Adjustable balancing circuit for an adaptive hybrid and method of adjusting the same

[0002] This invention relates generally to hybrid circuits, and more particularly to adaptive hybrid circuits for use with two-conductor transmission lines in full-duplex communications.

[0003] In full-duplex communications over a two-conductor transmission line, it is desirable to remove from a composite receive signal that portion due to reflection of a local transmit (Tx) signal, to thereby yield a corrected receive (Rx) signal which closely approximates the true signal broadcasted from a remote transmitter. Such correction is generally accomplished using a signal separator, or “hybrid,” and a subtractor.

[0004] FIG. 1 shows an abbreviated electrical block diagram of a telecommunications system 100, which includes a telecommunications apparatus 102 operative to communicate with a telecommunications apparatus 104 over a transmission line 106. Telecommunications apparatus 102, which may be a modem, typically includes a transceiver having a transmitter 108 and a receiver 110, a conventional hybrid circuit 112, a subtractor 114, and a line interface 116. Telecommunications apparatus 104 includes the same or similar components, at least a transceiver with a transmitter 118 and a receiver 120 coupled to a line interface 122.

[0005] Conventional hybrid circuit 112 typically includes a conventional (2-port) balancing network 124, a transmit source impedance (Z_S) 126, which is typically resistive, and a transmission line input impedance (Z_i). Circuitry of FIG. 1 is diagramed in an unbalanced (or ground-referenced) form for simplification, although balanced (or ungrounded) circuitry is frequently employed in practice.

[0006] Conventional balancing network 124 seeks to develop a voltage V_Bwhich is equal to, in magnitude and phase over the frequency range of interest, a reflected transmit signal component of voltage V_A. Thus, a subtractor 114 may yield a subtractor output voltage V_C=V_A−V_Bwhich is equal to the incoming receive signal. For this to work, the voltage transfer function V_B/V_Txof conventional balancing network 124 must be equal to that of the Z_S-to-Z_ivoltage divider (Z_i/Z_S+Z_i). This implies that conventional balancing network 124 must include elements that emulate the magnitude and phase variation behavior of Z_iover the frequency range of interest.

[0007] Typically, transmission line 106 is a twisted pair of insulated conductors whose impedance magnitude as a function of frequency behaves as shown in a graph 500 of FIG. 5. More particularly, graph 500 is representative of a twisted pair of insulated copper wires having AWG gauges 19 through 26, typically used in cables for telephony, Integrated Services Digital Network (ISDN), Digital Subscriber Lines (xDSL), and related communication formats. A line impedance curve 502 (shown in solid) represents a long (e.g., 10,000 feet (3048 meters)), unimpaired, terminated transmission line. A line impedance curve 504 (shown in dotted) represents a transmission line impaired by a long bridged tap located near the line's input. These bridged taps are often installed on the main lines in anticipation of line sharing and are typically inaccessible and of unknown length. The phase behavior of such transmission lines is closely and predictably related to the impedance-magnitude behavior and, for brevity, is not separately shown. Conventional hybrids have been designed to accommodate these monotonically-decreasing impedance versus frequency behaviors at low frequencies, such as voiceband frequencies (generally about 300-3000 hertz (Hz) and low data rate communication frequencies (tens of kilohertz (kHz) and below, as in modems operative at 14.4 through 56 kilobits per second (kbps)).

[0008] FIGS. 2, 3, and 4 show various conventional balancing networks 200, 300, and 400, respectively, for use in conventional hybrid circuit 112 of FIG. 1. Each of conventional balancing networks 200, 300, and 400 includes an impedance structure 202 and an impedance structure 204, with nodes 128 and 130 for coupling within conventional hybrid circuit 112 as indicated in FIG. 1.

