Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02254375 1998-11-13
INTERMODULATION CORRECTION FOR
COMBINED ANALOG- AND DIGITAL-FORMAT
RF TRANSMITTERS
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to intermodulation correction for digital-format
radio
frequency (RF) transmitters. More specifically, the invention relates to
intermodulation
correction especially suitable for digital television (DTV) transmitters.
2. Related Art
It is recognized that, when linearly-amplified signals have more than one
carrier
signal, intermodulationproducts are produced at frequencies that are the sum
and difference
of the frequencies of the original Garner signals. FIG. 1 illustrates this
phenomenon in the
context of an analog-formattelevision signal (for example, NTSC at
intermediate frequency
(IF)). A channel of analog-format TV at IF occupies a 41-47 MHz band. Included
in the
analog-format signal are three carriers:
the aural carrier (herein designated fA 41.25 MHz),
i s the color subcarrier (herein designated f~=42.171 MHz), and
-1-
CA 02254375 1998-11-13
the visual carrier (herein designated f~ 45.75 MHz).
Second-order intermodulation components arise from the mixing of any two of
these three
original or "parent" components. For example, when the aural and visual
carriers are mixed,
second-order intermodulation components arise at f~ ~ fA MHz. Specifically:
f~ - fA = 4.5 MHz (I)
f~ + fA = 87.0 MHz (II)
as shown in FIG. 1. Similarly, when the color subcarrier and aural carrier are
mixed,
second-order intermodulation components arise at f~ ~ fA MHz. Specifically:
f~ - fA = 0.921 MHz (III)
to f~ + fA = 83.421 MHz (IV)
For clarity of illustration, the remaining (f~ t f~) second-order
intermodulation components
are not specifically shown in FIG. 1.
The second-order intermodulation components are outside of the 41-47 MHz band
occupied by the analog TV signal, and thus are easily filtered. However, the
same cannot
be said of third-order intermodulationproducts. Third-order intermodulation
products arise
when second-order intermodulation products are mixed with the other "parent"
signal. For
example, when the second-order intermodulation product from the color and
aural carriers
(at frequency f~ - fA = 0.921 MHz) is mixed with the remaining "parent" signal
(the visual
carrier at frequency f~ = 45.75 MHz), the following third-order
intermodulationproducts are
2o produced:
-2-
CA 02254375 1998-11-13
f~ - (f~ - fA) = 44.829 MHz (V)
f~ + (f~ - fA) = 46.671 MHz (VI)
Undesirably, these third-order intermodulation products lie within the 41-47
MHz band
occupied by the analog TV signal, and cannot be filtered. These
intermodulation
components must be removed or compensated for.
FIG. 2 illustrates a known arrangement for cancellation of the 44.829 MHz
third-
order intermodulationcomponent. In FIG. 2, the visual carrier at f~ = 45.75
MHz at a high
level appropriate for the mixer of choice (for example, +10 dBm) and the aural
carrier at
fA = 41.25 and low level are input to a mixer 21. Mixer 21 provides its output
signal to a
1 o band pass filter 22, which passes only the frequency-sum signal at f~ + fA
= 87.0 MHz.
The frequency-sum signal output by the band pass filter is level-adjusted by
level
adjuster 23 and amplified by amplifier 24 to a proper low level (for example, -
10 dBm) for
input to a mixer 25. As recognized by those skilled in the art, the choice of
high (or low)
level inputs to the mixers determines whether the mixers are (or are not)
forced into
switching mode.
In addition to the 87 MHz signal, mixer 25 receives the color carrier at 42.17
MHz
and at a high level appropriate for the mixer of choice (for example, +10
dBm). Mixer 25
provides an intermodulationcancellation signal having a component at (f~ + f,~
- fc = 44.829
MHz. The 44.829 MHz frequency of the intermodulation cancellation signal
output by
-3-
CA 02254375 1998-11-13
mixer 25 corresponds to the frequency of the third-order intermodulation
component formed
in the main IF signal according to Equation V, above.
The intermodulation cancellation signal from mixer 25 differs in timing
(delay),
in phase, and in amplitude from the intermodulation component present in the
main IF
signal. To compensate for delay in the corrector, a delay element 26 is
provided at the input
for the main IF signal. The delay element drives the first input of a summing
circuit 29. To
compensate for the differences in phase and amplitude, a phase adjuster 27 and
an amplitude
adjuster 28 are provided. These adjustment elements adjust the corresponding
attributes of
the intermodulationcancellation signal from mixer 25 to correspond to those of
the delayed
to intermodulation signal in the uncorrected main IF signal. This adjustment
involves
adjustment to the same amplitude but opposite in phase (180°).
To accomplish the correction for the 44.829 MHz intermodulation component,
summing circuit 29 adds the adjusted intermodulation cancellation signal
(provided by
adjustment elements 27 and 28) to the uncorrected main IF signal (properly
delayed by delay
element 26) . As a result, summing circuit 29 outputs a sum signal whose
44.829 MHz
intermodulation component has been cancelled.
