ADDRESS TRANSTTTON DETECTOR FOR FAST FLASH MEMORY DEVICE

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims priority to United States Patent Application Serial Number 11/303,863, filed December 16, 2005, the entirety of which is hereby incorporated by reference herein.

BACKGROUND OF THE INVENTION

1. Field of the Invention

The present invention relates to Flash memory devices. More particularly, the present invention relates to an address-transition detector for a Flash memory device.

2. The Prior Art

The consistency of the pulse width in flash memory devices that perform read operations at the speed of 20ns, or equivalently in the 50MHz range is extremely important. Existing designs are not suitable for read speed at this frequency, because the timing variation becomes rather significant in this new environment.

United States Patent No. 4,633,102 discloses a scheme in which the output pulse width varies with the rise time and fall time of the input signal. Tt is believed that this design is unable generate a constant timing pulse.

United States Patent No. 4,742,247 discloses an address-transition-detector that has good temperature compensation.

United States Patent No. 4,922,461 discloses a pulse generator driving a clock generator.

United States Patent No. 4,982,117 discloses an address-transition-detector that ends its pulse width whenever a transition occurs at the output.

United States Patent No.US05057712 discloses an addrcss-transition-dctcctor having a pulse width that suffers from Vcc, temperature and process variations.

Since flash memory products perform read operations at less than 50MHz, there is no alternative design or method used in the industry. Thus, a new design that serves this need is desired to meet such a design requirement.

BRIEF DESCRIPTION OF THE INVENTION

The present invention relates to a address transition detector timing pulse generator designed for rapid read operation. This circuit has been designed to provide constant pulse width in various operational conditions. In the read operation, two

timing pulses define the duration of pre-equalization and the actual sensing time. At the end of the sensing operation, a one shot signal is issued to latch the output of the sense amplifier.

An address transition detector circuit includes an input node, an output node, a bandgap reference node, and Pbias and Nbias nodes having voltages derived from the bandgap reference node. First through fifth cascaded inverters are each powered by a p-channel and n-channel MOS bias transistors having their gates coupled respectively to the Pbias node and the Nwas node. The input of the first inverter is coupled to the input node. First and second capacitors arc coupled respectively to ground from the outputs of the first and fourth cascaded inverters. A NAND gate has a first input coupled to the input node, a second input coupled the output of the fifth cascaded inverter, and an output coupled to the output node.

BRIEF DESCRIPTION OF THE DRAWING FIGURES

FIG. 1 is a schematic diagram of an address-transition detector according to the present invention.

FIG. 2 is a schematic diagram of a circuit that may be used to generate bias voltages from a bandgap reference for use in a circuit such as shown in FIG. 1 DETAILED DESCRIPTION OF THE INVENTION

Persons of ordinary skill in the art will realize that the following description of the present invention is illustrative only and not in any way limiting. Other embodiments of the invention will readily suggest themselves to such skilled persons.

Although the term "flash" memory often refers to non-volatile memories that are bulk erased on a page-by-page, sector-by-sector, or entire array basis, the term is generally used in the art to refer to any electrically erasable programmable (and reprogrammable) non-volatile memory technology, regardless of the particular erase scheme. Current flash memories typically use floating gate transistors, though other types of transistor-based flash memory technologies are known in the art, such as, for example, floating trap devices such as silicon nanocrystal MOS transistors, SONOS, and MONOS transistors.

The present invention relates to a timing pulse generator designed for rapid read operations. This circuit provides a near-constant pulse width in various operational conditions. In the read operation, two timing pulses are needed to define the duration of pre-equalization and the actual sensing time. At the end of the sensing, a one shot signal is issued to latch the output of a sense amplifier.

The consistency of the pulse width is extremely important in environments such as where read operations are designed to be performed at speeds such as 20ns, or equivalently 50MHz. Existing design are not believed suitable for read speed at this frequency, because the timing variation becomes rather significant in this new environment. Thus, implementing a new design that serves this need becomes important to meet the requirement of design specifications requiring fast read times.