[0009] Conventional balancing network 300 of FIG. 3 provides a fixed one-pole, one-zero transfer function, and has been proposed for 784 kbps 2B1Q High-data-rate Digital Subscriber Lines (HDSL) transceivers where the frequencies of primary interest extend to about 200 KHz. See “Generic Requirements for High-Bit-Rate Digital Subscriber Lines,” Bellcore Technical AdvisoryTA-NWT-001210, Issue 1, October 1991; and W. Y. Chen et al., “High Bit Rate Digital Subscriber Line Echo Cancellation,” IEEE Journal on Selected Areas in Communications, Vol. 9, No. 6, August 1991. Conventional balancing network 400 of FIG. 4 implies adjustability of these pole and zero locations, which is described in detail in U.S. Pat. No. 4,096,362 (Crawford). Conventional balancing network 400 may include a magnitude-scaling component to accommodate the presence of bridged-tap line impairments, since the low frequency impact of such impairments simply uniformly scales the impedance magnitude over frequency. Multiple-pole, multiple-zero networks are also available to provide an arbitrarily close match to a given monotonically-decreasing impedance characteristic. Analysis and measurements show that such conventional balancing networks, whether fixedly or adaptively configured, are somewhat useful for unimpaired lines or for impaired lines with long bridged taps.

[0010] When one or more short bridged taps are present near a transmission line input, however, a conventional hybrid with one of these conventional balancing networks is found to be highly inadequate—or even counterproductive. In this context, a short bridged tap is one whose length is between roughly one-sixteenth wavelength at the maximum frequency of interest and several thousand feet. For example, short tap lengths would range from about two hundred to several thousand feet for 784-kbps HDSL.

[0011] FIG. 6 is a diagram which illustrates such a transmission line environment. A bridged tap 604 is tapped between transmission line sections 602 and 606 of a main line terminated by an impedance 608. Here, bridged tap 604 has a length Y₂located at distance Y₁from the line's input. A distance Y_T=Y₁+Y₂is the total distance from the line input to the end of the tap. While only a single local tap is shown, multiple local taps may also be present in such an environment.

[0012] FIG. 7 is a graph 700 showing some line impedance versus frequency behaviors of the environment shown in FIG. 6. The numerical impedance values shown in graph 700 represent AWG #26 twisted-pair lines. A line impedance curve 702 (shown in solid) represents the behavior of a long unimpaired line; a line impedance curve 704 (shown in dotted) represents the behavior of a long line having a bridged tap of length 800 feet (244 meters) near the line's input; and a line impedance curve 706 (shown in dashed) represents a long line having a bridged tap of length 1600 feet (488 meters) near the line's input. These tap-impaired impedance behaviors are non-monotonic and oscillatory, and are not adequately emulated by conventional hybrids. The frequencies of impedance minima and maxima depend on tap length. Thus, a conventional hybrid designed for an 800-foot tap would badly mismatch a 1600-foot tap.

[0013] FIG. 8 is a graph 800 revealing a further complication in impedance matching. Line impedance curves 802, 804, and 806 of FIG. 8 each correspond to a combined distance of 800 feet (244 meters) from the input to the end of the tap (distance Y_Tin the preceding discussion), but with different tap lengths and locations: line impedance curve 802 (shown in solid) represents an 800-foot (244-meter) tap at the input; line impedance curve 804 (shown in dotted) represents a 600-foot (183-meter) tap located 200 feet (61 meters) from the input; and line impedance curve 806 (shown in dashed) represents a 400-foot (122-meter) tap located 400 feet (122 meters) from the input. While the frequencies of the impedance extremes are essentially the same, the extreme impedance values are not. Further analysis and measurements have shown this to be generally true for all bridged tap lengths and locations of interest.

[0014] Accordingly, what is needed is a hybrid to accommodate these and similar problems.

[0015] The nature, objects, and advantages of the invention will become more apparent to those skilled in the art after considering the following detailed description in connection with the accompanying drawings, in which like reference numerals designate like parts throughout, and wherein:

[0016] FIG. 1 is an abbreviated electrical block diagram of a telecommunication system, which includes a telecommunications apparatus having a conventional hybrid circuit.

