The present invention relates to communications systems, and, in particular, to digital communications systems.
Typical current digital communication systems often use non-constant envelope modulation schemes, e.g. the new system EDGE using 3π/8-8PSK modulation. This means that some part of the information lies in the amplitude (envelope) of the transmitted signal and some part lies in the phase of the transmitted signal. In other words, this is a combination of Amplitude Modulation (AM) and Phase Modulation (PM).
To deal with amplitude modulation, an output Power Amplifier (PA) in the radio transmitter has to be linear, i.e. the relationship between the output power of the PA (Pout,PA) and the input power of the PA (Pin,PA) has to be linear for all possible power levels. Otherwise the result will be AM-to-AM distortion, i.e. the gain of the PA changes with the input amplitude.
To deal with the phase modulation, the phase-shift (Δφ) through the PA has to be constant for all possible power levels. Otherwise the result will be AM-to-PM distortion, i.e. the phase-shift of the PA changes with the input amplitude.
The consequences of using a PA with non-constant gain and/or non-constant phase-shift, will be amplitude distortion and/or phase distortion in the transmitted signal. This distortion leads to spectrum broadening, which results in an increased adjacent channel disturbance. The amplitude/phase distortion (vector distortion) in the transmitter also affects the performance of the communications system. For example, an increased BER (Bit Error Rate) in the communication system, will lead to a decreased signal quality (e.g. degraded audio quality in a voice application).
Therefore, linearity is crucial for a transmitter used in a digital modulation system with non-constant amplitude modulation. Moreover, high linearity requirements often lead to poor power efficiency. To attain good linearity and good power efficiency, some linearization method and/or some efficiency enhancement method are often used. A problem that often arises is then poor time alignment between the “information parameters” (or “information components”), i.e. gain and phase (polar representation), alternatively I and Q (cartesian representation).
There are several known ways to attain linearity and/or power efficiency in RF (Radio Frequency) transmitters for digital modulation systems with non-constant amplitude modulation, for example:
One embodiment of the present invention can compensate for time delay between amplitude and phase-information. Alternatively, compensation for time delay between the in-phase component (I) and the quadrature component (Q) can be obtained. The timing problem is transferred to the digital baseband domain, where it can be solved. The method could be used in different linearization configurations, such as “Cartesian Feedback”, “Polar Loop Feedback” and “Envelope Elimination and Restoration with Linearization”. Since the time delay compensation as well as the adaptive linearization takes place in the digital baseband domain, the invention is a form of “Adaptive Time-alignment of Information Components”. As will be shown, the invention also gives increased flexibility in the choice of circuit configuration in the feedback part of the linearizer.
The invention can be applied both in TDMA (Time Division Multiple Access) systems or in CDMA (Code Division Multiple Access) systems. An example of a system in a TDMA category is EDGE (Enhanced Data rates for GSM Evolution). In the CDMA category we have, for example, Wideband CDMA or UMTS.
The invention presented in this report reduces time miss-alignment between the amplitude and the phase-information, alternatively between I and Q, in a radio transmitter. The invention can be applied in TDMA (Time Division Multiple Access) systems, or in CDMA (Code Division Multiple Access) systems. An example of a system in the TDMA category is EDGE (Enhanced Data rates for GSM Evolution), another is UMTS. In the CDMA category we have for example W-CDMA.
A block diagram illustrating a first embodiment of the invention with compensation for time delay between the phase φ and the amplitude (envelope) r, is presented in
The system of
In an embodiment of the present invention, an output detector unit 3 is provided which serves to monitor the power amplifier output signal and to produce detected phase and amplitude (φ4, r4) signals. A local oscillator (LO) 5 is provided in order to enable the output detector unit 3 to convert the RF power amplifier output signal to the digital baseband frequency of the circuit. The RF signal is mixed down to the digital baseband frequency. This operation can be performed by a mixer having one input from the RF signal and another input from the local oscillator 5. The mixer multiplies the two signals to produce a signal having one component having a frequency equal to the local oscillator frequency plus the RF frequency, and another component having a frequency equal to the difference in LO and RF frequencies. The LO+RF frequency is filtered out, leaving a baseband frequency signal. The system also incorporates a signal generator 7 which receives digital data D and operates to produce phase and amplitude information (φ1, r1) for supply to the RF circuitry 1.
In an embodiment of the present invention, the phase information (φ1) produced by the signal generator 7 is supplied to a delay element 81. The delay element 81 operates to delay the signal φ1 by an amount of time controlled by a controller 91. The output of the delay unit 81 (i.e. a delayed φ1) is subtracted by a combining unit 101, from the detected phase signal (φ4) of the output detector unit 3. The delay controller 91 operates to modify the delay introduced by the delay unit 81 such that the magnitude of the difference between the detected phase value (φ4) and the delayed generated phase value (φ3) is minimised. The result of this control, signal d1 is a measurement of how much the phase signal φ is delayed in the RF circuitry 1.
