System and method for improved time-interleaved analog-to-digital converter arrays

• US 20060279445A1
• Filed: 06/13/2005
• Published: 12/14/2006
• Est. Priority Date: 06/13/2005
• Status: Active Grant

• Current Assignee: Texas Instruments, Inc.
• Original Assignee: Texas Instruments, Inc.
• Inventors: Kinyua, Martin Kithinji, Bright, William J.
• Granted Patent: US 7,292,170 B2
• Time in Patent Office: Days
• US Class Current: 341/155
• CPC Class Codes: H03M 1/0624 by synchronisation
• CPC Class Codes: H03M 1/1215 using time-division multiplexing

Abstract

System and method for improved time-interleaved analog-to-digital converter arrays which reduces sampling mismatch distortion found in prior art arrays. There may be two causes of non-uniform sampling mismatch in a TI-ADC array, a mismatch due to skew and a mismatch due to clock jitter. To minimize non-uniform sampling mismatch, the mismatch due to skew can be addressed. A preferred embodiment comprises adjusting a delay imparted on the sampling clock by an adjustable delay in each channel of a plurality of channels in the TI-ADC array to minimize skew and randomly switching between two delays that span a zero-skew delay to reduce residual skew in each channel and thus eliminate (or reduce) frequency domain tones caused by non-uniform sampling mismatch.

Claims

1. A method for reducing non-uniform sample mismatch in a time-interleaved analog-to-digital converter (TI-ADC) array, the method comprising:

adjusting a sampling clock delay imparted by an adjustable delay in each channel of a plurality of channels in the TI-ADC array in which the channels are utilized in a constant cyclic order to reduce a measured sample mismatch; and

randomly switching between two sampling clock delays imparted by the adjustable delay in each channel to reduce a residual skew, wherein the two delays span an actual delay that would bring the measured sample mismatch substantially to zero.

2. The method of claim 1, wherein the adjusting comprises:

fixing the delay at one channel to a maximum delay;

for each remaining channel,

setting the delay to a first delay value;

measuring a first sample mismatch;

incrementing the delay value;

measuring a second sample mismatch;

incrementing the delay value and repeating the second measuring and the incrementing if successive measured sample mismatch values continue to decrease; and

saving the delay for the channel to be equal to the delay value immediately prior to a current delay value if there is an increase in the measured sample mismatch of the current delay value compared to the measured sample mismatch of the delay value immediately prior to the current delay value.

3. The method of claim 2, wherein the maximum delay is a delay selected to exceed a propagation delay seen by any of the channels in the TI-ADC.

4. The method of claim 2, wherein the two delays comprises a first delay and a second delay, wherein the first delay is the saved delay value, and wherein the second delay is a delay value immediately succeeding the saved delay value.

5. The method of claim 1, wherein the random switching comprises:

for each channel in the TI-ADC array,

randomly generating a value; and

imparting one of the two delays, wherein the delay imparted is based upon the randomly generated value.

6. The method of claim 5, wherein the randomly generated value has one of two values, and wherein when the randomly generated value is of a first value, then a first delay is imparted by the channel, and when the randomly generated value is of a second value, then a second delay is imparted by the channel.

7. The method of claim 6, wherein the first delay and the second delay spans a delay range, wherein there exists a zero-mismatch delay in the delay range with which a measured sample mismatch is substantially equal to zero when the adjustable delay is set to the zero-mismatch delay value.

8. The method of claim 1, wherein the delay imparted by the adjustable delay is determined by a magnitude of a current provided to the adjustable delay.

9. The method of claim 1, wherein the adjusting occurs during an initial configuration operation.

10. The method of claim 1, wherein measuring a sample mismatch comprises measuring a magnitude of tones located at integer multiples of a channel'"'"'s sampling frequency.

11. A time-interleaved analog-to-digital converter (TI-ADC) array comprising:

a plurality of channels, each channel coupled to a signal input, wherein each channel comprises

a sample and hold circuit (SHC) coupled to the signal input, the SHC configured to capture a signal at the signal input and provide the captured signal at an output;

an analog-to-digital converter (ADC) coupled to the SHC, the ADC configured to convert an analog signal at an input into a digital representation of the analog signal;

the TI-ADC array further comprising,

a plurality of adjustable delays coupled to a clock signal source, each adjustable delay coupled to a SHC of a channel in the plurality of channels, each adjustable delay configured to impart a delay onto a signal at a first input, wherein the delay imparted by the adjustable delay is based upon a signal at a second input and wherein the clock signal source activates the channels in a constant cyclic order; and

a multiplexer having an input coupled to an ADC from each channel in the plurality of channels, the multiplexer configured to selectively couple an output from an ADC to a digital signal output of the TI-ADC array.

12. The TI-ADC array of claim 11, wherein the signal at the second input of an adjustable delay is a current, and wherein a magnitude of the current determines the delay imparted by the adjustable delay.

13. The TI-ADC array of claim 11, wherein the clock signal source provides a clock signal that is substantially equal to a desired sampling rate of the TI-ADC array divided by M, where M is a number of channels in the plurality of channels.

14. The TI-ADC array of claim 11 further comprising a plurality of built-in randomizers, each built-in randomizer coupled to an adjustable delay, the built-in randomizer configured to produce a random two-valued signal that is used to switch the delay imparted by the adjustable delay.

15. The TI-ADC array of claim 14, wherein the delays imparted by the adjustable delays are determined by a magnitude of a current provided at the second input of each adjustable delay, and wherein the built-in randomizer provides one of two currents to the second input of the adjustable delay, wherein a magnitude of the current provided by the built-in randomizer is based upon a value of the random two-valued signal.

16. The TI-ADC array of claim 15, wherein the built-in randomizer comprises:

a pseudo-random number generator being configured to generate the random two-valued signal;

a first current source;

a second current source; and

a switch having a first input coupled to the first current source, a second input coupled to the second current source, and a third input coupled to the pseudo-random number generator, the switch configured to couple either the first input or the second input to an output based upon a value of the random two-valued signal.

17. A system for reducing non-uniform sample mismatch in a time-interleaved analog-to-digital converter (TI-ADC) array, the system comprising:

a means for reducing a clock skew between clock signals in each channel of a plurality of channels in the TI-ADC array in which the channels are utilized in a constant cyclic order; and

a means for randomly switching between delays imparted to the clock signal in each channel.

18. The system of claim 17, wherein the means for reducing clock skew reduces a clock skew for a channel with respect to a clock of a reference channel.

19. The system of claim 17, wherein the means for reducing clock skew comprises an adjustable delay means for inserting a configurable delay to the clock signal of a channel.

20. The system of claim 19, wherein the means for randomly switching between delays comprises:

a random value means for generating a random two-valued sequence; and

a delay switching means for providing a delay value based on a value generated by the random value means to the adjustable delay means.