[0017] FIG. 2 is a first conventional balancing network of the conventional hybrid circuit of FIG. 1.

[0018] FIG. 3 is a second conventional balancing network of the conventional hybrid circuit of FIG. 1.

[0019] FIG. 4 is a third conventional balancing network of the conventional hybrid circuit of FIG. 1.

[0020] FIG. 5 is a graph showing line impedance versus frequency behaviors, which may be matched by the conventional hybrid circuit of FIG. 1.

[0021] FIG. 6 is an electrical diagram showing a transmission line impaired by a short bridged tap.

[0022] FIG. 7 is a graph showing line impedance versus frequency behaviors as a result of short bridged taps (such as that shown in FIG. 6).

[0023] FIG. 8 is a graph showing additional line impedance versus frequency behaviors as a result of short bridged taps (such as that shown in FIG. 6).

[0024] FIG. 9 is a schematic diagram of an adjustable balancing circuit of an adaptive hybrid, shown in an unbalanced form.

[0025] FIG. 10 is a schematic diagram of circuitry of an adjustable balancing circuit of an adaptive hybrid in a balanced form.

[0026] FIG. 11 is a schematic diagram of circuitry of an adjustable balancing circuit of an adaptive hybrid in another balanced form.

[0027] FIG. 12 is a schematic block diagram of a telecommunications system, which includes a telecommunications apparatus utilizing the adaptive hybrid.

[0028] FIG. 13 is a flowchart of a method of adjusting a signal transfer function the adaptive hybrid.

[0029] FIG. 14 is a graph showing impedance versus frequency behaviors of the adaptive hybrid.

[0030] FIG. 15 is a graph comparing cancellation effectiveness (CE) versus frequency behaviors of the adaptive hybrid and a conventional hybrid.

[0031] FIG. 16 is a schematic diagram of an first extended embodiment of an adjustable balancing circuit of an adaptive hybrid.

[0032] FIG. 17 is a schematic diagram of circuitry of a second extended embodiment of an adjustable balancing circuit of an adaptive hybrid.

[0033] The limitations of conventional hybrid circuits for bidirectional high frequency analog or high data rate digital communications over two-conductor transmission lines are overcome by using the principles of the invention to be described herein. In a preferred example, an adjustable balancing circuit of an adaptive hybrid includes a first impedance structure, a second impedance structure, a plurality of N resistor-inductor-capacitor (RLC) networks, a first plurality of N switch circuits, and a second plurality of N switch circuits. The first plurality of N switch circuits switchably cascadingly couples the first plurality of N RLC networks to a first end of the first impedance structure. The second plurality of N switch circuits switchably couples a first end of the second impedance structure to a first end of each RLC network. When properly configured, the adjustable balancing circuit helps generate signals which closely match reflected transmit signals caused by the presence of short bridged taps.

[0034] FIG. 9 is a schematic diagram of an adjustable balancing circuit 900 of an adaptive hybrid, which is suitable for use with a two-conductor transmission line in full-duplex communications. Adjustable balancing circuit 900 includes an impedance structure 902, an impedance structure 904, and additional circuitry 906 which includes a plurality of N RLC networks 908 (where N is a natural number greater than or equal to two) and switching circuitry 910. Adjustable balancing circuit 900 is shown in an unbalanced or grounded form, although balanced or ungrounded forms are understood and described below in relation to FIGS. 10 and 11.

[0035] FIG. 9 shows the plurality of N RLC networks 908 where N is any natural number greater than or equal to two, but also the specific case where N=3. Here, the plurality of N RLC networks 908 includes an RLC network 916, an RLC network 918, and an RLC network 920. Each RLC network of the plurality of N RLC networks 908 includes at least one resistive element (such as a resistor), at least one inductive element (such as an inductor), and at least one capacitive element (such as a capacitor). For example, RLC network 916 includes an inductor 922, a resistor 924, and a capacitor 926 coupled as shown.