To cancel the 46.671 intermodulationcomponentproducedaccordingto Equation VI,
above, the positions of the color and aural carrier are reversed in the FIG. 2
circuit.
Operation of the circuit is the same in principle as that described above.
-4-
CA 02254375 1998-11-13
The circuit of FIG. 2 has been useful in cancellation of intermodulation
components
in NTSC television signals because of the presence of discrete carrier
frequencies in the
original signal (the aural Garner at fA 41.25 MHz, the color subcarner at
f~=42.171 MHz,
and the visual carrier at f~=45.75 MHz).
Significantly,however, digital television (DTV) signals cannot be
characterized by
a small number of discrete-frequency component signals. Rather, DTV signals
are
essentially noise-like signals for which there is no identifiable set of
"parent" signals from
which to derive signals that can cancel intermodulationproduct signals that
are present in the
DTV signals. In DTV signals, there are no Garner signals or other definable
signal that can
l0 be fed into a mixer to generate a cancellation signal. Accordingly, when
attempting to solve
intermodulationproblems in DTV signals, one must adopt an approach that is
distinct from
that used in the NTSC intermodulation circuit of FIG. 2.
In addition to the technical desirability of removing or compensating for in-
band
intermodulationcomponents,the United States Federal
CommunicationsCommission(FCC)
has mandated out-of band signal suppression for digital television that is
stricter than for
analog (NTSC) television format signals. This mandate helps to ensure that
television
signals from adjacent channels do not interfere with each other.
The FCC Sixth Report and Order (April 1997), as modified by the Memorandum
Opinion and Order on Reconsiderationof the Sixth Report and Order (M.O.&O., 47
C.F.R.
~ 73.622(h)) (February 23,1998), has mandated an emissionmask accordingto the
following
-5-
CA 02254375 1998-11-13
suppression requirements. For the 0.5 MHz immediately surrounding the channel,
the
suppression S (relative to the average in-band transmitted power) must be:
S z47 dB (of < 0.5 MHz)
For frequencies distant from the channel by more than 0.5 MHz, the emission
mask requires
suppression according to the formula:
Sz 11.5(of+3.6)dB (0.5<of<6MHz)
where:
S is the suppression requirement (in dB) relative to the average in-band
power; and
of is the distance in frequency (in MHz) from the channel edges.
1o Of course, just inside the channel band there should be zero suppression
for optimum
transmission of the channel of interest.
Out-of channel degradation in DTV signals occurs due to high-power
amplification
of the DTV signal in the DTV transmitter. A typical Class AB, high-power,
linear amplifies
in a DTV transmitter exhibits a regeneration of out-of channel signal
components to about
37 dB below the in-channel signal power.
A band pass filter may be designed to provide about 10 dB of additional
suppression
immediately outside the band's edges without excessive (uncorrectable) group
delay. Thus,
if a single-channel (6 MHz) band pass filter were used around an isolated DTV
channel, the
total expected out-of channel suppression to be expected from the transmitter
with the band
2o pass filter would be 37+10=47 dB, meeting the FCC minimum requirement.
-6-
CA 02254375 1998-11-13
However, the United States Federal Communications Commission (FCC) has
allocated television channels so that many broadcasters have their DTV
channels assigned
immediately adjacent their analog (NTSC) channels in the frequency spectrum.
In this
adjacent-channel scenario, the out-of channel DTV intermodulationproducts that
lie outside
the DTV channel, but inside the adjacent NTSC channel, must somehow be
corrected.
In a transmitter in which both DTV and analog (e.g., NTSC) television signals
are
amplified by a single wideband (12 MHz) amplifier, a single-channel (6 MHz)
band pass
filter cannot be used on the DTV channel alone, because it would adversely
affect the signal
in the immediately-adjacent NTSC channel. Further, a dual-channel (12 MHz)
band pass
1 o filter cannot provide the desired suppression of DTV
intermodulationproducts that lie in the
NTSC channel. Yet, it is desired to provide the additional 10 dB of
suppression that a band
pass filter would otherwise enable, to ensure that the shoulder of the DTV
signal is
suppressed to below the threshold of visibility (TOV) of the analog signal
viewer.
It is to fulfill this need for additional suppression of out-of channel DTV
intermodulation products that the present invention has been developed.
SUMMARY OF THE INVENTION
The invention provides an arrangement especially suitable for use in
correcting out-
2o of channel intermodulation distortion in a combined analog and digital
television (DTV)
CA 02254375 1998-11-13
transmitter. The invention envisions that such intermodulation correction of a
DTV signal
at intermediate frequency (IF) provides the desired additional out-of channel
suppression in
a way that better protects an adjacent analog (for example, NTSC) channel
signal. The
invention is particularly useful in applications in which adjacent-channel DTV
and analog
television signals are amplified by a single, high-power amplifier, a scenario
in which a
single-channel band-pass filter cannot be used. The inventive intermodulation
corrector
suppresses the out-of DTV-channel DTV intermodulation products to below the
threshold
of visibility (TOV) of the analog channel viewer.
