 |
Description  |
|
|
BACKGROUND OF THE INVENTION
This invention relates generally to radio receivers for frequency modulated
(FM) signals and, more particularly, to FM receivers incorporating direct
conversion to baseband.
DESCRIPTION OF THE PRIOR ART
Several approaches to FM receiver design are known in the art. The commonly
used superheterodyne receiver converts an incoming radio signal to one or
more intermediate frequencies at which amplification and frequency
selection can more readily be performed than at the frequency of the
received signal. Conversion to an intermediate frequency (IF) is
accomplished by mixing the received signal with a locally generated
oscillator (LO). After filtering and amplification at a first IF, the
signal may be immediately demodulated in a detector circuit, which
constitutes a "single conversion" receiver. Alternatively, dual and triple
conversion designs are known having successive conversion to additional
IF's followed by demodulation of the signal.
An undesirable feature of conventional superheterodyne receivers has been
the difficulty in applying microelectronic techniques to achieve
miniaturization. To fabricate amplifier circuits in monolithic form has
long been known, but it has been difficult to integrate the high-Q crystal
or ceramic bandpass filters generally used for selectivity at the
intermediate frequencies. A solution known in the art is to convert an
incoming signal directly to baseband in what is known as a "direct
conversion" or "zero-IF" receiver. The local oscillator frequency is made
equal to the carrier frequency of the received signal, and the spectrum
occupied by the modulation is translated directly to baseband. The
necessary sharp selectivity is then achieved through lowpass, rather than
bandpass, filtering. Low frequency lowpass filters are readily fabricated
in monolithic form, allowing a much greater degree of miniaturization.
Direct conversion receivers for FM signals may include a technique used for
single-sideband (SSB) reception known as the "third-" or "Weaver-method,"
disclosed in Proc. IRE 44 (1956), pages 1703-1705, and the subject of U.S.
Pat. No. 2,928,055. As applied to FM detection, the purpose of the
technique is to distinguish the modulation information carried by positive
and negative frequency excursions about the carrier. When an FM signal is
mixed with a down-conversion oscillator to translate it to baseband, equal
positive and negative frequency excursions about the carrier result in the
same difference frequency, and the polarity of the modulation can no
longer be determined without some phase reference. The Weaver method
provides two substantially identical paths in which the signal is
down-converted to baseband, lowpass filtered to remove the sum products of
mixing as well as undesired adjacent channel signals, and up-converted to
an output frequency. The down-and up-conversion oscillators for one path
are in phase quadrature with their counterparts in the other path. The
process of two successive frequency conversions produces phase inversions
between the sidebands of the signals in the two paths. When the outputs of
the two paths are added, cancellation of the sidebands occurs in such a
manner that the modulation polarity of the original input signal is
retained, though translated to a new, predetermined output frequency. In
effect, the received signal is translated from an incoming frequency to
baseband, filtered to remove interfering adjacent channel signals, and
retranslated to an output frequency at which conventional FM demodulation
can take place. Such a circuit following the Weaver method may be termed a
"translating bandpass filter" (TBPF).
Prior art FM receivers following the direct conversion approach suffer from
a number of shortcomings. For example, in a direct conversion receiver,
the LO frequency equals the received signal carrier frequency, and if no
RF preamplifier is used, there is very little reverse isolation to prevent
LO energy from reaching the antenna and causing interference with other
receivers. Furthermore, noise and DC offsets make it difficult at baseband
to achieve the low noise amplification and high gain required for adequate
sensitivity.
In a superheterodyne FM receiver, amplitude limiting is generally used to
reduce noise and to improve signal capture. However, to maintain fidelity
of the modulation in the baseband zero-IF application, both instantaneous
amplitude and phase must be preserved. Limiting is not practical. Strong
adjacent channel signals that overload the IF can cause distortion if the
circuits are not designed for wide dynamic range.
Finally, if there is an offset in frequency between the incoming signal and
the down-conversion oscillator in the zero-IF, several undesired results
may occur. If the baseband paths are imperfectly matched, the cancellation
of mixing products will be incomplete, and a beat-note will oocur. The
beat-note has a primary component at twice the offset frequency but can
have distortion products at other harmonics of the offset frequency. If
the beat note is in the audible range, it can interfere with demodulated
audio output. Furthermore, the lowpass filter bandwidth necessary to pass
an FM modulated signal increases by the amount of the offset, and it
becomes difficult to obtain narrow selectivity with other than negligible
offsets. It is known in the design of FM receivers for broadcast
reception, for example, that a low-level beat-note slightly above the
audible band (approx. 20 kHz), caused by a 10 kHz offset, does not
interfere with reception. That is because the baseband lowpass filters
already have cutoff frequencies on the order of 100 kHz to pass a
broadcast FM signal, and they can accommodate a 10 kHz offset without
significantly distorting the modulation. However, in other applications,
say, for example, the land-mobile service in which channel spacing may be
as close as 12.5 kHz, the baseband filtering must be narrower than
one-half the spacing, or about 6 kHz, in order to separate the adjacent
channel signal from the desired. A frequency offset higher than a few tens
of Hertz would require that the baseband filters be widened to accommodate
the modulation swing plus offset, and this would degrade selectivity.