[0036] Preferably, the resistive, inductive, and capacitive elements are simple, lumped, passive components. Also preferably, the components and component values of each RLC network are substantially the same as the components and the component values of each other RLC network. Each RLC network preferably has an impedance that is substantially the same as an impedance of each other RLC network.

[0037] Each of impedance structures 902 and 904 may simply be or include a resistive element, such as a resistor. Alternatively, impedance structures 902 and 904 may be or include impedance structures 202 and 204, respectively, as described and shown in relation to FIGS. 2-4.

[0038] As shown in FIG. 9, switching circuitry 910 switchably cascadingly couples the plurality of N RLC networks to a first end of impedance structure 902. Switching circuitry 910 also switchably couples a first end of impedance structure 904 to a first end of each one of RLC networks 916, 918, and 920. As indicated in FIG. 9, the first end of impedance structure 902 is coupled to a node 903 and a second end of impedance structure 902 is coupled to a node 901, where adjustable balancing circuit 900 is suitable for use in an adaptive hybrid circuit 1216 of FIG. 12 as described further below.

[0039] Here, switching circuitry 910 includes a plurality of N switch circuits 912 and a plurality of N switch circuits 914. In the embodiment shown, the plurality of N switch circuits 912 includes three switch circuits, switch circuits 928, 930, and 932, and the plurality of N switch circuits 914 includes three switch circuits, switch circuits 934, 936, and 938. Each switch circuit of switching circuitry 910 may be or include any suitable switch or switches, such as a solid-state switch or a relay switch.

[0040] As shown in FIG. 9, switch circuit 928 switchably couples the first end of RLC network 916 to the first end of impedance structure 902, switch circuit 930 switchably couples the first end of RLC network 918 to a second end of RLC network 916, and switch circuit 932 switchably couples the first end of RLC network 920 to a second end of RLC network 918. In addition, switch circuit 934 switchably couples the first end of impedance structure 904 to the first end of RLC network 916, switch circuit 936 switchably couples the first end of impedance structure 904 to the first end of RLC network 918, and switch circuit 938 switchably couples the first end of impedance structure 904 to the first end of RLC network 920. For N>3, additional RLC networks and switch circuits may be coupled in the same or similar cascading fashion. It is understood and shown that switch circuits 934, 936, and 938 may switchably couple the first end of impedance structure 904 to the first end of each RLC network through switch circuits 928, 930, 932.

[0041] Adjustable balancing circuit 900 preferably has at least two RLC networks that are switchably cascadingly coupled. For example, adjustable balancing circuit 900 at least includes impedance structure 902, RLC network 916 having its first end switchably coupled to the first end of impedance structure 902 (whether switchably coupled directly thereto or through a preceding RLC network), RLC network 918 having its first end switchably coupled to a second end of RLC network 916, and impedance structure 904 having its first end switchably coupled to the first ends of RLC networks 916 and 918. Here, at least four useful switch configurations are available (as will be described below in more detail) to provide at least four unique adjustments of adjustable balancing circuit 900.

[0042] FIGS. 10 and 11 show switchably coupled RLC networks in balanced or ungrounded forms (unlike that shown in FIG. 9), both of which may be utilized in adjustable balancing circuit 900 of FIG. 9. FIG. 10 shows more particularly a plurality of N RLC networks 1002 switchably cascadingly coupled through a plurality of N switch circuits 1004. In this embodiment, three RLC networks are shown, RLC networks 1006, 1008, and 1010. Each one of the plurality of RLC networks 1002 here includes at least two inductive elements, at least two resistive elements, and at least one capacitive element. For example, RLC network 1006 includes inductors 1012 and 1014, resistors 1016 and 1018, and a capacitor 1020 coupled as shown. The plurality of N switch circuits 1004 include switch circuits 1022, 1024, and 1026.