The arrangement has a first mixer for receiving the input DTV signal at both a
first
1 o input and a second input, and for providing a first mixer output signal to
a second mixer.
The second mixer also receives the input DTV signal, and provides a broadband
intermodulation cancellation signal. Finally, the input DTV signal is delayed
in time so that
it is aligned with the cancellation signal, and then is combined with it (such
as by summing)
so as to provide a corrected DTV signal that provides for substantial
elimination of the out-
of band intermodulation distortion.
The inventionfurrherprovidesthatthe intermodulationcancellationsignal is
adjusted
in phase and amplitude before being combined (summed) with the input DTV
signal to more
precisely eliminate the intermodulationdistortion. In a particularpreferred
embodiment, the
timing and amplitude of the two signals are equal, but 180° (opposite)
in phase.
_g_
CA 02254375 1998-11-13
Other objects, features and advantages of the invention will be apparent to
those
skilled in the art upon reading the following detailed description with
reference to the
accompanying drawings.
BRIEF DESCRIPTION OF THE DRAWINGS
The invention is better understood by reading the following Detailed
Description of
the Preferred Embodiments with reference to the accompanying drawing figures,
in which
like reference numerals refer to like elements throughout, and in which:
l0 FIG. 1 is a frequency-domaindiagram of an analog (for example, NTSC)
television
signal format, further illustrating the phenomena of second- and third-order
intermodulation
components that lie inside and outside the television channel.
FIG. 2 illustrates a known arrangement for cancellation of the 44.829 MHz
third-
order in-band intermodulation component of an NTSC-format television signal.
FIG. 3 illustrates a preferred embodiment of an out-of channel intermodulation
correction circuit for digital-format RF transmitters according to a preferred
embodiment of
the present invention.
FIGS. 4A-4E show frequency spectra of signals at various points in the circuit
shown
in FIG. 3.
-9-
CA 02254375 1998-11-13
FIG. 5 illustrates a combined NTSC/DTV transmitter in which the
intermodulation
corrector according to FIG. 3 may be included in processor 330.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
In describing preferred embodiments of the present invention illustrated in
the
drawings, specific terminology is employed for the sake of clarity. However,
the invention
is not intended to be limited to the specific terminology so selected, and it
is to be understood
that each specific element includes all technical equivalents that operate in
a similar manner
to accomplish a similar purpose. Moreover, components and design procedures
that are
1o readily known to those skilled in the art are omitted for the sake of
clarity.
Briefly, FIG. 3 illustrates a preferred embodiment of an intermodulation
correction
circuit 100 for digital-format RF transmitters according to a preferred
embodiment of the
present invention. Referring to FIG. 3 , the circuit includes a first mixer
110, an amplifier
120 and a second mixer 140 that are connected in series. First mixer 110
receives a DTV IF
input signal on both its inputs, while second mixer 140 receives the DTV IF
signal at its
remaining input. At the output of the second mixer is a cancellation signal
adjustment circuit
150 that includes a phase adjuster 154 and an amplitude adjuster 156. A delay
element 152
receives the DTV input signal and provides a delayed DTV input signal.
Finally, the circuit
includes a summing circuit 190 that receives an output of the cancellation
adjustment circuit
150 as well as the delayed DTV IF television signal from delay element 152.
-10-
CA 02254375 1998-11-13
The FIG. 3 embodiment is preferably applied to the DTV signal at intermediate
frequency (IF) (41-47 MHz) before it is converted to an RF frequency (for
example, in the
VHF or UHF band). (See the discussion of FIG. S, element 330, that is provided
below.)
The present discussion refers to intermodulation correction, but it should be
noted that the
goal in the preferred application is to provide a corrected signal at the
output of the RF
transmitter (element 1150 in FIG. 5) and not at the immediate output of the
FIG. 3 circuit
itself (element 330 in FIG. 5).
The correction provided at IF by the FIG. 3 circuit is embodied in the signal
that is
amplified and translated to RF for transmission. It is the high-power
transmitted (RF) signal
to that reflects the substantial elimination of the out-of channel
intermodulation products.
Thus, the invention of FIG. 3 may be considered to be a pre-correction circuit
in the
illustrated application, which performs correction on IF DTV signals for the
benefit of RF
transmitted DTV signals. The invention of FIG. 3 is especially suitable for
correcting
intermodulation products that lie outside the DTV channel but within an
adjacent analog
channel.
In operation, a digital television (DTV) signal at intermediate frequency (IF)
is input
to the correction circuit on a path 102. A first level adjuster 104 (such as
an attenuator)
provides the same DTV signal at a lower power level on path 102'. A second
level adjuster
106 (such as an attenuator) provides the same DTV signal at a lower power
level on
2o path 102".
