Abstract: Disclosed is a system and method for improving the linearity of a clock and data recovery (CDR) circuit. In one embodiment, a data stream is received, and the phase of a clock signal is adjusted using two interpolators. The phase of the output signal of the second interpolator is adjusted simultaneously with, and complementary to, adjusting the phase of the first interpolator. The first interpolator's output signal is injected into a first delay cell in a delay loop having a plurality of delay cells, and the output of the second interpolator is inactivated. When the maximum phase of the first interpolator's output signal is reached, the second interpolator's output signal is injected into another one of the delay cells, and the first interpolator's output signal is inactivated. The data stream is then recovered using the output of the delay loop as a clock signal.

Abstract: In described embodiments, a communication system employing, for example, clock and data recovery (CDR) detects and applies correction for DC offset in an input data stream signal, termed as “DC offset calibration”. DC offset calibration applies static calibration for DC offset in input circuits, such as an input amplifier and detection latches, without input data, and applies statistical calibration for DC offset during operation with an input data stream to correct for dynamic shifts in DC level. Such DC offset calibration employs data eye measurements of the input data stream for detection of DC offset, and applies an opposite DC offset to maintain a relatively balanced data eye during live traffic.

Abstract: In described embodiments, a communication system employing, for example, clock and data recovery (CDR) detects and applies correction for DC offset in an input data stream signal, termed as “DC offset calibration”. DC offset calibration applies static calibration for DC offset in input circuits, such as an input amplifier and detection latches, without input data, and applies statistical calibration for DC offset during operation with an input data stream to correct for dynamic shifts in DC level. Such DC offset calibration employs data eye measurements of the input data stream for detection of DC offset, and applies an opposite DC offset to maintain a relatively balanced data eye during live traffic.

Abstract: Disclosed is a system and method for improving the linearity of a clock and data recovery (CDR) circuit. In one embodiment, a data stream is received, and the phase of a clock signal is adjusted using two interpolators. The phase of the output signal of the second interpolator is adjusted simultaneously with, and complementary to, adjusting the phase of the first interpolator. The first interpolator's output signal is injected into a first delay cell in a delay loop having a plurality of delay cells, and the output of the second interpolator is inactivated. When the maximum phase of the first interpolator's output signal is reached, the second interpolator's output signal is injected into another one of the delay cells, and the first interpolator's output signal is inactivated. The data stream is then recovered using the output of the delay loop as a clock signal.

Abstract: Disclosed is a system and method for improving the linearity of a clock and data recovery (CDR) circuit. In one embodiment, a data stream is received, and the phase of a clock signal is adjusted using two interpolators. The phase of the output signal of the second interpolator is adjusted simultaneously with, and complementary to, adjusting the phase of the first interpolator. The first interpolator's output signal is injected into a first delay cell in a delay loop having a plurality of delay cells, and the output of the second interpolator is inactivated. When the maximum phase of the first interpolator's output signal is reached, the second interpolator's output signal is injected into another one of the delay cells, and the first interpolator's output signal is inactivated. The data stream is then recovered using the output of the delay loop as a clock signal.

Abstract: Methods and apparatus are provided for determining a position of an offset latch employed for decision-feedback equalization. The position of an offset latch is determined by obtaining a plurality of samples of a data eye associated with a signal, the data eye comprised of a plurality of trajectories for transitions out of a given binary state; determining an amplitude of at least two of the trajectories based on the samples; and determining a position of an offset latch based on the determined amplitudes. The initial position of the offset latch can be placed, for example, approximately in the middle of the determined amplitudes for at least two of the trajectories. The initial position of the offset latch can be optionally skewed by a predefined amount to improve the noise margin.

Abstract: Disclosed is a system and method for improving the linearity of a clock and data recovery (CDR) circuit. A data stream is received and the phase of a clock signal is adjusted using two interpolators. The data stream is then recovered using the clock signal.

Abstract: Methods and apparatus are provided for controlling at least one of a rise time and a fall time of a signal. A plurality of time shifted clock signals are generated; and a received data signal is sampled using a plurality of parallel data paths, where each of the data paths are controlled by a corresponding one of the plurality of time shifted clock signals. The plurality of time shifted clock signals can be generated, for example, by at least one delay element. The plurality of parallel data paths can be substantially identical and comprise, for example, at least one latch or at least one flip flop. Compensation can optionally be provided for variations in, for example, process corner, supply voltage, aging and operating temperature.

