阅读说明：本技术 使用变容二极管的具有可调节高频增益的放大器 (Amplifier with adjustable high frequency gain using varactor diodes ) 是由 苏哈斯·拉坦 凯拉什·加里杜 于 2019-06-10 设计创作，主要内容包括：一种装置,包括：差分对(100),用于接收差分输入信号,并且在一对输出节点上响应生成放大输出信号；以及可调节电容,用于调节施加至所述差分输入信号上的高频放大作用,该可调节电容包括：与所述差分对连接的一对串联连接变容二极管,用于接收共用节点上的控制信号,以调节所述高频放大作用的频率；以及与所述一对串联连接变容二极管(133,134)并联的固定电容器(132),该固定电容器(132)的电容大于所述串联连接变容二极管(133,134)的电容,并用于抵消所述串联连接变容二极管(133,134)的沟道电荷密度的变化。(An apparatus, comprising: a differential pair (100) for receiving a differential input signal and generating an amplified output signal in response on a pair of output nodes; and an adjustable capacitance for adjusting a high frequency amplification applied to the differential input signal, the adjustable capacitance comprising: a pair of series connected varactors coupled to the differential pair for receiving a control signal on a common node to adjust the frequency of the high frequency amplification; and a fixed capacitor (132) connected in parallel with the pair of series-connected varactors (133,134), the fixed capacitor (132) having a capacitance greater than a capacitance of the series-connected varactors (133,134) and being configured to counteract a change in channel charge density of the series-connected varactors (133, 134).)

1. An apparatus, comprising:

a differential pair for receiving a differential input signal and generating an amplified output signal in response on a pair of output nodes;

an adjustable capacitance for adjusting a high frequency amplification applied to the differential input signal, wherein the adjustable capacitance comprises:

a pair of series connected varactors coupled to the differential pair for receiving a control signal on a common node to adjust the frequency of the high frequency amplification; and

a fixed capacitor in parallel with the pair of series-connected varactors, wherein a capacitance of the fixed capacitor is greater than a capacitance of the series-connected varactors and is to counteract a change in channel charge density of the series-connected varactors.

2. The apparatus of claim 1, further comprising an isolation transistor connected with the pair of output nodes.

3. The apparatus of claim 1, further comprising an adjustable load impedance connected with the pair of output nodes.

4. The apparatus of claim 3, the adjustable load impedance comprising a switch controlled resistor bank, wherein the switch controlled resistor bank is to receive a load impedance control signal and to selectively enable one or more resistors within the switch controlled resistor bank.

5. The apparatus of claim 1, further comprising a source impedance connected in parallel with the adjustable capacitance.

6. The apparatus of claim 5, wherein the source impedance is an adjustable source impedance.

7. The apparatus of claim 6, the adjustable source impedance comprising a switch controlled resistor bank, wherein the switch controlled resistor bank is to receive a source impedance control signal and to selectively enable one or more resistors within the switch controlled resistor bank.

8. The apparatus of claim 5, the source impedance comprising a transistor biased by a source impedance control signal, wherein the source impedance control signal is to selectively set the transistor in an open or shorted configuration.

9. The apparatus of claim 5, the source impedance comprising a plurality of transistors connected in parallel via a plurality of parallel circuits, wherein each transistor is biased by a respective one of a set of source impedance control signals.

10. The apparatus of claim 9, wherein each of the plurality of parallel circuits comprises an adjustable capacitance.

11. The apparatus of claim 10, wherein the adjustable capacitance within each of the plurality of parallel circuits is controlled by a respective control signal.

12. The apparatus of claim 9, wherein the set of source impedance control signals comprises a plurality of bits, wherein each bit is to fully enable or disable the biased transistor.

13. The apparatus of claim 9, the set of source impedance control signals being analog control signals for biasing the transistor in a linear region.

14. The apparatus of claim 13, wherein the analog control signal is used to select an equalization range.

15. The apparatus of claim 13, further comprising a digital-to-analog converter to generate the analog control signal.

16. The apparatus of claim 15, wherein the digital-to-analog converter is a resistive ladder digital-to-analog converter.

17. The apparatus of claim 1, further comprising a digital-to-analog converter to generate the control signal.

18. The apparatus of claim 17, wherein the digital-to-analog converter is a metal oxide silicon ladder digital-to-analog converter.

