Pipelining in a memory

Pipelining overlapping programming pulses addresses inefficiencies in memory cell programming by minimizing ramp-up times, improving write performance and throughput in memory devices.

DE112014001203B4Active Publication Date: 2024-12-05INFINEON TECHNOLOGIES LLC
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
DE112014001203
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Priority Date
2013-03-08
Filing Date
2014-03-06
Publication Date
2024-12-05
Estimated Expiration
2034-03-06

AI Technical Summary

Technical Problem

Existing memory cell programming methods suffer from inefficient programming performance due to significant ramp-up times caused by metallic interference, supply voltage limitations, and power consumption constraints, leading to poor write performance and reduced throughput.

Method used

Implement pipelining by partially overlapping programming pulses to minimize ramp-up duration, considering metallic interference, supply voltage limitations, and power consumption, allowing simultaneous programming of multiple memory cells while avoiding structural damage.

Benefits of technology

Enhances programming efficiency and increases throughput by reducing ramp-up times, enabling faster and more efficient programming of memory cells while maintaining control over power consumption and avoiding interference.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

A system that includes the following: a memory cell arrangement (100, 400) that includes a plurality of memory cells (W1-1, ... W1-N); and a programmer (401) configured to generate multiple consecutive programming pulses (P1, P2, P3) configured to program target memory cells from the plurality of memory cells (W1-1, ... W1-N), wherein the multiple consecutive programming pulses (P1, P2, P3) overlap during an overlap duration, the overlap duration being configured to be adjusted based on a performance parameter of the cell array (W1-1, ... W1-N), wherein, during the overlap period, a ramp-up period of a bit line (B3, B9, B15) associated with programming a subsequent target memory cell (W1-3, W1-9, W1-15) substantially overlaps a steady-state period of a bit line (B1, B7, B13) associated with programming a current target memory cell (W1-1, W1-7, W1-13), the system also includes: a ramp control circuit (410) configured to receive a variety of high-ramp voltages (V1, V2, V3) and, during the high-ramp period, to drive a variety of high-ramp stages of the bit line (B3, B9, B15) associated with programming the subsequent target memory cell (W1-3, W1-9, W1-15) by applying a respective variety of high-ramp voltages (V1, V2, V3), wherein the ramp control circuit (410) is configured to receive the plurality of high ramp voltages (V1, V2, V3) from respective independent voltage supplies (421, 422, 423), each independent voltage supply (421, 422, 423) having a respective controller.
Need to check novelty before this filing date? Find Prior Art

Description

GENERAL STATE OF THE ART

[0001] The present disclosure generally relates to pipelining in memory that allows pre-charging or partial operation at a next target location in memory while operating at a current target location in memory. Operation at a target location may include programming data to the target location or erasing data from the target location. For example, as discussed below, the above pipelining improves the throughput associated with the memory.

[0002] US 20020041527 A1 discloses a non-volatile memory with a shortened total write time, capable of stably writing data by maintaining a constant write current while reducing fluctuations in a voltage generated by a booster circuit. In a non-volatile memory, such as a flash memory, the data is determined at the time of the write operation. While skipping a bit corresponding to write data with a logic "1" (or logic "0"), the write operation is performed sequentially for bits corresponding to write data with a logic "0" (or logic "1").

[0003] US 20060291291 A1 discloses a non-volatile semiconductor memory device comprising a memory cell array and a sense amplifier, wherein the device is internally controlled to execute a write sequence comprising write pulse applications and write verification operations that are repeated to write a set of memory cells selected in the memory cell array. Here, the sense amplifier performs a write speed check to detect the write speed of a plurality of memory cells to be written to a specific data state after a specific write pulse application at the beginning of the write sequence, thereby obtaining discrimination data for classifying the plurality of memory cells into first and second cell groups.After the write speed check, the first and second cell groups are alternately written under different write conditions with reference to the discrimination data. SHORT OVERVIEW

[0004] Systems, methods, and computer program products for pipelining are described herein. The present invention relates to a system according to claim 1 and a method according to claim 7. Advantageous embodiments may include features of dependent claims.

[0005] In one embodiment of the present disclosure, a system includes a first memory cell, a second memory cell, and a programming device. The programming device is configured to generate respective first and second programming pulses to program the first and second memory cells. The first and second programming pulses overlap for an overlap duration, wherein the overlap duration is adjustable based on a performance parameter of the first and second memory cells.