-11-
CA 02254375 1998-11-13
An adjustable delay element 152 delays the input DTV signal and provides it on
path
102 "' to a summing circuit 190. The delay element compensates for the delay
that the
correction process imposes on the cancellation signal 158 (described below).
For most purposes, the signals on paths 102, 102', 102" and 102"' can be
considered
to be substantially the same signal, because their respective frequency
spectra are identical
for practical purposes. Path 102' carries a low-level version of the DTV input
signal to both
inputs of first mixer 110. Path 102" carries a low-level version of the DTV
input signal to
one input of second mixer 140. Path 102"' carries the selectively-delayed DTV
input signal
to a first input of summing circuit 190, by-passing the correction circuit.
to Conventionally, mixers have had two inputs, including a LO (local
oscillator) port
and an RF (radio frequency) port. Conventionally, for a mixer, the LO port has
accepted a
high-power-level (+7 dBm, or switching-level, appropriate for the particular
mixer
chosen) sinusoidal signal, and the RF port that has accepted an information-
bearing signal
of a much lower power level.
However, in the illustrated embodiment, the inputs to the first and second
mixers are
all of a lower power level, typically -10 dBm, to essentially allow the mixers
to perform
time-domain multiplication; no sinusoidal local oscillator signal, as such, is
input to them.
First mixer 110 essentially multiplies the input signal by itself in the time
domain, and
generates linear sum and difference frequency components in the frequency
domain,
2o preserving the amplitude information of the input signals.
-12-
CA 02254375 1998-11-13
Because the broadband, noise-like DTV signal is input to both inputs of mixer
110,
the sum and difference components at the output of mixer 110 are correlated
with the original
DTV signal on paths 102, 102' in both amplitude and phase. However, the
components are
far from each other in frequency because they are second-order intermodulation
components
(the result of only one mixing--see the discussion in the Background of the
Invention).
These intermodulation components are similar to those that are generated by
amplifiers, such as those used in DTV transmitters. Generally, when such
amplifiers amplify
a noise-like input signal, regenerated sideband components ("shoulders") arise
at the
amplifier output that surround the frequency spectrum of the signal that is
input to the
1o amplifier. Such sideband components are also noise-like and are correlated
with the input
signal, a fact used to advantage in the present invention.
Referring again to FIG. 3, the IF output of first mixer 110 is fed through an
amplifies
120, and the resulting amplified signal is fed on path 122 to the input port
of second mixer
140. Second mixer 140 thus receives both the original DTV signal on path 102"
and the
mixed signal on path 122. The sum and difference frequency components in the
mixed
signal on path 122 are correlated with the original DTV signal on paths 102,
102', 102",
102"'.
The components in the mixed signal on path 122 are similar to the sideband
components that are generated by the amplifiers in DTV transmitters, and they
are correlated
with the original signal. The invention envisions that such components on path
122 are
-13-
CA 02254375 1998-11-13
useful in cancelling the undesirable sideband components. Accordingly, second
mixer 140
outputs an intermodulationcancellation signal that may be adjusted for use in
cancelling the
undesirable out-of DTV-channel intermodulation components that are generated
by
high-power amplifiers in DTV transmitters. The inventive technique may be
thought of as
achieving high-power-levelRF output correction by means of pre-correctingthe
DTV signal
at low power levels at IF.
The phase and amplitude of a cancellation signal on path 142 must be properly
adjusted in order to achieve proper cancellation of the intermodulation
components at the
output of the high-power amplifier in the DTV transmitter. Adjustment circuit
150 includes
1o phase adjuster 154 and amplitude adjuster 156 to perform these adjustments.
Adjustment
circuit 150 outputs a phase- and amplitude-adjusted cancellation signal on
path 158 to
summing circuit 190. Adjustable delay element 152 is adjusted to ensure that
the delayed
signal on path 102"' is adjusted to compensate for the time delay in
delivering the
cancellation signal on path 158 caused by the correction circuit. In the
preferred
embodiment, the amplitude of the two signals is adjusted to be the same, but a
phase
adjustment is performed at IF so that the phase difference between the signals
is 180°
(opposite in phase) at the RF output of the high-power amplifier.
Summing circuit 190 receives the phase- and amplitude-adjusted cancellation
signal
on path 158 and adds it to the uncompensated time-adjusted DTV IF television
signal on
2o path 102"'. The sum signal output on path 192 has the out-of band
intermodulation
-14-
CA 02254375 1998-11-13
pre-correctionprovided by the inventive intermodulationcancellation circuit.
In a preferred
application of the invention, this intermodulation correction at IF causes
intermodulation
cancellation to occur at the output of the RF high-power amplifier (see FIG.
5).
To illustrate operation of this embodiment, FIGS. 4A-4E show frequency spectra
of
signals at various points in the circuit shown in FIG. 3.
FIG. 4A shows the frequency spectrum of an original, uncorrected intermediate-
frequency (IF) DTV signal that is input to the circuit. The input DTV signal's
energy is
concentrated in the 41-47 MHz band, with the pilot being near the upper end of
the band.
