The present invention relates to a balanced circuit arrangement such as a balanced or double-balanced amplifier or multiplier and to a method for linearizing such an arrangement e.g. in order to attenuate spurious signals and envelope distortions in radio receivers and transmitters.
In radio reception, the linearity of the receiver is an essential requirement. The linearity performance of a receiver, in general, is usually dominated by the downconversion mixer circuits. Such mixer circuits are used to translate or convert received high frequency signals down to a lower or intermediate frequency (IF). This conversion is achieved by mixing the received signals with a locally generated oscillator signal. By choosing the local oscillator signal to be a constant amount away from a selected or wanted signal in a first frequency band, the selected or wanted signal always appears at the same frequency in the intermediate frequency band. Thereby, selection of the selected or wanted signal may be facilitated by a fixed-tuned IF filter.
In homodyne or direct conversion receivers, the chosen intermediate frequency band is DC or zero frequency. The local oscillator then has a zero frequency separation from the selected or wanted signal. Any modulation on the selected or wanted signal that causes spectral components both above and below the nominal signal frequency becomes folded at the mixer output, as a component below the signal frequency or above the signal frequency will appear at the intermediate frequency above the nominal of zero. To allow for resolution of such folded components, two mixers are provided in a direct conversion receiver using local oscillator signals that are phase offset by 90 degrees. The components above and below the nominal signal frequency then appear folded as a sum signal at one mixer output and a difference signal at the other mixer output where they may be separated if desired. Such direct conversion receiver operations are described in more detail in document U.S. Pat. No. 5,241,702.
However, due to non-linearities, spurious responses will be generated in the direct conversion receiver, the worst being modulation-frequency interference at the receiver's mixer output caused by a strong amplitude-modulated signal of another transceiver. This will appear even if the frequency of the interfering signal considerably deviates from the receiving frequency. These interferences are mainly caused by the second-order distortion component which contains a variable-level DC component proportional to the amplitude of the interference-causing signal. The variable amplitude signal produces at the mixer output a signal which comprises a variable DC component and the frequency of which is identical with the variation of the amplitude. The spurious frequencies may corrupt the radio reception by blocking the following signal processing stages or deteriorating the detection of the desired signal which is overwhelmed by distortion.
The spurious frequencies can be categorized to exist due to the odd- and even-order non-linearities. The even-order mixing results are suppressed by using balanced or double-balanced mixer topologies. Ideally, the even-order spurious frequencies are cancelled in balanced and double-balanced constructions. However, in practice, the canceling is imperfect. The reason is the imperfect balance due to the mismatch of respective components in the differential branches, i.e. manufacturing tolerances.
In radio receivers utilizing a direct conversion architecture or a significantly low IF, the spurious signals cannot be removed by selecting an optimal IF. Due to the existence of the even-order distortions and imbalance in the circuitry, a variable DC component proportional to the signal level and amplitude modulation depth of the interfering signal occurs. Moreover, envelope distortions are detected, the amplitude of which is also proportional to the amplitude modulation depth of the interfering signal, and the frequency of which equals to the variation of the amplitude. Thus, not only a DC offset but also a low frequency disturbance may be generated to corrupt the desired reception band. This is a particular concern in down-conversion mixers of direct conversion receivers.
Several solutions for reducing distortions in radio receivers with low IF have been proposed. Document U.S. Pat. No. 5,749,051 suggests compensating unwanted terms caused by second-order intermodulation by feeding instantaneous power measurements to a signal processing unit along with the complex baseband signals. The signal processing unit then determines a complex compensation coefficient by correlating the power signal with the complex baseband signals. The complex compensation co-efficient is then employed to subtract a weighted amount of the power signal from the complex baseband signal in order to cancel the unwanted second-order intermodulation distortion terms. Furthermore, document EP 0 951 138 discloses a method for attenuating spurious signals in mixer circuits by setting variable-level bias voltages and/or currents to transistors in the mixer circuits. Additionally, document GB 2 346 777 suggests switching a DC offset correction in or out of the circuitry according to the received signal strength or signal-to-noise ratio.
