Buffer circuit

This application is based on Japanese Patent Application 2000-044863, filed on Feb. 22, 2000, the entire contents of which are incorporated herein by reference.

A) Field of the Invention

This invention relates to a buffer circuit having an output terminal through which data can be transferred to and from an external device.

B) Description of the Related Art

In the art of a semiconductor integrated circuit, it is important to raise the accumulation of the circuit while reducing the power consumption. It is effective for reducing the power consumption of an integrated circuit to lower the voltage of a power supply. In a transition period from a conventional 5V power supply to a new 3.3V power supply, for example, a multi-voltage circuit is used, wherein some parts of the circuit are designed to be driven by the 5V power supply while others to be driven by the 3.3V power supply. When a signal is output from a 5V part to a 3.3V part in that kind of multi-voltage circuit, some problems may occur. When a higher voltage than the voltage of a power source is applied to an input terminal, a Silicon Controlled Rectifier (SCR) of p-n-p-n structure with PMOS and NMOS is turned on, resulting in a formation of a current leak path or a latch up. In some case, a large amount of electric current flows through the SCR, resulting in over heating of the semiconductor integrated circuits.

Japanese Patent Publication H07-79232 discloses a driver circuit shown in FIG. 5 that solves those problems. Power source voltage VDD of this driver circuit is 3.3V, and when a high level (H-level) enabling signal EN is supplied to an enabling terminal 10 , data D supplied to a data input terminal 28 could be output from a data output terminal 24 . On the other hand, when an L-level (0V) enabling signal EN is supplied to an enabling terminal 10 , output impedance of the data output terminal 24 will be a state of high impedance. Therefore, a 5V signal can be supplied to a bus connected to the output terminal 24 . Moreover, this driver circuit is built in a P-type silicon substrate whereas N-channel transistors are formed on an N-well formed on the P-type silicon substrate. Especially N-channel transistors 30 , 32 , 36 and 38 are formed on the same N-well that is floating.

First, a case when an enabling signal is high level will be considered. In this case, an N-channel transistor 12 is turned on and so an N-channel transistor 34 is also turned on; therefore, gate voltage of a P-chancel transistor 32 will be at the L-level. The P-chancel transistor 32 will be turned on. Because gate voltage of an N-channel transistor 26 is maintained to be at the power source voltage VDD, the transistor 26 is also turned on. On the other hand, gate voltages of both of a P-channel transistor 30 and N-channel transistor 22 are inversion of the data D. Therefore, when the data D is at the high level, voltage of the data output terminal 24 will be at the high level, and when the data D is at the L-level, voltage of the data output terminal 24 would be at the L-level.

Next, a case when an enabling signal is at the L-level will be considered. In this case, an N-channel transistor 12 is turned off. The gate voltage of the N-channel transistor 22 will be at the L-level, and the N-channel transistor 22 will be turned off. Moreover, the gate voltage of the P-channel transistor 30 will be at the high level, and the P-channel transistor 30 will be turned off. Therefore, output impedance of the data output terminal 24 would be in a state of high impedance.

At that time, assumingly a switch 44 is turned on, and an output signal S with 0V L-level and 5V high level is supplied from an external device 42 driven at 5V to the driver circuit. If the threshold voltage of the P-channel transistor 30 is 0.7V and voltage of the signal S is 5V, the P-channel transistor 30 would be turned on. Then, voltage of a node B becomes 5V whereas the gate voltage of the P-channel transistor 36 is 0V; therefore, the transistor 36 will be turned on. Also, the P-channel transistor 32 will be turned off. By that, current flow to the side of a voltage source 28 (VDD) can be prevented.

Also, N-wells of the P-channel transistors 30 , 32 and 36 are auto-biased by parasitic diodes formed between their drains and the N-wells. Therefore, current feedback by the parasitic pnp transistors including N-wells and the P-type silicon substrate will be vanished.

