adjusting the PS1, PS2 and DE values ​​of the received stress signal. measuring the equalization parameters of PS1, PS2 and DE; repeating the steps from selecting expected values ​​to measuring the equalization parameters of PS1, PS2 and DE N times to generate N combinations of measured equalization parameters of PS1, PS2 and DE from the N combinations of expected equalization parameters of PS1, PS2 and DE; applying an estimation process to estimate coefficients of dependent variables of the multivariate model that defines the relationship between the expected and measured equalization parameters of PS1, PS2 and DE; and generating setting values ​​for the PS1, PS2 and DE equalization parameters for calibrating the signal generator using the multivariate model including the estimated coefficients of the dependent variables.

[0067] Example 22 is the method according to example 21, in which the signal generating device includes a bit error rate tester (BERT).

[0068] Example 23 is the method according to example 21, further comprising: testing a device under test (DUT) using the calibrated signal generator.

[0069] The above description of the present invention is set forth merely to illustrate the present invention and is not intended to be limiting. Since modifications of the disclosed embodiments incorporating the gist of the present invention can be thought of by those skilled in the art, the invention should be construed as including all modifications within the scope of the present invention.

[0070] Although the above-described versions of the presently disclosed subject matter have many advantages that have been described or that will be apparent to those skilled in the art, not all of these advantages or features are required in every version of the disclosed devices, systems, or methods.

[0071] Additionally, the description in this application refers to specific features. All features disclosed in this specification, including the claims, abstract, and drawings, and all steps in all disclosed methods or processes, may be combined in any combination, unless they are at least partially mutually exclusive. Each feature disclosed in this specification, including the claims, abstract, and drawings, may be replaced with an alternative feature serving the same, equivalent, or similar purpose, unless otherwise specified.

[0072] Furthermore, when this application refers to a method having two or more defined steps or processes, these defined steps or processes may be performed in any order or simultaneously, unless the circumstances preclude this possibility.

[0073] Although specific embodiments of the invention have been illustrated and described for purposes of illustration, it will be appreciated that various modifications can be made therein without departing from the spirit and scope of the invention. Accordingly, the invention should not be limited except as by the appended claims.

Claims

1. selecting expected values ​​for a set of two or more equalization parameters of the stress signal generated by the signal generator; setting values ​​of the set of two or more equalization parameters in the signal generator based on a multivariate model that defines a relationship between the expected values ​​and the measured values ​​of the two or more equalization parameters and the expected values; providing the stress signal generated by the signal generator to at least one test and measurement device; measuring the set of two or more equalization parameters of the stressed signal with the at least one test and measurement instrument; a process of repeating the steps from the process of selecting an expected value to the process of measuring the set of two or more equalization parameters N times to generate N combinations of the set of two or more measured equalization parameters from the N combinations of expected values ​​of the set of two or more equalization parameters; applying an estimation process to estimate coefficients of dependent variables of a multivariate model that defines the relationship between an expected set and a measured set of two or more equalization process parameters; generating settings for a set of two or more equalization processing parameters for calibrating the signal generator using a multivariate model including estimated coefficients for dependent variables; A method comprising:

2. 2. The method according to claim 1, wherein the set of two or more equalization parameters includes preshoot 1 (PS1), preshoot 2 (PS2), and de-emphasis (DE) equalization parameters.

3. 10. The method according to claim 1, further comprising: using the calibrated signal generator to test a device under test (DUT).

4. 2. The method according to claim 1, wherein testing the DUT using the calibrated signal generator comprises performing a PCIe 6.0 receiver (Rx) test on the DUT.

5. 2. The method according to claim 1, wherein said multivariate model is a linear model.

6. The set of two or more equalization parameters is equalization parameters of preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE), and the multivariable model is given by a three-variable model defined by the following equation: [Equation 1] 6. The method according to claim 5, wherein PS1(set), PS2(set) and DE(set) are the values ​​of the equalization processing parameters PS1, PS2 and DE set in the signal generator, and PS1(measure), PS2(measure) and DE(measure) are the corresponding measured values ​​of these equalization processing parameters in the at least one test measurement instrument.

7. The set of two or more equalization parameters is a preshoot 1 (PS1) and a de-emphasis (DE) equalization parameter, and the multivariable model is given by a two-variable model defined by the following equation: [Equation 3] 6. The method according to claim 5, wherein PS1(set) and DE(set) are the values ​​of the equalization processing parameters PS1 and DE set in the signal generator, and PS1(measure) and DE(measure) are the corresponding measured values ​​of these equalization processing parameters in the at least one test and measurement instrument.

8. 2. The method according to claim 1, wherein the estimation process is a minimum mean square error (MMSE) operation.