[0006] The present disclosure also provides a method including generating respective first and second programming pulses to program the first and second memory cells, wherein the first and second programming pulses overlap for an overlap duration that can be adjusted based on a performance parameter of the first and second memory cells, and programming the first and second memory cells using the first and second programming pulses.

[0007] Further features and advantages of the disclosure, as well as the structure and operation of various embodiments of the invention, are described in detail below with reference to the accompanying drawings. It should be understood that the disclosure is not limited to the specific embodiments described herein. Such embodiments are presented here for illustrative purposes only. Those skilled in the relevant art(s) will recognize additional embodiments based on the teachings contained herein. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] The accompanying drawings, which are incorporated herein and constitute a part of the specification, illustrate embodiments of the present invention and, together with the description, further serve to explain the principles of the invention and to enable any person skilled in the relevant art(s) to understand and use embodiments of the invention. Fig. 1 illustrates an exemplary memory sector including memory cells in the form of non-volatile memory cells according to an embodiment of the present disclosure. Fig. Figure 2 illustrates programming pulses used to program identified memory cells in a memory. Fig. 3 illustrates the programming of the identified memory cells. Fig. 4 illustrates an electrical system used for pipelining during programming of memory cells, according to an embodiment of the present disclosure. Fig. 5 illustrates an exemplary bitline driver according to an embodiment of the present disclosure. Fig. 6 illustrates exemplary programming of identified memory cells using pipelining according to an embodiment of the present disclosure. Fig. 7 further illustrates exemplary programming of the identified memory cells using pipelining according to an embodiment of the present disclosure. Fig. 8 illustrates a charge pump according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0009] This specification discloses one or more embodiments incorporating the features of this invention. The disclosed embodiment(s) merely exemplify the invention. The scope of the invention is not limited to the disclosed embodiment(s). The invention is defined by the claims appended hereto.

[0010] The described embodiment(s) and references in the specification to "one embodiment," "an exemplary embodiment," etc., indicate that the described embodiment(s) may include a particular feature, structure, or characteristic, but not every embodiment necessarily includes that particular feature, structure, or characteristic. Furthermore, such terms do not necessarily refer to the same embodiment. Further, it should be understood that if a particular feature, structure, or characteristic is described in connection with one embodiment, it is within the skill of one skilled in the art to effect such feature, structure, or characteristic in connection with other embodiments, whether expressly described or not.

[0011] Fig. 1 illustrates an exemplary memory sector 100 according to an embodiment of the present disclosure, including memory cells, e.g., in the form of non-volatile memory cells. The memory sector 100 includes a plurality of word lines W1...WN, where N is an integer. The word lines W1...WN are connected to respective gates of M-number of memory cells in the form of non-volatile memory cells, where M is an integer. For example, word line W1 is connected to the respective gates of the non-volatile memory cells W1-1, W1-2...W1-M, word line W2 is connected to the respective gates of the non-volatile memory cells W2-1, W2-2...W2-M, and so on. In this way, each word line is connected to M-number of memory cells. Furthermore, the memory sector 100 includes a plurality of bit lines B1...BM connected to respective sources and drains of N-number of the non-volatile memory cells.For example, bit line B1 is connected to respective drains of floating-gate transistors W1-1, W2-1...WN-1 and is connected to respective sources of non-volatile memory cells W1-2, W2-2...WN-2. Bit line B2 is connected to respective drains of non-volatile memory cells W1-2, W2-2...WN-2 and is connected to respective sources of non-volatile memory cells W1-3, W2-3...WN-3, and so on. In one embodiment, the non-volatile memory cells may include floating-gate transistors.

[0012] In one example, operations on the memory cells included in memory sector 100 include programming data to the memory cells and / or erasing data from the memory cells. In a given sector, data can only be programmed to, or data can only be erased from, the memory cells that are connected to a single word line at a given time. The pipelining proposed here is applicable to programming data to, and also to erasing data from, memory cells.