[0043] In FIG. 11, a plurality of N RLC networks 1102 is switchably cascadingly coupled through a plurality of switch circuits 1104 in a quasi-balanced form. Here, as in FIG. 10, three RLC networks are shown, RLC networks 1106, 1108, and 1110. Each one of the plurality of RLC networks 1102 includes at least one inductive element, at least one resistive element, and at least one capacitor. For example, RLC network 1106 includes an inductor 1112, a resistor 1114, and a capacitor 1116 coupled as shown. The plurality of switch circuits 1104 include switch circuits 1118, 1120, and 1122 switchably coupling RLC networks 1106, 1108, and 1110 as shown. Here, the series RL elements are alternately disposed in upper and lower lines of the cascaded sections.

[0044] As described above in relation to FIG. 9, the resistive, inductive, and capacitive elements of FIGS. 10 and 11 are preferably simple, lumped, passive components. Also preferably, the components and component values of each RLC network are substantially the same as the components and the component values of each other RLC network. Each RLC network preferably has an impedance that is substantially the same as an impedance of each other RLC network.

[0045] Referring back to FIG. 9, switching circuitry 910 is set to a particular switch configuration by control circuitry (shown and described below in relation to FIG. 12) during an adjustment phase of adjustable balancing circuit 900, preferably using a method shown and described below in relation to FIG. 13. The particular switch configuration determined during the adjustment phase should remain fixed during subsequent telecommunications operation, so long as operating conditions (e.g., operating frequency, tap location, etc.) remain substantially the same. Preferably, in response to a change in conditions, another adjustment phase occurs where the switch configuration may change.

[0046] Here, adjustable balancing circuit 900 has a number of useful switch configurations and, more particularly, 1+N(N+1)/2 useful switch configurations. Rules for determining these useful switch configurations of adjustable balancing circuit 900 include: (1) only one of switch circuits 914 may be closed at a time; (2) 0 through N of switch circuits 912 may be closed at a time; (3) if one of switch circuits 912 is closed, then those of switch circuits 912 that precede it must also be closed (i.e., switch circuits 912 are closed such that RLC networks 908 cascadingly connect starting at RLC network 916); and (4) an nth switch circuit of switch circuits 914 may be closed only if the corresponding nth switch circuit of switch circuits 912 is also closed.

[0047] As an example, useful switch configurations where N=2 include those described in TABLE 1 below.

[0048] As another example, useful switch configurations where N=3 those described in TABLE 2 below.

[0049] For N>3, the useful switch configurations are similarly limited as described and are readily apparent to those skilled in the art.

[0050] FIG. 12 is a schematic block diagram of a telecommunications system 1200. Telecommunications system 1200 includes a telecommunications apparatus 1202 operative to communicate with a telecommunications apparatus 1204. Most importantly, telecommunications apparatus 1202 has an adaptive hybrid circuit 1216 that includes adjustable balancing circuit 900 of FIG. 9 for full-duplex communications over a transmission line 1206 which may be impaired by a short bridged tap.

[0051] Telecommunications apparatus 1202, which is a modem or similar apparatus, includes a transceiver having a transmitter 1208 and a receiver 1210, a processor 1214, a subtractor 1218, a line interface 1220, and adaptive hybrid circuit 1216. In this embodiment, adaptive hybrid circuit 1216 includes adjustable balancing circuit 900, an impedance structure 1222 (Z_S), control circuitry 1224, and measuring circuitry 1228. Preferably, control circuitry 1224 has access to a memory 1230 for storage of data.