Furthermore, the use of a dynamic matching procedure is described by E. Bautista et al., “Improved Mixer IIP2 through Dynamic Matching”, in the Digest of ISSCC 2000, pp. 376-377. According to this procedure, any undesirable second-order intermodulation distortion product generated in the mixer circuit is modulated to a frequency where it can easily be filtered off. This can be achieved by applying a periodic signal to input switches of the mixer circuit in order to modulate the received input signal. If the periodic signal is replaced by a pseudo-random signal, the undesirable second-order intermodulation distortion products can be spread over a wide range of frequencies to achieve a desired second-order input intercept point (IIP2) performance.
The second-order distortion phenomena itself and its undesired products, i.e. DC offset and envelope distortions, have not been thoroughly investigated so far. Due to lack of proper analysis of this topic, most of the solutions have been focused on the removal of the DC offset. However, even if the DC offset at the output of the mixer circuit is reduced to zero, the circuit may still be in an imbalanced condition, due to the fact that the envelope distortion itself causes a DC term which is related to other DC offsets in a complex manner.
It is therefore an object of the present invention to provide a balanced circuit arrangement and method for linearizing such an arrangement, by means of which spurious signals and envelope distortions in radio receivers and transmitters can be reduced.
This object is achieved by a balanced circuit arrangement comprising:
Furthermore, the above object is achieved by a method for linearizing a balanced circuit arrangement, comprising the steps of:
Additionally, the above object is achieved by a method for calibrating a balanced circuit arrangement, comprising the steps of:
The input signal can be either connected or disconnected from the mixer input. In the former case, it should be assumed that the average level of the AM distortion in the input signal varies only a little during the calibration.
Accordingly, envelope distortions and even-order spurious signals are controlled by controlling or changing DC offset balance errors in the balanced circuit arrangement. The balance errors occur due to the mismatch of components, i.e. component manufacturing tolerances. Thereby, the balanced circuit arrangement can be linearized in terms of even-order non-linearity by introducing a controlled imbalance in the load of the balanced circuit arrangement. In particular, the invention can be applied to any balanced circuit arrangement in radio receivers and/or transmitters requiring even-order linearity, such as in mobile communications devices.
Thus, the second-order input intercept point (IIP2) can be maximized by adjusting the loads of balanced circuit arrangement into a slight imbalance. This can be achieved by creating a controllable extraneous imbalance between the output loads of the balanced circuit arrangement. According to the measurements and simulations, the IIP2 performance of the balanced circuit arrangement can be enhanced up to 30 dB. The imbalance in the DC voltage or lowpass filtered and averaged signal, generated between the output branches can then be cancelled at a later stage after the signals causing envelope or other distortions have been filtered out. The effect on other significant performance parameters is negligible due to the fact that the required artificial mismatch is relatively small compared to the absolute values of the load components.
According to an advantageous development, the linearity control means may be arranged to perform the adjustment by selectively switching load elements to at least one of the first and second load means. Preferably, the load elements may be weighted load elements. Thereby, the amount of load imbalance may easily be controlled by a switching control function based on corresponding control inputs. In case of a use of weighted load elements, a control based on binary control words can be implemented. Thus, the linearity control means may comprise at least one input terminal for inputting a control signal.
The balanced circuit arrangement may be a single- or double-balanced circuit arrangement. In general, the present invention is applicable to any mixer circuit comprising a balanced circuit arrangement, e.g. a Gilbert-cell multiplier. Furthermore, the present invention may be applied to any modulator and/or demodulator circuit, such as an IQ modulator and/or demodulator, or to any receiver and/or transmitter circuit comprising e.g. a direct conversion receiver, in which a balanced circuit arrangement can be used.
According to another advantageous development, the adjusting step may be performed by selectively switching load elements of the controllable load means.