[0013] The pipelining proposed here with respect to programming data into memory cells will now be discussed. The programming of memory cells may be performed using a method known as the hot electron injection method. In one example, memory cells may be programmed or written sequentially using the hot electron injection method. The memory cells may be divided into groups, and the groups may be programmed sequentially. For example, a controller (not shown) may identify the memory cells to be programmed and may divide the identified memory cells into groups. Once the groups are divided, the controller may determine the locations of all memory cells within the groups.For example, a first group G1 may include memory cells W1-1, W1-7 (not shown), and W1-13 (not shown); a second group G2 may include memory cells W1-3, W1-9 (not shown), and W1-15 (not shown); and a third group G3 may include memory cells W1-5, W1-11 (not shown), and W1-17 (not shown). To program the identified memory cells of a given group, a voltage associated with the word line connected to the identified target memory cells may initially be ramped up to a first voltage. Then, the voltages associated with the bit lines connected to the drains of the target memory cells are ramped up to a second voltage level, while the bit lines connected to the sources of the target memory cells are maintained at a very low voltage level (e.g., near ground).

[0014] The above voltage levels on the word line and bit lines enable the operation of the associated non-volatile memory cells. When the memory cells included in the first group G1, the second group G2, and the third group G3 are to be programmed, the controller determines the locations of the memory cells included in the first group G1, the second group G2, and the third group G3.

[0015] Fig. Figure 2 illustrates conventional programming pulses P1, P2, and P3 used to program respective memory cells, e.g., memory cells in the first group G1, the second group G2, and the third group G3. The programming pulses P1, P2, and P3 are generated sequentially to enable programming of the memory cells in the first group G1, the second group G2, and the third group G3 in succession. Specifically, the programming pulse P1 is activated at time t0 to enable programming of memory cells W1-1, W1-7, and W1-13 included in the first group G1, and deactivated at time t1 when memory cells W1-1, W1-7, and W1-13 have been programmed.The programming pulse P2 is then activated at time t1 to enable programming of memory cells W1-3, W1-9, and W1-15 included in the second group G2, and deactivated at time t2 when memory cells W1-3, W1-9, and W1-15 have been programmed. Finally, the programming pulse P3 is activated at time t2 to enable programming of memory W1-5, W1-11, and W1-17, and deactivated at time t3 when memory cells W1-5, W1-11, and W1-17 have been programmed. In this way, the identified memory cells in groups G1, G2, and G3 are sequentially enabled to be programmed one after the other.

[0016] Fig. 3 illustrates the conventional programming of the identified memory cells in further detail compared to Fig. 2. The description of Fig. 3 will again focus on the elements in Fig. 1. The voltage associated with word line W1 is initially ramped up to the first voltage level. Then, at time t0, voltages associated with bit lines B0, B6, and B12 (not shown), which are respectively connected to the sources of target memory cells W1-1, W1-7, and W1-13, are maintained at a very low voltage level, and programming pulse P1 is activated to ramp up the voltages associated with bit lines B1, B7 (not shown), and B13 (not shown), which are respectively connected to the drains of target memory cells W1-1, W1-7, and W1-13. When the voltages associated with bit lines B1, B7, and B13 reach a predetermined steady-state level (e.g., V w1-1), memory cells W1-1, W1-7, and W1-13 are programmed. The voltages associated with bit lines B1, B7, and B13 are then discharged. At time t1, after the target memory cells W1-1, W1-7, and W1-13 have been programmed, the programming pulse P1 can be deactivated. In addition, at time t1, voltages associated with bit lines B2, B8 (not shown), and B14 (not shown), which are respectively connected to the sources of the target memory cells W1-3, W1-9, and W1-15, are maintained at a very low voltage level, and the programming pulse P2 is activated to program target memory cells W1-3, W1-9, and W1-15 to ramp up the voltages associated with bit lines B3, B9 (not shown), and B15 (not shown), which are respectively connected to the drains of the target memory cells W1-3, W1-9, and W1-15. When the voltages associated with bit lines B3, B9, and B15 reach a predetermined steady-state level (e.g., V w1-3), memory cells W1-3, W1-9, and W1-15 are programmed. The voltages associated with bit lines B3, B9, and B15 are then discharged. At time t2, after the target memory cells W1-3, W1-9, and W1-15 have been programmed, the programming pulse P2 may be deactivated. The target memory cells W1-5, W1-11, and W1-17 included in group G3 are sequentially programmed in a manner similar to that described above with respect to the target memory cells included in groups G1 and G2.