Referring now to FIG. 1, an address-transition detector circuit 10 according to the present invention includes a NAND gate 12 and a chain of serial inverters 14, 16, 18, 20, and 22. The positive power connections of inverters 14, 16, 18, 20, and 22 arc connected, serially to P-channel MOS bias transistors 24, 26, 28, 30, and 32, whose sources are driven by a stable supply voltage indicated in FIG. 1 as Vscc shown at reference numeral 34, which is generated from a bandgap reference voltage. Generation of a bandgap reference voltage is known in the art. For example, in a device with a supply voltage (Vcc) that varies between approximately .96 volts and 1.6 volts, a stable supply voltage Vscc of 0.9 volts may be supplied that does not fluctuate, even if the Vcc of the device fluctuates.

The gates of P-channel MOS bias transistors 24, 26, 28, 30, and 32 are driven by a Pbias signal shown at reference numeral 36. The negative power connections of inverters 14, 16, 18, 20, and 22 are connected serially to N-channel MOS bias transistors 38, 40, 42, 44, and 46, whose sources are coupled to ground. The gates of N-channel MOS bias transistors 38, 40, 42, 44, and 46 are driven by an Nbi _as signal shown at reference numeral 48. Bandgap reference circuits are known in the art. Both the Pbias and Nbias signals are derived from the bandgap reference. Since the bandgap reference generates both bias signals, the current capabilities of inverters 14, 16, 18, 20, and 22 are more consistent across temperature and process corner variations. This is the first attribute of the circuit of the present invention that contributes to providing constant timing.

In integrated circuits where the power supply voltage Vcc ranges from 0.96v to 1.6v, a stabilized, bandgap reference voltage can be provided at 0.9v by using the output voltage of a bandgap reference (e.g., 0.6v) and traditional analog buffer design. The supply of inverters 14, 16, 18, 20, and 22 is coupled to V _BGR to remove the timing variation caused by the Vcc variation. This is the second attribute to provide constant timing.

Inverters 14, 16, 18, 20, and 22 are identical inverters that merely invert the input signal. To create more delay throughout the inverter chain, capacitors 50 and 52 are coupled to ground from the outputs of inverters 14 and 20 and are typically identical in value. Typical values for these capacitors are in the range of between about 0.05pF and about 0.2pF. Inverters 14 and 20 are identical devices as are inverters 16 and 22. Inverters 14 and 20 and capacitors 50 and 52 create the delay and inverters 16 and 22 just buffer the output. Inverters 16 and 22 are relatively stronger than inverters 14 and 20. In the curve of Vout/Vin of any inverter, the trip point of the inverter varies from case to case. In addition, the MOS bias transistors 24, 30, 38, and 44 coupled to inverters 14 and 20 are downsized relative to the MOS bias transistors 26, 28, 32, 40, 42, and 46, driving inverters 16, 18, and 22 to lower their current capability. According to the present invention, the size differences between the downsized transistors and the other bias transistors is about 1 :4.

For purposes of explaining the operation of the circuit of FIG. 1 , the output of inverter 14 will be identified as node 54, the output of inverter 16 will be identified as node 56, the output of inverter 18 will be identified as node 58, the output of inverter 20 will be identified as node 60, and the output of inverter 22 will be identified as node 62.

The rising delay "T _r " of node 54 and falling delay "T _f " of node 62 depend on the trip points of the inverters. If the n-channel MOS bias transistors are stronger than the p-channel MOS bias transistors, the capacitors 50 and 52 will both discharge faster than they charge, since the n-channel MOS bias transistors 38 and 44 have more discharge current carrying capacity than the relatively weaker p-channel MOS bias transistors 24 and 30 have charging current capacity. Therefore the rise time T ₁ of node 54 would be relatively shorter than the fall time Tf of node 62. This occurs because capacitor 48 is discharged through n-channcl MOS bias transistor 38 causing the output of inverter 16 to rise quickly and because capacitor 50 is charged by the relatively weaker p-channel MOS bias transistor 30. If, however, the p-channel MOS bias transistors are stronger than the n-channel MOS bias transistors, then the rise time T ₁ of node 54 would be relatively longer than the fall time Tf of node 62. In either case, the quantity T _r +T _f will remain relatively constant.

