Compensation system and methods for a linear power amplifier

• US 5,990,738 A
• Filed: 12/17/1998
• Issued: 11/23/1999
• Est. Priority Date: 06/19/1998
• Status: Expired due to Term

• Current Assignee: Maxlinear Asia Singapore Pte Ltd. (Maxlinear Incorporated)
• Original Assignee: Datum Telegraphic, Inc. (Microchip Technology Incorporated)
• Inventors: Wright, Andrew S., Bennett, Steven J.
• Primary Examiner(s): Pascal, Robert
• Assistant Examiner(s): NGUYEN, KHANH V
• Application Number:
  US09/216,091
• Time in Patent Office: 341 Days
• Field of Search: 330/2, 330/124 R, 330/136, 330/149, 330/295
• US Class Current: 330/149
• CPC Class Codes: H03F 1/0294 using vector summing of two or more constant amplitude phase-modulated signals
• CPC Class Codes: H03F 1/3247 using feedback acting on predistortion circuits H03F1/3264 takes precedence
• CPC Class Codes: H03F 2201/3233 Adaptive predistortion using lookup table, e.g. memory, RAM, ROM, LUT, to generate the predistortion

Abstract

A LINC amplifier of a radio frequency transmitter provides substantially linear amplification from two nonlinear amplifiers by decomposing the original signal into two constant amplitude envelope, phase varying signals, which, when combined, constructively and destructively interfere to re-form the original signal. The output of the LINC amplifier, which is to be transmitted via an antenna, is an amplified form of the original signal. The LINC amplifier uses a digital control mechanism to control and adapt a digital compensation network that directly compensates for the imperfections of the analog RF environment, including the amplifiers. The mechanism monitors the combined amplifier output and adjusts the signal components in order to precisely compensate for any differences in the characteristics of the separate signal paths which would cause the combination not to accurately represent the original signal. The mechanism also corrects the component signals using information which can be applied to the amplifiers independent of the signal to be transmitted.

Claims

1. In a LINC power amplifier in which an input transmission signal is separated into component signals that are amplified along respective amplification chains and then combined by a signal combiner to form an output transmission signal, and in which circuitry is provided for monitoring the output transmission signal, a method of generating a numeric model of the amplification chains and combiner so that imperfections therein can be corrected during normal operation of the amplifier, the method comprising:

(a) providing the numeric model with an initial set of model parameters;

(b) applying a stimulation sequence to the amplification chains to generate an observed output sequence at an output of the combiner;

(c) applying the stimulation sequence to the model to calculate a predicted output sequence, and comparing the predicted output sequence to the observed output sequence; and

(d) iteratively refining the model parameters and repeating step (c) until a difference between the predicted and observed output sequences is substantially minimized.

2. The method of claim 1, further comprising repeating steps (b), (c) and (d) using at least one other stimulation sequence to identify a set of model parameters which substantially minimizes a difference between the predicted and observed output sequences for a plurality of different stimulation sequences.

3. The method of claim 1, wherein step (d) comprises using a Least Mean Squares, Recursive Least Squares, Kalman, or Extended Kalman algorithm to generate updates to the model parameters.

4. The method of claim 1, wherein step (a) comprises using a default set of model parameters.

5. The method of claim 1, wherein step (a) comprises generating the initial set of model parameters by:

(a1) applying a known stimulation sequence to at least one the amplification chains and recording a resulting output sequence of the combiner;

(a2) correlating the known stimulation sequence with the output sequence recorded in step (a1) to measure at least one of the following amplification chain characteristics;

amplitude gain, propagation delay and phase rotation; and

(a3) generating an initial set of model parameters which incorporates measurements made in step (a2).

6. The method of claim 5, wherein step (a1) comprises applying a first stimulation sequence to a first amplification chain while concurrently applying a second stimulation sequence which is orthogonal to the first stimulation sequence to a second amplification chain.

7. The method of claim 5, wherein step (a1) comprises applying a first stimulation sequence to a first amplification chain while concurrently applying a second stimulation sequence which is in antiphase with the first stimulation sequence to a second amplification chain.

8. The method of claim 1, wherein the numerical model implements a finite impulse response digital filter.

9. The method of claim 8, wherein the numerical model further implements a quadrature modulator imperfection circuit.

10. The method of claim 1, further comprising using a set of model parameters generated using steps (a)-(d) to generate a set of compensation parameters to be used by a compensation processor to modify the component signals during normal operation of the amplifier.

11. The method of claim 10, wherein generating the set of compensation parameters comprises using the compensation processor in combination with the numeric model to simulate the operation of the amplifier while adaptively adjusting the compensation parameters.