By forming a P-channel transistor 38 , when the voltage of the data output terminal 24 is at the L-level, an N-well is always biased by the power supply voltage VDD. The formation of the P-channel transistor 38 minimizes the possibility that the parasitic pnp transistors turn on during the transition of the signal S from the L-level to the high level.

As said in the above, in the driver circuit shown in FIG. 4, because there are no paths for the current flow to the semiconductor substrate, the problem of latch up can be solved.

By the way, in the above-described driver circuit, chip sizes of the P-channel transistors 32 and 30 and the N-channel transistors 22 and 26 will be large if large output current needed when the data D is output from the data output terminal 24 because gate widths of the transistors have to be increased to derive large current from the transistors.

In a practical circuit, a plurality of transistors need to be connected in parallel to form each of the P-channel transistors 32 and 30 and the N-channel transistors 22 and 26 .

However, there are problems in the enlargement of the chip sizes because the large chip sizes increase cost of manufacturing and also decrease the productivity due to the large number of the elements needed for manufacturing.

It is an object of the present invention to provide a driver circuit that is small in chip size while eliminating paths for the current flowing to a semiconductor substrate.

According to one aspect of the present invention, there is provided a buffer circuit comprising: a data input terminal to which a data signal is input; an enabling terminal to which an enabling signal is input; an output terminal from which output data are output; a first power source terminal to which high potential voltage is supplied; a second power source terminal to which low potential voltage lower than said high potential voltage is supplied; a first N-channel transistor that is connected between said output terminal and said second power source terminal; a common bulk P-channel transistors group that comprises a first, a second, a third, a fourth and a fifth transistors formed on a common bulk region, wherein the first P-channel transistor that is connected between said first power source terminal and said output terminal, the second P-channel transistor that is formed between said output terminal and one node and comprises a gate electrode connected to said first power source terminal, the third P-channel transistor that comprises a first current terminal connected to said output terminal, a second current terminal connected to a gate electrode of said first P-channel transistor, and a gate electrode connected to said first power source terminal, the fourth P-channel transistor that is formed between said first power source terminal and said gate electrode of said first P-channel transistor and comprises a gate electrode of which is supplied an inverted signal of the enabling signal, and the fifth P-channel transistor that comprises a drain electrode connected to said bulk region, a source electrode connected to said first power source terminal and a gate electrode connected to said one node; a second N-channel transistor that is formed between said one node and said second power source terminal and comprises a gate electrode supplied an inverted signal of an enabling signal; and a logic circuit that inputs an inverted signal of said input signal to said gate electrodes of said first P-channel transistor and said first N-channel transistor when status of said enabling signal is in one state, or inputs a signal keeping said first P-channel transistor turned off to the gate electrode of said first P-channel transistor when the status of said enabling signal is in another state.

In this buffer circuit, when said enabling signal is in one state, a signal based on the input data signal input from the data input terminal will be output from said output terminal.

On the other hand, when said enabling signal is in another state, said P-channel transistor will be kept turned off, and so said output terminal will be a state of high impedance.

As said in the above, according to this invention, in a buffer circuit wherein an output terminal can be controlled to be a state of high impedance, the number of transistors for output can be reduced while preventing current leak and latch up when voltage induced to the output terminal is higher than that of power supply.

FIG. 1 is a circuit diagram showing a structure of a buffer circuit according to a first embodiment of the invention.

FIG. 2 is a cross sectional view showing a part of the buffer circuit according to the first embodiment.

FIG. 3 is a circuit diagram showing a structure of a NAND circuit according to the first embodiment.

FIG. 4 is a circuit diagram showing a structure of a buffer circuit according to a second embodiment of the invention.

FIG. 5 is a circuit diagram showing a structure of a buffer circuit according to a prior art.

A first embodiment of this invention will be explained below with reference to FIGS. 1 to 3 .

