Classifications

• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/10—Calibration or testing
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/002—Provisions or arrangements for saving power, e.g. by allowing a sleep mode, using lower supply voltage for downstream stages, using multiple clock domains or by selectively turning on stages when needed
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/10—Calibration or testing
  • H03M1/1009—Calibration
  • H03M1/1014—Calibration at one point of the transfer characteristic, i.e. by adjusting a single reference value, e.g. bias or gain error
  • H03M1/1023—Offset correction
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/14—Conversion in steps with each step involving the same or a different conversion means and delivering more than one bit
  • H03M1/16—Conversion in steps with each step involving the same or a different conversion means and delivering more than one bit with scale factor modification, i.e. by changing the amplification between the steps
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/34—Analogue value compared with reference values
  • H03M1/36—Analogue value compared with reference values simultaneously only, i.e. parallel type
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/34—Analogue value compared with reference values
  • H03M1/38—Analogue value compared with reference values sequentially only, e.g. successive approximation type
  • H03M1/44—Sequential comparisons in series-connected stages with change in value of analogue signal
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/1205—Multiplexed conversion systems
  • H03M1/121—Interleaved, i.e. using multiple converters or converter parts for one channel
  • H03M1/1215—Interleaved, i.e. using multiple converters or converter parts for one channel using time-division multiplexing
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/1235—Non-linear conversion not otherwise provided for in subgroups of H03M1/12
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/14—Conversion in steps with each step involving the same or a different conversion means and delivering more than one bit
  • H03M1/16—Conversion in steps with each step involving the same or a different conversion means and delivering more than one bit with scale factor modification, i.e. by changing the amplification between the steps
  • H03M1/164—Conversion in steps with each step involving the same or a different conversion means and delivering more than one bit with scale factor modification, i.e. by changing the amplification between the steps the steps being performed sequentially in series-connected stages
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/34—Analogue value compared with reference values
  • H03M1/36—Analogue value compared with reference values simultaneously only, i.e. parallel type
  • H03M1/361—Analogue value compared with reference values simultaneously only, i.e. parallel type having a separate comparator and reference value for each quantisation level, i.e. full flash converter type
• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M1/00—Analogue/digital conversion; Digital/analogue conversion
  • H03M1/12—Analogue/digital converters
  • H03M1/34—Analogue value compared with reference values
  • H03M1/38—Analogue value compared with reference values sequentially only, e.g. successive approximation type
  • H03M1/44—Sequential comparisons in series-connected stages with change in value of analogue signal
  • H03M1/445—Sequential comparisons in series-connected stages with change in value of analogue signal the stages being of the folding type