[0052] Transmitter 1208, receiver 1210, processor 1214, subtractor 1218, and line interface 1220 may be, for example, typical components found in modems. The transceiver is operative for communications in accordance with xDSL or HDSL standards. Control circuitry 1224 may be or include common switch drivers to drive switching circuitry 910. Preferably, control circuitry 1224 may be or include common suitable software-driven circuitry, such as a microcontroller or microprocessor. Control circuitry 1224 and memory 1230 may include separate components or may be combined as a single component. Also preferably, some or all components of control circuitry 1224, memory 1230, and subtractor 1210 are pre-existing components of telecommunications apparatus 1202 which adaptive hybrid circuit 1214 reuses. As examples, subtractor 1218 may be included in a pre-existing subtractor of telecommunications apparatus 1202; control circuitry 1224 and memory 1230 may be included in (pre-existing) processor 1214; and measuring circuitry 1228 may be included in (pre-existing) receiver 1210.

[0053] As shown in FIG. 12, transmitter 1208 has a transmitter output coupled to line interface 1220 through impedance structure 1222 and to adjustable balancing circuit 900 (at node 901 or the second end of impedance structure 902 of FIG. 9). Subtractor 1218 has a subtractor input coupled to line interface 1220 and another subtractor input coupled to adjustable balancing circuit 900 (at node 903 or the first end of impedance structure 902 of FIG. 9). Subtractor 1218 also has a subtractor output coupled to a receiver input of receiver 1210 and to a measuring input of measuring circuitry 1228. Measuring circuitry 1228 is coupled to an input of control circuitry 1224, which is coupled to adjustable balancing circuit 900. More particularly, control circuitry 1224 is coupled to switching circuitry 910 (FIG. 9) through switch control lines 1226 and is operative to configure switching circuitry 910 in various open and closed positions.

[0054] Preferable general telecommunications operation of telecommunications apparatus 1202 is described. During an adjustment or calibration phase of adaptive hybrid circuit 1216, control circuitry 1224 configures adjustable balancing circuit 900 so that an impedance that most closely matches the input impedance Z_iis created. To do this, control circuitry 1224 sets switching circuitry 910 (FIG. 9) to the useful switch configurations as described above.

[0055] During full-duplex telecommunications operation, transmitter 1208 and receiver 1210 (as well as a transmitter and receiver of telecommunications apparatus 1204) are active and operating. At one subtractor input, subtractor 1218 receives a combined signal (V_A) which includes a receive signal (broadcasted from telecommunications apparatus 1204) undesirably mixed with a reflected transmit signal. At the other subtractor input, subtractor 1218 receives a local transmit signal (V_B) modified by adaptive hybrid circuit 1216 to closely resemble the reflected transmit signal in magnitude and phase. Subtractor 1218 generates the difference between the combined signal and the modified local transmit signal in an attempt to eliminate the reflected transmit signal and generate a corrected receive signal at its subtractor output. The corrected receive signal is fed into the receiver input of receiver 1210 for reception.

[0056] Referring back to FIG. 9, a set of fixed component values for the plurality of N RLC networks 908 is provided for illustrative and comparative purposes. In this example, R=19.44 ohms, L=34.8 microhenries, C=3.14 nanofarads, and Z_b′=110 ohms. These RLC values are optimized to emulate line impedance versus frequency behavior for an AWG #26 twisted-pair line, with a resolution of 200 feet (61 meters) of line length. To emulate line impedance over a range of wire gauges, the L and C values may remain fixed while a compromise R value is chosen. In a practical application, the component values would likely be impedance-scaled in consideration of the non-zero output impedance of a transmit line driver and the finite input impedance of the subtractor.

[0057] FIG. 14 shows the magnitude of input impedance Z_i′ (that of the Z_b′+cascade network) versus frequency using the component values of the preceding paragraph, where N≧4. Impedance curves 1402, 1404, and 1406 of FIG. 14 were obtained by configuring the adaptive hybrid to switch configurations 1, 2, and 3, respectively, shown in TABLE 3 below. Switch configurations 1, 2, and 3 are a subset of eleven useful switch configurations of the adaptive hybrid having four RLC networks. Switch circuits Sx1 correspond to upper switch circuits (e.g., switch circuits 928, 930, 932, etc. of FIG. 9), and switch circuits Sx2 correspond to lower switch circuits (e.g., switch circuits 934, 936, 938, etc. of FIG. 9) of the adaptive hybrid.

