Search

DE-102025145713-A1 - MAXIMUM LIKELIHOOD SEQUENCE ESTIMATION ON REAL-TIME OSCILLOSCOPES

DE102025145713A1DE 102025145713 A1DE102025145713 A1DE 102025145713A1DE-102025145713-A1

Abstract

A real-time oscilloscope comprises one or more interfaces for connecting to a device under test (DUT) to establish a communication channel between the oscilloscope and the DUT; one or more analog-to-digital converters (ADCs) for sampling a signal received from the DUT to convert the signal into a waveform; one or more processors for receiving and sampling a time-of-flight signal, including actual symbols, from the DUT to generate a time-of-flight waveform; resampling the time-of-flight waveform to locate samples in the center of each unit interval; using a characterization of the communication channel, the samples in the center of each unit interval, a symbol constellation for the time-of-flight waveform, and a traceback length as inputs for a maximum likelihood sequence estimation (MLSE) procedure to estimate symbols in the time-of-flight waveform; and comparing the estimated symbols with the actual symbols to obtain a symbol error ratio.

Inventors

  • Kan Tan

Assignees

  • TEKTRONIX, INC.

Dates

Publication Date
20260513
Application Date
20251106
Priority Date
20251105

Claims (18)

  1. A real-time oscilloscope comprising: one or more interfaces for connecting to a device under test (DUT) and establishing a communication channel between the oscilloscope and the DUT; one or more analog-to-digital converters (ADCs) for sampling a signal received from the DUT to convert the signal into a waveform; one or more processors configured to execute code to cause the one or more processors to: receive and sample a time-of-flight signal, including actual symbols, from the DUT to generate a time-of-flight waveform; resample the time-of-flight waveform to locate samples in the center of each unit interval; Using a characterization of the communication channel, the samples in the middle of each unit interval, a symbol constellation for the time-of-flight waveform, and a traceback length formation as inputs for a maximum likelihood sequence estimation (MLSE) procedure to estimate symbols in the time-of-flight waveform; and comparing the estimated symbols with the actual symbols to obtain a symbol error ratio.
  2. The real-time oscilloscope according to Claim 1 , wherein the one or more processors are further configured to execute code to cause the one or more processors to characterize the communication channel.
  3. The real-time oscilloscope according to Claim 2 , wherein the code that causes the one or more processors to characterize the communication channel includes code that causes the one or more processors to receive the communication channel characterization from a vector network analyzer (VNA) or from a time domain reflectometer (TDR).
  4. The real-time oscilloscope according to Claim 2 or 3 , wherein the code that causes the one or more processors to characterize the communication channel comprises code that causes the one or more processors to: put the oscilloscope into a training mode; apply a training signal with a known pattern to the communication channel; receive and sample the training signal to generate a training waveform; resample the training waveform to generate a resampled training waveform with a predetermined number of samples per unit interval; extract a linearly fitted impulse response (LFPR) from the resampled training waveform; derive an impulse response from the LFPR; and select samples at unit interval midpoints around a peak value of the impulse response to estimate the characterization of the communication channel.
  5. The real-time oscilloscope according to Claim 4 , wherein the code that causes the one or more processors to derive the impulse response comprises code that causes the one or more processors to: construct a matrix X from a vector of an ideal impulse response; and unfold the matrix X from the LFPR.
  6. The real-time oscilloscope according to Claim 5 , wherein the code that causes the one or more processors to unfold the matrix X from the LFPR includes code that causes the one or more processors to unfold the matrix X by normalization.
  7. The real-time oscilloscope according to one of the Claims 4 until 6 , wherein the code that causes the one or more processors to characterize the communication channel comprises code that causes the one or more processors to: determine that the communication channel has a longer impulse response duration; and apply an equalizer to the samples that shortens the duration of the impulse response before deriving the impulse response.
  8. The real-time oscilloscope according to Claim 7 , wherein the code that causes one or more processors to determine that the channel has a long impulse response duration includes code that determines that a period of time in which the impulse response stabilizes includes a time exceeding a predetermined number of unit intervals.
  9. The real-time oscilloscope according to Claim 7 or 8 , wherein the code that causes one or more processors to apply an equalizer to the samples includes code that causes one or more processors to apply a forward feedback equalizer (FFE) or to use a linear real-time equalizer (CTLE).
  10. A method comprising: Samples a time-of-flight signal received from a device under test (DUT) via a communication channel as a time-of-flight waveform; Resample the time-of-flight signal to locate samples in the center of each unit interval; Use a characterization of the communication channel, the samples in the center of each unit interval, a symbol constellation for the waveform, and a traceback length as inputs for a maximum likelihood sequence estimation (MLSE) procedure to generate estimated symbols in the time-of-flight signal; and Compare the estimated symbols with actual symbols to determine a symbol error rate.
  11. The procedure according Claim 10 , which further includes characterizing the communication channel on a real-time oscilloscope using a training signal with a known or captured pattern to generate the characterization of the communication channel.
  12. The procedure according Claim 11 , wherein characterizing the communication channel includes receiving a channel characterization from a vector network analyzer (VNA) or a time domain reflectometer (TDR).
  13. The procedure according Claim 11 or 12 , wherein characterizing the communication channel includes: placing the oscilloscope into a training mode; applying the training signal to the communication channel; receiving and sampling the training signal to generate a training waveform; resampling the training waveform to obtain a predetermined number of samples per unit interval; extracting a linearly fitted impulse response (LFPR) from the resampled training waveform; deriving an impulse response from the LFPR; and selecting samples at unit interval midpoints around a peak value of the impulse response to estimate the characterization of the communication channel.
  14. The procedure according Claim 13 , where deriving the impulse response includes: constructing a matrix X from a vector of an ideal impulse response; and unfolding the matrix X from the LFPR.
  15. The procedure according Claim 14 , where unfolding matrix X from the LFPR includes unfolding matrix X by normalization.
  16. The procedure according to one of the Claims 13 until 15 , wherein characterizing the communication channel includes: determining that the communication channel has a long impulse response duration; and applying an equalizer to the samples to shorten the duration of the impulse response before deriving the impulse response.
  17. The procedure according Claim 16 , wherein determining that the communication channel has a long impulse response duration includes determining that a time span required for the impulse response to settle includes a time exceeding a predetermined number of unit intervals.
  18. The procedure according Claim 16 or 17 , where applying an equalizer to the samples includes applying a forward feedback equalizer (FFE) or a real-time linear equalizer (CTLE).