Definitions

• the present invention generally relates to an analog-to-digital converter wherein a binary search is used.
• Some applications such as hard disk read channels or wideband wireless standards, require a low-resolution (for example, approximately 6 bit), high- speed (for example, greater than 1 Giga-samples per second (GS/s)) analog-to- digital converter (ADC).
• a low-resolution for example, approximately 6 bit
• high- speed for example, greater than 1 Giga-samples per second (GS/s)
• ADC analog-to- digital converter
• ADCs time-interleaved Successive Approximation Register (SAR) converters and flash converters.
• SAR Successive Approximation Register
• flash converters Single-channel SAR converters typically operate at sampling frequencies of a few hundred megasamples per second (MS/s) (for example, approximately 300 MS/s). As a result, a large number of channels would need to be interleaved, yielding a large input capacitance.
• a time-interleaved SAR architecture for the same specifications could have an input capacitance 10-20 times larger than a pipelined binary search ADC.
• Flash converters on the other hand would be severely limited by quantized power, as for each conversion 63 comparisons (6 bit) would have to be made at low noise/offset. The power requirement for similar specifications with a calibrated flash converter would be 10 times larger than the power consumption in a pipelined binary search ADC.
• Pipelined analog-to-digital converters have become popular for sampling rates from a few megasamples per second up to 100 megasamples per second. Dynamic pipelined conversion enables low power quantization at high speed with low input capacitance but requires calibration.
• US patent application US2005/0062635 introduces a pipelined analog- to-digital converter that follows a non-linear scale and allows operation at frequencies of 2 GHz and more.
• the pipelined ADC comprises a number of comparator stages where the thresholds of the comparator stages are adjusted in accordance with the digital conversion results from the previous stage.
• an architecture and method are proposed in this document to provide a pipelined ADC with a programmable characteristic so even a non-linear scale may be implemented.
• the output signals are processed via linear signal processing, using linear amplifiers.
• the present invention aims to provide for an analog-to-digital converter with reduced power consumption (low-resolution, high-speed).
• a pipelined analog-to-digital converter for converting an analog input signal into a digital signal comprises a plurality of comparing means having tuneable thresholds for comparing an input signal with, whereby at least two of said given thresholds are different, and a plurality of amplifying circuits.
• the plurality of comparing means is configured to form a hierarchical tree structure, having a plurality of hierarchical levels. At least one of the hierarchical levels is associated with at least one amplifying circuit of the plurality of amplifying circuits. The at least one amplifying circuit generates the input of at least one comparing means at the next hierarchical level.
• the method for converting an analog input signal into a digital output is preceded by a calibration step.
• the threshold of at least one comparing means is tuned during a calibration period and from that point onwards the comparator has a given threshold.
• the calibration step preferably comprises tuning the at least one amplifying circuit associated with the at least one of the hierarchical levels by means of a variable capacitance.
• the step of comparing yields an output signal that is fed to a amplifying circuit/DAC, implementing a successive approximation process. A binary code is determined.
• Fig. 1 illustrates a general block diagram of 3 bits of 1 bit per stage pipelined A/D converter.
• FIG. 2 illustrates an example of a hybrid ADC.
• Fig. 4 shows a block diagram of 3 bits of 1 bit per stage pipelined ADC converter.
• FIG. 5 shows a schematic of a folding front-end stage.
• Fig. 6 the waveforms of a folding front-end stage.
• Fig. 7 shows the timing of the clock signals of a folding stage.
• Fig. 8 shows an example of a possible comparator-T/H circuit.
• top, bottom, over, under and the like in the description are used for descriptive purposes and not necessarily for describing relative positions.
• the terms so used are interchangeable under appropriate circumstances and the embodiments of the invention described herein can operate in other orientations than described or illustrated herein.
• This disclosure provides a pipelined analog-to-digital converter with non-linear sig nal processing (th is is eq u ivalent to resid ue generation and amplification) where each ADC threshold is implemented with a different calibrated comparator. This calibration can then compensate for nonlinearity as well as random offset due to device mismatch. Furthermore, only the comparators and amplifiers required for a binary search are being activated whereby low power consumption is achieved.
• the pipelined converter further comprises a folding front-end.
• the pipelined converter further comprises an n-bit flash analog-to-digital converter.
• the plurality of comparing means is configured to form a hierarchical tree structure having a plurality of hierarchical levels (36), (37), wherein at least one of said hierarchical levels is associated with at least one amplifying circuit of the plurality of amplifying circuits.
• the plurality of hierarchical levels comprises means for setting the tuneable thresholds in accordance to the output (comparison result) of previous hierarchical level so that non-linear distortion of the preceding hierarchical level is removed.
• ADC is presented comprising an interleaved structure, this structure comprising a folding front-end, a PL ADC and a flash ADC.