[0017] In this example, as in Fig. As shown in Figure 3, each bit line B1, B3, B7, B9, B13, and B15 connected to the drains of the memory cells being programmed goes through three stages. These three stages include a ramp-up stage, a steady-state stage, and a discharge stage. During the ramp-up stage, the bit lines connected to the drains of the memory cells being programmed are charged to raise a voltage associated with the bit lines. Charging of the bit lines stops when the voltages reach a steady-state voltage after the steady-state stage begins. The memory cells are programmed during / at the end of the steady-state stage. The voltages associated with the bit lines are discharged during the discharge stage.

[0018] As in the conventional example Fig. As shown in Figure 3, the duration of a bit line's ramp-up stage extends over a significant portion of the associated programming pulse. This significant duration of the ramp-up stage can result in poor programming / write performance, as the programming pulses must be of a longer duration to ensure sufficient time for the steady-state stage to properly program the memory cells. It is desirable to minimize the duration of the ramp-up stage so that the programming pulses can be shorter and the memory cells can therefore be programmed more quickly. The duration of the ramp-up stage cannot always be minimized because this duration is based on inherent metallic noise effects of the non-volatile memory cells, limitations of the supply voltages used to charge the bit lines, and core cell technology requirements to avoid noise from the memory cell being programmed to other memory cells.Ideally, multiple memory cells would be programmed in parallel, but the available power supply capacities and control over the power supplies are very difficult to implement. Additional challenges in programming memory cells include the required address management of the multiple memory cells and the power consumption constraints associated with the device containing the memory. For all of the above reasons, sequential programming of the identified memory cells can be inefficient and may result in poor programming / writing performance of the memory cells.

[0019] Embodiments of the present disclosure describe pipelining during programming of the memory cell, for example, to improve the program / write performance and increase the throughput associated with the memory. In one embodiment, the pipelining includes partially overlapping the programming pulses, while simultaneously taking into account the inherent metallic interference effects of the non-volatile memory cells, limitations of the supply voltages used to charge the bit lines, power consumption constraints associated with the device incorporating the memory, and core cell technology requirements. The exemplary pipelining, whereby cells are connected to the same word line for programming, will now be described with reference to Fig. 4-6 described.

[0020] Fig. Figure 4 illustrates an electrical system 400 used for pipelining during programming of memory cells according to an embodiment of the present disclosure. In the example shown, the electrical system 400 includes the memory cells W1-1, W1-2, W1-3, W1-4...W1-M, a controller 401, a ramp control circuit 410, power supplies 421, 422, and 423, and respective bitline drivers 431, 432, 433, 434, and 435 connected to the respective bitlines of the memory cells to be programmed.

[0021] Power supplies 421, 422, and 423 provide respective voltages V1, V2, V3 to ramp control circuit 410 and / or each of bitline drivers 431-435. The relationship between voltages V1, V2, and V3 can be described as V1 < V2 < V3, where V3 can be substantially equal to the steady-state voltage associated with the bitlines of the memory cells to be programmed. Controller 401 provides control signals to ramp control circuit 410, which in turn provide ramp and discharge signals (e.g., low ramp, mid ramp, high ramp, bitline discharge) for each of bitline drivers 431-435 based on the control signals. The bit line drivers 431-435 charge and discharge the respective bit lines associated with the memory cells to be programmed based on the received voltages V1, V2 and V3 and the ramp and discharge signals.As described in further detail below, the ramp-up stage for each bit line includes three charging sections (e.g., a low-voltage section, a medium-voltage section, and a high-voltage section), with each section driven by a respective power supply. For example, Power Supply 1 421 drives / charges the bit line via the low-voltage section to voltage V1 (e.g., 1.5 V), Power Supply 2 422 drives / charges the bit line via the medium-voltage section to voltage V2 (e.g., 3 V), and Power Supply 3 423 drives / charges the bit line via the high-voltage section to voltage V3 (e.g., 4.5 V).

