Blind and decision directed multi-level channel estimation

1. A system for blind channel estimation, comprising:

an interface configured to:

receive a read value that is read from a multi-level storage device;

receive a set of bins having bin ranges; and

for each of the bins in the set, receive a corresponding portion of read values which fall into that particular bin; and

a processor configured to adjust one or more of the bin ranges such that the received portions of read values remain substantially the same after adjustment and after assignment of the read value to one of the set of bins after adjustment, wherein:

the processor is configured to adjust a bin range by incrementing or decrementing a previous bin range by k*μ, where k is a positive integer and μ is associated with a step size; and

a first read value is read prior to a second read value and the processor is further configured to set k to a larger value for the second read value compared to the first read value.

2. The system of claim 1 , wherein the processor is further configured to determine whether, prior to adjustment, one or more of the portions of read values would remain substantially the same based at least in part on which pre-adjustment bin in the set the read value falls into.

3. The system of claim 1 further comprising a channel corrector configured to perform channel correction on the read value using at least some of the one or more bin ranges that are adjusted.

4. The system of claim 1 , wherein:

the multi-level storage device has M levels;

there are M+1 bins and M bin ranges corresponding to centers of the M levels; and

the portion of read values which fall into a lowest bin and a highest bin in the set is substantially equal to 1/(2M) and the portion of read values which fall into other bins in the set is substantially equal to 1/M.

5. The system of claim 1 , wherein the processor is further configured to determine a value for μ by sweeping μ and selecting μ based at least in part on error rate.

6. A system for blind channel estimation, comprising:

an interface configured to:

receive a first read value that is read from a multi-level storage device;

receive a set of bins having bin ranges;

for each of the bins in the set, receive a corresponding portion of read values which fall into that particular bin; and

receive a second read value which is read from a reference cell in the multi-level storage device;

a processor configured to:

adjust one or more of the bin ranges such that the received portions of read values remain substantially the same after adjustment and after assignment of the read value to one of the set of bins after adjustment; and

adjust one or more of the bin ranges such that the portions of read values remain substantially the same after adjustment and after assignment of the second read value to at least some of the one or more bin ranges that are adjusted, wherein:

the processor is configured to adjust, for the first read value, by incrementing or decrementing a previous bin range by k*μ 1 , where k is a positive integer and μ 1 is associated with a first step size; and

the processor is configured to adjust, for the second read value, by incrementing or decrementing a previous bin range by k*μ 2 , where k is a positive integer, μ 2 is associated with a second step size and μ 2 >μ 1 .

7. The system of claim 1 , wherein the processor is further configured to determine a number of read values to obtain from the multi-level storage device based at least in part on a number of levels stored by the multi-level storage device.

8. The system of claim 1 , wherein the processor is further configured to scramble the read value to obtain a scrambled value and determine and adjust using the scrambled value.

9. A method for blind channel estimation, comprising:

receiving a read value that is read from a multi-level storage device;

receiving a set of bins having bin ranges;

for each of the bins in the set, receiving a corresponding portion of read values which fall into that particular bin; and

adjusting, using a processor, one or more of the bin ranges such that the received portions of read values remain substantially the same after adjustment and after assignment of the read value to one of the set of bins after adjustment, wherein:

adjusting a bin range using the processor includes incrementing or decrementing a previous bin range by k*μ, where k is a positive integer and μ is associated with a step size; and

a first read value is read prior to a second read value and the processor is further configured to set k to a larger value for the second read value compared to the first read value.

10. A method for blind channel estimation, comprising:

receive a first read value that is read from a multi-level storage device;

receive a set of bins having bin ranges;

for each of the bins in the set, receive a corresponding portion of read values which fall into that particular bin; and

receive a second read value which is read from a reference cell in the multi-level storage device;

using a processor to:

adjust one or more of the bin ranges such that the received portions of read values remain substantially the same after adjustment and after assignment of the read value to one of the set of bins after adjustment; and

adjust one or more of the bin ranges such that the portions of read values remain substantially the same after adjustment and after assignment of the second read value to at least some of the one or more bin ranges that are adjusted, wherein:

the processor is configured to adjust, for the first read value, by incrementing or decrementing a previous bin range by k*μ 1 , where k is a positive integer and μ 1 is associated with a first step size; and

the processor is configured to adjust, for the second read value, by incrementing or decrementing a previous bin range by k*μ 2 , where k is a positive integer, μ 2 is associated with a second step size and μ 2 >μ 1 .

11. The method of claim 10 further comprising determining whether, prior to adjustment, one or more of the portions of read values would remain substantially the same based at least in part on which pre-adjustment bin in the set the read value falls into.

12. The method of claim 10 further comprising performing channel correction on the read value using at least some of the one or more bin ranges that are adjusted.

13. The method of claim 10 , wherein:

the multi-level storage device has M levels;

there are M+1 bins and M bin ranges corresponding to centers of the M levels; and

the portion of read values which fall into a lowest bin and a highest bin in the set is substantially equal to 1/(2M) and the portion of read values which fall into other bins in the set is substantially equal to 1/M.

14. The method of claim 10 further comprising determine a value for at least one of μ 1 or μ 2 by sweeping at least one of μ 1 or μ 2 and selecting at least one of μ 1 or μ 2 based at least in part on error rate.

15. The method of claim 10 further comprising determining a number of read values to obtain from the multi-level storage device based at least in part on a number of levels stored by the multi-level storage device.

16. The method of claim 10 further comprising scrambling the read value to obtain a scrambled value, wherein determining and adjusting use the scrambled value.