This application claims benefit under 35 USC 119, of Italian Application no. MI2004A 001904, filed on Oct. 7, 2004.
The present invention relates to a programming approach for a nonvolatile electronic device, such as flash memory, that ensures a desired final destination for a cell is reached even when a cell may not be in an ideal distribution.
Logical information associated with a flash memory cell is related to its physical size. In a single level device, a cell is said to be erased (logic level “1”) when few electrons are stored in the floating gate. Vice versa, a cell is said to be programmed (logic level “0”) when ‘enough’ electrons are stored in the floating gate. Depending on how many electrons are trapped in the floating gate, the gate threshold of the cell changes (i.e., when more electrons are trapped, the threshold is higher). Endurance and retention are two parameters often used to describe the quality of a flash memory cell. Endurance refers to the capability of the cell to maintain stored information after erase/program/read cycling, while retention refers to the capability of a cell to keep stored information over time.
In single level flash memory, each cell after a modify operation has a threshold higher than a program voltage threshold, Vtp, or lower than an erase voltage threshold, Vte. By comparing the current sunk from a read cell and a reference driven in the same condition, the information of the selected cell is established. A light endurance problem exists if some programmed cells lose a little charge from the floating gate. However, if these cells assume a threshold slightly lower than Vtp but higher enough of Vt_read, flash functionality is still good. If a cell is in this condition and programming is required on it, the program verify is better with a reference with a threshold Vtp. In this case, the cell will be verified as a logical “1” (it has a threshold lower than the reference Vtp used), and so it will be programmed. After this operation, the cell will have a threshold higher of Vtp and this means that the cell has been recovered. This operation is not user visible. In fact, if, before programming, the user reads the cell, the cell is evaluated as programmed. However, by using Vtp as reference of the verify, the cell is recovered and the recovery has no drawbacks.
In a multilevel flash device, the situation is not the same, and the choice of reference for verify before programming is more complex and can cause wrong operations in the case of cells with a threshold not exactly inside target distributions.
The threshold situation for multilevel flash memory cells is represented by the graph 100 shown in
For the physical situation of distribution in a multilevel flash memory, a program algorithm can “move” the threshold of a cell towards higher values. Thus, it is physically possible to move a cell from distribution “11” to “10” or from “10” to “01” or from “11” to “00”. It is not physically permitted to move from “01” to “10”, for example. A possible program algorithm can follow the physics of the memory. Thus, if a user gives a pattern “01” on a cell in “10” distribution, the operation is physically correct and the final destination of the selected cell will be “01”. On the other hand, if a cell is in “01” and the user gives as a desired pattern “10”, this is not a physically correct request and the algorithm will not move the threshold of the cell.
Thus, for a physical approach to programming of a cell, the final destination of the cell will coincide always with the pattern (data to program) desired if the transition is physically possible. Otherwise, the cell will not be programmed at all. This is true for single level cells, as well. Under this approach, since the final destination coincides with the pattern required if a modify occurs, the first verify step before programming has to establish the absolute pattern on the selected cell. The best way to establish this position is to use Vtp0, Vtp1, and Vtp2. In this way, if a cell is slightly out of distribution, it will be correctly recovered.
For physical approaches, software written for use with single level cells has to be changed if multilevel cells are substituted. As an example, if a user gives a pattern “01” in a single level flash device, this refers to two cells of memory with each cell being treated separately as single level cells. If these two cells are in “10” condition, to reach a pattern “01,” a pattern of “1” is required on a cell “0”, which is not physically possible. The transition of the cell is left at its original position of “0”. On the other cell, the pattern required is “0” on a “1” cell, which is possible. Thus, the final destination will be “0”. The overall result is that with a starting situation “10” and a required pattern “01”, the actual final destination is “00”. In contrast, for a multilevel cell in “10”, the required pattern “01” will be the final destination, as this is a possible transition. A graph 200 of
Another approach to programming permits the management of single level and multilevel flash by the same software. This approach is referred to as a “logical” approach. Under this approach, even if flash has a multilevel structure, the user can use it as a single level one. If a pattern is given to program a bit at “1”, the corresponding bit does not need to be modified. If a pattern is given to program a bit to “0”, the corresponding bit has to be “0” after programming. Thus, if a cell is in distribution “10” and the pattern to program is “01”, the requested final distribution of the cell is “00”. If the cell is in distribution “01” and the pattern to program is “10”, the final position of the cell has to be “00”. Graph 210 of
Many different methods to program cells are known in the prior art that try to move cell thresholds with a high precision and with a relatively fast execution time. As mentioned previously, physical approaches require dedicated software for each of single level and multilevel devices. Logical approaches use the same programming for single and multilevel devices. However, these prior art logical approaches assume that cells are in their ideal distributions (i.e., there is no retention problem). When there is a light charge losing and some cells are lightly out of their distributions, the device can work correctly but programming on these cells can fail.