The inverter 18 is used to invert the incoming signal, so that the falling delay of node 54 and rising delay of node 62, respectively are timed using capacitors 50 and 52. The reason for doing this is to balance out the delay-time difference caused by the

shift of the trip points of the inverters. The trip point of the inverter shifts because of process variations (which cause different driving capability of n-channel MOS transistors and p-channel MOS transistors, respectively). Temperature variations also have a similar effect since the mobility of electrons and holes varies in a different ratio.

The signal "ENf" goes high when any address transition occurs. By definition this is regarded as an initialization of reading a new address. Once the signal "ESf" goes high, "OUT" will go high through inverter 64 since node 62 at the output of inverter 22 remains high, until the "IN" signal transition propagates through the inverter chain to drive node 62 to low. The time it takes to toggle node 62 to a low state is the width of the actual pulse generated by the ATD circuit.

An advantage of the address-transition detector circuit of the present invention is that it removes the need for an analog comparator which would introduce more timing variation. Also, the present invention shows great consistency under various supply voltages and temperature fluctuations, as well as physical differences in circuit components caused by process issues.

Referring now to FIG. 2, a representative circuit 70 is shown that may be used to generate the V _BGR , P _bias and N _bias signals from a bandgap reference. Two sense amplifiers 72 and 74 are used in circuit 70. Their main feature is to equalize the voltages at their positive and negative inputs by feedback from their outputs to the circuits that drive these input voltages.

Sense amplifier 72 has its non-inverting input coupled to the common connection of the drain of p-channel MOS transistor 76 and the emitter of PNP transistor 78. The base of PNP transistor 78 is coupled to its collector. The gate of p- channel MOS transistor 76 is coupled to the output of sense amplifier 72. The inverting input of sense amplifier 72 is coupled to the drain of p-channel MOS transistor 80, which is, in turn, coupled to the emitter of PNP transistor 82 through resistor 84. The base of PNP transistor 80 is coupled to its collector. The gate of p- channel MOS transistor 80 is coupled to the output of sense amplifier 72.

Sense amplifier 74 has its non-inverting input coupled to the common connection of the drain of p-channel MOS transistor 76 and the emitter of PNP transistor 78. The inverting input of sense amplifier 72 is coupled to the common connection of the drain of p-channel MOS transistor 86 and resistor 88, which is, in

turn, coupled to ground. The output of sense amplifier 74 is coupled to the gate of p- channel MOS transistor 86.

The output of sense amplifier 72 is also coupled to the gates of p-channel MOS transistors 90 and 92. The drain of p-channel MOS transistor 90 is coupled to ground through resistor 94. The common connection of the drain of p-channel MOS transistor 90 and resistor 94 is the output node V _BGR . The drain of p-channel MOS transistor 92 is coupled to the drain of n-channel MOS transistor 96. N-channel MOS transistor 98 forms a current mirror with n-channel MOS transistor 96. The common connection of the gate and drain of n-channel MOS transistor 96 and the gate of n- channel MOS transistor 98 is the N _bias node, which is the output for the voltage N _bias . The drain of n-channel MOS transistor 98 is coupled to the gate and drain of p- channel MOS transistor 100 and forms the P _bias node, which is the output for the voltage P _bias .