Furthermore, the adjusting step of the calibrating method may be an iterative step and the difference may be monitored by an analog or by a digital signal processing routine.
In the following, the present invention will be described in greater detail based on preferred embodiments with reference to the accompanying drawings, in which:
The present invention will now be described on the basis of a single-balanced and double-balanced mixer circuit, as an examples for a balanced circuit arrangement according to the present invention, which may be used in a direct conversion receiver for WCDMA (Wideband Code Division Multiple Access) base station applications in 3rd generation mobile communication networks.
According to the present invention, the load values of the load impedances ZLa and/or ZLb are controlled by the controller 4 so as to introduced a load imbalance required to maximize the IIP2 performance of the multiplier or mixer circuit. It is noted that the load impedances ZLa and ZLb may be any load means or circuitry which provides a mechanism for adjusting the effective load value introduced into the respective output branch. Thereby, the IIP2 performance of the mixer circuit can be improved, since the DC and low-frequency effects of the envelope distortions can be compensated by the introduced load imbalance.
In general, any transceiver, receiver or transmitter circuit can be calibrated by properly adjusting the load imbalance of an included balanced circuit arrangement. A corresponding calibration method is described in the following.
To obtain a reference or test signal, a transmission local oscillator (Tx LO) 50 is connected via an amplifier circuit 9 and a second switching element 14 to the input of the mixer 6. Thus, either a received radio signal or the generated test signal or both can be switched to the bandpass filter 20. The test signal can be any locally or externally generated signal.
After the start of the method in step S100, the switching element 11 is closed to connect the LNA 10 while no input signal is supplied to the mixer 6. In step S101, it may be considered or checked whether the receiving local oscillator 5 is switched on or off, since this may lead to different measuring results. Then, the output DC or lowpass filtered signal level of the mixer 6 is measured in step S102 by the test monitoring circuitry 12 at the respective one of the test nodes 13 and the measuring result is supplied to the DSP 8 where it may be stored. In step S103, the second switching element 14 is closed to connect to the output terminal of the amplifier circuit 9 so as to supply a sample of the test signal, e.g. the signal from the transmission local oscillator 50 in a receiver calibration, to the mixer 6, wherein the amplifier circuit 9 is arranged to provide an interface equivalent to that of the LNA 10. Alternatively, an attenuated test signal can be connected directly-to the input of the LNA 10, while the second switching element 14 remains open or can be dispensed with. Due to the second-order non-linearity and imbalance in the mixer 6, a DC error voltage is generated at the output of the mixer 6. This DC error is proportional to the amplitude of the even-order spurious signal, and is measured and may be stored in the DSP 8.
Based on the measured DC outputs, the DSP 8 provides a control to the controller 4 so as to adjust the load imbalance and thereby minimize the increment or increase in the DC voltage or in the lowpass filtered output signal at the output of the mixer 6 due to the DC error. Thus, the receiver circuit can be linearized by this DC level set control. According to
Thus, an automatically controlled calibration of receiver, transmitter or transceiver circuits can be provided.
The load elements ZL1 to ZLn may be arranged to provide weighted loads i.e. their load values may be generated by multiplying a basic load value with an integer number. As an example, the load value of an individual load element may be generated based on the following equation:
ZLi=2(i-1)·k·ZL
wherein ZL denotes a basic load value or unit load value, and k denotes an integer (1. . . n) selected according to the desired tuning range. Due to the fact that the above equation corresponds to the weighting of a dual number system, a direct relation between the binary control word applied to the control terminals ctrl1 to ctrlN and the generated load value can be obtained, while the control terminal ctrl corresponds to the most significant bit (MSB) and the control terminal ctrlN corresponds to the least significant bit (LSB). The adjustment circuitry comprises weighted fingers of load elements connected in parallel to a basic or original load. Each finger can be selected using the respective switch. Thus, in case resistor loads are used, the control terminal ctrlN connects the largest parallel resistor finger in parallel to the basic load ZL to thereby obtain a minimum load change. The resolution of the load adjustment control can be selected to achieve a desirable range, e.g. the resolution may be 0.1%. As determined on the basis of simulations and measurements, the total tuning range should preferably cover a range of±5% of the basic or original load value.