FIG. 1 is a circuit diagram showing a structure of a buffer circuit according to the first embodiment. FIG. 2 is a cross sectional view showing an essential part of the buffer circuit. The buffer circuit 100 has P-channel transistors P 1 to P 6 , P 11 and P 12 , N-channel transistors N 1 , N 2 , N 5 and N 6 , a nand circuit NAND, a nor circuit NOR, an inverter circuit INV 1 , input terminal T 1 , an enabling terminal T 2 , and input/output (IO) terminal T 3 . A parasitic diode D 1 is formed between a drain electrode of the P-channel transistor P 1 and a bulk.

The buffer circuit 100 also has a first and second power source terminal. Through the first power source terminal VDD is supplied lower power source voltage VDD on the high potential side. Through the second power source terminal GND is supplied power source voltage GND on the low potential side. The voltage VDD, for example, is 3.3V. An enabling signal OE activated at the L-level (Low level) is supplied to the enabling terminal T 2 , and a first input data Din 1 is supplied to the input terminal T 1 . Logic level voltage of the enabling signal OE and the first input data Din 1 at the L-level is the GND while that at the H-level (High level) is the VDD.

Also, output data Dout is output from the IO terminal T 3 when the enabling signal OE is at the L-level. Logical level voltage of the output data Dout at the L-level is GND while that at the high level is VDD. On the other hand, when the enabling signal OE is at the high level, output impedance of the IO terminal T 3 will be a state of high impedance. At the same time, a second input data Din 2 is supplied from an external circuit D 2 . Logical level voltage of the second input data Din 2 at the L-level is GND while that at the high level is VDD′ (higher power source voltage). The voltage VDD′, for example, is 5V. That is, this buffer circuit 100 has two modes: an output mode wherein the output data Dout is output from the IO terminal T 3 by buffering the first input data Din 1 ; an input mode wherein the second input data Din 2 is taken in through the IO terminal T 3 .

The P-channel transistor P 1 and the N-channel transistor N 1 are output transistors for amplifying current, and so their occupying areas are wide because each of them is formed with a parallel connection of a plurality of transistors on an integrated circuit.

As shown in FIG. 2, the P-channel transistors P 1 to P 6 and the P-channel transistors P 7 to P 10 forming the later described nand circuit NAND has a common bulk region 103 a . In this embodiment, a silicon substrate 101 is a P-type substrate, and the bulk region 103 a shown as solid lines in FIG. 1 is an N-well formed on the P-type silicon substrate 101 . Provided that a N-type substrate is used for the silicon substrate 101 , the bulk region 103 a will be the N-type silicon substrate. Further, the common bulk region 103 a has no terminals for supplying power supply voltage VDD or earth voltage GND, so that the common bulk region 103 a is floating. Voltage of the common bulk region 103 a will be hereinafter called Vddf.

In detail, a P-well 103 and an N-well 105 are formed on an N-type silicon substrate 101 . The P-well 103 and the N-well 105 are, for example, separated by a silicon oxide layer formed by the local oxidation of silicon (LOCOS) method.

The P-well 103 has many regions separated by the silicon oxide layer 107 . In FIG. 2, a first P-well region and a second P-well region 103 b are illustrated, and the first P-well region hereinafter called a common bulk region 103 a.

In the common bulk region 103 a are formed a first P-channel transistor P 1 and a second to sixth P-channel transistors P 2 to P 6 .

In addition, on the common bulk region 103 a , a seventh to a tenth P-channel transistors P 7 to P 10 that are included in the later-described output buffer circuit OB are formed.

Each transistor is formed of silicon oxide and has a gate electrode G formed across a gate insulating film 111 , a source electrode S and a drain electrode D.

A high dense n-type region 113 is formed in contact with the drain electrode D of the sixth P-channel transistor P 6 . A common electrode 115 is formed on the drain electrode D and the high dense n-type region 113 .

A first N-channel transistor N 1 is included in the N-well 105 . A drain electrode D of the first N-channel transistor N 1 and the first P-channel transistor P 1 are connected to form an IO terminal T 3 . The common bulk region may consist of a plurality of regions connected with each another.