[0022] Fig. Figure 5 illustrates an exemplary bitline driver 500 according to an embodiment of the present disclosure. Each bitline driver 500 may include four transistor switches 501, 502, 503, and 504 for driving a given bitline through the ramp-up, steady-state, and discharge stages. Transistor switches 501, 502, and 503 may be used to charge the bitline during the ramp-up stage and to maintain the steady-state stage.For example, the drain of transistor 501 may be connected to power supply 1 421, and the source of transistor 501 may be connected to the bit line to be driven through the low-voltage portion of the ramp-up stage. The drain of transistor 502 may be connected to power supply 2 422, and the source of transistor 502 may be connected to the bit line to be driven through the medium-voltage portion of the ramp-up stage. The drain of transistor 503 may be connected to power supply 3 423, and the source of transistor 503 may be connected to the bit line to be driven through the high-voltage portion of the ramp-up stage and to be maintained above the steady-state voltage during the ramp-up stage. Finally, the drain of transistor 504 may be connected to the bit line in, with the source of transistor 504 connected to ground to discharge the bit line.

[0023] In an exemplary operation and with reference to Fig. 4 and Fig. 5, when all transistors are off, at the initiation of the ramp-up stage, the ramp control circuit 410 activates the low ramp signal to turn on transistor 501, so that the voltage V1 supplied by power supply 1 421 is used to drive the bit line through the low voltage section. When the bit line reaches a first predetermined voltage, the ramp control circuit 410 deactivates the low ramp signal to turn off transistor 501 and activates the mid ramp signal to turn on transistor 502, so that the voltage V2 supplied by power supply 2 422 is used to drive the bit line through the mid voltage section.When the bit line reaches a second predetermined voltage, the ramp control circuit 410 deactivates the mid-ramp signal to turn off transistor 502 and activates the high-ramp signal to turn on transistor 503, so that voltage V3 supplied by power supply 3 423 is used to drive the bit line through the high-voltage section. When the bit line reaches the steady-state voltage, the steady-state stage begins, and transistor 503 is held in the ON state, thus programming the memory cell. The ramp control circuit 410 then deactivates the high-ramp signal to turn off transistor 503 and activates the discharge signal to turn on transistor 504, allowing the bit line to be discharged.

[0024] Fig. 6 illustrates exemplary programming of the memory cells, e.g., memory cells W1-1, W1-7 (not shown), and W1-13 (not shown), included in a first group G1; memory cells W1-3, W1-9 (not shown), and W1-15 (not shown), included in a second group G2; and memory cells W1-5, W1-11 (not shown), and W1-17 (not shown), included in a third group G3, using pipelining according to an embodiment of the present disclosure. Initially, the controller 401 determines the locations of the above memory cells to be programmed. The controller 401 then generates programming pulses P1, P2, and P3 to enable programming of the identified memory cells in the respective groups G1, G2, and G3, respectively.It should be noted that although each of the above groups G1, G2, and G3 is disclosed as including three memory cells to be programmed, each of the above groups may include any number of memory cells to be programmed. Furthermore, in one embodiment, at least one of the above groups may include a different number of memory cells to be programmed with respect to at least one other of the above groups.

[0025] In contrast to Fig. 3 illustrated programming are the ones in Fig. 6 are staggered such that each programming pulse overlaps at least one other subsequent programming pulse for a predetermined overlap duration. For example, programming pulse P1, which enables programming of memory cells W1-1, W1-7, and W1-13, overlaps for a first predetermined overlap duration with a subsequent programming pulse P2, which enables programming of memory cells W1-3, W19, and W1-15. Furthermore, pulse P2, which enables programming of memory cells W1-3, W19, and W1-15, also overlaps for a second predetermined overlap duration with a subsequent programming pulse P3, which enables programming of memory cells W1-5, W1-11, and W1-17. In one embodiment, the duration of the first predetermined overlap duration and the second predetermined overlap duration are the same, while in other embodiments, this overlap duration may be of different durations.

[0026] In one example, the programming pulse P1 is activated at time t1 to enable programming of memory cells W1-1, W1-7, and W1-13, and deactivated at time t1' when these memory cells have been programmed. The programming pulse P2 is activated at time t2 to enable programming of memory cells W1-3, W1-9, and W1-15, and deactivated at time t2' when these memory cells have been programmed. As shown in Fig. As can be seen in Figure 6, the programming pulse P2 is activated at time t2 before time t1'. The programming pulses P1 and P2 are therefore staggered and overlap for the first predetermined overlap period. Similarly, the programming pulse P3 is activated at time t3 to enable the programming of memory cells W1-5, W1-11, and W1-17, and deactivated at time t3' when these memory cells have been programmed. As shown in Fig. As can be seen in Figure 6, the programming pulse P3 is activated at time t3 before time t2'. The programming pulses P2 and P3 are therefore staggered and overlap for the second predetermined overlap period. In this way, the identified memory cells included in each of the groups G1, G2, and G3 are programmed.

