DIGITAL COMPENSATION FOR A NON-LINEAR ANALOG RECEIVER

• US 20160087604A1
• Filed: 09/22/2014
• Published: 03/24/2016
• Est. Priority Date: 09/22/2014
• Status: Active Grant

• Current Assignee: NanoSemi, Incorporated (Maxlinear Incorporated)
• Original Assignee: NanoSemi, Incorporated (Maxlinear Incorporated)
• Inventors: Kim, Helen H., Megretski, Alexandre, Li, Yan, Chuang, Kevin
• Granted Patent: US 9,564,876 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: H03H 17/0261 Non linear filters

Abstract

Aspects and embodiments are directed to non-linear systems including a digital compensator structure, a method of digital compensation, and methods for designing digital compensator structures for analog receivers. A digital compensator is configured to substantially reduce the one or more nonlinear distortion components in the sampled digital output signal from the analog receiver to provide an output signal achieving a receiver linearity requirement for the combination of the analog receiver and a digital compensator.

Claims

1. A digital compensator comprising:

at least one fixed, discrete-time (DT), linear time invariant (LTI) filter that provides an output signal z[n] and that is configured based on the known non-linear aspects of an analog receiver to a continuous time input signal x(t);

a reduced-complexity, discrete-time, nonlinear compensator that follows the fixed, discrete-time (DT), linear time invariant (LTI) filter to receive the output signal z[n] and provides a compensated output signal u[n];

wherein the digital compensator is configured to substantially reduce the one or more nonlinear distortion components in a sampled digital output signal of the analog receiver y[n] to provide the output signal u[n] achieving a receiver linearity requirement for the combination of the analog receiver and the digital compensator.

2. The digital compensator as claimed in claim 1, wherein the digital compensator is configured to provide a sum of products of filtered and time-delayed analog receiver samples combined and mapped through memoryless nonliearities to represent the sampled harmonic and intermodulation terms output by the analog receiver.

3. The digital compensator as claimed in claim 1, wherein the digital compensator comprises a cascade of one or more stages of compensator elements (1-Δ

, or 1-Δ

₁ and 1-Δ

₂), that replicate an added distortion (Δ

) component from the analog receiver, that subtracts the replicated discrete time distortion (Δ

) component so as to substantially reduce the added distortion (Δ

) component to yield an distortion component of the output signal u[n] approximately proportional to Δ

².

4. The digital compensator as claimed in claim 1, wherein the digital compensator includes a plurality of fixed, discrete-time (DT), linear time invariant (LTI) filters in series with a respective plurality of reduced-complexity, discrete-time, nonlinear compensators, with each such cascaded LTI filters and nonlinear compensator disposed in parallel and configured to account for varying harmonics.

5. The digital compensator as claimed in claim 1, wherein the fixed, discrete-time (DT), linear time invariant (LTI) filter comprises IIR filters that directly target certain harmonic resonances.

6. The digital compensator as claimed in claim 1, which comprises fixed, discrete-time (DT) linear time invariant (LTI) filters that are different for different Nyquist ranges.

7. The digital compensator as claimed in claim 1, wherein the reduced-complexity nonlinear compensator comprises a memoryless (NL) compensator that compensates the signal z[n] from the at least one fixed, discrete-time (DT), linear time invariant (LTI) filter based on coefficients of compensation to provide the compensated output signal u[n].

8. The digital compensator as claimed in claim 7, wherein the memoryless nonlinear (NL) compensator comprises fixed coefficients that are different for different Nyquist ranges.

9. The digital compensator as claimed in claim 8, wherein the coefficients comprise coefficients of nonlinear maps that change with changes in external parameters μ

, and wherein the digital compensator further comprises a calibration unit that computes and provides the coefficients of the nonlinear maps to the memoryless (NL) compensator as a function of the value of the external parameters μ

.

10. The digital compensator as claimed in claim 9, wherein the calibration unit computes and provides the coefficients of the nonlinear maps to the memoryless (NL) compensator based at least one of known operating characteristics of the analog receiver, a sensed operating voltage, a sensed temperature, and a Nyquist sampling rate, Nyquist range (1^st or 2^nd) input termination impedance type, input from aging sensors, and process variation.