Accordingly, a need exists for a programming approach that ensures a desired final destination for a cell is reached even when a cell may not be in an ideal distribution. The present invention addresses such a need.
Aspects for programming a nonvolatile electronic device are described. The aspects include performing an initial verify step of a programming algorithm with an initial type of reference voltage value, and performing one or more subsequent verify steps in the programming algorithm with a second type of reference voltage value. Further included is utilizing a read reference voltage for the initial verify step.
Through the present invention, a straightforward and effective logical approach to programming is achieved that ensures desired programming for a cell that falls out of ideal distribution. These and other advantages of the aspects of the present invention will be more fully understood in conjunction with the following detailed description and accompanying drawings.
The present invention relates to a programming approach for a nonvolatile electronic device. The following description is presented to enable one of ordinary skill in the art to make and use the invention and is provided in the context of a patent application and its requirements. Various modifications to the preferred embodiment and the generic principles and features described herein will be readily apparent to those skilled in the art. Thus, the present invention is not intended to be limited to the embodiments shown but is to be accorded the widest scope consistent with the principles and features described herein.
In accordance with the present invention, a programming approach for a nonvolatile electronic device, such as flash memory, is provided that ensures a desired final destination for a cell is reached even when a cell may not be in an ideal distribution. In order to perform the programming, a microcontroller is used in a flash memory to manage modify operations, including programming. Referring to
In a traditional approach to programming, a user read access is made by read reference voltage values (Vr0, Vr1, Vr2), but first verify is made by setting program reference voltage values (Vp0, Vp1, Vp2). The information evaluated after these two accesses will coincide if all cells are within ideal distributions. In this case, if a cell in fact has, for example, a threshold higher than Vr0, the threshold will also be higher than Vp0. However, if a cell is not in an ideal distribution, e.g., it has lost retention of a little charge, the user evaluation can differ from the microcontroller first verify, and the final destination could be different from that expected by a user.
The following example shows a wrong programming if not all distributions are ideal, i.e., if some cells lose a little charge, and the traditional approach is used. Referring to
When the first verify is made using Vp0, Vp1, and Vp2, the cell is, in fact, evaluated to be in “11” distribution as its threshold is lower than Vtp0. As pattern chosen by user is “01”, the final destination of the cell will be “01” and not “00” as hoped by the user. The wrong transition is shown by arrow 410 in
In accordance with the present invention, the first verify in the programming algorithm stored on device 210 (
It should be appreciated that the example described for
Although the present invention has been described in accordance with the embodiments shown, one of ordinary skill in the art will readily recognize that there could be variations to the embodiments and those variations would be within the spirit and scope of the present invention. Accordingly, many modifications may be made by one of ordinary skill in the art without departing from the spirit and scope of the appended claims.
Number | Date | Country | Kind |
---|---|---|---|
MI2004A1904 | Oct 2004 | IT | national |
Number | Name | Date | Kind |
---|---|---|---|
5497119 | Tedrow et al. | Mar 1996 | A |
5509134 | Fandrich et al. | Apr 1996 | A |
5546042 | Tedrow et al. | Aug 1996 | A |
5615165 | Tanaka et al. | Mar 1997 | A |
5757699 | Takeshima et al. | May 1998 | A |
5765184 | Durante | Jun 1998 | A |
5809541 | Fandrich et al. | Sep 1998 | A |
5937424 | Leak et al. | Aug 1999 | A |
6134140 | Tanaka et al. | Oct 2000 | A |
6282145 | Tran et al. | Aug 2001 | B1 |
20060028875 | Avraham et al. | Feb 2006 | A1 |
Number | Date | Country | |
---|---|---|---|
20060077714 A1 | Apr 2006 | US |