19. The apparatus of claim 17, wherein the digital-to-analog converter comprises transmission logic gates arranged in an R-2R ladder.

20. The apparatus of claim 1, further comprising an adjustable current source for adjusting the amplitude of the high frequency amplification.

21. A method, comprising:

receiving a differential input signal by the differential pair and generating an amplified output signal in response on a pair of output nodes;

adjusting, by an adjustable capacitance, a high frequency amplification applied to the differential input signal, wherein adjusting the high frequency amplification comprises:

receiving a control signal on a common node by a pair of series connected varactors connected to the differential pair, wherein the pair of series connected varactors adjusts a frequency of the high frequency amplification; and

variations in channel charge density of the series varactor are offset by a fixed capacitor in parallel with the pair of series connected varactors, wherein a capacitance of the fixed capacitor is greater than a capacitance of the series connected varactors.

22. The method of claim 21, wherein an adjustable load impedance is connected with the pair of output nodes, wherein the method further comprises: a load impedance control signal is received to selectively enable one or more resistors within a switch controlled resistor bank.

23. The method of claim 21, further comprising: a source impedance control signal is received by a source impedance connected in parallel with the adjustable capacitance.

24. The method of claim 23, wherein the source impedance control signal selectively enables one or more resistors within a switch controlled resistor bank of the source impedance.

25. The method of claim 23, wherein the source impedance control signal selectively sets the transistors in an open or shorted configuration by biasing the transistors.

26. The method of claim 23, wherein the source impedance comprises a plurality of transistors connected in parallel via a plurality of parallel circuits, wherein each transistor is biased by a respective one of the portions of the source impedance control signal.

27. The method of claim 26, wherein each of the plurality of parallel circuits comprises an adjustable capacitance.

28. The method of claim 27, wherein the adjustable capacitance within each of the plurality of parallel circuits is controlled by a respective control signal.

29. The method of claim 26, wherein the source impedance control signal comprises a plurality of bits, wherein each bit is to fully enable or disable a biased transistor.

30. The method of claim 26, wherein the source impedance control signal comprises a set of analog control signals for biasing the transistor in a linear region.

31. The method of claim 30, wherein the set of analog control signals is used to select an equalization range.

32. The method of claim 30, further comprising: the set of analog control signals is generated by a digital-to-analog converter.

33. The method of claim 32, wherein the digital-to-analog converter is a resistive ladder digital-to-analog converter.

34. The method of claim 21, further comprising: the control signal is generated by a digital-to-analog converter.

35. The method of claim 34, wherein the digital-to-analog converter is a metal oxide silicon ladder digital-to-analog converter.

36. The method of claim 34, wherein the digital-to-analog converter includes transmission logic gates arranged in an R-2R ladder.

37. The method of claim 21, further comprising: the amplitude of the high frequency amplification is adjusted by an adjustable current source.

Disclosure of Invention

An amplifier circuit with configurable frequency compensation functionality is described that is suitable for use as a continuous-time linear equalizer (CTLE) for a communications receiver input signal. Among the elements of the design, the configurable DAC elements are tightly integrated with the analog devices they control, thereby facilitating compactness of the circuit layout.

Drawings

Fig. 1 is a circuit diagram of a CTLE amplifier employing NMOS transistors according to some embodiments.

Fig. 2 is a circuit diagram illustrating how multiple instances of the embodiment, as shown in fig. 1, may be combined to provide greater configurability.

Fig. 3A is a circuit diagram of a digital-to-analog converter using PMOS transistors suitable for generating analog control signals according to some embodiments.

Fig. 3B is a circuit diagram of a transmission logic gate implementation including the transistors shown in fig. 3A.

Fig. 4 shows an alternative embodiment of the circuit of fig. 1 using PMOS transistors.

Fig. 5 is a spectrum of an adjustable high frequency peaking CTLE according to some embodiments.

Fig. 6 is a frequency spectrum illustrating CTLE gain with or without a fixed capacitor connected in parallel with a variable capacitor, according to some embodiments.

Fig. 7 is a block diagram of a digital-to-analog converter (DAC) circuit that generates a control signal for setting an amplifier equalization gain range according to some embodiments.