The input DTV signal with this frequency spectrum is input at low power level
to both inputs
1 o of first mixer 110.
Essentially, first mixer 110 multiplies each infinitesimally narrow frequency
component in the 41-47 MHz band with every other infinitesimally narrow
frequency
component in the same band. This time-domain multiplication causes a frequency
addition
that results in a spectrum that is twice as wide as the original, and whose
boundaries are at
twice the respective frequencies of the original boundaries. FIG. 4B shows the
frequency
spectrum of the resulting signal that is output by first mixer 110.
The energy of the FIG. 4B signal is concentratedbetween 82 MHz and 94 MHz,
with
the spectrum also reflecting the effect of the pilot carrier multiplied by
itself near the upper
end of the band. The frequency spectrum of the signal output by first mixer
110 is
-15-
CA 02254375 1998-11-13
essentially retained through amplification by element 120. The FIG. 4B signal
is input at
low power level (such as -10 dBm) to second mixer 140.
Second mixer 140 multiplies time-domain components in the 41-47 MHz band from
signal on path 102" with components in the 82-94 MHz band from the signal on
path 122.
Thus time-domain multiplication causes a frequency subtraction that results in
a signal with
a spectrum that extends from 3 5-53 MHz, with more energy focused in the 41-47
MHz band.
This frequency spectrum is illustrated in FIG. 4C.
FIG. 4C illustrates how the portions 441, 442 (in the outlying frequency bands
35-41 MHz and 47-53 MHz, respectively) are lower than the portion in the
center portion
l0 443 (in the 41-47 MHz frequency band). This signal, provided on path 142
(FIG. 3), may
be considered to be the "unadjusted" intermodulation cancellation signal that
is input to
adjustment circuit 150.
Delay circuit 152, which may simply be a suitable length of coaxial cable,
delays the
signal on path 102 timewise to produce signal 102"', so that portions 441, 442
(FIG. 4C) line
up timewise with the corresponding "shoulders" of the uncorrected signal.
Phase adjuster 154 may simply be a variable-length cable much shorter than the
length of cable forming element 152. Phase adj uster 154 shifts the phase of
the cancellation
signal without introducing adversely affecting time delay. This phase
adjustment allows
opposite phase matching of portions 441, 442 with the corresponding
"shoulders" of the
2o uncorrected signal in FIG. 4A.
-16-
CA 02254375 1998-11-13
Finally, amplitude adjuster 156 may be a suitable variable attenuator that
adjusts the
amplitude of the cancellation signal that has just been correctly aligned in
time and phase by
elements 152 and 154. FIG. 4D illustrates the frequency spectrum of the
adjusted
cancellation signal output by adjustment circuit 150 on path 158 to match that
of the
uncorrected signal. The signal whose spectrum is shown in FIG. 4D is
180° out of phase
with the signal whose spectrum is shown in FIG. 4A (the amplitude-versus-
frequency
diagrams cannot explicitly illustrate phase difference). In this manner, the
adjusted
cancellation signal that is output by amplitude adjuster 156 precisely cancels
the "shoulders"
of the uncorrected signal. This cancellation is accomplished by the addition
performed by
to summing circuit 190.
FIG. 4E shows the frequency spectrum that results when signals on paths 102"'
and
158, with respective frequency spectra in FIGS. 4A and 4D, are added. Portions
491, 492
of the spectrum of the corrected signal are shown to be at a lower level than
the portions 401,
402 (shown in dotted lines in FIG. 4E) of the uncorrected signal on path 102.
Summing circuit 190 achieves intelmodulation cancellation by adding portions
401
and 451 (representing signals that are aligned in time and amplitude but 180
° out of phase),
and likewise by adding portions 403 and 453 (likewise representing signals
that are aligned
in time and amplitude but 180 ° out of phase). The summation performed
by summing
circuit 190 does not cause a significant subtractive effect between portion
103 (FIG. 4A) and
-17-
CA 02254375 1998-11-13
portion 453 (FIG. 4D). Thus, the desired effect of out-of channel
intermodulationreduction
is achieved without substantial alteration of the in-channel RF signal.
FIG. 4E shows how the FIG. 3 circuit produces an output on path 192 that
cancels
shoulders at the output of the high-power RF amplifier (see FIG. 5 amplifier
1145). The
FIG. 3 circuit is preferably used at IF, before the corrected DTV signal is
amplified by the
main DTV amplifier 1145. The amplitude, phase and timing delay adjustments (in
FIG. 3
elements 152, 154,156) are made with the goal of achieving cancellationat the
output of the
RF transmitter (FIG. 5). These adjustments may be different than the
amplitude, phase and
timing delay settings that would achieve cancellation at the output of the
FIG. 3 circuit as
such.
