Patent No. 6587528 System and methods for extracting and digitizing phase and frequency information from an analog signal
Patent No. 6587528
System and methods for extracting and digitizing phase and frequency information from an analog signal (Wynn, Jul 1, 2003)
Abstract
Systems for periodically determining and digitally representing the phase and/or frequency of an input signal are disclosed which include a frequency discriminator, a phase discriminator and a control network. The frequency and phase discriminators may be used to periodically sample and estimate, respectively, the frequency and phase of an input analog signal, and these estimates are provided to a control network which calculates an error estimate and adjusts the estimate of the frequency provided by the frequency discriminator based on this estimate of the error. The systems of the present invention may further include means for digitally representing the phase and/or frequency of the input signal, and may also include a frequency-to-phase converter coupled to the output of the control network which converts the adjusted output of the frequency discriminator into a phase value. Associated methods for extracting and digitizing phase and frequency information from an analog input signal are also disclosed.
Notes:
FIELD
OF THE INVENTION
The present invention relates generally to extracting and converting into a
digital representation the information-bearing component of a signal and, more
particularly, to systems and methods for extracting and digitizing phase and/or
frequency information contained in an analog signal.
BACKGROUND OF THE INVENTION
In many electronic applications, and in particular, in radio communication systems,
the phase or frequency of an analog electrical signal carries information which
may be detected and used by an end user. To facilitate such detection and use,
it may be advantageous to sample the analog signal, extract phase or frequency
information, and represent that information in a digital representation. For
example, such extraction and conversion to a digital representation allows for
demodulation of phase or frequency modulated radio signals using small, low
cost, and low power digital signal processors. Use of such digital processing
techniques may even be more preferred if additional numerical processing of
the received signal is required such as, for example, de-interleaving or decoding
operations. Moreover, even if extensive downstream processing is not required,
digital processing, in many instances, may be more accurate and/or easier to
implement than corresponding analog processing methods. Accordingly, systems
and circuits for extracting and digitizing the frequency or phase of a signal
are utilized in a wide variety of electronic applications.
As mentioned above, systems for extracting and digitizing frequency and/or phase
information are particularly useful in a wide variety of radio communications
applications, as the information carried by the radio signal is often encoded
in the signal by modulating the phase or frequency of the signal. Thus, by extracting
and digitizing phase or frequency information, digital demodulation of a phase
or frequency modulated signal may be accomplished.
Typically, phase or frequency information is extracted from the received radio
frequency ("RF") signal after it has been downconverted at the receive terminal
to an intermediate ("IF") frequency (e.g., 450 kHz). However, it may also be
extracted directly from the RF carrier, or at some other frequency. Conventionally,
the frequency or phase of the received signal is represented relative to the
frequency or phase of the RF or IF carrier signal, although other reference
signals may be used.
A variety of systems and circuits for extracting and/or digitizing frequency
and phase information are known in the art. For instance, a conventional method
to extract and digitize the phase of an alternating current electrical signal
is to apply the signal, along with a known reference signal, to a phase comparator.
In response to these inputs, the phase comparator produces an output voltage
or current which is proportional to the phase difference between the input signal
and the reference signal. This analog representation of the phase of the input
signal may then be applied to an analog-to-digital ("A/D") converter, which,
at some specified rate, samples and outputs in digital format the values of
the phase of the input signal. Devices or systems that extract and digitize
the phase of an analog signal, such as the combination of a phase comparator
and an A/D converter, are commonly referred to as phase digitizers. Moreover,
as will be understood by those of skill in the art, phase and frequency have
a close mathematical relationship in that frequency is the time derivative of
phase. Accordingly, if frequency, as opposed to phase, information is desired,
the frequency information may be obtained from the phase samples via differentiation.
As the phase information is in a digital representation at the output of the
phase digitizer, a digital representation of the frequency information may conveniently
be obtained via numerical differentiation of the phase samples using modulo
2.pi. (circular) arithmetic subtraction.
While phase digitizers typically perform well in conventional phase modulated
digital communication systems which use widely spaced phase modulations (such
as BPSK or QPSK), their performance may be less than satisfactory in communication
systems which use finely spaced phase modulations such as 32-PSK or 64-PSK.
This problem may arise because the quantization of the phase from the received
IF signal is a function of the ratio of the clock rate of the phase digitizer
as compared to the IF frequency. By way of example, if 16-bit phase resolution
is desired (as opposed to the 5 or 6 bits of phase resolution typically available
in BPSK and QPSK demodulators), the digitizer clock would need to operate at
2.sup.16 times the IF frequency. Thus, to provide 16 bits of phase quantization
when operating at a typical IF frequency of 450 kHz, the digitizer clock would
need to run at nearly 29 GHz, which may be prohibitive both in terms of component
costs and the current drain on the terminal.
In addition to limiting the use of phase digitizers to widely digital phase
modulations, the above-described quantization noise problem may render the use
of phase digitizers unsuitable for use in many analog systems, such as frequency
modulated ("FM") analog radio systems. This occurs because the noise introduced
by a low quantization level of the analog signal causes the demodulated signal
to noise ratio to decrease noticeably. As such, phase digitizers which operate
at clock rates at least ten times the 19.2 MHz rate of conventional phase digitizer
clocks are likely required to meet the link hum and noise tolerances specified
in many analog FM applications. Accordingly, quantization noise concerns typically
limit the use of conventional phase digitizers with both high order digital
phase modulations (such as 32- or 64-PSK) and analog FM systems.