11. A non-linear digital compensation system, comprising:

an analog receiver having a plurality of analog circuit components and an analog-to-digital converter, at least some of the analog circuit components having known nonlinear response characteristics to a continuous time input signal x(t), the analog receiver configured to receive the continuous time input signal x(t) and provide a sampled digital output signal y[n], which has one or more nonlinear distortion components; and

a digital compensator coupled to an output of the analog receiver, the digital compensator comprising at least one fixed, discrete-time (DT), linear time invariant (LTI) filter that provides an output signal z[n] and that is configured based on the known non-linear aspects of the analog receiver to a continuous time input signal x(t), followed by at least one reduced-complexity, discrete-time, nonlinear compensator that provides a compensated output signal u[n], the digital compensator configured in combination with the analog receiver to substantially reduce the one or more nonlinear distortion components in the sampled digital output signal of the analog receiver y[n] to provide an output signal u[n] achieving a receiver linearity requirement for the combination of the analog receiver and a digital compensator.

12. The non-linear digital compensation system as claimed in claim 11, wherein the one or more nonlinear distortion components comprise sampled harmonic and intermodulation terms, and wherein the digital compensator is configured to provide a sum of products of the filtered and time-delayed analog receiver samples combined and mapped through memoryless nonliearities to represent the sampled harmonic and intermodulation terms output by the analog receiver.

13. The non-linear digital compensation system as claimed in claim 11, wherein the digital compensator comprises a cascade of one or more stages of compensator elements (1-Δ

, or 1-Δ

₁ and 1-Δ

₂), that replicate a discrete time distortion (Δ

) component from the analog receiver and that subtracts the replicated discrete time distortion (Δ

) component so as to substantially reduce the added distortion (Δ

) component provided by the analog receiver to yield an distortion component of the output signal u[n] approximately proportional to Δ

².

14. The non-linear digital compensation system as claimed in claim 11, wherein the digital compensator includes a plurality of fixed, discrete-time (DT), linear time invariant (LTI) filters in series with a respective plurality of reduced-complexity, discrete-time, nonlinear compensators, with each such cascaded LTI filters and nonlinear compensator disposed in parallel and configured to account for varying harmonics.

15. The non-linear digital compensation system as claimed in claim 11, wherein the fixed, discrete-time (DT), linear time invariant (LTI) filter comprises IIR filters that directly target certain harmonic resonances.

16. The non-linear digital compensation system as claimed in claim 11, which comprises fixed, discrete-time (DT) linear time invariant (LTI) filters that are different for different Nyquist ranges.

17. The non-linear digital compensation system as claimed in claim 11, wherein the reduced-complexity nonlinear compensator comprises a memoryless (NL) compensator that compensates the signal z[n] from the at least one fixed, discrete-time (DT), linear time invariant (LTI) filter based on coefficients of compensation to provide the compensated output signal u[n].

18. The non-linear digital compensation system as claimed in claim 17, wherein the memoryless nonlinear (NL) compensator comprises fixed coefficients that are different for different Nyquist ranges.

19. The non-linear digital compensation system as claimed in claim 17, wherein the coefficients for compensation comprise coefficients of nonlinear maps that change with changes in external parameters μ

, and wherein the digital compensator further comprises a calibration unit that computes and provides the coefficients of the nonlinear maps to the memoryless (NL) compensator as a function of the value of the external parameters μ

.

20. The non-linear digital compensation system as claimed in claim 19, wherein the calibration unit computes and provides the coefficients of the nonlinear maps to the memoryless (NL) compensator based at least one of known operating characteristics of the analog receiver, a sensed operating voltage, a sensed temperature, and a Nyquist sampling rate, Nyquist range (1^st or 2^nd), input termination impedance type, input from aging sensors, and process variation.

21. The non-linear digital compensation system as claimed in claim 11, wherein the analog receiver comprises M-way interleaved Analog-to-Digital Converters (ADCs), and wherein the digital compensator comprises a shared, fixed LTI system that receives M output signals y[n] provided at the output of the interleaved analog receiver, that partially compensates the M output signals y[n] for known nonlinear response characteristics to a continuous time input signal x(t) to provide M output signals z[n] to M banks of programmable memoryless nonlinearities, which further compensate the M output signals z[n] based on coefficients of compensation to provide M output signals, and a summer that sums the M output signals to provide a compensated output signal u[n].

22. The non-linear digital compensation system as claimed in claim 21, wherein the a bank of programmable memoryless nonlinearities further comprises a respective first multiplier that multiplies one of M output signals from a bank with a sin function and a second multiplier that multiplies the one of M output signals with a cos function to provide a shift of each of M output signals.

23. The non-linear digital compensation system as claimed in claim 21, wherein the fixed LTI system comprises a combination of delay lines and reconstruction filters, followed by a partially-populated addition matrix that produces L combinations of z[n] outputs from the fixed LTI system.