[0027] Fig. 7 illustrates programming memory cells using pipelining according to one embodiment of the present disclosure. The voltage associated with word line W1 is initially ramped up to the first high voltage level. Then, at time t1, when pulse P1 is activated, the voltage associated with bit lines B0, B6, and B12, which are connected to the sources of target memory cells W1-1, W1-7, and W1-13, is maintained at a very low voltage level. Furthermore, at time t1, to enable programming of target memory cells W1-1, W1-7, and W1-13, programming pulse P1 is activated to ramp up the voltages associated with bit lines B1, B7, and B13, which are connected to the sources of target memory cells W1-1, W1-7, and W1-13.

[0028] For example, as mentioned above with reference to Fig. 5, bitline driver 431 activates the Low Ramp1 signal to enable power supply 1 421 to charge / drive bitlines B1, B7, and B13 through their respective low-voltage portions. When the voltages associated with bitlines B1, B7, and B13 reach, for example, V1, bitline driver 431 deactivates the Low Ramp1 signal and activates the Mid Ramp1 signal to enable power supply 2 422 to charge / drive bitlines B1, B7, and B13 through their respective medium-voltage portions. When the voltages associated with bitlines B1, B7, and B13 reach, for example, V2, bitline driver 431 deactivates the Mid Ramp1 signal and activates the High Ramp1 signal to enable power supply 3 423 to charge / drive bitlines B1, B7, and B13 through their respective high-voltage portions. For example, if the voltages associated with bit lines B1, B7 and B13 exceed the steady-state level Vw1-1 the ramp-up stage is completed.

[0029] At this point, steady state begins, and bitline driver 431 maintains the Ramp-Up1 signal to allow memory cells W1-1, W1-7, and W1-13 to be programmed. Now, at time t2, pulse P2 is activated, enabling programming of memory cells W1-3, W1-9, and W1-15. Also, at time t2, bitline driver 432 activates the Ramp-Down3 signal to allow power supply 421 to charge / drive bitlines B3, B9, and B15 through their respective low-voltage portions. For example, when the voltages associated with bit lines B3, B9, and B15 reach V1, bit line driver 431 deactivates the Low Ramp3 signal and activates the Mid Ramp3 signal to allow power supply 2 422 to charge / drive bit lines B3, B9, and B15 through their respective mid-voltage portions.For example, when the voltages associated with bitlines B3, B9, and B15 reach V2, bitline driver 431 deactivates the MidRamp3 signal and activates the HighRamp3 signal to allow power supply 3 423 to charge / drive bitlines B3, B9, and B15 through their respective high-voltage portions. For example, when the voltages associated with bitlines B3, B9, and B15 reach the steady-state level V. W1-3 the ramp-up stage is completed.

[0030] Furthermore, at time t1', bitline driver 431 deactivates the RampUp1 signal and activates the Bitline_Discharge1 signal to discharge the voltage associated with bitlines B1, B7, and B13. At time t1', memory cells W1-1, W1-7, and W1-13, included in group G1, have been programmed. At time t2', bitline driver 431 deactivates the RampUp3 signal and activates the Bitline_Discharge3 signal to discharge the voltage associated with bitlines B3, B9, and B15. At time t2', memory cells W1-3, W1-9, and W1-15, included in group G2, have been programmed.

[0031] Programming additional memory cells, e.g., the memory cells included in group G3, may be performed in a manner similar to that described above with respect to programming the memory cells included in groups G1 and G2. In one example, it is within the scope of the present disclosure to program groups of memory cells using respective programming pulses and / or to program individual memory cells using respective programming pulses (as discussed above), as long as the programming pulses are staggered as suggested.