Frequency offset must therefore be tightly controlled
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a direct conversion FM
receiver that nevertheless overcomes the foregoing deficiencies.
It is a more particular object to provide a receiver having direct
conversion to baseband to enable the selectivity function to be performed
with low frequency lowpass filtering and to thereby facilitate
miniaturization through microelectronic techniques.
It is a further object of the invention to provide a zero-IF system with
sharp selectivity to be used to receive FM signals transmitted in a narrow
channel spacing environment.
It is also an object to avoid the detrimental effects of mismatch in a
zero-IF system and to avoid a beat note by providing for precise frequency
control of the frequency conversion sources in the IF.
A further object is to include in a direct conversion FM receiver the
advantageous placement of the amplification and limiting functions at a
non-zero intermediate frequency as in a superheterodyne receiver.
The invention lies in the use of a baseband zero-IF for providing the
ultimate selectivity, followed by up-conversion to an intermediate
frequency at which amplification and limiting take place. Frequency
control by phase lock technique is used to precisely center the zero-IF
signal. The phase lock loop also demodulates the received signal.
BRIEF DESCRIPTION OF THE DRAWINGS
The features of the present invention that are believed to be novel are set
forth with particularity in the appended claims. The invention itself,
however, together with further objects and the advantages thereof, may be
best understood by reference to the following description when taken in
conjunction with the accompanying drawings, in which:
FIG. 1 is a block diagram of an FM receiver which has been constructed in
accordance with the present invention;
FIG. 2 is a more detailed schematic and block diagram of the translating
bandpass filter incorporated in the receiver of FIG. 1;
FIG. 3 is an alternate embodiment of the present invention which includes
additional front end structure; and
FIG. 4 discloses an additional realization of frequency control arrangement
as an alternative to that illustrated in FIG. 1.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
Referring now to the drawings, an FM receiver 10 is shown in FIG. 1, which
receiver has been constructed in accordance with the present invention. An
incoming signal 12 is amplified in preamplifier 14, a major function of
which is to provide low noise amplification that establishes the
sensitivity of the receiver. The preamplifier 14 additionally provides
reverse isolation that reduces the escape of radiation from voltage
controlled oscillator (VCO) 40. The output of preamplifier 14 feeds a
translating bandpass filter 30 (TBPF). Filter 30 may be of a type known in
the art for accomplishing direct conversion that is frequently referred to
as a "Weaver" circuit. A detailed description of this prior art circuit
may be found in Weaver, D. K. Jr. "A Third Method of Generation and
Detection of Single-Sideband Signals," Proc. IRE, vol. 44, pp. 1703-1705,
December 1956. This circuit translates a signal from a known input
frequency to a predetermined output frequency while simultaneously
effecting appropriate bandpass filtering. A more detailed description of
filter 30 is set forth in FIG. 2 and will be described hereinafter.
The input signal 31 from preamplifier 14 is applied to each of two
parallel, substantially identical, paths 32a-33a-34a and 32b-33b-34b.
Elements 32a and 32b are like down-conversion mixers that translate the
incoming signal to essentially baseband. A down-conversion frequency is
supplied at line 35 in quadrature to both mixers 32a and 32b using phase
shifter 36 or the equivalent to provide two signals in phase quadrature.
The outputs of mixers 32a and 32b are fed to two identical lowpass
filters, 33a and 33b, having a cutoff frequency on the order of one-half
the bandwidth of the desired passband. The design of these lowpass filters
affects the frequency response of the demodulated signal and will be
treated below in discussion of the phase lock loop components. The
respective outputs of these lowpass filters are coupled to up-conversion
mixers 34a and 34b, which are likewise supplied in phase quadrature with
frequency f.sub.up, at 37, through phase-shifter 38 or similar apparatus
for providing quadrature signals. The up-converted signals from mixers 34a
and 34b are summed in element 39 to provide a bandpass filtered output at
a predetermined frequency. The circuit elements of the TBPF require good
matching, and it is advantageous to fabricate them in monolithic form.
Referring again to FIG. 1, it may be seen that the incoming signal is
translated from frequency f.sub.down to frequency f.sub.up while being
filtered to a desired selection bandwidth. The output 41 of TBPF 30 is
then applied in sequence to filter 60, amplifier 70, and then to limiter
80. The purpose of filter 60 is to attenuate harmonics of f.sub.up and
their sidebands, which may be generated in mixers 34a and 34b in the TBPF.