Corresponding circuit elements are provided for the generated amplitude signal r1. The amplitude signal r1 is delayed by a delay unit 82 which is itself controlled by a delay controller 92. A combining unit 102 subtracts the delayed generated amplitude signal r3 from the detected amplitude signal r4. The delay controller 92 operates to minimize the magnitude of the difference between the detected and delayed generated amplitude signals (r4, r3). As before, the delay control signal d2 for the amplitude circuit is a measurement of how much the amplitude signal r is delayed by the RF circuitry 1.
An embodiment of the present invention includes a delay calculation unit 12 which receives the outputs from the delay control units 91 and 92 (signals d1 and d2). The delay calculation unit 12 determines the difference between the two input signals and produces control outputs dφ control and dr control. The control outputs dφ, dr from the calculation unit 12 are used as inputs to a phase controller 14 and an amplitude controller 16 respectively. The phase controller 14 operates to adjust the generated phase signal φ1 for supply (φ2) to the RF power amplifier circuitry, and the amplitude controller 16 operates to adjust the generated amplitude signal r1 for supply (r2) to the power amplifier circuitry. The phase and amplitude controllers 14 and 16 operate to compensate for the actual detected time delay between the phase and the amplitude detected by the output detector unit 3.
I+j·Q=r·ej·φ (1)
In the following, x and y are used to represent parameters that, from the above-described embodiments would be a polar or Cartesian parameter. The block Delay 1 Control 91 changes the delay control parameter d1 (i.e. the delay value of Delay unit 81) until the difference Δx between x3 and x4 has been minimised. The difference between x3 and x4 could for example (however other possibilities exist) be calculated as the “Least-Mean-Square”-value (LMS) given by equation (2):
where m is the number of samples over which the LMS-value is calculated. The value d1 is the number of samples which x1 is delayed in order to form X4. When min {Δx} has been found, the “final” value of d1 has also been found.
In the same way, delay control 92 changes the delay parameter d2 (i.e. the control delay value of delay unit 82) until the difference Δy between y3 and y4 has been minimised. This means that d2 is obtained by minimising Δy in the expression (equation (3)):
After d1 and d2 have been found, we can calculate dx and dy, which are the two parameters used for achieving time-alignment between x and y. Since d1 and d2 tell us how much the signals x respectively y are delayed in the system, the time delay between x and y can be found by calculating Δxy=d1−d2. If Δxy>0, i.e. if d1>d2, then x2 should be sent Δxy samples before y2. Use for example dx=0 and dy=Δxy.
Correspondingly, if Δxy<0, i.e. if d1<d2, then x2 should be sent Δxy samples after y2. Use for example dx=Δxy and dy=0.
If Δxy=0, no correction is needed. Use for example dx=dy=0.
Benefits of embodiments of the invention are listed below:
As mentioned, embodiments of the invention can be very flexible. It could be used in several types of system:
1) Systems with different types of modulation principles
Number | Date | Country | Kind |
---|---|---|---|
0025436.7 | Oct 2000 | GB | national |
This application claims priority under 35 U.S.C. §§ 119 and/or 365 to 0025436.7 filed in the United Kingdom on Oct. 17, 2000 and to 60/241,796 filed in The United States of America on Oct. 20, 2000; the entire content of which is hereby incorporated by reference.
Number | Name | Date | Kind |
---|---|---|---|
5148448 | Karam et al. | Sep 1992 | A |
5598436 | Brajal et al. | Jan 1997 | A |
5644268 | Hang | Jul 1997 | A |
5672974 | Turner | Sep 1997 | A |
5748678 | Valentine et al. | May 1998 | A |
5903611 | Schnabl et al. | May 1999 | A |
5909471 | Yun | Jun 1999 | A |
6014570 | Wong et al. | Jan 2000 | A |
6081698 | Moriyama et al. | Jun 2000 | A |
6091941 | Moriyama et al. | Jul 2000 | A |
6246286 | Persson | Jun 2001 | B1 |
6288610 | Miyashita | Sep 2001 | B1 |
6614854 | Chow et al. | Sep 2003 | B1 |
6647073 | Tapio | Nov 2003 | B1 |
6771708 | Suga et al. | Aug 2004 | B1 |
20020131523 | Nagasaka et al. | Sep 2002 | A1 |
Number | Date | Country |
---|---|---|
0 377 519 | Jul 1990 | EP |
0 907 276 | Apr 1999 | EP |
1089428 | Apr 2001 | EP |
WO9730521 | Aug 1997 | WO |
WO9919990 | Apr 1999 | WO |
WO0031881 | Jun 2000 | WO |
WO 0072438 | Nov 2000 | WO |
Number | Date | Country | |
---|---|---|---|
20020058486 A1 | May 2002 | US |
Number | Date | Country | |
---|---|---|---|
60241796 | Oct 2000 | US |