[0032] The overlap period between times t2 and t1' is the first predetermined overlap period during which the programming pulse P1, which programs memory cells W1-1, W1-7, and W1-13, and the programming pulse P2, which programs memory cells W1-3, W1-9, and W1-15, are staggered. Note that the illustration of a predetermined overlap period is exemplary, and the duration of a predetermined stress period may be shorter or longer as long as the steady-state stages of two staggered pulses do not overlap. For example, the maximum overlap period between times t2 and t1' is equal to the duration of the ramp-up stage of memory cells W1-3, W1-9, and W1-15.

[0033] Additionally, or optionally, additional subsequent programming pulses may be activated to program the identified memory cells. In this embodiment, the controller 401 may verify whether all memory cells to be programmed have actually been programmed. For example, the controller 401 may verify whether the respective threshold voltages associated with the transistors of the respective memory cells to be programmed have reached their respective predetermined threshold voltages. If the controller 401 determines that all respective threshold voltages associated with the transistors of the respective memory cells to be programmed have reached their respective predetermined threshold voltages, the controller 401 may not need to activate the additional subsequent programming pulses to program the memory cells.

[0034] However, if the controller 401 determines that a threshold voltage associated with a transistor of a given memory cell to be programmed has not reached its predetermined threshold voltage, the controller 401 may then execute the complete sequence of programming pulses P1, P2 and P3 as described above with reference to Fig. 6. Alternatively, the controller 401 may subsequently re-enable only the programming pulse associated with the given memory cell whose threshold voltage is determined not to have reached its predetermined threshold voltage. For example, if the controller 401 determines that the threshold voltage associated with the transistor of memory cell W1-7 has not reached its predetermined threshold voltage, the controller 401 may subsequently re-enable only pulse P1 to ensure that the threshold voltage associated with the transistor of memory cell W1-7 reaches its predetermined threshold voltage. After re-enablement of the relevant subsequent programming pulses, the controller 401 may once again verify whether the respective threshold voltages associated with the transistors of the respective memory cells to be programmed have reached their respective predetermined threshold voltages.

[0035] In one example, it is within the scope of the present disclosure to program groups of memory cells using respective programming pulses and / or to program individual memory cells using respective programming pulses (as discussed above), as long as the programming pulses are staggered as suggested.

[0036] In another example, the proposed pipelining during memory cell programming improves program / write performance and increases the throughput associated with the memory. For example, partially (or substantially completely) overlapping the programming pulses, while simultaneously considering the inherent metallic spurious effects of non-volatile memory cells, limitations of the supply voltages used to charge the bit lines, power consumption constraints associated with the device incorporating the memory, and core cell technology requirements, allows for faster and more efficient programming of the memory cells. For example, some memory architectures only allow a given voltage delta at a given time between voltages applied to the memory cells to avoid structural damage to the memory.

[0037] In another example, the predetermined overlap duration may be dynamically adjusted or adapted based on the behavior of the non-volatile memory cells, for example, during the ramp-up and steady-state stages. Additionally or alternatively, the predetermined overlap duration may be dynamically adjusted based on ambient operating conditions of the system, such as ambient temperature. Additionally or alternatively, the predetermined overlap duration may be dynamically adjusted based on a threshold voltage distribution of the respective transistors associated with the memory cells and / or a predetermined power consumption by a memory including the memory cells to be programmed.

[0038] Fig.Figure 8 illustrates a charge pump 810 according to one embodiment of the present disclosure. As an alternative to using three separate power supplies, Power Supply 1, Power Supply 2, and Power Supply 3, a single power supply 800 can be used together with a charge pump 810 to generate the voltages V1, V2, and V3 that drive the bit lines of the memory cells to be programmed. The charge pump regulates the single power supply 800 simultaneously and individually based on the required output voltages V1, V2, and V3.

[0039] The present invention has been described above using functional building blocks that illustrate the implementation of specific functions and relationships thereof. The boundaries of these functional blocks have been arbitrarily defined here for convenience of description. Alternative boundaries may be defined as long as the specified functions and relationships are adequately implemented.

[0040] The breadth and scope of the present invention should not be limited in any way by the exemplary embodiments described above, but should be defined only in accordance with the following claims and their equivalents.