Fig. 8 is an illustration of equalization gain ranges for different coding schemes according to some embodiments.

Detailed Description

Continuous Time Linear Equalization (CTLE) circuits are well known in the art. One common design is based on the use of a pair of matched transistors with their source loads sharing a common drain connection to a fixed current sink. By dividing the current sink into two parts and respectively corresponding to different transistor drains, the drains can be cross-coupled with frequency-dependent impedance elements such as a parallel RC network, so that the originally basically flat gain-frequency characteristic of the basic differential amplifier is changed into the characteristic with completely different high and low frequency gains.

In communication system receivers, such CTLE circuits are typically used to provide greater high frequency gain to equalize or compensate for the high frequency loss inevitable for most communication media. In some embodiments, careful configuration of the amplitude and equalization functions is required to facilitate accuracy of subsequent circuit signal detection and/or clock recovery. In some embodiments, both the gain characteristic and the frequency break of the above-described frequency-dependent compensation of a CTLE circuit allow for adjustment and configuration.

Fig. 1 shows an embodiment of a configurable CTLE circuit using NMOS transistors as gain elements. Fig. 4 shows an equivalent embodiment using PMOS transistors. The following description will refer to the circuit of fig. 1, but this is not meant to be limiting.

It is noted that such embodiments are intended for use in an integrated circuit environment where very high frequency signals need to be processed with minimal power consumption. The available power rails Vdd and Vss typically provide an operating voltage of 1 volt or less, and thus microampere-level currents represent thousands or even millions of ohms of impedance in the path. Since resistors of this magnitude may require significant surface area in some integrated circuit processes, active circuit elements such as transistors may be preferred over passive circuit elements.

In FIG. 1, the inductor L₀And L₁The formed inductive load is subjected to parallel peaking at high frequency, and the frequency peaking amount of more than 10dB is realized. The corresponding load impedance is generated by generating a load impedance R_L0And R_L1Is provided. As shown, each resistor array may receive a multi-bit thermometer code load resistance control input R_L<n-1:0>And optionally connecting successive resistors in the parallel network to set the load resistance RL. In some embodiments, the adjustment of the load resistance may facilitate adjustment of the common mode of subsequent processing stages (e.g., variable gain amplifiers). Alternatively, the amount of high frequency peaking and the location of the responsive second pole may also be adjusted by adjusting the load resistance, as will be described in further detail below with reference to FIG. 5. As shown in FIG. 1, circuit portion 100 includes transistors 112 and 122 as matched differential pairs for receiving input signals Vin + and Vin-, respectively, and generating output signals Vout-and Vout +, respectively. The cascode transistors 111 and 121 may be used to receive the gate voltage of Vcasc and may help reduce the input miller capacitance of the input terminal by its inherent properties while isolating the input terminal from the output terminal. Identical current sources 113 and 123 are used to set the allowable current and, as described below, through adjustment of these two current sources, adjustment of the amount of high frequency peaking can be achieved. The source resistance Rs provided by the transistor 131 and the capacitances provided by the fixed capacitor 132 and the variable capacitors 133 and 134 in combination determine the frequency-dependent gain characteristic of the equalizer.

In some embodiments, the voltage Vsw allows configuration to achieve impedance adjustment of the transistor 131. In other embodiments, the voltage Vsw is a fixed voltage, and the voltage and the physical size of the transistor channel together determine the final impedance.

In another embodiment, the voltage Vsw may be set to one of two different preset values (i.e., binary selection), as described below. In one such embodiment, when Vsw causes transistor 131 to turn on (e.g., low impedance), circuit 100 enters a first ("flat") mode of operation that minimizes the frequency domain nulls generated by capacitors 132, 133,134 and reduces the DC equalization and peak equalization correlation. Conversely, when Vsw turns off transistor 131 (high impedance), this impedance, in conjunction with capacitors 132, 133,134, causes circuit 100 to enter a second (high frequency "peaking") mode of operation that increases the correlation of DC equalization and peak equalization.