[0058] Comparison of impedance curves 1402, 1404, and 1406 with line impedance curves 802, 804, and 806 respectively of FIG. 8 shows good emulation of the three short bridged-tap-impaired line impedance behaviors.

[0059] FIG. 15 compares typical cancellation effectiveness (CE) versus frequency results of adaptive hybrid circuit 1216 and a conventional hybrid in the presence of a short bridged tap near the transmission line's input. A hybrid's CE, in decibels, is defined in terms of voltages V_Aand V_Bdeveloped in the hybrid as CE=−20 log [magnitude of (1−V_B/V_A)]. This ultimate performance parameter is also referred to in the art as transhybrid loss or echo return loss. Here, adaptive hybrid circuit 1216 uses the values and switch configuration 1 of TABLE 3 above, with Z_Sand Z_a′ both being resistive and equal to 130 ohms. The conventional hybrid uses balancing network 300 of FIG. 3 with R_A=130 ohms, R_B=2500 ohms, R_C=115 ohms, and C_A=8.7 nanofarads, a representative set of values for a conventional fixed hybrid. Here, the transmission line is impaired by an 800-foot (244-meter) bridged tap at its input, as in the remote-terminal end of HDSL Carrier Servicing Area (CSA) Type Test Loop #7 of Bellcore Technical AdvisoryTA-NWT-001210, supra. As shown, adaptive hybrid circuit 1216 provides some 15 dB of improvement in CE over the frequency range shown. In a digital communications transceiver design, this would reduce the dynamic-range requirement of the modem's analog-to-digital (A/D) converter by 2.5 bits and/or enable communication over a substantially longer transmission line.

[0060] FIG. 13 is a flowchart of a method of adjusting or calibrating a signal transfer function of adaptive hybrid circuit 1216 for telecommunications operation. FIGS. 9 and 12 will be referred to in combination with FIG. 13. The calibration process may commence with instructing signals from processor 1214 to transmitter 1208 and control circuitry 1224. Starting at a step 1300, transmitter 1208 generates a transmit signal (step 1302) which is available at its transmitter output. Preferably, the transmit signal is a pulse sequence chosen to represent, in a suitable-weighted fashion, the frequency range of interest or operation. Control circuitry 1224 sets switching circuitry 910 of adjustable balancing circuit 900 to one of the 1+N(N+1)/2 switch configurations described above (step 1304).

[0061] When the transmit signal is being generated and switching circuitry 910 is in one of the switch configurations, subtractor 1218 generates a difference signal between signals at its subtractor inputs (step 1306). This difference signal is available at a subtractor output of subtractor 1218 and is fed into measuring circuitry 1228. Measuring circuitry 1228 generates a performance indication based on this difference signal over a time period (step 1308). Measuring circuitry 1228 may obtain such a performance indication, for example, by rectifying and integrating the difference signal over the time period using a rectifier and an integrator (both not shown), respectively, or by using many other suitable measurement methods and means known in the art.

[0062] If, based on this performance indication, it is unnecessary to configure switching circuitry 910 to other switch configurations (step 1310), control circuitry 1224 selects this switch configuration as an appropriate switch configuration for telecommunications operation (step 1312). This will occur, for example, if the performance indication meets or exceeds a threshold performance indication, or if the performance indication is better than a previously obtained performance indication or all other previously obtained performance indications. Thus, at step 1310, control circuitry 1224 may compare the performance indication to a threshold performance indication, or may compare the performance indication to a previously obtained performance indication (e.g., the best previously obtained performance indication) or indications. Using these techniques, control circuitry 1224 may store and read digitized performance indications into and from memory 1230.

[0063] If other switch configurations are to be tested at step 1310, the flowchart repeats steps 1304 through 1308, where a different one of the 1+N(N+1)/2 switch configurations is set in step 1304. Preferably, at step 1312, control circuitry 1224 selects the switch configuration that provides the best performance indication for telecommunications operation.

