Eye center determination system and method
View Patent ↗A system and method for bit eye center determination is provided. In general, the system samples an incoming data stream to determine where transitions in the data stream occur, selectively offsets the selected samples based on state criteria and the number of transitions in each set of samples, accumulates the offset samples and averages the result to determine the center of the bit eye. The system and method also provides the ability to locate the eye center even in the case of noise in the system, whether the noise is random or deterministic, including odd/even noise.
1. An eye center trainer, the eye center trainer comprising:
a sampler, the sampler receiving a data stream and taking a plurality of samples of the data stream;
a controller, the controller receiving the plurality of samples, the plurality of samples comprising a plurality of sample sets, each sample set corresponding to samples in a bit time in the data stream, the controller locating transitions identified in the plurality of samples, selecting tap numbers corresponding to the located transitions, and offsetting the selected tap numbers based on a number of transitions in each sample set of the plurality of sample sets; and
an accumulator, the accumulator receiving the offset selected tap numbers and accumulating the offset selected tap numbers to determine an accumulated value, and wherein higher order bits in the accumulated value are used to determine which of the plurality of samples in the data stream correspond to an eye center of bits in the data stream.
2. The eye center trainer of claim 1 wherein the plurality of sample sets comprises a first sample set, a second sample set, a third sample set, and a fourth sample set, and wherein the controller offsets a first tap number in the first sample set by an offset T, and wherein the controller offsets a second tap number by an offset 0 if the second tap number is in the first sample set, and wherein the controller offsets the second tap number by the offset T if the second tap number is in the second sample set, and wherein the controller offsets the second tap number by an offset 2T if the second tap number is in the third sample set, and wherein the controller offsets a third tap number by the offset 0 if the third tap number is in the second sample set, and wherein the controller offsets the third tap number by the offset T if the third tap number is in a third sample set, and wherein the controller offsets the third tap number by the offset 2T if the third tap number is in the fourth sample set, and wherein the controller offsets a fourth tap number by the offset 0 if the fourth tap number is in the third sample set, and wherein the controller offsets the fourth tap number by the offset T if the fourth tap number is in the fourth sample set, and wherein the controller offsets the fourth tap number by the offset 2T if the fourth tap number is in a fifth sample set, and where T comprises a width of the bit time in the data stream based on a number of taps sampling each bit in the data stream.
3. The eye center trainer of claim 1 wherein the accumulator outputs the accumulated value and wherein the accumulator is seeded with an initial value selected to shift the accumulator output to identify the plurality of samples in the data stream corresponding to the eye center of bits in the data stream.
4. An eye center trainer, the eye center trainer comprising:
a sampler, the sampler receiving a data stream and taking a plurality of samples of the data stream;
a controller, the controller receiving the plurality of samples, the plurality of samples comprising a plurality of sample sets, each sample set corresponding to samples in a bit time in the data stream, the controller locating transitions identified in the plurality of samples, selecting tap numbers corresponding to the located transitions, and offsetting the selected tap numbers based on a number of transitions in each sample set of the plurality of sample sets, wherein the controller offsets tap numbers that correspond to early transitions by 0, offsets tap numbers that correspond to on time transitions by T, and offsets tap numbers that correspond to late transitions by 2T; and
an accumulator, the accumulator receiving the offset selected tap numbers and accumulating the offset selected tap numbers to determine which of the plurality of samples in the data stream correspond to an eye center of bits in the data stream.
5. The eye center trainer of claim 4 wherein T comprises width of the bit time in the data stream based on a number of taps sampling each bit in the data stream.
6. The eye center trainer of claim 4 wherein the accumulator averages the accumulated offset tap numbers to determine which of the plurality of samples in the data stream correspond to the eye center of bits in the data stream.
7. An eye center trainer, the eye center trainer comprising:
a sampler, the sampler receiving a data stream and taking a plurality of samples of the data stream;
a controller, the controller receiving the plurality of samples, the plurality of samples comprising a plurality of sample sets, each sample set corresponding to samples in a bit time in the data stream, the controller locating transitions identified in the plurality of samples, selecting tap numbers corresponding to the located transitions, and offsetting the selected tap numbers based on a number of transitions in each sample set of the plurality of sample sets, wherein the controller offsets each tap number by an offset equal to (S−X+1)T, where S is a sample set number for the tap number, X is a transition number for the tap number, and T comprises a width of the bit time in the data stream based on a number of taps sampling each bit in the data stream; and
an accumulator, the accumulator receiving the offset selected tap numbers and accumulating the offset selected tap numbers to determine which of the plurality of samples in the data stream correspond to an eye center of bits in the data stream.
8. The eye center trainer of claim 7 wherein the accumulator outputs an accumulated value, and wherein higher order bits in the accumulated value determine which of the plurality of samples in the data stream correspond to the eye center of bits in the data stream.
9. An eye center trainer, the eye center trainer comprising:
a sampler, the sampler receiving a data stream including a plurality of transitions, the sampler taking a plurality of samples of the data stream;
a controller, the controller receiving a plurality of sample sets from the plurality of samples, each sample set corresponding to a bit time in the data stream, the controller locating transitions identified in the plurality of sample sets, the controller selecting tap numbers corresponding to the located transitions in the plurality of sample sets, the controller determining if the located transitions is a first transition in a sample set or a second transition in a sample set, the controller offsetting the tap numbers corresponding to the located transitions based on whether the corresponding located transition is on time, early or late as determined by the located transition's sample set; and
an accumulator, the accumulator receiving the offset tap numbers and accumulating the offset tap numbers, the accumulator outputting an accumulated value, and wherein higher order bits in the accumulated value locates an eye center of bits in the data stream to determine which of the plurality of samples in the data stream are used to recover data from the data stream.
