WIRELESS TRANSCEIVER WITH TX/FBRX SEQUENTIAL QMC CALIBRATION USING SEPARATE/SHARED PLLS

Abstract

A direct conversion wireless transceiver is configured for TX/FBRX sequential QMC calibration (coefficient generation) using separate/shared PLLs. A TX path includes a TX LO driving upconversion, and an FBRX path includes an RX LO driving downconversion. TX/RX digital compensators include TX/RX QMC compensators that perform QMC compensation to compensate for IQ mismatch based on TX/RX QMC filter coefficients, and QMC calibration to calibrate the TX/RX QMC filter coefficients based on a QMC calibration procedure. The TX LO signal source is a TX PLL, and the RX LO signal source is selectively the TX PLL or a separate RX PLL. A QMC controller perform QMC calibration to generate calibrated TX/FBRX QMC filter coefficients, including: disconnecting the TX PLL from, and connecting the FBRX PLL to, the RX LO; generating calibrated TX QMC filter coefficients; generating calibrated FBRX QMC filter coefficients; disconnecting the FBRX PLL from, and connecting the TX PLL to, the RX LO; generating re-calibrated FBRX QMC filter coefficients. The TX digital compensator can be configured to perform DPD compensation (after QMC compensation).

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

Priority is claimed under USC §119(e) to U.S. Provisional Application 62/035248 (Docket TI-75217PS), filed 2014 Aug. 08.

BACKGROUND

1. Technical Field

This Patent Document relates generally to direct-conversion wireless transceiver design including IQ mismatch compensation.

2. Related Art

In wireless transceivers, direct conversion can be used for the transmitter (TX) and/or receiver (RX). Direct conversion (zero/low IF) wireless architectures, based on IQ modulation/demodulation and RF upconversion/downconversion, have a number of advantages over IF (heterodyne) architectures.

For wireless base-station applications, the transceiver must meet stringent requirements on out-of-band emission in the transmitter. For this reason, direct conversion transmitter designs use digital compensation for TX non-linearities and IQ mismatch, and include a feedback receiver (FBRX) that captures data required for such compensation.

TX non-linearities are compensated by digital pre-distortion (DPD). IQ mismatch is compensated by digital filtering (IQ mismatch compensation or QMC). To reduce phase noise distortion for DPD processing, the TX LO (local oscillator) and FBRX LO can be generated using the same PLL (phase locked loop).

One approach to achieving acceptable TX DPD and QMC compensation, is to use an IF (rather than direct conversion) FBRX. Such an approach avoids introducing FBRX IQ imbalance into the TX path (i.e., avoiding the requirement for FBRX QMC).

While this Background information references wireless base station application, the Disclosure in this Patent Document is not limited to such applications, but is more generally directed to direct conversion wireless architectures.

BRIEF SUMMARY

This Brief Summary is provided as a general introduction to the Disclosure provided by the Detailed Description and Drawings, summarizing aspects and features of the Disclosure. It is not a complete overview of the Disclosure, and should not be interpreted as identifying key elements or features of, or otherwise characterizing or delimiting the scope of, the disclosed invention.

The Disclosure describes a direct conversion wireless transceiver configured for TX/FBRX sequential QMC calibration (coefficient generation) using separate/shared PLLs. According to aspects of the Disclosure, the wireless transceiver includes a TX path includes a TX LO driving upconversion, and an FBRX path includes an RX LO driving downconversion. TX/RX digital compensators include TX/RX QMC compensators that perform QMC compensation to compensate for IQ mismatch based on TX/RX QMC filter coefficients, and QMC calibration to calibrate the TX/RX QMC filter coefficients based on a QMC calibration procedure. The TX LO signal source is a TX PLL, and the RX LO signal source is selectively the TX PLL or a separate RX PLL. A QMC controller perform QMC calibration to generate calibrated TX/FBRX QMC filter coefficients, including: disconnecting the TX PLL from, and connecting the FBRX PLL to, the RX LO; generating calibrated TX QMC filter coefficients; generating calibrated FBRX QMC filter coefficients; disconnecting the FBRX PLL from, and connecting the TX PLL to, the RX LO; generating re-calibrated FBRX QMC filter coefficients. The TX digital compensator can be configured to perform DPD compensation (after QMC compensation).