• the architecture offers a power consumption proportional to the sampling frequency.
• a 4x interleaved 6bit ADC is presented.
• Each conversion channel comprises a 1 bit folding front-end (81 ), 3 bits of pipelined conversion (82) and 2 bits of flash conversion (83) as illustrated in Figure 2.
• the folding front-end samples the input signal, removes common-mode component of the input signal and rectifies the differential signal of the input while determining the polarity of the input signal.
• the disclosure further provides for a method for converting an analog input signal into a digital output.
• the method implements a binary search algorithm.
• the number of active comparators is reduced and therefore the power consumption .
• At least one comparing means of a first hierarchical level is further arranged for controlling at least one other comparing means of a subsequent hierarchical level.
• controlling is meant that a comparing means is arranged for selecting a path in the structure based on the comparison result of the previous step, whereby the structure is formed by the plurality of comparing means. This path is illustrated in Figure 2.
• the PL ADC determines 3 bits of the conversion and generates a residue on only one of 4 outputs (output of (94), (95), (96) or (97)).
• bits are determined via a parallel search, requiring a lot of power consuming comparators.
• a binary search instead of a parallel one, the number of active comparators and therefore the power consumption is reduced.
• This PL ADC uses dynamic nonlinear amplifiers for low power and high speed. Linearity requirements in these amplifiers are avoided by activating a different dynamic comparator for each ADC threshold and calibrating the corresponding comparator threshold to a desired input-referred value, cancelling errors both from non-linear signal processing and offsets in the comparators. This is achieved by a tree of comparators combined with amplifying circuits each of which are calibrated ind ivid ual ly. Threshold cal ibration corrects for ampl ifier and comparator imperfections. Each stage of the tree can be combined with one amplifying circuit (as illustrated in Figure 1 ) or each comparator in a stage can be combined with an amplifying circuit (as illustrated in Figure 4).
• these amplifying circuits can be built-in track-and-hold amplifiers (comparator/track-and-hold amplifying in one).
• the "to be converted" signal is being sent through the chain (or cascade). There is no signal dependent routing or path selection as provided in this disclosure.
• the signal recombination consists of two stages of two-input multiplexers, which combine the different interleaved channel outputs into 6 full- speed bit streams. While static linearity with relaxed linearity and matching constraints is improved using threshold calibration, tinning calibration is avoided due to its complexity.
• the clock generation generates two sets of clock signals, both of which run at a fourth of the sampling frequency.
• the first set of signals is a low precision differential set of quadrature signals used to control non-critical timing instances in the ADCs (in each channel) and to synchronize the different channels with respect to the others. These signals have large fan-out and hence large drive strength.
• the folding front-end samples the input signal and first removes its common-mode component.
• the polarity of the resulting differential signal is determined by a comparator (1 ) and using a chopper controlled by this comparator (1 ) the signal is rectified to be in range of a succeeding ADC.
• the primary reason a folding stage is implemented in the complete converter (hybrid ADC) is to limit the calibration complexity.
• the schematic of the folding stage is shown in Figure 5.
• the bottom plate nodes will have a common mode voltage of zero, and a differential voltage which depends on the ratio
• the first threshold of the next stage is calibrated in the positive and negative half of the ADC range. These two values are compared, and the mean is assumed "correct” and set. Then based on this threshold, the C par values can be adapted.
• FIG. 5 shows a simplified schematic of the folding front-end and Figure 6 its waveforms. It samples and rectifies the input signal while removing its common-mode.
• the Si switches When the Si switches are closed the input voltages are tracked across C s . At a falling ⁇ i edge the charges on C s are fixed. Their bottom plates are at ground and the top plates at their sampled input voltage, neglecting charge injection. Closing S2 shorts the top plates and generates the differential voltage at the bottom plates with some loss due to stray capacitance.
• the folding stage comparator is then activated, and based on its decision closes one set of switches in the chopper (at ti), sharing the charge on the bottom plates with the next stage such that the differential output voltage is always positive.
• the common-mode output is independent of the common-mode input which fixes the common-mode voltage for the ADC back-end and significantly improves the common-mode input range. Moreover, the applied common-mode voltage may differ in calibration and normal operation. Pipelined binary search
• a pipelined binary search (PLBS) converter consists of a cascade of non-linear multiplying digital to analog converters (NLMDACs) and a tree of comparators, as shown in Figure 1 for 3 bits of 1 bit per stage PLBS.
• NLMDACs non-linear multiplying digital to analog converters
• the NLMDACs purpose is to sample its input signal, to amplify it and to subtract/add some value from the output to bring it closer to zero.
• the linearity requirement on the first stage MDAC is equal to the overall desired linearity.