Fig. 5 illustrates a frequency response condition of a CTLE providing high frequency peaking according to some embodiments. Fig. 5 includes four points of interest: first zero point A (w)_zero) (ii) a First pole B (w)_p0) (ii) a Peak frequency C (w)_peak) (ii) a Second pole D (w)_p1). The value of each point of interest is shown in the following formulas 1 to 4:

where Rs is the source resistance provided by transistor 131 and Cs is the source capacitance. The source capacitance may be equal to C_Dom+C_vc0/1Wherein, C_DomTo fix the dominant capacitance of the capacitor 132, C_vc0/1Is the capacitance of one of the variable capacitance diodes 133 or 134. Here and hereinafter, consistent with common practice in half-circuit analysis, the designator "0/1" may correspond to a value of "0" or "1" for an element in the group of 0, 1. In at least one embodiment, C_DomHas a total capacitance of about 100fF (e.g., 100C)_DomThe total capacitance of 20 out of 5 fF), and the capacitance of the variable capacitance diodes 133 and 134 is approximately atA range of 80-400 fF (e.g., 20 total capacitances out of 100 variable capacitors of 4 fF-20 fF based on Vctrl-0-800 mV). The capacitance of the fixed capacitor and the variable capacitor may be designed accordingly according to the application thereof. For example, a fixed capacitor with a large capacitance and a plurality of variable capacitors with a small capacitance and used for capacitance trimming can be designed. Alternatively, the fixed capacitor having a large capacitance may be divided into a plurality of individual capacitive elements.

First pole w_p0Can be calculated as follows:

peak frequency w_peakCan be calculated as follows:

second pole w_p1Can be calculated as follows:

in some embodiments, capacitors 133 and 134 are implemented by a variable voltage capacitor, which may include a variable capacitance diode or other PN junction diode, and a voltage-dependent bulk capacitance of a MOS transistor device whose channel is a non-linear channel that may vary and may vary over time depending on the manufacturing process used. The charge density in active devices varies with time and is particularly significant in short channel devices. As shown, variable capacitors 133 and 134 are connected back-to-back to minimize the modulation effect that the signal voltage exerts on the final capacitance, while the analog control voltage Vctrl is used to adjust the total capacitance. The use of the variable capacitor can greatly reduce the occupation of the chip area as compared with a switched capacitor bank or the like. Furthermore, the variable capacitor may also increase bandwidth and speed because it introduces less parasitic capacitance than existing capacitor arrays. In addition, since the variable capacitors 133 and 134 can reduce the adjustment range to an appropriate amount, by connecting the fixed capacitor 132 in parallel with the variable capacitor, the signal distortion effect can also be reduced. Fig. 6 shows two frequency spectrums representing the variation of CTLE gain with respect to frequency. Wherein the upper graph shows the case where the first pole frequency is adjusted by adjusting the variable capacitors 133 and 134 when there is no parallel fixed capacitor 132, and the lower graph shows the case where the first pole frequency is adjusted by adjusting the variable capacitors 133 and 134 when there is a parallel fixed capacitor 132. As shown in the figure, when the fixed capacitor 132 is not provided, it causes a very large non-linearity between the adjustment steps of the variable capacitor, and when the fixed capacitor 132 is provided, it causes a large increase in the linearity of the corresponding curve. Further, it is to be noted that when the fixed capacitor 132 is provided, the resultant peak value is kept more constant. In addition, by providing a fixed capacitor 132, the contribution of the zero is more constant and closer to the desired 20 dB/dec. When the fixed capacitor 132 is not provided, the zero contribution amount changes in a slope with respect to the frequency.

As another advantage, fixing the capacitance of capacitor 132 may allow for the use of a MOS ladder DAC (e.g., the MOS ladder DAC shown in fig. 3A). Compared with a resistor ladder type DAC and the like, the MOS ladder type DAC can further reduce power consumption and chip area. While a MOS ladder DAC may suffer from Differential Nonlinearity (DNL) and Integral Nonlinearity (INL) problems, the fixed capacitor 132 helps to eliminate such problems in a similar manner as the fixed capacitor eliminates variable capacitance diode nonlinearity. In addition, the fixed capacitor may be divided into two. The large size variable capacitor (or capacitors) occupying the larger area can be replaced by source degeneration capacitance. In such cases, half of the capacitance may be responsible for the variable capacitor, while the other half of the capacitance may be responsible for the metal-insulator-metal (MIM) capacitor or metal-on-metal (MOM) capacitor. Thus, the MIM/MOM capacitor can be vertically stacked in layers, where the bottom silicon layer contains the variable capacitor and the top metal layer contains the MIM/MOM capacitor. Such capacitors can be used in applications where the zero is shifted in the low frequency direction with a large capacitance.

