Determining soft data from a hard read
Apparatuses and methods involving the determination of soft data from hard reads are provided. One example method can include determining, using a hard read, a state of a memory cell. Soft data is determined based, at least partially, on the determined state.
1. A method, comprising:
determining, using a hard read, a state of a memory cell; and
determining soft data based, at least partially, on the determined state of the memory cell;
wherein the soft data includes:
a first log-likelihood ratio (LLR) if a digit at a position in a data value corresponding to the determined state has a same value as a digit at the same position in a data value corresponding to a neighboring state; and
a second LLR if the digit at the same position in the data value corresponding to the neighboring state does not have the same value; and
wherein the first LLR is different than the second LLR.
2. The method of claim 1 , further comprising refining the soft data based, at least in part, on a condition of a memory array including the memory cell.
3. The method of claim 1 , wherein determining the state of the memory cell comprises sensing the state of the memory cell.
4. The method of claim 1 , wherein determining the state of the memory cell comprises reading a data value stored by the memory cell.
5. The method of claim 1 , wherein determining the state of the memory cell comprises determining which state the memory cell is currently in.
6. The method of claim 1 , wherein determining the state of the memory cell comprises determining which group of states the memory cell is currently in.
7. The method of claim 1 , wherein determining the state of the memory cell comprises determining which states a combination of memory cells are currently in.
8. The method of claim 1 , wherein the soft data is based, at least partially, on whether the digit at the position in the data value corresponding to the determined state has the same value as the digit at the same position in the data value corresponding to the neighboring state.
9. The method of claim 8 ,
wherein the first LLR is greater than the second LLR, and
wherein the signs of the first and second LLRs are determined from the hard read of the memory cell.
10. The method of claim 8 , wherein determining the soft data comprises weighting the soft data corresponding to the value of the digit in the data value corresponding to the determined state and the value of the digit in the data value corresponding to the neighboring state.
11. The method of claim 10 , wherein the soft data corresponding to the value of the digit in the data value corresponding to the determined state is more heavily weighted than the soft data corresponding to the value of the digit in the data value corresponding to the neighboring state.
12. The method of claim 10 , wherein determining the soft data comprises weighting the soft data based, at least partially, on a condition of the memory cell.
13. The method of claim 1 , wherein the soft data is data value probabilities.
14. The method of claim 1 , wherein the soft data is based, at least partially, on digits at the same position in respective data values corresponding to neighboring states.
15. The method of claim 14 , wherein the soft data is based, at least partially, on whether the digits at the same position in the respective data values corresponding to the neighboring states have the same value as a digit at the same position in a data value corresponding to the determined state.
16. The method of claim 15 , wherein the soft data includes:
a first log-likelihood ratio (LLR) if the digit at the same position in the data value corresponding to each of the neighboring states have the same value; and
a second LLR if the digit at the same position in the data value corresponding to at least one of the neighboring states has the same value as the digit at the position in the data value corresponding to the determined state and the digit at the same position in the data value corresponding to at least one of the neighboring states does not have the same value as the digit at the position in the data value corresponding to the determined state,
wherein the first LLR is greater than the second LLR, and
wherein the signs of the first and second LLRs are determined from the hard read of the memory cell.
17. The method of claim 16 , wherein the soft data includes a third LLR if the digit at the same position in the data value corresponding to each of the neighboring states do not all have the same value,
wherein the first LLR is greater than the third LLR, and the third LLR is greater than the second LLR.
18. The method of claim 17 , wherein the soft data includes a null LLR if the data value of the determined state is determined without comparison of the digits at positions in the data value of the determined state to the digits at the same positions in the data value corresponding to each of the neighboring states, wherein the first LLR is greater than the null LLR, the null LLR is greater than the third LLR, and the third LLR is greater than the second LLR.
19. The method of claim 1 , wherein determining the soft data comprises retrieving the soft data from a data structure based, at least partially, on the determined state of the memory cell.
20. The method of claim 19 , wherein retrieving the soft data from a data structure comprises retrieving the soft data from a look-up table.
21. A method, comprising:
receiving a data value sensed from a memory cell, wherein the data value corresponds to a sensed state of the memory cell; and
determining soft data from the data value, where the soft data is based, at least partially, on whether a digit at a position in the data value has a same value as a digit at a same position in the data value corresponding to a neighboring state of the sensed state;
wherein the soft data includes:
a first log-likelihood ratio (LLR) when the digit has the same value as the digit corresponding to the neighboring state; and
a second LLR when the digit does not have the same value as the digit corresponding to the neighboring state; and
wherein the first LLR is different than the second LLR.
22. The method of claim 21 , wherein determining soft data includes retrieving the soft data from a data structure based, at least partially, on the determined state of the memory cell.
23. The method of claim 22 , wherein retrieving the soft data includes retrieving the soft data from a look-up table.
24. The method of claim 21 , wherein determining soft data includes computing the soft data using a controller.
25. The method of claim 21 , wherein receiving the data value sensed from the memory cell includes receiving only the sensed state of the memory cell via a hard read.
26. The method of claim 21 , wherein receiving the data value sensed from the memory cell includes receiving the sensed state of the memory cell and receiving sensed soft data with the sensed state.
27. The method of claim 26 , wherein determining soft data from the data value includes adjusting the sensed soft data based, at least in part, on the determined soft data.