Abstract: Methods and apparatus are provided for sigma-delta delay control in a Delay-Locked-Loop that employs a delay line to generate a clock signal based on a reference signal. A first value is generated if a clock signal has a time lead relative to a reference signal; and a second value is generated if a clock signal has a time lag relative to a reference signal. The first and second values are accumulated to generate an N bit digital word; and the N bit digital word is reduced to an M bit digital word, where M is less than N. Thereafter, the M bit digital word can be converted to an analog bias signal. The reducing step can be performed, for example, by a sigma-delta modulator. The high frequency quantization noise generated by the sigma-delta modulator can be filtered using a low pass filter. The converting step can be performed by a digital-to-analog converter, such as a master/slave digital-to-analog converter.

Abstract: An integrated circuit includes a phase-locked loop (PLL) in which the loop bandwidth of the PLL is proportional to the input frequency of the PLL. The PLL includes a phase/frequency detector (PFD), a charge pump, a loop filter, and a voltage-controlled oscillator (VCO) that generates the PLL output clock. The VCO includes a current scaling block that scales the sum of a variable current, which is proportional to the loop filter voltage, and a fixed current. The frequency of the PLL output clock is a function of the current output from the current scaling block. Since the same scaling factor is applied to both the fixed current and the variable current, the gain from the loop filter voltage to the PLL output frequency is proportional to the PLL output frequency, and thus the loop bandwidth of the PLL is proportional to the PLL input frequency.

Abstract: Methods and apparatus are provided for improved digital-to-analog conversion. The disclosed digital-to-analog converter comprises a master digital-to-analog converter that generates a master analog value, and a slave digital-to-analog converter that generates a slave analog value that is based on the master analog value. The slave analog value can be, for example, substantially proportional to the master analog value. The master D/A converter can be varied during a coarse tuning mode, while the input to the slave D/A can be fixed, for example, to an approximately mid-range value until the master analog value satisfies one or more predefined conditions. Thereafter, during a fine tuning mode, the slave D/A converter can be varied, while the master D/A converter is fixed, so that the output Y is equal to a desired value (within a specified tolerance).

Abstract: A delay loop (e.g., a voltage-controlled delay loop) has (at least) two devices (e.g., interpolators) for generating clock signals for injection into the delay elements of the delay loop in a leap-frog manner, in which, while one interpolator is generating the clock signal currently selected for injection, the other interpolator can be controlled to generate the next clock signal to be selected for injection. This leap-frog technique can provide more settling time for generating injected clock signals than implementations that rely on a single interpolator.

Abstract: A bias circuit for providing at least first and second bias signals for biasing a cascode current source and/or a cascode current sink includes a resistive element and first, second and third transistors, each transistor having first and second source/drain terminals and a gate terminal. The first source/drain terminal of the first transistor is coupled to the gate terminal, the first bias signal being generated at the first source/drain terminal in response to receiving a first reference current at the first source/drain terminal. A first end of the first resistive element is coupled to the second source/drain terminal of the first transistor. The gate terminal of the second transistor is coupled to the gate terminal of the first transistor, the second bias signal being generated at the first source/drain terminal of the second transistor in response to receiving a second reference current at the first source/drain terminal of the second transistor.

Abstract: Embodiments of the invention include a voltage-controlled oscillator (VCO) circuit in which the gain of the VCO is proportional to the output or operating frequency of the VCO. The VCO circuit includes a voltage-controlled oscillator, a current scaling block, and a summing node. The summing node couples a VCO control current and a VCO reference current into the current scaling block. The current scaling block scales the sum of the currents by a scaling factor. The output of the current scaling block, which is coupled to the input of the VCO, provides a bias current to the VCO, which bias current adjusts the oscillation frequency of the oscillator. The VCO control current and the VCO reference current are scaled by the same scaling factor, thus allowing the gain of the VCO to be proportional to but not dependent on the output frequency of the VCO.

Abstract: A system for a backplane that employs i) an adjustment of positive-to-negative (P-N) driver skew of a transmit signal of a relatively high-speed differential driver to reduce far-end crosstalk, ii) a high-speed differential subtraction circuit combining a gain-adjusted replica of at least one transmit signal with a received signal to reduce near-end crosstalk, and iii) a phase-locked loop (PLL) synchronization circuit to align timing events between a set of near-end and far-end high-speed interfaces.