In complex receiver, transmitter or transceiver arrangements, each mixer, modulator or other balanced circuit arrangement can be adjusted separately. The calibration sequence may include counting of digital codes for adjusting the load imbalance, wherein measurements are swapped until the accepted level is reached.
According to the embodiment of
As can be gathered from
ZLb=ZLa·(1−0.5 Δz),
wherein Δz indicates the tuning range of the load adjustment circuit 41 in percentages relative to the actual load. Thereby, an adjustment of the load imbalance is possible by the same amount in both directions.
It is noted that both single-balanced and double-balanced mixers can be linearized by trimming in either both output branches or one of the branches
It is noted that the present invention is not restricted to the concrete circuit arrangements described in the preferred embodiments. The load adjustment may be performed by any control means or elements which are suitable to change or control load values in at least one of the respective output branches. In particular, the load adjustment may be realized by active elements such as bipolar or unipolar transistors, diodes or other semiconductor elements. Furthermore, the balanced circuit arrangement may be provided in IQ modulators, multi-carrier or multi-frequency receiver and/or transmitter systems. Furthermore, the calibration signals used for adjusting the load adjustment circuit can be obtained by using modulated carriers to allow for a greater flexibility of calibration. The measuring of the DC level or the lowpass filtered test signal may be performed solely by the DSP 8, such that the test monitoring circuitry 12 can be dispensed with. The above preferred embodiments may thus vary within the scope of the attached claims.
This is a Continuation of application Ser. No. 11/346,162 filed Feb. 3, 2006 now U.S. Pat. No. 7,236,761, which is a Continuation of application Ser. No. 10/474,848 filed Oct. 15, 2003 (now U.S. Pat. No. 7,031,687 B2, issued Apr. 18, 2006), which is a 371 of PCT/EP01/04414 filed Apr. 18, 2001. The disclosure of the prior application(s) is hereby incorporated by reference herein in its entirety.
Number | Name | Date | Kind |
---|---|---|---|
3519954 | Parkyn | Jul 1970 | A |
3727078 | Wollesen | Apr 1973 | A |
3911358 | Shalyt et al. | Oct 1975 | A |
4095180 | Brown | Jun 1978 | A |
4313089 | Predina | Jan 1982 | A |
4569084 | Takahama | Feb 1986 | A |
4636663 | Jongepier et al. | Jan 1987 | A |
4862485 | Guinea et al. | Aug 1989 | A |
4984293 | Cummings et al. | Jan 1991 | A |
5060298 | Waugh et al. | Oct 1991 | A |
5237332 | Estrick et al. | Aug 1993 | A |
5241702 | Dent | Aug 1993 | A |
5270824 | Dobrovolny | Dec 1993 | A |
5521545 | Terry et al. | May 1996 | A |
5524281 | Bradley et al. | Jun 1996 | A |
5548840 | Heck | Aug 1996 | A |
5584066 | Okanobu | Dec 1996 | A |
5590411 | Sroka et al. | Dec 1996 | A |
5630228 | Mittel | May 1997 | A |