9. 2. The method according to claim 1, wherein said signal generator comprises a Bit Error Rate Tester (BERT).

10. 10. The method according to claim 1, wherein said at least one test and measurement instrument comprises at least one oscilloscope.

11. 11. The method according to claim 10, wherein said at least one oscilloscope includes a master oscilloscope and an expansion oscilloscope.

12. 12. The method of claim 11, wherein a lane of a bus has a first differential signal coupled to the master oscilloscope and a second differential signal coupled to the extension oscilloscope.

13. a signal generator configured to generate a stress signal having characteristics defined by parameters including a set of two or more equalization parameters; at least one test and measurement device coupled to the signal generator to receive the stress signal and configured to capture the stress signal, the test and measurement device having a processor; Equipped with The processor: selecting expected values ​​for a set of two or more equalization parameters of the stress signal generated by a signal generator; setting values ​​of the set of two or more equalization parameters in the signal generator based on a multivariate model that defines a relationship between the expected values ​​and the measured values ​​of the two or more equalization parameters and the expected values; providing the stress signal from the signal generator to the at least one test and measurement instrument; measuring a set of two or more equalization parameters of the captured stress signal; applying an estimation process to estimate coefficients of dependent variables of a multivariate model that defines relationships between setpoints and measurements of two or more sets of equalization process parameters; generating values ​​of the set of two or more equalization processing parameters for calibrating the signal generator using the multivariate model including the estimated coefficients of the dependent variables; A test and measurement system configured to perform a calibration process to:

14. 14. The test and measurement system of claim 13, wherein the set of two or more equalization parameters includes a preshoot 1 (PS1) equalization parameter, a preshoot 2 (PS2) equalization parameter, and a de-emphasis (DE) equalization parameter.

15. 14. The test and measurement system of claim 13, further comprising testing a device under test (DUT) using the calibrated signal generator.

16. 16. The test and measurement system of claim 15, wherein the test of the DUT is a test of a PCIe 6.0 receiver (Rx) of the DUT.

17. 14. The test and measurement system of claim 13, wherein the multivariable model is a linear model.

18. The set of two or more equalization parameters is equalization parameters of preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE), and the multivariable model is given by a three-variable model defined by the following equation: [Equation 1] 18. The test and measurement system of claim 17, wherein PS1(setting), PS2(setting) and DE(setting) are values ​​of the equalization processing parameters PS1, PS2 and DE set in the signal generator, and PS1(measurement), PS2(measurement) and DE(measurement) are corresponding measured values ​​of these equalization processing parameters in the at least one test and measurement instrument.

19. The set of two or more equalization parameters is a preshoot 1 (PS1) and a de-emphasis (DE) equalization parameter, and the multivariable model is given by a two-variable model defined by the following equation: [Equation 3] 18. The test and measurement system of claim 17, wherein PS1(set) and DE(set) are values ​​of the equalization processing parameters PS1 and DE set in the signal generator, and PS1(measurement) and DE(measurement) are corresponding measured values ​​of these equalization processing parameters in the at least one test and measurement instrument.

20. 14. The test and measurement system of claim 13, wherein the estimation process is a minimum mean square error (MMSE) calculation.

21. selecting expected values ​​of preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE) equalization parameters of the stress signal generated by the signal generator; setting values ​​of preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE) equalization parameters in a signal generator based on a multivariate model that defines a relationship between the expected values ​​and measured values ​​of the PS1, PS2 and DE equalization parameters and the expected values; receiving, from the signal generator, a stress signal generated based on equalization processing parameters of preshoot 1 (PS1), preshoot 2 (PS2) and de-emphasis (DE) set in the signal generator; measuring the PS1, PS2 and DE equalization parameters of the received stressed signal; a process of repeating the steps from the process of selecting an expected value to the process of measuring the equalization processing parameters PS1, PS2, and DE N times, and generating N combinations of measured values ​​of the equalization processing parameters PS1, PS2, and DE from N combinations of set values ​​of the equalization processing parameters PS1, PS2, and DE; applying an estimation process to estimate coefficients of dependent variables of the multivariable model defining the relationship between the set values ​​of the PS1, PS2 and DE equalization process parameters and the measurements; and using the multivariate model including the estimated coefficients of the dependent variables to generate settings for the PS1, PS2 and DE equalization parameters for calibrating the signal generator for any expected combination of the PS1, PS2 and DE equalization parameters.

22. 22. The method of claim 21, wherein the signal generator comprises a Bit Error Rate Tester (BERT).

23. 22. The method of claim 21, further comprising: testing a device under test (DUT) using the calibrated signal generator.

Images