FIG. 5 illustrates a transmitter in which the intermodulation corrector
according to
the present invention may be employed. Briefly, the transmitter transmits
analog (e.g.,
NTSC) and DTV signals using a common antenna 1150 after amplifying them with a
single
wideband (12 MHz) amplifier 1145. The top portion of FIG. 5 shows the NTSC
portion, and
the bottom portion of FIG. 5 shows the DTV portion. Such a circuit was
disclosed in
co-pending U.S. Application No. 09/050,109, which is incorporated herein by
reference.
Refernng to FIG. 5, a video processor 210 manipulates an input video signal to
compensate, in advance, for distortion that is expected to occur at
intermediate frequency
(IF) and broadcast radio frequency (RF) to the modulated visual signal. Video
processor 210
provides such compensation functions as differential gain compensation and
differential
-18-
CA 02254375 1998-11-13
phase compensation and luminance non-linearity compensation, to achieve
linearity of
response.
An NTSC modulator 220 receives the compensated video signal from the video
processor 210, along with an associated audio signal. Essentially, the
modulator modulates
the video and audio signals onto an intermediate frequency carrier from a
phase lock loop
(PLL) 225 that is phase locked to common frequency reference 100. The
modulator may
comprise, for example, a vestigial side band (V SB) filter implemented with
surface acoustic
wave (SAVE technology. NTSC modulator 220 outputs an intermediate-frequency
signal
with NTSC-standard visual and aural carriers at 45.75 MHZ and 41.25 MHZ
carrier
1 o frequencies, respectively.
Frequency reference 100 may include, for example, an oscillator 101 that
provides
a stable-frequency reference signal (such as 10 MHZ) to various circuit
components via a
signal splitter 102. Oscillator 101 may be implemented as a temperature-
controlled crystal
oscillator, an oven-controlledcrystal oscillator, a GPS (global positioning
system) reference
signal, and the like. Signal splitter 102 may be any circuit that fans out the
reference signal,
ensuring a constant phase relationship throughout the circuits it drives.
The NTSC IF signal from modulator 220 is pre-distorted by an NTSC IF processor
230 to compensate for distortion that is expected to occur at radio frequency
(RF) to the
modulated NTSC signal. NTSC IF processor 230 may compensate for such
undesirable
2o phenomena as intermodulationdistortion, cross-modulationdistortion, and
incidental carrier
-19-
CA 02254375 1998-11-13
phase modulation distortion, and the like, resulting in a compensated, purely
amplitude-
modulated signal that is desired. The pre-compensatedNTSC IF signal is
provided to IF-to-
broadcast-frequency converter 240.
IF-to-broadcast-frequency converter 240 also receives a sinusoidal carrier of
a
frequency determined by the desired broadcast frequency of the particular
broadcast
channel "N", such as in the UHF range, that is allocated to the broadcast site
involved. The
carrier is provided by a phase lock loop (PLL) 250, which is phase-lockedto an
output from
the frequency reference 100. IF-to-broadcast-frequencyconverter 240 includes a
mixer that
provides a low-power (for example, one watt) modulated NTSC signal at Channel
N's
1 o broadcast frequency. Converter 240 reverses the frequency order of the
aural and visual
components of this low-power, broadcast-frequency, modulated NTSC signal, so
that the
visual component carrier is now below the aural component carrier, in
accordance with
broadcast standards.
A series of amplifiers, shown by exemplary intermediate power amplifier (IPA)
260
and driver amplifier 270, amplify the low-power, broadcast-frequency,
modulated NTSC
signal from converter 240 to a power level closer to broadcast power levels.
For example,
driver amplifier 240 may output a signal of 2.5 kW peak average power at sync,
with 125 W
aural power. This signal is provided to the first input of the combiner 1140.
Preferably, the signal output to the combiner is subject to automatic gain
control
2o (AGC). For this purpose, one example (not shown) of an AGC feedback path is
provided
-20-
CA 02254375 1998-11-13
from the output of driver amplifier 270 to the IF-to-broadcast-frequency
converter 240.
Gain control circuitry that may be of conventional design, and located within
converter 240,
ensures that an NTSC signal of substantially constant power level is provided
to the
combiner.
Like the description of the NTSC signal modulator, the following description
omits
conventional elements known to those skilled in the art, with the
understanding that
commercially-available products perform the same overall function. Further,
the present
description is abbreviated because the functionsperfolmed by elements 320,
325, 330, 340,
350, 360, 370, and 377 perform functions that are analogous to the functions
performed by
to elements 220, 225, 230, 240, 250, 260, 270, and 277, respectively.
Modulator 320 receives a Garner frequency signal that is phase locked by PLL
325
to a reference Garner from frequency reference element 100. Modulator 320
further encodes
a 19.39 MHZ, SMPTE 310M-compliantMPEG bit stream, and modulates a pilot
carrier at
46.69 MHZ in accordance with (for example) the 8-VSB standard accepted by the
Federal
Communications Commission for terrestrial broadcast. Modulator 320 outputs an
intermediate-frequency analog signal with a pilot carrier at 46.69 MHZ, at the
upper edge
of the 41-47 MHZ band allocated for television signals at IF.