The output of sense amplifier 74 is coupled to the gates of p-channel MOS transistors 102 and 104. The drain of p-channel MOS transistor 102 is coupled to the drain of p-channel MOS transistor 90 and the drain of p-channel MOS transistor 104 is coupled to the drain of p-channel MOS transistor 92. The voltage at. resistor 94, V _BGR , may be used to generate the stable supply voltage V _scc by employing error amplifier 106 with a voltage divider including resistors 108 and 110 deriving the feedback voltage to its inverting input. In the example shown in FlG.2, resistor 110 is twice the value of resistor 108. If V _BGR is 0.6, the output voltage V _scc at the output of amplifier 106 (a stable voltage supply node) will be 0.9V as is understood in the art.

In the circuit of FIG. 2, when the system reaches equilibrium, the voltage at the non-inverting input of sense amplifier 72 and the inverting input of sense amplifier 74 are both equal to VBEl. The outputs of both sense amplifiers 72 and 74 are tied to numerous PMOS transistors as current supplies for each division.

Current mirrors are formed by tying together the gates of the p-channel MOS transistors 76, 80, 90, And 92 connected to Vcc, and the n-channel MOS transistors 96 and 98 connected to ground. When all transistors are operated in saturation mode, the driving current of these transistors will be approximately the same, regardless of the level of drain voltage.

In the circuit depicted in FIG. 2, the gates of p-channcl MOS transistors 76, 80, 90, And 92 are connected together. As well as two other groups of current mirrors

formed from p-channel MOS transistors 86, 102, and 104 and n-channel MOS transistors 96 and 98. Also, the sizes of these transistors in each group are the same. The function of the current mirrors is to duplicate the current from one leg to another, and to provide additional current to other legs. Thus, the currents through p-channel MOS transistors 76, 80, 90, and 92 are equal. The currents through n-channel MOS transistors 96 and 98 are equal. The current through n-channel MOS transistor 96 equals the sum of the currents through p-channel MOS transistors 92 and 104. The current through resistor 94 equals the sum of the currents through p-channel MOS transistors 90 and 102.

The constant voltage output V _BGR is generated by adding two quantities (the outputs of amplifiers 72 and 74) with opposite temperature coefficients with proper weighting. For example, the typical base-emitter voltage in 0.13 micron processes is around 40OmV, with negative temperature dependence. With a PNP transistor VBE = 30OmV and T = 300 ⁰ K, 3V _BE /δT ~ -1.5mV/ ⁰ K. V _BE is the negative temperature coefficient. PNP transistor 82 is formed from three transistors, each the size of PNP transistor 78 in order to provide three times the current through PNP transistor 82 than through PNP transistor 78.

After proper sizing of resistors 84, 88, and 94, a constant voltage VBG _R is generated. The sizing of the resistors can be understood by examining the following equation: VBG _R — ATRg ₄ ZRg ₄ + VB _E /RPVRSS- The first term of the equation is a positive coefficient and the second term of the equation is a negative coefficient. The constant "A" is a constant derived from the familiar Vt expression KT/Q multipled by In (3) from the fact that transistor 82 is composed of three paralleled bipolar transistors, and "T" is the absolute temperature in degrees Kelvin. As is apparent from the equation, any resistance variation as a function of temperature is canceled out.

Since the current through resistor 94 equals the current through n-channcl MOS transistor 96, by providing n-channel MOS transistor 96 with both the gate and drain connected to the drain of p-channel MOS transistor 92, the "Nbia _s " signal is generated as the bias voltage for the n-channel MOS transistors. By further tying the gate of n-channel MOS transistor 96 to the gate of n-channel MOS transistor 98, and connecting the drain of n-channel MOS transistor 98 to the drain and gate of p- channel MOS transistor 100, the "Pbias" signal is generated as the bias voltage for the p-channel MOS transistors.

While embodiments and applications of this invention have been shown and described, it would be apparent to those skilled in the art that many more modifications than mentioned above are possible without departing from the inventive concepts herein. The invention, therefore, is not to be restricted except in the spirit of the appended claims.