Other aspects and features of the invention claimed in this Patent Document will be apparent to those skilled in the art from the following Disclosure.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates, for a direct conversion transceiver TX/RX/FBRX, an example embodiment digital DPD and FBRX/TX QMC (IQ imbalance) compensation is performed prior to DPD.

FIG. 2 illustrates an example functional embodiment of a direct conversion transceiver TX/RX/FBRX architecture, with LO generation using a shared TX PLL during normal TX/RX (after QMC and DPD compensation), such as for the purpose of reducing phase noise distortion for DPD compensation processing.

FIG. 3 illustrates an example functional embodiment of a direct conversion transceiver TX/RX/FBRX architecture, with separate LO generation using separate TX and FBRX PLLs during TX/FBRX sequential QMC (coefficient) calibration.

FIG. 4 illustrates an example embodiment of the methodology for TX/FBRX sequential QMC calibration using separate/shared TX/FBRX PLLs, including calibration of TX and FBRX QMC filter coefficients for normal TX operation with DPD compensation based on TX and FBRX signal path IQ modulation/demodulation using a shared TX PLL.

FIGS. 5A-5E illustrate an example embodiment of the methodology for TX/FBRX sequential QMC calibration using separate/shared TX/FBRX PLLs, including sequential calibration/re-calibration of TX and FBRX QMC filter coefficients prior to DPD compensation: (FIG. 5A) illustrates reconfiguring the transceiver for separate TX/FBRX PLLs, including generation of TX/FBRX images based on IQ imbalance; (FIGS. 5B-5C) illustrate sequential FBRX and TX QMC calibration using separate TX/FBRX PLLs, to suppress TX/FBRX images; (FIG. 5D) illustrates re-configuring the transceiver to use the shared TX PLL for FBRX and TX LO generation, and FBRX QMC re-calibration to suppress the FBRX image caused by the transition to the shared TX PPL; and (FIG. 5E) illustrates transceiver TX/FBRX signal path operation with the shared TX PLL for TX/FBRX LO generation, and including QMC compensation (with updated FBRX/TX QMC filter coefficients) and DPD adaptation.

DETAILED DESCRIPTION

This Description and the Drawings constitute a Disclosure for a direct conversion wireless transceiver architecture with TX/FBRX sequential QMC calibration using separate/shared PLLs, including example embodiments, and including various technical features and advantages.

As used in this Disclosure, direct conversion refers to zero and low IF (intermediate frequency) conversion based on quadrature (IQ) modulation and demodulation, with RF (radio frequency) upconversion and downconversion.

In brief overview, the Disclosed direct conversion wireless transceiver is configured for TX/FBRX sequential QMC calibration (coefficient generation) using separate/shared PLLs. A TX path includes a TX LO driving upconversion, and an FBRX path includes an RX LO driving downconversion. TX/RX digital compensators include TX/RX QMC compensators that perform QMC compensation to compensate for IQ mismatch based on TX/RX QMC filter coefficients, and QMC calibration to calibrate the TX/RX QMC filter coefficients based on a QMC calibration procedure. The TX LO signal source is a TX PLL, and the RX LO signal source is selectively the TX PLL or a separate RX PLL. A QMC controller perform QMC calibration to generate calibrated TX/FBRX QMC filter coefficients, including: disconnecting the TX PLL from, and connecting the FBRX PLL to, the RX LO; generating calibrated TX QMC filter coefficients; generating calibrated FBRX QMC filter coefficients; disconnecting the FBRX PLL from, and connecting the TX PLL to, the RX LO; generating re-calibrated FBRX QMC filter coefficients. The TX digital compensator can be configured to perform DPD compensation (after QMC compensation).

FIG. 1 illustrates, for a direct conversion transceiver TX/RX/FBRX, an example embodiment digital DPD and FBRX/TX QMC (IQ imbalance) compensation is performed prior to DPD.

For a transmission with a high output power, it is required for a DPD algorithm to linearize the PA for higher linearity. Before DPD, it is necessary to idealize both transmission path and feedback receive path so that no other impairments from TX and RX affects the DPD adaptation. These impairments include IQ imbalance of TX and RX paths. Hence, QMC (IQ imbalance) compensation is always run before DPD adaptation.

For the example embodiment, QMC compensation is implemented in a DSP (digital signal processor). Periodic QMC calibration generates filter coefficients for QMC compensation processing.

