Pulse or digital communications – Transmitters – Antinoise or distortion
Reexamination Certificate
2000-06-20
2003-10-28
Vo, Don N. (Department: 2631)
Pulse or digital communications
Transmitters
Antinoise or distortion
C330S107000, C455S126000
Reexamination Certificate
active
06639950
ABSTRACT:
BACKGROUND OF THE INVENTION
The invention relates to amplifiers and in particular to a method and arrangement for linearizing power amplifiers.
Linearized amplifiers are needed, for instance, for modern digital wireless communication systems, because of the requirement that the spectrum of a signal to be transmitted must not spread outside the actual useful band. Spreading of the spectrum results from non-linearity of amplifiers and it causes interference to adjacent channels, for instance. The linearity of amplifying stages depends on how they are biased and they can be classified according to linearity: a class A amplifier is the most linear but its efficiency is poor, whereas a class C amplifier's efficiency is good but it is at the same time highly non-linear. Good efficiency is an important property in power amplifiers, and this is particularly emphasized in wireless communication devices whose battery capacity is limited. Therefore, amplifiers that have good efficiency but are non-linear and need to be linearized are used.
One known method of linearizing a non-linear radio-frequency power amplifier is Cartesian feedback. In broad outline, its operational principle is as follows: data to be transmitted is included in baseband signals I and Q. These signals are applied to an I/Q modulator, in which the signals are combined and modulated directly to a final frequency. The final-frequency signal is amplified by one or more non-linear radio-frequency power amplifiers and passed to an antenna. The amplified radio-frequency signal is sampled after the last amplifying stage by a directional coupler, for instance. The sample signal is applied to the I/Q modulator, in which it is demodulated to the baseband, and the I and Q signals are separated therefrom. The baseband I and Q sample signals are finally summed to the actual I and Q signals. This causes predistortion in the I and Q signals, thanks to which the non-linearity produced in the power amplifiers is at least partly cancelled.
In the above-described arrangement, both the I/Q modulator and the I/Q demodulator receive a local oscillator signal from the same source. On the other hand, the power amplifiers cause delay and the sampled I and Q signals return in a wrong phase. This distortion can be compensated by adjusting the phase of the local oscillator signal applied to the I/Q modulator or I/Q demodulator considering the delay produced in the power amplifiers. In order that the phase error could be corrected, its magnitude has to be measured first. EP 0 706 259 A1 discloses one method of measuring and correcting a phase error in a Cartesian feedback loop. According to the method, the loop is broken for the duration of measuring and the measuring is performed by feeding excitation signals to the I and Q inputs of the loop and by measuring the resultant signals at the outputs of the I/Q demodulator, and further, by calculating a phase error from the measurement results. The method has a drawback that the loop has to be broken for the duration of the measuring, and consequently switches are required for both feedback branches. Further, the transmitter has to be turned off when operating the switches, in order for the transmission spectrum not to spread. Moreover, amplification of an open loop is typically much higher than amplification of a corresponding closed loop, whereby the effect of DC offsets of baseband operational amplifiers and noise on the accuracy achieved with phase measuring is rather high. Amplification of the open loop also varies much more than amplification of the closed loop, and consequently it is difficult to carry out phase measuring at a given power level. U.S. Pat. No. 5,175,879 discloses phase adjustment of a loop on a continuous basis the loop being in a normal closed state. The method employs a phase detector between the input of a linearization loop and the output of an I/Q demodulator. From the output of the phase detector, a phase difference signal is further generated by means of an integrator for a phase shifter which corrects the phase. One drawback of the method is, for instance, that while turning the amplifier on, the linearization loop may jitter and cause spectrum spreading. Discontinuities of the phase shifter would also cause problems to the implementation.
BRIEF DESCRIPTION OF THE INVENTION
The object of the invention is thus to provide a method to the effect that the above-described problems can be solved. This is achieved by a method for correcting a phase error in a linearization loop of a power amplifier, the loop comprising an I/Q modulator, one or more delay-causing power amplifiers to be linearized, an I/Q demodulator for producing I and Q feedback signals from the amplifier output signal, difference means of the I and Q branches for producing I and Q difference signals from the I and Q feedback signals and the I and Q input signals, the I/Q modulator and the I/Q demodulator receiving an oscillator frequency from the same local oscillator, and a phase shifter, the method comprising determination of a phase error resulting from delay produced in the linearization loop, which determination comprises feeding excitation signals to the I and Q inputs of the linearization loop, measuring the signals resulting from the excitation signals and calculating the phase error by means of the measured signals and excitation signals, and correcting the phase error by adjusting a phase of a local oscillator signal passing to the I/Q modulator or I/Q demodulator by means of the phase shifter, whereby the method is characterized in that when determining the phase error, the signals resulting from the excitation signals are measured from the I and Q difference signals or from the I and Q input signals of the I/Q modulator and that the phase error determination is performed the linearization loop being closed.
The invention is based on the idea that when using DC excitation signals the phase error can be readily calculated from the I and Q baseband signals of the linearization loop. A deviation of the measured resultant vector angle from the input vector angle indicates directly the phase error. By performing the measuring with a plurality of I and Q input signal combinations, the phase errors of the branches can be averaged. When the excitation signals are applied to the I and Q inputs of the linearization loop and the I and Q difference signals or I/Q modulator input signals are used as measuring signals, the phase error can be measured and compensated the linearization loop being closed.
The method of the invention has an advantage that the phase of the linearization loop can be adjusted in a normal mode on an accurately defined power level the loop being closed. The event of measuring and adjusting will not make the spectrum spread harmfully to adjacent channels. At the modulator input, signal levels are typically high, whereby interference signals do not cause considerable errors in measuring.
The invention also relates to a linearization arrangement of a power amplifier, the arrangement comprising difference means, which form I and Q difference signals out of the actual I and Q input signals and I and Q feedback signals of the linearization loop, an I/Q modulator, in which the data-containing baseband I and Q difference signals received from the difference means are combined and modulated to a final frequency, one or more delay-causing power amplifiers to be linearized by which the final-frequency signal is amplified whereafter it is applied to an antenna to be transmitted, a sampling arrangement, by which a sample signal is taken from the amplified final-frequency signal prior to the antenna, an I/Q demodulator, to which said sample signal is applied and in which the sample signal is demodulated to the baseband and the I and Q sample signals which form said I and Q feedback signals are separated therefrom, a local oscillator, from which a local oscillator signal is applied to the I/Q modulator and I/Q demodulator and a phase shifter, by which the phase of the local oscillator signal pass
Lagerblom Niklas
Thomasson Kristian
Nokia Networks Oy
Pillsbury & Winthrop LLP
Vo Don N.
LandOfFree
Method and arrangement for correcting phase error in... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and arrangement for correcting phase error in..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and arrangement for correcting phase error in... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3167071