Frequency response correction for a receiver having a frequency translation device

Abstract

A method of correcting the frequency response of an RF receiver having a frequency translation device and a fixed IF frequency section uses calibration data representing frequency responses over a wide frequency range for multiple center frequencies. The calibration data are based upon a multi-dimensional model for the frequency translation device that includes at least center frequency and offset frequency. The calibration data are stored in the RF receiver during factory calibration. A run-time normalization of the RF receiver produces a normalization frequency response at a reference frequency, and a relative frequency response at a desired center frequency is derived from the calibration data. The normalized and relative frequency responses are combined to produce an overall frequency response that is used by the RF receiver to configure an inverse filter to correct channel linear distortion produced by the frequency response of the frequency translation device at the desired center frequency.

Description

BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWING

FIG. 1 is a graphic view of a wide IF bandwidth communication system having multiple channel center frequencies over a wide frequency range.

FIG. 2 is a generic block diagram for a typical RF receiver that includes frequency response correction according to the present invention.

FIG. 3 is a generic block diagram view of a typical RF receiver having multiple signal paths that includes frequency response correction according to the present invention.

DETAILED DESCRIPTION OF THE INVENTION

For the following discussion “calibration” refers to the factory calibration of an RF receiver using external test equipment, and “normalization” or “alignment” refers to RF receiver self-calibration using an internal reference test signal.

Referring now to FIG. 1 a typical RF system has multiple channels over a wide frequency range, each channel having a unique center frequency, ω_c, and a specified channel bandwidth, ω_b. The RF system covers a frequency range from a start center frequency, ω₁, to a stop center frequency, ω₂. A channel frequency response for the RF receiver is modeled as a two-dimensional function of center frequency and offset frequency as follows:

H(ω, ω_c)=H_R(ω+ω_c)H_M(ω, ω_c)H_IF(ω)

where H is the overall channel frequency response, H_R is the RF frequency response preceding a mixer, H_M is the RF mixer frequency response, H_IF is the combined IF frequency response following the mixer and ω is the frequency offset from the channel center frequency, ω_c. The frequency response may be expressed as a complex function that contains both magnitude and phase information. As a result the magnitude of the frequency response may be expressed as an absolute value of the complex frequency response. By recognizing that the mixer and its interacting circuits are essentially not separable, the combined frequency response is characterized during factory calibration at each of a plurality of center frequencies and may be used as a correction factor. Another factor that also enters into linear distortion in a signal path having a frequency translation device is the local oscillator (LO) drive level, which affects conversion loss, harmonic inter-modulation, reflection across ports and leakage. Therefore the calibration data may also include this “third” dimension to the multi-dimensional frequency translation device model.

Referring now to FIG. 2 the RF receiver has a low-pass image rejection filter 24 followed by an RF mixer 26 and multiple IF conversion stages generally represented as a single block or section 28, since the IF stages normally operate at fixed IF frequencies and the linear distortion introduced in the IF stages may be readily measured by the normalization procedure. The signal from the IF stages 28 is digitized by an analog-to-digital converter (ADC) 30. Digital signal processing may be implemented in digital hardware, such as a field programmable gate array (FPGA), and in software within a data processor 32. A normalization source 34, which is generally an internal signal generation module, generates a normalization stimulus signal with a specified magnitude and phase response during run-time.

The calibration data obtained during the calibration process at the factory is stored in the data processor 32 The calibration data may be obtained as described in co-pending U.S. patent application Ser. No. [Attorney docket DF8199] entitled CHARACTERIZATION OF A FREQUENCY RESPONSE FOR A FREQUENCY TRANSLATION DEVICE. Alternatively any characterization data that represents the mixer as a multi-dimensional function may be used. In the following equation the frequency response is a complex function of two variables (expressed in linear scale, not dB). The complex frequency response at the calibration frequency, ω_c, is:

H _C =H(ω, ω_c)/H(ω, ω_r)={H_R(ω+ω_c)H_M(ω, ω_c)}/{H_R(ω+ω_r)H_M(ω, ω_r)}

After performing a run-time normalization, i.e., measuring the results from the normalization source 34 at the reference frequency, T_r, the channel response at the normalization frequency is:

H _n(ω, ω_r)=H_R(ω+ω_r)H_M(ω, ω_r)H_IF(ω)

To determine the frequency response at any desired frequency, ω_d, the calibration data is interpolated to obtain a relative frequency response at ω_d