The harmonics must be suppressed so that limiter 80 does not respond to
them. It should be noted that filter 60 in fact is unnecessary with
certain types of up-conversion mixers. Amplifier 70 and limiter 80 provide
the bulk of the receiver gain. Gain in the zero-IF path is intentionally
avoided so as to prevent overload, which would result in distortion of the
demodulated signal. Limiter 80 functions both to reject amplitude noise
variations and to maintain a constant amplitude signal for phase detector
90. This keeps the phase detector gain constant, which is necessary for
control of the overall loop gain and closed loop frequency response.
Up-conversion oscillator 50, phase detector 90, loop filter 100, and
voltage controlled oscillator (VCO) 40 form a phase lock loop (PLL) that
controls the zero-IF down-conversion frequency. The use of a phase lock
loop for frequency control and the components constituting the loop are
well known in the prior art. For example, see Gardner, F. M., Phaselock
Techniques, New York: Wiley, 2d ed. 1979, or Blanchard, A., Phase Locked
Loops, New York: Wiley, 1976.
The output of limiter 80 is applied to phase detector 90, for which it is
advantageous to have a wide range phase capability (.+-.2Pi radians). The
other input to the phase detector is the in-phase component of f.sub.up,
37. The phase detector output is applied through loop filter 100 to the
control line of oscillator 40, which may be a voltage controlled
oscillator (VCO) or voltage controlled crystal oscillator (VCXO). Loop
filter 100 is a lowpass filter used to prevent spurious harmonics of the
phase detector reference frequency from modulating the controlled
oscillator. If a low feedthrough sample-and-hold phase detector is used,
filter 100 may be unnecessary. The signal at the oscillator control line,
110, is the demodulated audio of the incoming signal
When the loop locks, both inputs to the phase detector are at the same
frequency, f.sub.up. This happens when the frequency of VCO 40 exactly
equals the carrier frequency of signal 31 applied to the TBPF 30. Then,
the signals in the quadrature baseband paths are precisely centered about
zero frequency, and the up-converted output signal 41 equals f.sub.up.
Under these conditions, no beat-note will exist in the baseband paths.
Several design details of the components comprising the translating
bandpass filter and the phase lock loop are noteworthy. The phase lock
loop not only controls the down-conversion frequency, but it also recovers
the modulation of the input signal. The amplitude versus modulation
frequency response of the demodulated output is determined by the closed
loop response of the oscillator control signal versus deviation frequency.
The closed loop frequency response depends on the open loop pole locations
and the overall loop gain. To obtain a desired closed loop response, one
design technique is to work back to find the appropriate open loop filter
response and open loop gain to give the closed loop response.
The dynamic elements of the PLL include lowpass filters 33A and 33B,
bandpass filter 60, and loop filter 100. With so many poles of filtering
in the loop, it is difficult to maintain stability unless the loop gain
falls below unity at a frequency below which 180.degree. of phase shift is
reached. It is generally found that the loop gain necessary to satisfy the
dual requirements of stability and frequency response is low compared with
typical loop gains in phase lock loop demodulators having equivalent
bandwidth. This occurs because of the greater number of dynamic filtering
elements in the system here. If the phase detector is designed for low
spurious output, filtering of its output by loop filter 100 may be
avoided, reducing the overall number of poles in the loop and simplifying
the design.
Further considerations in the choice of loop components relate to adjacent
channel selectivity. Adjacent channel selectivity depends primarily on the
open-loop attenuation characteristics of the loop. For best selectivity,
the cutoff frequency of the open loop response should be as low as
possible. The loop is designed to track the deviation of the input signal
up to a predetermined modulation cutoff frequency. For the loop to
properly track the input signal, the phase error between the two signal
inputs to the phase detector must not exceed the phase difference
capability of the detector. The maximum phase error depends on the loop
gain, the peak deviation and maximum modulation frequency of the input
signal, and the cutoff frequency of the open loop response. Because the
loop gain is otherwise made low for stability reasons, and the loop cutoff
frequency is low for selectivity reasons, the phase error may be large. It
is therefore advantageous to design the phase detector for a wide range
phase difference capability.
Because the signals in the translating bandpass filter 30 are centered
about zero frequency, DC offsets and carrier feedthrough can affect
receiver performance. The chief problem of carrier feedthrough at f.sub.up
is that it would tend to capture the limiter under weak- or no-signal
conditions, causing self-quieting and degrading sensitivity. Feedthrough
may be caused by poor carrier suppression in up-conversion mixers 34A and
34B and by DC offsets in the baseband circuits. DC offsets may be
minimized by the use of differential circuits in the zero-IF.