Fig. 2 shows a series of identical parallel CTLE circuits 100 for the above binary mode selection implemented with a controlled voltage Vsw. In some embodiments, the parallel circuits 100 may all share a common load inductance (e.g., inductor L)₀And L₁) And a load impedance (e.g., load impedance R)_L0And R_L1) Are connected. Although eight circuits 100 are shown in the figure, this is not meant to be limiting. Wherein each circuit is taken from Vsw<7:0>A single binary value (which can be considered as a single variable encoded as a "thermometer" code) or a one-bit count value taken from 0 (all low) to 7 (all high). When all transistors 131 are "open" under Vsw control, each circuit 100 will be in its first ("high frequency") mode of operation, in which the dc balance gain with respect to peak is maximized and the dc gain is minimized. The more Vsw values that are set to selectively "short" the transistor 131, the more the circuit 100 goes into the second ("wide band") mode of operation, so that the greater the resulting dc gain and the lower the dc balance gain relative to the peak. When all Vsw values "short" the respective transistors, all circuits 100 enter their second mode of operation, thereby maximizing the dc gain and minimizing the dc balance gain with respect to the peak. By controlling the dc balance gain with respect to the peak value in such a parallel configuration, broadband noise such as thermal noise can be reduced, and the parasitic capacitance introduced can be reduced.

In some embodiments, transistor 131 may be operated between "high frequency" and "wide band" modes of operation by operating transistor 131 in the linear region with source impedance control signal Vsw <7:0> as an input to NMOS transistor 131. FIG. 7 is a block diagram of a DAC 705 for providing multiple voltages within a range of voltages as source impedance control signals Vsw <7:0 >. In fig. 7, the voltage is supplied by 200mV in the range of 0mV to 1000mV by the resistor ladder DAC 710, but these numbers should not be considered as a limitation, and the DAC 705 can be designed to have any set of values. Furthermore, DAC 705 should not be considered limited to the illustrated resistive ladder DAC 710, as other types of DACs known to those skilled in the art may also be applied.

It is noted that when the gate voltage of the NMOS transistor is raised, for example, by the source impedance control signal Vsw <0> supplied to the NMOS transistor 131 in fig. 1, the channel resistance of the transistor 131 will be lowered, thereby lowering the equalization gain described above; vice versa, when the gate voltage is lowered, the channel resistance of the NMOS transistor 131 will be reduced. In contrast, when the voltage supplied to the PMOS transistor (e.g., transistor 431 of fig. 4) is reduced, the channel resistance of PMOS transistor 431 is reduced, thereby reducing the above-mentioned equalizing gain; vice versa, when the gate voltage is increased, the channel resistance of PMOS transistor 431 will be increased. In such embodiments, the voltage value output by the DAC to the transistor 131/431 is used for "coarse" adjustment in the equalization gain range setting, while the number of parallel circuits 100 that have been enabled is used for "fine" adjustment in the equalization gain range setting, i.e., fine adjustment of the equalization gain within that range.

Fig. 8 illustrates frequency responses of two equalization gain ranges according to some embodiments. As shown, fig. 8 contains equalization gain ranges 802 and 805. In the case of the NMOS implementation of fig. 1, DAC 705 may be used to output a relatively high voltage that is provided as a source impedance control signal Vsw <7:0> to NMOS transistor 131 in enabled parallel circuit 100, thereby placing the amplifier within the equalization gain range 802. In addition, DAC 705 may also be used to output a relatively low voltage that is provided as a control signal to transistor 131 in enabled parallel circuit 100, thereby placing the amplifier within the equalization gain range 805. In contrast, in a PMOS implementation, the DAC may be used to place the amplifier within the equalizing gain range 802 by outputting a source impedance control signal having a relatively high voltage to the PMOS transistor 431 in the enabled parallel circuit 100, and to place the amplifier within the equalizing gain range 805 by outputting a relatively low voltage to the enabled parallel circuit 100.