5640161 | Johnson et al. | Jun 1997 | A |
5749047 | Cabot | May 1998 | A |
5749051 | Dent | May 1998 | A |
5825231 | Chevallier et al. | Oct 1998 | A |
5826182 | Gilbert | Oct 1998 | A |
5844449 | Abeno et al. | Dec 1998 | A |
5862466 | Erickson | Jan 1999 | A |
5872531 | Johnson et al. | Feb 1999 | A |
5896063 | Marsh et al. | Apr 1999 | A |
5933771 | Tiller et al. | Aug 1999 | A |
5940759 | Lopez-Torres et al. | Aug 1999 | A |
6006081 | Moore | Dec 1999 | A |
6009318 | Freed | Dec 1999 | A |
6032188 | Mairs et al. | Feb 2000 | A |
6054889 | Kobayashi | Apr 2000 | A |
6073002 | Peterson | Jun 2000 | A |
6078802 | Kobayashi | Jun 2000 | A |
6088581 | Bickley et al. | Jul 2000 | A |
6094571 | Groe | Jul 2000 | A |
6140849 | Trask | Oct 2000 | A |
6157822 | Bastani et al. | Dec 2000 | A |
6167247 | Kannell et al. | Dec 2000 | A |
6198423 | Yu | Mar 2001 | B1 |
6230001 | Wyse | May 2001 | B1 |
6242963 | Su et al. | Jun 2001 | B1 |
6307836 | Jones et al. | Oct 2001 | B1 |
6316996 | Puotiniemi | Nov 2001 | B1 |
6329864 | Suematsu et al. | Dec 2001 | B2 |
6347091 | Wallentin et al. | Feb 2002 | B1 |
6348830 | Rebeiz et al. | Feb 2002 | B1 |
6388605 | Petz et al. | May 2002 | B1 |
6393260 | Murtojarvi et al. | May 2002 | B1 |
6396330 | Fujii | May 2002 | B1 |
6411801 | Kim et al. | Jun 2002 | B1 |
6480720 | Coan et al. | Nov 2002 | B1 |
6483805 | Davies et al. | Nov 2002 | B1 |
6504819 | Fowler et al. | Jan 2003 | B2 |
6590439 | Larson | Jul 2003 | B1 |
6650883 | Stephane et al. | Nov 2003 | B1 |
6665527 | Schiltz | Dec 2003 | B2 |
6668175 | Almgren et al. | Dec 2003 | B1 |
6674733 | Huusko | Jan 2004 | B1 |
6710632 | Tada | Mar 2004 | B2 |
6711396 | Bergsma et al. | Mar 2004 | B1 |
6759904 | Behzad | Jul 2004 | B2 |
6794941 | Wichern et al. | Sep 2004 | B2 |
6807407 | Ji | Oct 2004 | B2 |
6892062 | Lee et al. | May 2005 | B2 |
6931253 | Hartikainen et al. | Aug 2005 | B1 |
7031687 | Kivekas et al. | Apr 2006 | B2 |
7120115 | Laaksonen | Oct 2006 | B1 |
7123630 | Ueno | Oct 2006 | B1 |
7587192 | Barre | Sep 2009 | B2 |
20020004376 | Lee et al. | Jan 2002 | A1 |
20020065061 | Schiltz | May 2002 | A1 |
20020160738 | Allott et al. | Oct 2002 | A1 |
20020193089 | Hatcher et al. | Dec 2002 | A1 |
20030211858 | Coan et al. | Nov 2003 | A1 |
20040131127 | Nadiri et al. | Jul 2004 | A1 |
20050013272 | Hwang et al. | Jan 2005 | A1 |
20050170806 | Kim | Aug 2005 | A1 |
20060128344 | Kivekas et al. | Jun 2006 | A1 |
20060261845 | Kim et al. | Nov 2006 | A1 |
Number | Date | Country |
---|---|---|
0 316 999 | May 1989 | EP |
0 951 138 | Oct 1999 | EP |
0 998 025 | May 2000 | EP |
2 794 869 | Jun 1999 | FR |
2 243 965 | Nov 1991 | GB |
2 346 777 | Aug 2000 | GB |
11-122046 | Apr 1999 | JP |
WO 9707596 | Feb 1997 | WO |
Number | Date | Country | |
---|---|---|---|
20070066269 A1 | Mar 2007 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 11346162 | Feb 2006 | US |
Child | 11600520 | US | |
Parent | 10474848 | US | |
Child | 11346162 | US |