According to the TX/FBRX sequential QMC calibration methodology (using separate/shared), the first set of FBRX QMC coefficients and the TX QMC coefficients are updated using separate PLLs to generate the respective LO signals. After the RXFB and TX QMC coefficients are fixed, the FBRX LO signal is switched to the same PLL source as the TX LO. Because the new LO source can change the gain and phase mismatch generated by the FBRX mixer, introducing IQ mismatch, the FBRX QMC coefficients are then separately re-calibrated in this Shared-PLL configuration (such as using a least squares method).

An example approach for joint TX/FBRX QMC calibration using separate PLLS is described in U.S. Pat. No. 8,311,083 and U.S. application Ser. No. 14/814,197, the disclosures of which are incorporated by reference. In particular, Application '197 describes separate FBRX and TX QMC calibration using phase rotation (using separate PLLs outside the respective LO distribution paths). This example approach is referred to in this Disclosure as Joint TX/FBRX QMC Calibration

FIG. 2 illustrates an example functional embodiment of a direct conversion transceiver TX/RX/FBRX architecture, with LO generation using a shared TX PLL during normal TX/RX (after QMC and DPD compensation). To reduce phase noise distortion for the computation of the DPD, it is desirable that the FBRX LO that drives the FBRX mixer is generated by the same PLL as the TX LO that drives the TX mixer.

FIG. 3 illustrates an example functional embodiment of a direct conversion transceiver TX/RX/FBRX architecture, with separate LO generation using separate TX and FBRX PLLs during TX/FBRX sequential QMC (coefficient) calibration.

TX/FBRX LO generation using separate TX/FBRX PLLs can be used with the above referenced Joint TX/FBRX QMC Calibration methodology. As noted, separate TX/FBRX PLLs are allow phase rotation using PLL adjustments outside the LO distribution paths, separately driving the FBRX and TX mixers. This Joint TX/FBRX QMC Calibration approach relies on FBRX phase rotation to rotate the phase of FBRX LO without affecting the IQ mismatch in the FBRX path. Because of the requirement of the phase rotation, and that the phase rotation be accomplished outside the LO distribution paths, separate TX and FBRX PLL's are used for TX and FBRX QMC calibration.

However, as noted, the requirements on the performance of QMC compensation are stringent, and changing the LO waveform can change the gain and phase IQ mismatch, causing significant changes in the required QMC filtering if the PLL generating the LO signal is changed. Thus, FBRX QMC coefficients jointly generated with the TX QMC coefficients using separate FBRX and TX PLLs, such as the above referenced Joint TX/FBRX QMC Calibration technique (requiring phase rotation) will be degraded/invalid when the transceiver is re-configured with the TX PLL signal source used to generate the FBRX LO driving the FBRX mixer, because of degradation of image suppression performance (see, the description of FIGS. 5B and 5D).

FIGS. 4 and 5A-5E illustrate TX/FBRX sequential QMC calibration using separate/shared PLLs, using the transceiver TX/FBRX configuration of FIG. 3, with separate TX and FBRX PLLs for Joint TX/FBRX QMC Filter Calibration (coefficient generation), and then transitioning to the transceiver TX/FBRX configuration of FIG. 2 using a shared TX PLL as the signal source for both TX and FBRX LO generation for DPD adaptation (although, requiring FBRX QMC re-calibration, as described in connection with FIGS. 5D-5E).

FIG. 4 illustrates a sequence of TX/FBRX QMC calibration routines that are run before DPD adaptation/compensation based on TX and FBRX signal path IQ modulation/demodulation using a shared TX PLL. These routines implement the TX/FBRX sequential QMC calibration methodology using separate/shared TX/FBRX PLLs, including calibration of TX and FBRX QMC filter coefficients.

FIGS. 5A-5E (corresponding to FIG. 4, Steps A-D) illustrate the methodology for TX/FBRX sequential QMC calibration using separate/shared TX/FBRX PLLs, including a sequence of TX/FBRX QMC calibration/re-calibration routines (FIGS. 5B-5D) for computing TX/FBRX QMC filter coefficients prior to DPD adaptation/compensation.