[0064] As shown in FIG. 16, such an adjustable balancing circuit may readily be extended to accurately emulate the impedance behavior of a line having more than one bridged-tap impairment by nestingly cascadingly connecting these structures. For example, an adjustable balancing circuit 1600 includes an impedance structure 1602, an impedance structure 1604, and two switchably cascadingly coupled RLC networks, networks 1606 and 1608. As shown, each of networks 1606 and 1608 may be the same as that shown and described in relation to FIG. 9. Such an arrangement is optimal for accommodation of two local impairments, where there are [1+M(M+1)/2][1+N(N+1)/2] sets of useful switch settings.

[0065] As shown in FIG. 17, an adjustable balancing circuit 1700 may have improved resolution (with respect to line length emulation) by adding one or more non-identical switchable sections (as shown in circuitry 1702) to switchably cascadingly coupled RLC networks (as shown in circuitry 1704). Here, a plurality of P switchable binary-weighted (immittance-tapered) sections 1706 are added immediately preceding a first RLC network of a plurality of N RLC networks 1714 switchably cascadingly coupled with a plurality of N switch circuits 1720. Each of the P switchable binary-weighted sections 1706 may include electrical components having relative component values and coupling as shown in FIG. 17. Each of the P switchable binary-weighted sections 1706 may or may not be bypassed, depending on the setting of its associated single-pole, double-throw (SPDT) switch circuit. For binary-weighted component values, a factor of 2^Pimprovement in line-length resolution may be obtained via adaptive configuration of adjustable balancing circuit 1700. For example, where P=2 (a first section 1708 with component values L/2, R/2, and C/2, a second section 1710 with component values L/4, R/4, and C/4, preceding the switchably cascadingly coupled plurality of N RLC networks 1714 with component values R, L, and C), the line-length resolution may be improved by a factor of four. A combination of P tapered sections and N identical sections may be employed to require fewer circuit components and adjustment steps (for emulation of a given maximum line length with a given length resolution) than would be required if only identical networks were employed.

[0066] Thus, an adaptive hybrid circuit shown and described herein enhances full-duplex communication over a two-conductor transmission line. The adaptive hybrid circuit adaptively corrects a composite receive signal which is distorted by transmit signal reflections caused by the presence of bridged taps of unknown lengths at inaccessible locations on the transmission line. The adaptive hybrid circuit may employ direct cascading of a determined number of identical, fixed, passive, 2-port networks. The cascading process inherently accommodates the impact of bridged tap length and location on the line's impedance versus frequency behavior. In a digital communications application, such as a modem application, a configuration process of the adaptive hybrid circuit may utilize pre-existing processor and control circuitry.

[0067] Such an adjustable balancing circuit is uniquely amenable to hardware implementation, in that a plurality of fixed RLC networks are simply cascaded together and a switch configuration selected during the adaptive configuration process. No component values need be adjusted; impedance-versus-frequency behavior is adjusted via the switchable cascading process.

[0068] While the adaptively-adjustable, passive, RLC balancing-network structure is the preferred embodiment of the invention, it will be apparent to skilled practitioners of the art that modifications may be made thereto without departing from the spirit or scope of the appended claims. For example, one might adaptively switch-select one of a plurality of fixed passive RLC structures, each of which emulates a fixed bridged-tap length and location [by producing a non-monotonic or oscillatory voltage transfer ratio versus frequency characteristic], for developing a compensating voltage. More specifically, referring to FIG. 9, a first RLC structure may include RLC network 916, a second RLC structure may include RLC networks 916 and 918 cascadedly connected, a third RLC structure may include RLC networks 916, 918, and 920 cascadedly connected, etc. As another example, one might fixedly couple impedance structure 904 of FIG. 9 to a first end of one of RLC networks 908 (and eliminate switch circuits 934, 936, and 938) to reduce the number of switch configurations.