10. The eye center trainer of claim 9 wherein the plurality of samples comprises a selected subset of samples for a data stream having a first data rate, and wherein the plurality of samples comprises all samples for a data stream having a second data rate faster than the first data rate.
11. The eye center trainer of claim 9 wherein the eye center trainer further comprises a sample FIFO, the sample FIFO receiving the plurality of samples from the sampler and passing the plurality of samples to the controller.
12. The eye center trainer of claim 9 wherein the eye center trainer further comprises a data FIFO, the data FIFO receiving the plurality of samples and selecting one of the plurality of samples per bit time based on the higher order bits in the accumulated value.
13. The eye center trainer of claim 9 wherein the accumulator is seeded with an initial value selected to shift the accumulator output to the eye center of bits in the data stream and to compensate for a difference between adjacent samples in the plurality of samples.
14. The eye center trainer of claim 9 wherein the controller offsets each tap number that correspond to early transitions by 0, tap numbers that correspond to on time transitions by T, and late transitions by 2T, where T comprises a width of bit time in the data stream based on a number of taps sampling each bit in the data stream.
15. A method of locating an eye center of bits in a data stream, the method comprising:
sampling the data stream to generate a plurality of samples, the plurality of samples comprising a plurality of sample sets, each of the plurality of sample sets corresponding to a bit time in the data stream;
locating transitions identified in the plurality of sample sets;
selecting tap numbers corresponding to the located transitions;
offsetting the selected tap numbers based on a number of transitions in a corresponding sample set; and
accumulating the offset tap numbers to determine which of the plurality of samples in the data stream correspond to the eye center of the bits in the data stream by locating the eye center of bits in the data stream using higher order bits of an accumulated output.
16. The method of claim 15 wherein the plurality of sample sets comprises a first sample set, a second sample set, a third sample set, and a fourth sample set, and wherein the step of offsetting the selected tap numbers based on a number of transitions in a corresponding sample set comprises offsetting a first tap number in the first sample set by an offset T, and offsetting a second tap number by an offset 0 if the second tap number is in the first sample set, and offsetting the second tap number by the offset T if the second tap number is in the second sample set, and offsetting the second tap number by an offset 2T if the second tap number is in the third sample set, and wherein the step of offsetting the selected tap numbers based on a number of transitions in a corresponding sample set further comprises offsetting a third tap number by the offset 0 if the third tap number is in the second sample set, and offsetting the third tap number by the offset T if the third tap number is in a third sample set, and offsetting the third tap number by the offset 2T if the third tap number is in the fourth sample set, and wherein the step of offsetting the selected tap numbers based on a number of transitions in a corresponding sample set further comprises offsetting a fourth tap number by the offset 0 if the fourth tap number is in the third sample set, and offsetting the fourth tap number by the offset T if the fourth tap number is in the fourth sample set, and offsetting the fourth tap number by the offset 2T if the fourth tap number is in a fifth sample set, where T comprises a width of a bit time in the data stream based on a number of taps sampling each bit in the data stream.
17. A method of locating an eye center of bits in a data stream, the method comprising:
sampling the data stream to generate a plurality of samples, the plurality of samples comprising a plurality of sample sets, each of the plurality of sample sets corresponding to a bit time in the data stream;
locating transitions identified in the plurality of sample sets;
selecting tap numbers corresponding to the located transitions;
offsetting the selected tap numbers based on a number of transitions in a corresponding sample set, wherein offsetting the selected tap numbers based on a number of transitions in a corresponding sample set comprises offsetting each tap number that correspond to an early transition by 0, each tap number that corresponds to an on time transition by T, and each tap number that corresponds to a late transition by 2T; and
accumulating the offset tap numbers to determine which of the plurality of samples in the data stream correspond to the eye center of the bits in the data stream.
18. A method of locating an eye center of bits in a data stream, the method comprising:
sampling the data stream to generate a plurality of samples, the plurality of samples comprising a plurality of sample sets, each of the plurality of sample sets corresponding to a bit time in the data stream;
locating transitions identified in the plurality of sample sets;
selecting tap numbers corresponding to the located transitions;
offsetting the selected tap numbers based on a number of transitions in a corresponding sample set, wherein offsetting the selected tap numbers based on a number of transitions in a corresponding sample set comprises offsetting each tap number by an offset equal to (S−X+1)T, where S is a sample set for the tap number, X is a transition number for the tap number, and T comprises a width of a bit time in the data stream based on a number of taps sampling each bit in the data stream; and
accumulating the offset tap numbers to determine which of the plurality of samples in the data stream correspond to the eye center of the bits in the data stream.
19. The method of claim 18 wherein accumulating the offset tap numbers to determine which of the plurality of samples in the data stream correspond to the eye center of bits in the data stream comprises seeding with an initial value selected to shift an output to identify the plurality of samples in the data stream corresponding to the eye center of bits in the data stream.
20. The method of claim 18 wherein accumulating the offset tap numbers to determine which of the plurality of samples in the data stream correspond to the eye center of bits in the data stream comprises locating the eye center of bits in the data stream using higher order bits of an output.