Claims

[1] A system that includes: a memory cell array (100, 400) including a plurality of memory cells (W1-1, ... W1-N); and a programming device (401) configured to generate a plurality of continuous programming pulses (P1, P2, P3) configured to program target memory cells from the plurality of memory cells (W1-1, ... W1-N), wherein the plurality of continuous programming pulses (P1, P2, P3) overlap during an overlap period, wherein the overlap period is configured to be adjusted based on a performance parameter of the cell array (W1-1, ... W1-N), wherein, during the overlap period, a ramp-up period of a bit line (B3, B9, B15) associated with the programming of a subsequent target memory cell (W1-3, W1-9, W1-15) substantially overlaps a steady-state period of a bit line (B1, B7, B13) associated with the programming of a current target memory cell (W1-1, W1-7, W1-13), The system also includes: a ramp control circuit (410) configured to receive a plurality of ramp-up voltages (V1, V2, V3) and, during the ramp-up period, to drive a plurality of ramp-up stages of the bit line (B3, B9, B15) associated with programming the subsequent target memory cell (W1-3, W1-9, W1-15) by applying a respective one of a plurality of ramp-up voltages (V1, V2, V3), wherein the ramp control circuit (410) is configured to receive the plurality of ramp-up voltages (V1, V2, V3) from respective independent voltage supplies (421, 422, 423), each independent voltage supply (421, 422, 423) having a respective regulator. [2] The system of claim 1, wherein the power parameter includes a current required to precharge and program a target memory cell. [3] The system of claim 1, wherein the power parameter includes a current required to precharge and erase a target memory cell. [4] The system of claim 1, wherein the performance parameter includes a predetermined power consumption by a memory consisting of the memory cell array (100, 400). [5] The system of claim 1, wherein the ramp-up period of a bit line (B3, B9, B15) associated with programming the subsequent target memory cell (W1-3, W1-9, W1-15) includes a plurality of ramp-up stages, the plurality of ramp-up stages being driven by respective ramp-up voltages (V1, V2, V3). [6] The system of claim 1, wherein the ramp control circuit (410) is configured to sequentially apply the plurality of ramp-up voltages (V1, V2, V3) to drive the plurality of ramp-up stages. [7] A method in a system for programming memory cells (W1-1, ... W1-N), the method comprising: generating continuous first and subsequent programming pulses (P1, P2, P3) to enable programming of target memory cells, wherein the first and subsequent programming pulses (P1, P2, P3) overlap for an overlap duration that is adjustable based on a performance parameter of the system; and Programming to the target memory cells using the first and subsequent programming pulses (P1, P2, P3), wherein the programming substantially includes overlapping, during the overlap period, a ramp-up period of a bit line (B3, B9, B15) associated with programming a subsequent target memory cell (W1-3, W1-9, W1-15) with a steady-state period of a bit line (B1, B7, B13) associated with programming a current target memory cell (W1-1, W1-7, W1-13), receiving a plurality of ramp-up voltages (V1, V2, V3); and Driving, during the ramp-up period, a plurality of ramp-up stages of the bit line (B3, B9, B15) associated with programming the subsequent target memory cell (W1-3, W1-9, W1-15) by applying a respective one of the plurality of ramp-up voltages (V1, V2, V3), wherein receiving a plurality of ramp-up voltages (V1, V2, V3) includes receiving the plurality of ramp-up voltages from corresponding respective voltage supplies (421, 422, 423). [8] The method of claim 7, wherein the power parameter includes a current required to precharge and program a target memory cell. [9] The method of claim 7, wherein the power parameter includes a current required to precharge and erase a target memory cell. [10] The method of claim 7, wherein the power parameter includes a desired power consumption associated with programming the target memory cells. [11] The method of claim 7, wherein the ramp-up period of a bit line (B3, B9, B15) associated with programming the subsequent target memory cell (W1-3, W1-9, W1-15) includes a plurality of ramp-up stages, the plurality of ramp-up stages being driven by respective ramp-up voltages (V1, V2, V3). [12] The method of claim 7, wherein applying includes sequentially applying the respective ones of the plurality of ramp-up voltages (V1, V2, V3) to drive the plurality of ramp-up stages.

Citation Information

Patent Citations

  • Nonvolatile memory, semiconductor device and method of programming to nonvolatile memory

    US20020041527A1

  • Non-volatile semiconductor memory device

    US20060291291A1