The embodiment illustrated in FIG. 1 accomplishes the above mentioned
objects of the invention-while providing a particular realization of
direct conversion FM receiver. The translating bandpass filter 30 provides
a zero-IF system that is adaptable to microelectronic techniques. Through
precise frequency control as provided by the phase lock loop, the zero-IF
signals are centered at baseband, and no beat note will arise, even with
imperfect matching of the quadrature paths. Narrow bandpass filtering is
achieved through the use of low frequency lowpass filter elements, which
are readily fabricated in monolithic form. Signal amplification is
primarily achieved in a non-zero IF, which may be operated in limiting as
is known in the FM receiver art to be advantageous for good noise and
capture performance. Furthermore, the problem of achieving low noise gain
at baseband is avoided through the use of a low noise radio frequency
preamplifier that dominates the noise performance of the entire system. It
is the cooperation of circuit elements, in particular precise frequency
control along with up-conversion to a non-zero IF, that makes possible
achieving these objects
FIG. 3 shows an embodiment of the invention that illustrates how the
teachings of the invention may be used with a first, non-zero IF stage
instead of an RF preamplifier. This arrangement can offer improved
performance in an environment of strong signals that may cause
intermodulation distortion or in which the frequency band of signals to be
received is such that low noise amplification is difficult or uneconomical
to achieve.
Whereas FIG. 1 illustrates a receiver 10 in which the incoming signal 12 is
coupled through preamplifier 14 to the TBPF 30, FIG. 3 depicts receiver
10' used in conjunction with a first IF stage 14' for coupling the
incoming signal 12 to the TBPF 30. Stage 14' includes a preselector 201, a
first local oscillator 202 and its injection filter 203, mixer 204, first
IF filter 205, and IF amplifier 206. The preselector 201 is a bandpass
filter that blocks all but a selected band of radio channels from
appearing at mixer 204, protecting the mixer from overload from undesired
signals and preventing signals at the so-called "image" frequency from
being converted to IF. The oscillator 202 provides the frequency by which
incoming signals will be translated in mixer 204. Injection filter 203
prevents extraneous noise energy from oscillator 202 from degrading the
mixer noise performance. The output of mixer 204, which is at a
predetermined intermediate frequency, is filtered in first IF filter 205
to remove the undesired products of the frequency conversion process and
to couple the desired signal to first IF amplifier 206. The output of
amplifier 206 is coupled to receiver 20 as described earlier in the
description of FIG. 1.
In addition to providing traditional functions of an IF stage, the elements
of block 14' also protect the TBPF circuits from overload. Signal levels
are chosen so that in the presence of excessively strong on-channel
signals the circuits in stage 14' will saturate and thereby provide a
predetermined maximum output amplitude below the overload point of the
zero-IF circuits. Protection from limiting on strong off-channel signals
is provided by IF filter 205.
FIG. 4 illustrates an embodiment 10", constructed in accordance with the
invention, which includes alternative means for controlling the
down-conversion oscillator frequency. Instead of being converted to
exactly zero frequency, the translated input signal is provided with a
small offset from zero. The purpose of the frequency offset is to avoid
the problems of DC offsets yet retain the precise frequency control and
inherent demodulation capabilities of the phase lock loop, as in the
receiver of FIG. 1. The two input signals to phase detector 90 are now the
output signal from limiter 80 and a reference frequency 52 from oscillator
51. Reference 52 differs from up-conversion frequency 37 by a frequency
lower than the modulation frequencies to be received. The phase lock loop
locks with the output of limiter 80 equal to the frequency of reference
oscillator 51. For this to occur, the translated carrier frequency of the
signals in the baseband paths of translating filter 30 must equal the
difference between frequencies 37 and 52. This occurs when VCO 40 differs
from input 31 by the same offset amount. This offset frequency may be set
on the order of 10 to 100 Hz. It should be noted that there will be an
image response, that is, that input 31 may be offset above or below the
VCO. However, the separation between images will be much less than the
channel spacings, and no interfering signals should be present.
Because an unmodulated carrier at 31 produces no DC signals in the baseband
paths, DC coupling and the need to maintain low offsets may be avoided.
Any beat note that may arise because of imperfect matching will be below
the lowest modulation frequency and may be blocked from the demodulated
output signal by high-pass coupling techniques.
In all other respects, the receiver of FIG. 4 is identical with that of
FIG. 1 and is designed according to the same principles.
Although the present invention has been disclosed in connection with the
embodiments herein, it is understood that the novelty lies in the
particular combination of translating bandpass filter, up-conversion to an
intermediate frequency at which amplification and limiting are performed,
and precise automatic control of the translation frequencies.
Modifications and additional applications of the invention apparent to
those skilled in the art are included within the scope of the invention.
* * * * *
|
|
 |
|
 |
Description |
|