FIG. 5A illustrates reconfiguring the transceiver for separate TX/FBRX PLLs, including generation of TX/FBRX images based on IQ imbalance. Referring also to FIG. 4 (step A), the transceiver TX and FBRX paths are switched to the FIG. 2 transceiver configuration by disconnecting the TX PLL from the FBRX mixer (FBRX LO), and connecting the FBRX PLL as the signal source for the FBRX LO driving the FBRX mixer. Due to IQ imbalances of TX path and RX path, TX and RX images are overlapped on top of to each other.

FIGS. 5B-5C illustrate sequential FBRX and TX QMC calibration using separate TX/FBRX PLLs, to suppress TX/FBRX images. Referring also to FIG. 4 (steps B-C), this sequence of QMC filter calibration (coefficient generation) routines can be accomplished using the above referenced Joint TX/FBRX QMC Calibration methodology (including phase rotation).

FIG. 5B illustrates FBRX QMC filter calibration. Referring also to FIG. 4 (Step B), FBRX QMC calibration routine is run to update QMC filter coefficients for the FBRX path. This configuration routine eliminates image from RX path induced by FBRX LO sourced from FBRX PLL.

FIG. 5C illustrates TX QMC filter configuration. Referring also to FIG. 4 (Step C), TX QMC configuration routine is run to update TX QMC filter coefficients for the TX path. This configuration routine eliminates image from TX path induced by the TX LO sourced from the TX PLL.

FIGS. 5D-5E transition from the dual TX/FBRX PLL configuration of FIG. 3, to the signal TX PLL configuration of FIG. 2 to complete QMC calibration/re-calibration (coefficient generation) in preparation for DPD adaptation.

FIG. 5D illustrates re-configuring the transceiver TX/FBRX to use the shared TX PLL for both TX and FBRX LO generation. QMC filter coefficients, calibrated under FIG. 5B with the FBRX PLL as the signal source in RX path, become invalid, and an RX image re-appears.

FBRX QMC re-calibration is performed to suppress the FBRX image. FBRX QMC re-calibration can be accomplished using, for example, using adaptive least squares.

FIG. 5E illustrates transceiver TX/FBRX signal path operation with QMC compensation (using updated FBRX/TX QMC filter coefficients) to suppress TX/FBRX images, and DPD adaptation/compensation. TX/FBRX signal path IQ modulation/demodulation uses a shared TX PLL as the signal source for the TX/FBRX Los driving the TX/FBRX mixers.

The Disclosure provided by this Description and the Figures sets forth example embodiments and applications illustrating aspects and features of the invention, and does not limit the scope of the invention, which is defined by the claims. Known circuits, functions and operations are not described in detail to avoid obscuring the principles and features of the invention. These example embodiments and applications can be used by ordinarily skilled artisans as a basis for modifications, substitutions and alternatives to construct other embodiments, including adaptations for other applications.

Claims

1. A wireless transceiver circuit, comprising

a transmit (TX) signal chain including an IQ modulator, and an RF upconverter including a TX mixer driven by a TX local oscillator (TX LO), a TX digital compensator, including a TX QMC (IQ mismatch) compensator configured to in QMC compensation mode, compensate for IQ mismatch in the IQ modulator based on TX QMC filter coefficients, in QMC calibration mode, calibrate the TX QMC filter coefficients based on a QMC calibration procedure;

a feedback receive (FBRX) signal chain including an IQ demodulator, and an RF downconverter including an RX mixer driven by a FBRX local oscillator (RX LO), an RX digital compensator, including an QMC (IQ mismatch) compensator configured to in QMC compensation mode, compensate for IQ mismatch in the IQ demodulator based on RX QMC filter coefficients, in QMC calibration mode, calibrate the FBRX QMC filter coefficients based on a QMC calibration procedure;

a TX PLL signal source connected to the TX LO, and selectively connectable to the RX mixer;

an FBRX PLL signal source selectively connectable to the RX LO;

a QMC controller configured to perform the QMC calibration procedure to generate calibrated TX and FBRX QMC filter coefficients, including disconnecting the TX PLL from, and connecting the FBRX PLL to, the RX LO; generating calibrated TX QMC filter coefficients; generating calibrated FBRX QMC filter coefficients; disconnecting the FBRX PLL from, and connecting the TX PLL to, the RX LO; generating re-calibrated FBRX QMC filter coefficients.

2. The wireless transceiver circuit of claim 1, wherein the TX digital compensator further includes a DPD (digital pre-distortion) compensator.