Such embodiments for operating an amplifier in multiple equalization gain ranges may be adapted to various factors, including combinations of different coding scheme types, different cable/channel lengths, and/or various other factors. In a particular embodiment, a non-return-to-zero (NRZ) coding scheme may be configured to have a maximum equalization gain of about 7dB, while an overall non-return-to-zero (ENRZ) orthogonal differential vector signaling scheme may be configured to have a maximum equalization gain of about 10 dB. In this way, by selecting the output voltage of the DAC, the correct equalization gain range can be selected depending on the ENRZ or NRZ coding scheme used. At the same time, by enabling and disabling the parallel circuit 100, a fine adjustment of the desired equalization gain can be made within the selected equalization gain range.

This configuration of operating various numbers of substantially parallel amplifier circuits in either the first or second mode of operation can directly control the resulting gain difference between the low and high frequency regions of the overall gain-frequency curve of the system. In conjunction with the above-described control of the variable capacitance element of each amplifier circuit, the amplitude and corner frequency of the high frequency "peaking" can be independently configured. Such regulation may also be used in combination with other control methods, including varying the direct current of the circuit by regulating current sources 113 and 123, and by passing through R_L<n-1:0>Controlled parallel resistor network regulation R_L0And R_L1To change the effective load impedance. Furthermore, as shown in FIG. 2, by using multiple parallel circuits, an option is provided for separately and independently controlling the variable capacitors within each circuit to control the signal Vctrl with multiple bits<7:0>Greater fineness is achieved at the expense. However, in some embodiments, the same control signal Vctrl may also be provided to all circuits.

From equations 1 through 4 and the above regarding the variable capacitance diode for adjusting the source capacitance Cs and the multi-parallel circuit configuration for adjusting the source impedance Rs, it can be seen from the frequency response of fig. 5 that the circuits of fig. 1 and 2 can have multiple control dimensions. First zero point w_zeroControl may be achieved by controlling the variable capacitor voltage with Vctrl which regulates the regulated source capacitance Cs. The dc balance gain with respect to the peak can be selectively switched off (i.e. switched off)Open, high frequency mode) and the ratio of the number of parallel circuits 100 to the number of parallel circuits 100 that are conductive (i.e., "shorted," wide band mode) are controlled. Such adjustment operations are illustrated by 8 horizontal lines 505 in fig. 5. In 505, each horizontal line may correspond to one of the eight steps of the thermometer code described above in connection with FIG. 2. As shown in fig. 2, the lowest horizontal line in 505 (e.g., where the dc gain is minimal, and thus the dc equalization gain with respect to the peak is maximal) may correspond to a situation where transistor 131 is off (i.e., "open") in each circuit. Alternatively, a condition where each transistor 131 is on (i.e., "shorted") may correspond to the highest horizontal line in 505 and indicate that the dc gain is at a maximum, and thus the dc equalization gain with respect to the peak is minimized. Second pole w of frequency response_p1Can pass through a loaded resistor R_L0And R_L1The manner in which the effective load impedance is adjusted. In some embodiments, the effective load impedance may be set by a tunable resistor (not shown). Through the adjustment of the zero point and the pole point, the adjustment of the frequency band (width and position) required by high-frequency peaking and the high-frequency peaking gain can be realized. Such embodiments facilitate backward compatibility with certain conventional designs, as well as dynamic tuning in response to changes in channel response. Another control dimension may correspond to an adjustment to the bias current of current sources 113 and 123, which may control the peak amplitude by increasing or decreasing gm. Furthermore, the common mode provided to the subsequent signal processing stage may be controlled by controlling the load impedance R, for example with a resistor bank as shown in FIG. 1_L0And R_L1Is adjusted.