A DTV IF processor 330 processes the IF signal from the modulator, performing
pre-compensationand pre-conditioningfunctions generally analogous to that
performed by
2o processor 230 for NTSC signals. However, DTV IF processor 330 is preferably
-21 -
CA 02254375 1998-11-13
implemented as a digital signal processor (DSP) to perform the pre-
compensation and
pre-conditioning functions on a digital-content signal, using techniques (such
as finite
impulse response filters) that are better suited to processing of such
signals. In any event,
the DTV IF processor 330 provides a pre-compensated and pre-conditioned signal
to an IF-
to-broadcast-frequency converter 340.
As mentioned above, the inventive intermodulation correction device of FIG. 3
may
be part of processor 330 (FIG. 5). The FIG. 3 circuit is used at IF, before
the corrected DTV
signal is RF-converted, and amplified by the main DTV amplifier 1145. The
amplitude,
phase and timing delay adjustments (in FIG. 3 elements 152, 154, 156) are made
with the
1o goal of achieving cancellationat the FIG. 5 output 1150 of the transmitter.
These amplitude
phase and timing adj ustments may be different than the amplitude, phase and
timing settings
that would achieve cancellation at the output of processor 330 itself.
Referring again to FIG. 5, an IF-to-broadcast-frequency converter 340 converts
the
analog IF signal from the DTV IF processor 330 to a broadcast-frequency
signal. In the
1 s preferred application of the invention, in which the DTV channel is
immediately adjacent the
corresponding NTSC channel in the frequency spectrum in accordance with FCC
channel
assignments, two situations are encountered. The "N-1 " situation involves a
DTV channel
that is immediately below the NTSC channel, and the "N+1" situation involves a
DTV
channel that is immediately above the NTSC channel. The IF-to-broadcast-
frequency
2o converter 340 is therefore illustrated as providing a signal on Channel N-1
or Channel N+1,
-22-
CA 02254375 1998-11-13
where "N" is the channel assigned the corresponding NTSC channel. Essentially
including
a mixer, converter 340 receives a sinusoidal carrier signal from a phase lock
loop 350 that
is driven by frequency reference element 100 to be modulated by the DTV IF
signal.
The converter's operation results in a reversal of the frequency order of the
DTV
signal from the upper end of the channel (46.69 MHZ is near 47 MHZ) to the
lower end of
the channel at broadcast frequencies. It is noteworthy that, in the N+1
situation, this
placement of the DTV signal at the lower end of the DTV channel places it only
510 kHz
away from the deviated NTSC aural carrier.
Converter 340 provides a low-power, broadcast-frequency signal to a series of
Io amplifiers, shown as including an intermediate power amplifier (IPA) 360
and a driver
amplifier 370. Driver amplifier 370 provides to the combiner, an 8-VSB-
compliant DTV
signal, in Channel N-1 for the "n-1" channel allocation situation or in
Channel N+1 for the
"N+1 " channel allocation situation. The signal from driver amplifier 370 is
of a power level
sufficient to drive the high power amplifier 1145 to provide the desired
broadcast output
power. An AGC feedback path (not shown) may be provided from the drivers's
output back
to the converter 340, which ensures that the signal provided to the combiner
is of
substantially constant power level.
Combiner 1140 is preferably implemented as a conventional quadrature hybrid
combiner of the type discussed in detail in commonly-assigned U.S. Patent No.
4,804,931.
2o As is readily appreciated by those skilled in the art, hybrid combiners are
four-port devices
- 23 -
CA 02254375 1998-11-13
that have two outputs, each one of which receives half the signal power from
each of the
combiner's two inputs. Thus, an undesirable characteristic of hybrid combiners
is that they
halve the power of the sum signal. In the present use of hybrid combiners,
half the power
from each input signal is provided to the high-power amplifier 1145, while the
other half of
the power from each input signal is wasted through dissipation in a resistance
to ground.
Despite the power loss to resistance, the desirable linearity of the hybrid
combiner, and the
isolation of the input signals from each other to thereby avoid undesirable
mixing of the two
inputs, make it a preferred implementation for combiner 1140.
High-power amplifier 1145 fulfills the demand of flatness of response (less
than 1
to dB) across a two-channel-wide bandwidth (12 MHZ in the United States, 16
MHz in most
countries outside the U.S.), and the requirement for meaningful power to the
NTSC+DTV
signal s with minimal inter-channel interference. A tetrode-class device, and
especially a
diacrode amplifying device such as a Thomson TH-680, provide optimum
performance for
this application. Tetrode and diacrode implementationsare preferred because of
their ability
15 to operate with cavity sections tunable to wide (two-channel wide)
bandwidths, to exhibit
sufficient linearity so that cross modulation and intermodulationdistortion
may be corrected
with established methods, and to provide meaningful broadcast power levels. Of
course, the
scope of the invention should not be limited to tetrode and diacrode
solutions; alternative
implementations, such as those involving solid state amplifiers, also lie
within the
2o contemplation of the invention. The diacrode or tetrode power amplifier may
be replaced
-24-
CA 02254375 1998-11-13
by a suitable broadband solid state amplifier using an appropriate number of
power RF
transistors to get to the required power, and advantageously can operate in
both the UHF and
VHF bands.