Description

REFERENCE TO RELATED REGISTRATIONS This revelation is a non-provisional revelation and claims the primacy of the provisional revelations. US application no. 63/718870 entitled “MOST LIKELIHOOD SEQUENCE DETECTION ON REAL-TIME OSCILLOSCOPES”, filed on November 11, 2024, the disclosure of which is incorporated herein by reference in its entirety. AREA OF TECHNOLOGY This disclosure relates to test and measuring instruments, in particular symbol recognition in transmitted symbols, especially the reduction of errors in transmitted symbols. BACKGROUND For high-speed serial data standards such as PCIe and IEEE 802.3 Ethernet, PAM4 signal transmission was developed to replace NRZ signal transmission. The explicit solution is in the US Patent No. 11,765,002 (the “002 patent”) of September 19, 2023, for the decision equalizer taps (DFE) derived from the single extracted linearly fitted impulse response (LFPR). Tan et al., “Multiple Linear Fit Pulse Responses for PAM4 Measurements”, DesignCon 2024 Compiled Papers TR8-14, February 22, 2024, pp. 438–453, available at https://fliphtml5.com/mddwo/smbr/DesignCon_2024_Compiled_Papers_Tr_8-14/ , hereinafter referred to as “Tan et al.”, the contents of which are hereby incorporated into this disclosure by reference, a new method is presented with which multiple LFPRs can be extracted from a pattern waveform. Multiple LFPRs contain more information about the signal. The new DFE structure called Multiple Pulses Based DFE (MPDFE), which uses multiple LFPRs, is described in the US patent application no. 18/754,871 presented, which was submitted on June 26, 2024, and the contents of which are hereby incorporated into this disclosure by reference. As signal speed increases, equalizers are often used in transmitters and receivers to improve system performance. Equalizers include the linear real-time equalizer (CTLE), the forward feedback equalizer (FFE), the decision feedback equalizer (DFE), and the newer multi-impulse DFE (MPDFE). Among the equalizers, the DFE and MPDFE recognize symbols and then use the recognized symbols to improve the recognition of subsequent symbols. If a symbol is incorrectly recognized, the symbol error can occur because the incorrect symbol is used in the feedback loop, as shown in [reference to example]. 1 shown to cause a series of multiple symbol errors, which are referred to as error propagation. BRIEF DESCRIPTION OF THE DRAWINGS 1 shows an embodiment of a decision feedback equalizer.2 shows an embodiment of a real-time oscilloscope.3 shows an embodiment of a trellis structure used in the disclosed embodiments.4 shows an example of a 4-step pulse amplitude modulation waveform (PAM4) from a transmitted signal.5 shows an enlarged view of the waveform from 4 .6 shows a linearly fitted impulse response derived from the waveform of 4 was extracted with samples in the unit interval centers.7 shows the linearly fitted impulse response with samples in the unit interval midpoints around the impulse level.8 shows results of a Maximum Likelihood Sequence Estimation (MLSE) for the PAM4 signal.9 shows a flowchart of an embodiment of a method for using MLSD to detect symbols in a waveform.10 shows an example of a PAM4 waveform after a channel with higher loss.11 shows an impulse response for the channel from 10 .12 displays the MLSE results for the PAM4 signal from 10 . DETAILED DESCRIPTION The embodiments described here use Maximum Likelihood Sequence Estimation (MLSE) to estimate the symbols. MSLE is sometimes also referred to as Maximum Likelihood Sequence Detection (MLSD). As explained above, the equalizers DFE and MPDFE detect the symbols and then use the detected symbols to improve the detection of subsequent symbols. If a symbol is detected incorrectly, the symbol error can occur because the incorrect symbol is used in the feedback loop, as shown in 1 shown to cause a series of multiple symbol errors, which are referred to as error propagation. To achieve higher symbol error rate performance, the embodiments disclosed herein employ Maximum Likelihood Sequence Estimation (MLSE). MLSE offers the optimal solution for sequence estimation in signals with noise and intersymbol interference (ISI). MLSE does not suffer from error propagation problems like DFE. The embodiments disclosed herein include a method for simulating MLSE on real-time oscilloscopes. Real-time oscilloscopes generally digitize a signal in real time by capturing and displaying a signal waveform in a single, continuous acquisition, typically using one or more high-speed analog-to-digital converters (ADCs) and an independent clock. They differ from equivalent-time sampling oscilloscopes in that sampling oscilloscopes take multiple samples over several signal periods to "build up" the complete waveform over time. 2 Figure 1 shows an embodiment of a real-time oscilloscope 10. The real-time oscilloscope 10 comprises one or more terminals 12, which can be any electrical signal transmission medium. The t