28. A method, comprising:
determining, using a soft read, a state of a memory cell; and
determining soft data based, at least partially, on the determined state of the memory cell with respect to states neighboring the determined state;
wherein the soft data includes:
a first log-likelihood ratio (LLR) if a digit at a position in a data value corresponding to a determined state has the same value as at least one digit at the same position in a data value corresponding to at least one neighboring state of the states neighboring the determined state; and
a second LLR if the digit at the position in a data value corresponding to the determined state does not have the same value as the at least one digit at the same position in the data value corresponding to the at least one neighboring state; and
wherein the first LLR is different than the second LLR.
29. An apparatus, comprising
an array of memory cells; and
a controller configured to:
receive a determined state of a memory cell of the array of memory cells, and
determine soft data based, at least partially, on the determined state of the memory cell;
wherein the soft data includes:
a first log-likelihood ratio (LLR) if a digit at a position in a data value corresponding to a determined state has the same value as a digit at the same position in a data value corresponding to a neighboring state; and
a second LLR if the digit at the position in a data value corresponding to the determined state does not have the same value as the digit at the same position in the data value corresponding to the neighboring state; and
wherein the first LLR is different than the second LLR.
30. The apparatus of claim 29 , wherein the controller is an internal controller communicatively coupled to the array of memory cells.
31. The apparatus of claim 29 , wherein the controller is an external controller communicatively coupled to a memory device, the memory device including the array of memory cells.
32. The apparatus of claim 29 , wherein
the first LLR is greater than the second LLR, and
wherein the signs of the first and second LLRs are determined from the hard read of the memory cell.
33. The apparatus of claim 32 , wherein the soft data further includes a third LLR if the digit at the position of the data value corresponding to the determined state has the same value as the digit at the same position in the data values corresponding to a plurality of neighboring states,
wherein the third LLR is greater than the first LLR.
34. The apparatus of claim 33 , wherein the soft data further include a null LLR if the data value is determined without comparison of the digits at positions in the data value of the determined state to the digits at the same positions in the data value corresponding to the neighboring states, wherein the first LLR is greater than the null LLR, and the null LLR is greater than the second LLR.
35. A method for determining soft data, comprising:
determining a data value corresponding to a sensed state of a memory cell;
comparing a digit of the determined data value to a digit at a same position in data values of states neighboring the sensed state; and
computing a log-likelihood ratio (LLR) for the digit of the determined data value based on the comparison;
wherein computing the LLR includes computing a relatively larger LLR when the digit of the determined data value is in agreement with more digits at a same position in the data value corresponding to the at least one neighboring state than for the digit of the determined data value being in agreement with fewer digits at a same position in the data value corresponding to the at least one neighboring state.
36. The method of claim 35 , wherein
the sign of the LLR is determined from a hard read of the memory cell.
37. The method of claim 35 , wherein computing the LLR includes:
determining soft data as a largest LLR for the digit in the determined data value being the same as all digits at a same position in the data value corresponding to the at least one neighboring state;
determining soft data as an intermediate value LLR for the digit in the determined data value being the same as some but not all digits at the same position in the data value corresponding to the at least one neighboring state; and
determining soft data as a smallest LLR for the digit in the determined data value being different than all digits at the same position in the data value corresponding to the at least one neighboring state,
wherein the sign of the LLR is determined from a hard read of the memory cell.
38. The method of claim 37 , wherein the largest LLR is greater than an LLR for the digit in the determined data value prior to comparison to a digit at a same position in a data value corresponding to at least one neighboring state.
39. The method of claim 38 , wherein the largest LLR is greater than an LLR for a digit in a data value determined without completely determining a singular state of the memory cell.
40. A method, comprising:
determining, using a hard read, a state of a memory cell; and
determining soft data based, at least partially, on the determined state of the memory cell;
wherein:
the soft data is based, at least partially, on:
digits at the same position in respective data values corresponding to neighboring states; and
whether the digits at the same position in the respective data values corresponding to the neighboring states have the same value as a digit at the same position in a data value corresponding to the determined state;
the soft data includes:
a first log-likelihood ratio (LLR) if the digit at the same position in the data value corresponding to each of the neighboring states have the same value; and
a second LLR if the digit at the same position in the data value corresponding to at least one of the neighboring states has the same value as the digit at the position in the data value corresponding to the determined state and the digit at the same position in the data value corresponding to at least one of the neighboring states does not have the same value as the digit at the position in the data value corresponding to the determined state;
the first LLR is greater than the second LLR; and
the signs of the first and second LLRs are determined from the hard read of the memory cell.
41. A method for determining soft data, comprising:
determining a data value corresponding to a sensed state of a memory cell;
comparing a digit of the determined data value to a digit at a same position in data values of states neighboring the sensed state; and
computing a log-likelihood ratio (LLR) for the digit of the determined data value based on the comparison;
wherein computing the LLR includes:
determining soft data as a largest LLR for the digit in the determined data value being the same as all digits at a same position in the data value corresponding to the at least one neighboring state;
determining soft data as an intermediate value LLR for the digit in the determined data value being the same as some but not all digits at the same position in the data value corresponding to the at least one neighboring state; and
determining soft data as a smallest LLR for the digit in the determined data value being different than all digits at the same position in the data value corresponding to the at least one neighboring state; and
wherein the sign of the LLR is determined from a hard read of the memory cell.