Static information storage and retrieval – Read/write circuit – Including signal clamping
Reexamination Certificate
2003-08-26
2004-12-14
Dinh, Son T. (Department: 2824)
Static information storage and retrieval
Read/write circuit
Including signal clamping
C365S210130
Reexamination Certificate
active
06831866
ABSTRACT:
BACKGROUND OF INVENTION
The present invention relates generally to semiconductor memory devices and, more particularly, to a method and apparatus for read bitline clamping for three-transistor (3T) gain cell DRAM devices.
Dynamic Random Access Memory (DRAM) devices commonly employ a one-transistor memory cell arrangement that includes an access transistor and a storage capacitor. The information is stored in the storage capacitor in the form of an electrical charge that represents a logical level, 0 or 1. This information can be read out from or written to the cell by activating a wordline coupled to the access transistor via a word line. During a read operation, the electrical charge stored in the storage capacitor is transferred to a bitline and sensed by sense amplifier circuitry. During a write operation, electrical charge is added to or removed from the storage capacitor by the bitline.
Since the storage density increases from generation to generation of memory, the required surface area of the one-transistor memory cell must be reduced from generation to generation. This leads to fundamental technological and physical problems (particularly for performance), because the threshold voltage of an access transistor in a one-transistor memory cell is not reduced due to the device leakage. Yet, the operation voltage should be reduced to guarantee the device reliability.
This problem is avoided in one approach by an alternative DRAM cell arrangement wherein so-called “gain” cells are used as storage cells. Like the one-transistor/storage capacitor cell, data is also stored in gain cells in the form of an electrical charge. However, the electrical charge in a gain cell is not directly coupled to a bitline, but rather is stored in a gate electrode of a gain transistor (thereby controlling the conductivity of that transistor), for which purpose a very small amount of electrical charge is sufficient. Because gain cells do not require an electrical charge for signal storage, the capacitance of the cell can be reduced significantly. This results in a relatively short time constant as compared to the conventional one transistor memory cell with a large storage capacitance. In addition, as long as the storage node voltage is higher than the threshold voltage of the gain transistor, the device is turned on, further improving performance.
In a gain cell
100
having a three-transistor (3T) configuration such as shown in
FIG. 1
, the electrical charge is stored in a first gate electrode of a first transistor
114
(i.e., a gain transistor). The writing or storage of the electrical charge therein occurs with the aid of a second transistor
116
(i.e., a write access transistor). In particular, the gate electrode N of the gain transistor
114
is connected to one of the source/drain regions of the write access transistor
116
, while the other source/drain region of the write access transistor
116
is connected to a write bitline WBL. During a write operation to the gain cell
100
, the gate electrode of the write access transistor
116
is activated via an appropriate signal on a write wordline WWL. This allows the gate electrode N of the gain transistor
114
to be coupled to the write bitline WBL through transistor
116
. The amount of electrical charge stored in the node N (and thus the data bit information stored in the gain transistor gate electrode) is determined by the voltage present on the write bitline when the write wordline is activated. The gate capacitor of the gain transistor
114
keeps sufficient electrical charge at storage node N. Similar to a conventional one-transistor DRAM, the storage node N should be periodically refreshed to rewrite (maintain) the voltage at the node N before the charge has been lost completely.
The readout of stored information in a 3T gain cell occurs with the aid of a third transistor
112
(i.e., a read access transistor). One of the source/drain regions of the gain transistor
114
is connected to one of the source/drain regions of the read access transistor
112
, while the other of the source/drain regions of the gain transistor
114
is connected to either V
DD
or ground, depending upon the type of the cell transistors (PFET or NFET). In addition, the other of the source/drain regions of the read access transistor is connected to a read bitline RBL. The read bitline RBL is precharged to either V
DD
or ground, again depending upon the type of the cell transistors (PFET or NFET). For example, in a 3T gain cell having NFET transistors, the read bitline is precharged to V
DD
, and the other of the source/drain regions of the gain transistor is connected to ground, as shown in FIG.
1
. During a read operation, the gate electrode of the read access transistor
112
is actuated via a read wordline RWL. This allows the drain of gain transistor
114
to be coupled to the read bitline RBL through the transistor
112
. The read bit line is therefore determined by the conductance through the series combination of the read transistor
112
and gain transistor
114
coupling to ground.
In addition,
FIG. 1
illustrates aPFET pull-up transistor
118
configured as a precharge device that dynamically precharges the read bitline RBL to the supply voltage V
DD
. Thus, if the storage node N maintains a “1” data bit therein, the read bitline RBL is discharged down from V
DD
since both transistors
112
and
114
are conductive. On the other hand, if the storage node N maintains a “0” data bit therein, the read bitline RBL is not discharged from the precharge voltage since the gain transistor
114
is non-conductive. The presence of a RBL voltage swing (i.e., reading “1” data) is detected by appropriate sense amplifier circuitry (not shown), which is well known and therefore not discussed in further detail hereinafter. After the read operation of the cell data is completed, the RBL precharge device
118
is turned on (i.e., the signal on precharge line PRE goes low) to restore the read bitline voltage to V
DD
.
One problem associated with the above described cell configuration results during a read operation of a cell having a “1” stored therein (using NFET transistors in the cell). In this case, the activation of the read wordline, combined with the gate of the gain transistor being activated as a result of storing a “1”, causes the read bitline to be discharged toward ground. In order to provide desired fast access speed, a rapid RBL swing is needed to sense the “1” data (it is noted that no bitline swing occurs when reading a “0” since the gain transistor is not conductive). However, if the read bitline swing is too fast, the voltage on the read bitline goes to ground, thereby resulting a longer restore time to return the read bitline voltage back to V
DD
. As also mentioned previously, the circuit of
FIG. 1
provides a fast access speed for reading “1” data as a result of the fast RBL swing.
A typical known solution is to include a read bitline clamp device to limit a read bitline swing. Without this clamp device, the voltage on read bitline RBL (in worst case) goes to ground. This condition results in a longer restore time for returning the RBL voltage back to V
DD
. If the clamp device is implemented by substituting it for a dynamically controlled precharge device, the restore time is improved, but at the cost of a slower read bitline swing. This is due to the fact that a read access transistor needs to be able drive the read bitline RBL to ground. However, some of the current is also used for the clamp device coupling to V
DD
. In addition, the clamp device must be laid out for each column, which increases the silicon area. Furthermore, the conventional use of a separate clamping device does not allow for “tracking” of variations in certain memory cell transistor and voltage parameters. Thus, it is desirable to have a gain transistor DRAM cell configuration that has both a fast read bitline voltage and an improved restore time, while also allowing for device process tracking and without sacrificing device real estate area.
SUMMARY OF INVENTION
Th
Abate Joseph P.
Cantor & Colburn LLP
Dinh Son T.
International Business Machines - Corporation
LandOfFree
Method and apparatus for read bitline clamping for gain cell... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus for read bitline clamping for gain cell..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus for read bitline clamping for gain cell... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3320485