A second conventional method of extracting frequency (and/or phase) information
from an analog signal is to use an analog frequency discriminator to obtain
frequency samples from the signal. If digitization is required, the analog output
of the frequency discriminator may be digitized using an analog-to-digital ("A/D")
converter. Devices or systems that extract and digitize the frequency of an
analog signal, such as the combination of a frequency discriminator and an A/D
converter, are commonly referred to as frequency digitizers. Moreover, as phase
is the time integral of frequency, such a frequency digitizer may also be used
to extract and digitize phase information from an analog signal by re-integrating
the frequency samples.
While frequency discriminators and frequency digitizers may be acceptable for
various analog communication techniques such as conventional FM radio, in practice,
problems arise when frequency discriminators are used in digital communications
systems such as DQPSK. The principal difficulty arises in the integration step
which is necessary to convert the frequency samples into phase information,
as the accuracy of the resulting phase information is dependent on the accuracy
of the amplitude, or "level," of the output of the frequency discriminator.
This problem arises because the output signal level (i.e., the amplitude of
the frequency) of commercially available frequency discriminators may vary significantly
from unit-to-unit as a result of component variations, and the output may further
vary with temperature. As a result, analog frequency discriminators typically
are unable to extract phase information with sufficient accuracy to be of use
in many digital communication systems.
In light of the above-mentioned problems with existing frequency and phase discriminators
and digitizers, a need exists for phase/frequency discriminators and digitizers
that provide very accurate phase or frequency information that is generally
unaffected by quantization noise. Moreover, a need also exists for phase and
frequency discriminators and digitizers that work well with all modulation types.
SUMMARY OF THE INVENTION
In view of the above limitations associated with existing frequency discriminators
and phase digitizers, it is an object of the present invention to provide electronic
devices or systems which produce the digital value representation of the phase
angle and/or frequency of an AC signal which work well with both analog and
digital modulation techniques.
Another object of the present invention is to provide phase/frequency discriminators
and digitizers which accurately extract the phase and/or frequency of a received
signal.
It is still a further object of the present invention to provide phase/frequency
discriminators or digitizers which exhibit low quantization noise.
Additional objects, features and advantages of the present invention will become
apparent upon reading the following detailed description and appended claims
and upon reference to the accompanying drawings.
These and other objects of the present invention are provided by systems and
methods for extracting and optionally digitizing phase and frequency information
from an analog electrical signal which use a phase digitizer or discriminator
to sample and measure the amplitude of the analog signal, and then use this
information to correct frequency (or phase) estimates obtained from a conventional
frequency discriminator. In this manner, it is possible to reduce the error
in output frequency levels associated with currently available frequency discriminators,
while similarly avoiding the quantization noise effects which arise when conventional
phase digitizers or discriminators are used to extract phase or frequency information
from an analog signal. Thus, through the teachings of the present invention,
systems and methods for accurately extracting phase and frequency information
are provided which avoid both the accuracy and quantization noise problems associated
with prior art extraction techniques.
In one embodiment of the present invention, a system for periodically determining
the phase and/or frequency of an input signal is provided which includes a frequency
discriminator, a phase discriminator and a control network. In this embodiment,
the frequency and phase discriminators are used to periodically sample an analog
input signal and measure, respectively, the frequency and phase of that signal.
These frequency and phase samples, in turn, are provided to a control network,
which, responsive to these measurements, adjusts the estimates of the frequency
provided by the frequency discriminator. The systems of the present invention
may further include means for digitally representing the phase and/or frequency
of the input signal, and may also include a frequency-to-phase converter coupled
to the output of the control network for converting the adjusted output of the
frequency discriminator into a phase value. In a specific embodiment of the
present invention, the control network comprises an error estimator which produces
an error signal based on the outputs of the frequency discriminator and the
phase discriminator and adjusting means responsive to the error signal for adjusting
the amplitude of the output of the frequency discriminator. In this embodiment,
the system may further include a phase-to-frequency converter coupled to the
output of the phase discriminator and the error estimator may be used to periodically
calculate the difference between the outputs of the frequency discriminator
and the phase-to-frequency converter. This error estimator comprises means for
calculating the absolute value of the difference between the frequency estimates
provided by the phase and frequency discriminators. The error estimator may
further include a low pass filter or other averaging circuit or system for averaging
the calculated differences between the outputs of the frequency discriminator
and the phase discriminator.
In still another embodiment of the present invention, the frequency of an input
signal relative to a reference frequency is determined using a frequency discriminator
and a phase discriminator. The frequency is determined by refining the estimate
of the frequency of the input signal provided by the frequency discriminator
based on the estimate of the phase of the input signal provided by the phase
discriminator. In another embodiment, the phase and/or frequency of an input
signal relative to a reference signal are determined by (i) generating a digital
representation of a first estimate of the frequency of the input signal using
a frequency discriminator, (ii) generating a digital representation of a first
estimate of the phase of the input signal using the phase digitizer, (iii) converting
the digitized first estimate of the phase into a second estimate of the frequency
of the input signal, (iv) estimating an error in the first estimate of the frequency
based on the second estimate of the frequency, and (v) adjusting the first estimate
of the frequency based on the estimated error. Thus, it is possible to obtain
highly accurate estimates of the phase and frequency of an analog signal that
may be unaffected by quantization noise effects.
Comments