24. The non-linear digital compensation system as claimed in claim 23, wherein each of the M banks of programmable memoryless nonlinearities comprises N memoryless nonlinear (NL) compensators with different compensation coefficients.

25. A method for designing a non-linear compensator system, comprising:

characterizing nonlinearities in an analog receiver design, the analog receiver configured to receive a continuous time input signal x(t) and provide a sampled digital output signal y[n];

designing a digital compensator for the analog receiver design based on the nonlinearities, the digital compensator configured to substantially reduce the nonlinearities in a sampled digital output signal y[n] of the analog receiver to provide an output signal u[n], the digital compensator comprising at least one fixed, discrete-time (DT), linear time invariant (LTI) filter configured based on the non-linearities of the analog receiver to a continuous time input signal x(t) that provides an output signal z[n] followed by at least one reduced-complexity, discrete-time, nonlinear compensator that provides a compensated output signal u[n].

26. The method for designing a non-linear system as claimed in claim 25, wherein the analog receiver comprises M-way interleaved Analog-to-Digital Converters (ADCs), and wherein designing the digital compensator comprises configuring the fixed, discrete-time (DT), linear time invariant (LTI) filter as a shared, fixed LTI system that receives M output signals y[n] provided at the output of the M-way interleaved Analog-to-Digital Converters (ADCs), that partially compensates the M output signals y[n] for known nonlinear response characteristics to provide M output signals z[n], and configuring the reduced-complexity, discrete-time, filter as M banks of programmable memoryless nonlinearities that further compensate the M output signals z[n] based on coefficients of compensation to provide M output signals, and further providing a summer that sums the M output signals to provide a compensated output signal u[n].

27. The method for designing a non-linear system as claimed in claim 25, wherein the method further comprises compensating for one or more sampled harmonic and intermodulation terms with a sum of filtered and time-delayed analog receiver samples combined and mapped through memoryless nonlinearities to represent sampled harmonic and intermodulation terms output by the analog receiver.

28. The method for designing a non-linear system as claimed in claim 25, wherein the method further comprises modeling the analog receiver as an ideal sampler that provides an undistorted discrete-time representation of the continuous time input signal x(t) and an added distortion (Δ

) component.

29. The method for designing a non-linear system as claimed in claim 28, wherein the method further comprises implementing the digital compensator as a cascade of one or more stages of compensator elements (1-Δ

, or 1-Δ

₁ and 1-Δ

₂) that replicate the discrete time distortion (Δ

) component and that are configured to subtract the replicated discrete time distortion (Δ

) component so as to substantially reduce the added distortion (Δ

) component provided by the analog receiver to yield an distortion component of the output signal approximately proportional to Δ

².

30. The method for designing a non-linear system as claimed in claim 25, wherein the method further comprises implementing the digital compensator as a plurality of fixed, discrete-time (DT), linear time invariant (LTI) filters in series with a respective plurality of reduced-complexity, discrete-time, nonlinear compensators, with each such cascaded LTI filters and nonlinear compensator disposed in parallel and configured to account for varying harmonics.

31. The method for designing a non-linear system as claimed in claim 25, wherein the method further comprises implementing the fixed, discrete-time (DT), linear time invariant (LTI) filter with IIR filters that directly target certain harmonic resonances.

32. The method for designing a non-linear system as claimed in claim 25, wherein the method further comprises implementing fixed, discrete-time (DT) linear time invariant (LTI) filters that are different for different Nyquist ranges.

33. The method for designing a non-linear system as claimed in claim 25, wherein the method further comprises implementing the reduced-complexity nonlinear compensator as a memoryless (NL) compensator that compensates the signal z[n] from the at least one fixed, discrete-time (DT), linear time invariant (LTI) filter based on coefficients of compensation to provide the compensated output signal u[n].

34. The method for designing a non-linear system as claimed in claim 33, wherein the method further comprises implementing the memoryless (NL) compensator with nonlinear maps that change with changes in external parameters μ

, and designing the digital compensator further comprises providing a calibration unit that computes and provides the coefficients of the nonlinear maps to the memoryless (NL) compensator as a function of the value of the external parameters μ

.

35. The method for designing a non-linear system as claimed in claim 34, wherein the method further comprises determining the coefficients based on at least one of known operating characteristics of the analog receiver, a sensed operating voltage, a sensed temperature, and a Nyquist sampling rate, Nyquist range (1^st or 2^nd), input termination impedance type, input from aging sensors, and process variation.