The various configurable elements described above may be adjusted by various control elements such as digital-to-analog converters (DACs), but these control elements increase system power consumption and circuit layout area. Fig. 3A shows an embodiment of a PMOS ladder DAC optimized to have minimal integrated circuit area and low current consumption. In some embodiments, the PMOS ladder DAC is highly suitable for the PMOS circuit of FIG. 4 because it can use the same substrate as the circuit 100, resulting in a significant savings in total circuit area. Since the fixed value resistor may occupy a significant layout area in some integrated circuit processes, the channel resistance of the PMOS transistors, i.e., 351-358, is used instead. In a practical implementation, 351-358 are identical PMOS transistors designed with channels narrow enough to provide the required series resistance ("R") value in the existing R-2R ladder structure. In such embodiments, the width-to-length ratio (W/L) of the transistors 351-358 may be twice the W/L ratio of the transistor 313. Since the higher the width-to-length ratio, the lower the resistance, the former has a resistance "R" half that of the latter. In some embodiments, each transistor applies a gate bias voltage Vss to ensure that it is in a resistive channel state.

Each driving element 300 of the ladder structure is illustrated as a multiplexer consisting of two identical MOS transistors 311 and 312, and corresponds to a switch-selective voltage source connected in series with transistors 351-358 with a resistance value R, but with a resistance value of 2R for transistor 313, where transistor 311 is used to select a voltage Vrefh and transistor 312 is used to select a voltage Vrefl, which are high and low voltage values required for the DAC output range. In some embodiments, the channel resistance of the transistor 313 is twice that of the transistors 351-358. In some embodiments, Vrefh and Vrefl can be in the range of 700-900 mV. In the embodiment using the NMOS transistor, the voltage range may be lower, such as 0-200 mV, because the NMOS transistor is more suitable for transmitting low voltage.

Transistors 311 and 312 are driven by complementary control signals, shown in the figure as derived from a binary control input Vc7, one value of which is taken from the control word Vc<0:7>The other value is an inverted form of VC7In the example of fig. 3A, eight taps of the ladder structure are driven using eight instances of 310, so the binary control input for each instance is derived from the control codeword Vc, respectively<0:7>Wherein Vc7 controls the most significant bit and Vc0 controls the least significant bit of said ladder structure. In some embodiments, control codeword Vc<0:7>Is a differential value, so, for example, the control value Vc7 is itself the same asClocked by Vc7 and complementaryTwo forms exist which can be used directly to control transistors 311 and 312. In other embodiments, each instance of 310 contains a buffer/inverter logic gate for providing the appropriate drive signals to 311 and 312 with a single digital control value.

In one particular integrated circuit implementation, the DAC 300 used to generate the analog output Vctrl is small enough to be physically collocated with or located near the variable capacitance diodes 133 and 134 it controls, thereby minimizing the introduction of unnecessary parasitic loads into the analog circuit 100.

Although fig. 3A shows the use of PMOS transistors, NMOS transistors may be used in an equivalent implementation. For NMOS designs, such single transistor implementations may be used for outputting narrow voltage ranges (e.g., 0-200 mV) around Vss, while for PMOS designs, such single transistor implementations may be used for outputting narrow voltage ranges (e.g., 700-900 mV) around Vdd. In the above embodiments, the variable capacitor may respond particularly well to control voltages near Vdd, but such embodiments should not be considered as limiting. Depending on the relative voltages of Vrefh and Vrefl with Vdd and Vss and the specific transistor gate thresholds, it may be desirable to specifically choose an NMOS implementation, a PMOS implementation, or even a hybrid implementation where multiplexer structure 310 is implemented by transmission logic gates. Such an implementation of multiplexer element 310 and transistor elements 313 and 351-358 as transmission logic gates is shown in FIG. 3B. Multiplexer element 310 of fig. 3B employs transmission logic gates in which inputs Vc7 and Vc are selectedControlling NMOS and PMOS transistors connected to Vrefh, respectively, and controlling PMOS and NMOS transistors connected to Vrefl, respectively. In the transistor circuit 313, an NMOS connected to Vdd and a PMOS connected to Vss are connected in parallel. In miningIn such embodiments with transmission logic gates, NMOS transistors are adapted to pass low voltages, while PMOS transistors are adapted to pass high voltages. Such embodiments may increase the voltage range provided by Vrefh and Vrefl. The MOS ladder structure only containing PMOS can receive the voltage in the range of 700-900 mV, and the implementation mode of the transmission logic gate can receive the voltage in the range of 0-900 mV.

Although a MOS ladder DAC may provide significant advantages in terms of power savings and chip area savings, it should be noted that other DACs, such as an R-2R ladder structure with resistors instead of transistors, may also be used to provide various control signals.