• significant nonlinearity is allowed by using a different comparator with tunable threshold for each PL ADC threshold.
• the comparator threshold can be tuned so it cancels the nonlinear effects of the cascade of preceding NLMDACs for the desired ADC threshold. Since the only requirement on the NLMDAC is that it is monotonic, power savings are possible.
• this converter's input range is not symmetrical around 0 V differential. Because the input signal is rectified in the previous stage (folding front end stage), only positive differential signals should be converted. By subtracting half of the input range from the input signal of the first stage, the succeeding stages can be made roughly differential around 0.
• each NLMDAC (44) in the chosen implementation is in parallel with a comparator with tunable threshold. As such these were merged into a single structure called a comparator/track-and-hold amplifier (CTHA). All CTHAs except those in the penultimate stage then have a load of two CTHAs, with the convention that the CTHA in the earlier stage is called “parent” and the CTHAs in the later stage are called "child”. The CTHAs in the penultimate stage are loaded simply by two comparators each in the last stage.
• CTHA comparator/track-and-hold amplifier
• the dynamic preamplifier and output driver combine to form a track and hold amplifier.
• transistor pairs N1 and N2 turn off, while P2 and P3 turn on.
• the nodes Dm and Dp are pulled from ground up to Vdd at a rate depending on the input voltage.
• transistors P5 are on and charge the nodes aOutp and aOutm.
• the P5 pair is turned off, and no more current flows in the circuit, so that the voltage on aOutp and aOutm is fixed by the amount of charge added to these nodes.
• the input voltage is thus converted into a time (the time the P5 pair is on) and then back into voltage (the charge added to the output capacitors). Since the output voltage depends on the input voltage, a track and hold function is achieved.
• the comparator threshold is zero and the input output relation is given by out ⁇ in x gain , with the gain being determined by the transistor sizes and the chosen values of Cd and Ca. If the dynamic preamplifier circuit is imbalanced in some way, the comparator threshold will change to a value V O ff Se t while the input output relationship changes to out ⁇ ⁇ in- V ⁇ et ) ⁇ gain . In other words: the output is roughly zero when the comparator is at its threshold (note that the comparator is formed by combining the latch and the dynamic preamplifier).
• the PLBS converter is supposed to have an input range from 0 to V
• the Cd capacitors are changed first to set the parent threshold (thereby calibrating comparator threshold of highest hierarchical level).
• the parent threshold is the threshold of the CTHA (41 ), (44) of the first level.
• the Ca capacitors are used for coarse threshold tuning for both child CTHAs.
• a first child CTHA is illustrated in Figure 4 and is a combination of (42) and (45), a second child CTHA is a combination of (43) and (46).
• the Ca capacitors are used to set the amplifier output close to the uncalibrated comparator thresholds of the next level when their corresponding ADC thresholds are applied.
• the results of the calibration process are illustrated in Figure 10.
• the top threshold (th top ) of the next stage is applied, and Ca at node aOutp is changed to bring the amplifier output in the calibration range of the comparator which implements this threshold.
• the bottom threshold (th bo ttom) of the next stage is applied and Ca at node aOutm is similarly changed.
• the next PLBS stage can then be calibrated using the same process: calibrating the comparator threshold first and the Ca capacitances next. This assumes that
• FIG. 12 An example of the timing of the different clock signals is shown in Figure 12.
• clkG ⁇ n> is a global clock shared by all CTHAs in the n th row.
• the signal clkG ⁇ n+1 > is always T C i k /4 delayed with respect to clkG ⁇ n>. Consequently, to ensure that each comparator has equal regeneration time, a comparator must decide in Tcik/4-toeiay- In the graph shown the center regeneration is slow. Consequently, the decision time for the next comparator is smaller. However this comparator will have a larger input signal, and therefore decide faster.
• each comparator must decide in T C ik/2-tDeiay-tA P erture, any two consecutive comparators must decide in 3.T C ik/4-2.tDeiay-tAperture, any three consecutive comparators must decide in Tcik-3.t De iay- tAperture, and so on.
• toeiay is the gate delay from the circuits in Figure 11 and t A perture is the aperture time of the next CTHA.
• the encoder (84) converts the comparator decisions into 3 bit gray code. It consists of precharge/discharge ROM lines controlled by the clkC outputs on each row. If metastability occurs in one of the rows, all bits starting from this row will be O.

Landscapes

• Engineering & Computer Science (AREA)
• Theoretical Computer Science (AREA)
• Analogue/Digital Conversion (AREA)

Priority Applications (4)

Applications Claiming Priority (4)

Publications (2)

Family

ID=42269365

Family Applications (1)

Country Status (6)

Cited By (3)

Families Citing this family (12)

Citations (3)

Family Cites Families (6)

• 2010
  • 2010-01-26 EP EP10151660A patent/EP2280486A1/en not_active Withdrawn
  • 2010-07-08 WO PCT/EP2010/059821 patent/WO2011003978A2/en not_active Ceased
  • 2010-07-08 EP EP10736647.8A patent/EP2452437B1/en active Active
  • 2010-07-08 JP JP2012518996A patent/JP5558566B2/ja active Active
  • 2010-07-08 CN CN201080031210.7A patent/CN102474262B/zh active Active
  • 2010-07-08 KR KR1020127002931A patent/KR20120062695A/ko not_active Withdrawn
  • 2010-07-08 US US13/382,735 patent/US8618973B2/en active Active

Patent Citations (3)

Also Published As

Similar Documents

Legal Events