In an exemplary embodiment, the TH-680 can provide 104 kW of peak envelope
s power, which may (as a non-limiting example) include the following
allocation of power
levels. To reduce interference with channels outside the adjacent-channel
pair, a suitable
two-channel-wide( 12 MHZ in the U. S.) band pass filter (BPF) 1146 is provided
at the output
to the amplifier 1145. To comply with broadcast power standards, the amplifier
1145 must
amplify the combined NTSC+DTV signal so that the BPF provides a signal 25 kW
average
1 o peak-of syncpower (NTSC),1.25 kW NTSC average aural power, and 2.5 kW
average DTV
power. Of course, variation of the above particulars in accordance with
commonly-known
principles lies within the ability of those skilled in the art.
As is readily appreciated by those skilled in the art, such an amplifier
involves a tube
that performs the power amplification, as well as a resonator cavity that
limits the frequency
15 range in which the tube amplifies signals. For any assigned adjacent-
channel pair (either
Channel N-1 through N, or Channel N through N+1 ), one skilled in the art,
upon reading this
specification,is readily capable, without undue experimentation,of
implementing a properly-
tuned amplifier using a suitable tetrode-class device and resonator cavity.
The
implementationis different for each adjacent-channel pair, but the design
principles remain
-25-
CA 02254375 1998-11-13
the same regardless of the particular assignment, and further details need not
be provided
here to illustrate the implementation and operation of the invention.
For many applications, high-power amplifier 1145 comprises a tetrode-class
device,
especially a Thomson TH-680 diacrode, available from Thomson Tubes
Electronique. It is
to be understood that the scope of the invention should not be limited to a
particular
component or to a specific set of signal types.
FIG. 5 emphasizes an implementation of the power level AGC feedback
arrangements that ensure that output power levels are maintained substantially
constant. In
FIG. 5, feedback paths 276 and 376 are shown leading from the broadcast signal
output by
to the band pass filter 1146, back to respective IF-to-broadcast-frequency
converters 240 and
340. Paths 276 and 376 are provided in lieu of paths (now shown) from
amplifiers 270, 370,
respectively. NTSC channel bandpass filter 277, and DTV channel band pass
filter 377, are
provided in feedback paths 276, 376, respectively, so that only in-channel
frequency
components are returned to converters 240, 340.
The feedback arrangements operate on similar principles of feedback control,
known
to those skilled in the art. When average power varies from a desired steady-
state power
level, either at the outputs of driver amplifiers 270, 370 or at the output of
BPF 1146,
feedback arrangements within converters 240, 340 act to correct the variation
to return the
power level at the sensed point back to the desired steady-state power level.
The gain factor
2o that converters 240, 340 apply to the feedback signals on paths 275, 375 or
276, 376 are
-26-
CA 02254375 1998-11-13
different, and are determined by the differences in magnitude of power between
the outputs
of driver amplifiers 270, 370 and of BPF 1146. However, the principles remain
the same.
Gain correction achieved locally (within the respective NTSC and DTV paths)
compensates only for variations that occur through amplification paths 260,
270 and 360,
370. However, the implementation shown in FIG. 5 achieves a more comprehensive
gain
correction over the entire path between the IF-to-broadcast-frequency
converters 240, 340
and the ultimate output of BPF 1146.
Those skilled in the art will recognize that, although the disclosed
embodiment is
designed especially for eliminating intermodulation distortion from broadband,
noise-like,
to DTV signals, the invention is equally applicable to eliminating
intermodulation distortion
from analog (e.g., NTSC) signals that essentially constitute a set of carrier
signals at discrete
frequencies. Accordingly, an embodiment with the structure of FIG. 3 may be
used to
implement part of element 230 (FIG. 5) as well as element 330 (FIG. 5). Thus,
it is to be
understood that the invention has utility in eliminating intermodulation
distortion in either
or both analog (e.g., NTSC) format signals and digital format (DTV) signals.
Modifications and variations of the above-described embodiments of the present
invention are possible, as appreciated by those skilled in the art in light of
the above
teachings. For example, the particular electrical and electronic components
used, the signals
on which the components operate, the power levels of the signals, the
particular frequency
2o band of the signals, the ordering of elements such as the adjustment
elements, and the like,
-27-
CA 02254375 1998-11-13
may be varied in accordance with principles known to those skilled in the art
without
departing from the scope of the present invention. It is therefore to be
understood that,
within the scope of the appended claims and their equivalents, the invention
may be
practiced otherwise than as specifically described.
-28-