H _c(ω, ω_d, ω_r)=H(ω, ω_d)/H(ω, ω_r)={H_R(ω+ω_d)H_M(ω, ω_d)}/{H_R(ω+ω_r)H_M(ω, ω_r)}

By multiplying the relative frequency response from the calibration data with the frequency response at the normalization frequency, an overall frequency response at the desired frequency is obtained:

H(ω, ωd)=H_R(ω+ω_d)H_M(ω, ω_d)H_IF(ω)=H_n(ω, ω_r)H_c(ω, ω_d, ω_r)

The data processor 32 then configures an inverse digital filter to compensate for the linear distortion with the following frequency response over the frequency range of the channel

H ⁻¹(ω, ω_d)=1/H(ω, ω_d)

In some RF receivers, such as those in spectrum analyzers, multiple signal paths are used to cover a much wider frequency range. As shown in FIG. 3 a low-band path has a lowpass filter 24_L and a corresponding mixer 26_L, while a high-band path has a bandpass filter 36 and a corresponding mixer 26_H. The high-band path covers a higher frequency range than the low-band path. The above-described alignment technique may still be applied to both or either of the high-band and low-band path signals as required to meet system requirements.

Thus the present invention provides a method of correcting the frequency response of an RF receiver having a fixed IF frequency section by using calibration data in the form of a complex frequency response based upon a multi-dimensional model of a frequency translation device to modify a run-time normalization response at a reference frequency to calculate the frequency response at a desired center frequency, from which an inverse digital filter is built to process the signal applied to the RF receiver for greater measurement accuracy by reducing channel linear distortion.

Claims

1. A method of correcting an RF system having a frequency translation device and a fixed-frequency IF section comprising the steps of: storing calibration data in the device that represents frequency responses as a complex function for the frequency translation device over a plurality of center frequencies, the calibration data representing a multi-dimensional model of the frequency translation device that includes at least center frequency and offset frequency;performing a run-time normalization of the device using a stimulus signal at a reference frequency to produce a normalization frequency response; andcombining a relative frequency response derived from the calibration data at a desired center frequency with the normalization frequency response to obtain an overall frequency response at the desired center frequency; andconstructing an inverse filter for the device from the overall frequency to compensate for linear distortion over a specified IF bandwidth for the desired center frequency.

2. The method as recited in claim 1 further comprising the step of repeating the combining and constructing steps for each change in the desired center frequency.

3. The method as recited in claim 1 further comprising the step of interpolating the calibration data to obtain the relative frequency response at the desired center frequency.

4. The method as recited in claim 3 further comprising the step of repeating the combining and constructing steps for each change in the desired center frequency.

5. An RF receiver comprising: an input section having a first frequency response characteristic for receiving an input signal;a frequency translation device having a second frequency response characteristic coupled to an output of the input section for converting power in the input signal at a first frequency to power at a second frequency;an IF section having a third frequency response coupled to the frequency translation device to receive the power at the second frequency and produce digitized data;a normalization signal source for selectively providing a test signal at a reference frequency to the input of the input section; anda data processor coupled to the IF section for receiving the digitized data, the data processor having calibration data representing the combination of the first, second and third frequency response characteristics as a complex function for a plurality of center frequencies, the calibration data being based upon a multi-dimensional model for the frequency translation device that includes at least center frequency and offset frequency for each of a plurality of center frequencies and the data processor (i) calculating a normalization frequency response for the digitized data produced by the test signal, (ii) determining a relative frequency response at a desired center frequency for the input signal from the calibration data, (iii) combining the normalization and relative frequency responses to produce an overall frequency response, and (iv) configuring an inverse filter from the overall frequency response for correcting the digitized data produced by the input signal.

6. The RF receiver as recited in claim 5 further comprising a plurality of signal paths, each signal path having an input section, a frequency translation device and an IF section, the input sections being selectively coupled to receive one of the input signal and the test signal and the IF sections providing digitized data selectively to the data processor according to the one of the plurality of signal paths selected so that the data processor configures the inverse filter according to the one of the plurality of signal paths selected.

7. The RF receiver as recited in claim 5 wherein the data processor further interpolates the calibration data in determining the relative frequency response at the desired center frequency.

8. The RF receiver as recited in claim 5 wherein the inverse filter is configured with a new overall frequency response by the data processor by repeating the determining and combining calculations when the desired center frequency is changed.