12. The method of claim 10, further comprising adaptively updating the compensation parameters during normal operation of the amplifier.

13. The method of claim 1, wherein steps (a)-(d) are implemented though a firmware program within a transmitter.

14. In a LINC amplifier system in which an input transmission signal is separated into component signals that are amplified along respective amplification chains and then combined by a signal combiner to form an output transmission signal, and in which the component signals are modified by compensation processors, based on sets of compensation parameters, to compensate for differences in the amplification chains, a method of adaptively adjusting the compensation parameters, comprising:

(a) generating a numeric model which represents the amplification chains and the combiner;

(b) applying a test sequence to a compensation module which includes or represents the compensation processors, and applying a resulting output of the compensation module to the numeric model, to simulate the operation of the amplifier system; and

(c) during step (b), adaptively updating the compensation parameters until an error in an output signal generated by the numeric model falls to within a predetermined threshold.

15. The method of claim 14, wherein steps (a)-(c) are performed while inhibiting transmissions of a signal.

16. The method of claim 14, wherein step (a) comprises:

(a1) applying a stimulation sequence to the amplification chains to generate an observed output sequence at an output of the combiner;

(a2) applying the stimulation sequence to the model to calculate a predicted output sequence, and comparing the predicted output sequence to the observed output sequence; and

(a3) iteratively refining parameters of the model and repeating step (a2) until a difference between the predicted and observed output sequences is substantially minimized.

17. The method of claim 14, wherein step (c) comprises using a Least Mean Squares, Recursive Least Squares, Kalman, or Extended Kalman algorithm to generate updates to the compensation parameters.

18. The method of claim 14, further comprising commencing a transmission using a set of compensation parameters computed from step (c).

19. The method of claim 14, wherein steps (b) and (c) are implemented using a firmware program of a transmitter.

20. The method of claim 14, wherein step (b) comprises replicating the compensation processors through software.

21. In a LINC amplifier system in which an input transmission signal is separated into component signals that are amplified along respective amplification chains and then combined by a signal combiner to form an output transmission signal, a method of compensating for a difference between the amplification chains, comprising, during a transmission event:

(a) separating the input transmission signal into first and second component signals of substantially constant amplitude and varying phase;

(b) modifying at least one of the first and second component signals using a compensation processor, based on a set of compensation parameters, to compensate for the difference;

(c) applying the first and second component signals as modified in step (b) to respective amplification chains;

(d) monitoring an observed output sequence at an output of the signal combiner;

(e) comparing the observed output sequence to a predicted output sequence, the predicted output sequence based on the input transmission signal; and

(f) adaptively adjusting the compensation parameters to reduce a difference between the observed output sequence and the predicted output sequence.

22. The method of claim 21, wherein step (e) comprises generating the predicted output sequence from the input transmission signal.

23. The method of claim 21, wherein step (e) comprises generating the predicted output sequence from the component signals.

24. The method of claim 21, wherein step (f) comprises using a Least Mean Squares, Recursive Least Squares, Kalman, or Extended Kalman algorithm to generate updates to the compensation parameters.

25. The method of claim 21, wherein step (f) comprises applying an interpolation function to successive sets of compensation parameters to thereby reduce spurious emissions during updates to the compensation parameters.

26. The method of claim 21, wherein steps (e) and (c) comprise processing signal samples that are stored in memory.

27. The method of claim 21, wherein step (f) is performed in a non-real-time mode.

28. A linear power amplifier system for radio frequency transmissions, comprising:

a signal separator circuit which separates an input transmission signal into first and second component signals;

first and second signal processing chains which process the first and second component signals, respectively, at least one of the signal processing chains including a compensation processor which modifies the respective component signal based on a set of compensation parameters to compensate for differences between the first and second processing chains;

a signal combiner circuit which combines the components signals for transmission following processing along the signal processing chains;

a signal monitoring circuit which captures and stores within a computer memory sequences of digital signal samples that represent an expected and an actual output of the signal combiner circuit during transmission events; and

a compensation estimation circuit which reads the samples from the computer memory and processes the samples in a non-real-time mode to generate updates to the compensation parameters.

29. The linear power amplifier system as in claim 28, wherein the compensation estimation circuit implements a Least Mean Squares, Recursive Least Squares, Kalman, or Extended Kalman algorithm in non-real-time to generate the updates to the compensation parameters.

30. The linear power amplifier system as in claim 29, wherein the compensation estimation circuit comprises a general-purpose DSP device that executes a computer program to implement the Least Mean Squares, Recursive Least Squares, Kalman, or Extended Kalman algorithm.

31. The linear power amplifier system in claim 28, wherein the signal separator circuit separates the input transmission signal into component signals of substantially constant amplitude and varying phase.