The P-channel transistors P 2 and P 3 are formed between a node X and the IO terminal T 3 and work as a switch. Especially, the P-channel transistor P 3 is turned on and supplies voltage Vt 3 to the node X when the voltage Vt 3 of the IO terminal T 3 is higher than that of the power supply voltage in said input mode, i.e., when the second input data Din 2 is at the high level. The transistor P 2 may be omitted.

The P-channel transistor P 5 is turned on in the output mode and can turn off the gate electrode of the P-channel transistor P 1 by impressing power source voltage VDD. Moreover, the P-channel transistor P 4 is turned on and supplies VDD when the voltage Vt 3 is higher than the power supply voltage VDD.

The P-channel transistor P 6 is turned on in the output mode and supplies power source voltage VDD to the common bulk region 103 a . On the other hand, in the input mode, the P-channel transistor P 6 is turned off and does not supply power source voltage VDD to the common bulk region 103 a . Moreover, the N-channel transistor N 2 is turned on in the out put mode and biases a node X at 0V.

FIG. 3 shows a structure of the nand circuit NAND. As shown in this drawing, the nand circuit NAND has P-channel transistors P 7 to P 10 and N-channel transistors N 3 and N 4 . In this nand circuit NAND the common bulk region 103 a is used as a bulk for the P-channel transistors P 7 to P 10 . Also, the nand circuit is characterized by having the P-channel transistors P 8 and P 10 .

Each gate electrode of the P-channel transistors P 8 and P 10 is supplied voltage Vx of the node X. Those transistors P 8 and P 10 are turned off when the voltage Vx is at the high level.

Transistors at an output stage are formed only with the P-channel transistor P 1 and the N-channel transistor N 1 . Supplying large current from the input terminal T 3 in the output mode can be realized only by enlarging sizes of the P-channel transistor P 1 and the N-channel transistor N 1 ; therefore, whole chip area of the buffer circuit 100 can be reduced.

Moreover, the common bulk region 103 a of the P-channel transistors P 2 to P 10 is floating. If the voltage Vt 3 at the input-output terminal becomes larger than the power source voltage VDD in the input mode, only the common bulk region 103 a would be biased via the parasitic diode D 1 . Therefore, there will be no chance to latch up by flowing large current.

As shown in FIG. 1, the buffer circuit 100 has the output buffer circuit OB connected to the 10 terminal T 3 . The output buffer circuit OB has a first CMOS inverter OB 1 consisting of an eleventh P-channel transistor P 11 and a fifth N-channel transistor N 5 . Also, the output buffer circuit OB has a second CMOS inverter OB 2 consisting of an twelfth P-channel transistor P 12 and a sixth N-channel transistor N 6 . The output buffer circuit OB amplifies an output signal.

Next, operation of the buffer circuit 100 according to the first embodiment will be described by classifying the operation into two modes: the input mode and the output mode.

In the input mode, the enabling signal OE will be H-level. The P-channel transistors P 1 to P 6 make up a p-n junction diode between VDD and Vddf. Therefore, the value of Vddf is lower than that of the VDD by forward on-voltage Vf of the diode.

By the way, because the enabling signal OE is H-level, an output signal na 1 of the nand circuit NAND will be H-level, and the voltage of the output signal na 1 should be the same as the power source voltage VDD. However, the voltage of the output signal na 1 is supplied from the power source voltage VDD via the P-channel transistors P 7 and P 8 or P 9 and P 10 . Therefore, the nand circuit NAND cannot make the gate voltage of the P-channel transistor P 1 agree with the power source voltage VDD only by itself. In order to solve this problem, the P-channel transistor P 5 is formed. That is, because an inverted enabling signal ENN is supplied to a gate electrode of the P-channel transistor P 5 , it will be turned on in the input mode. Therefore, it will be possible to make the voltage of the output signal na 1 agree with the power source voltage VDD, and that makes the P-channel transistor P 1 certainly turned off. By that, the buffer circuit 100 will operate correctly without unnecessary current leak.

Next, a case wherein the Vt 3 is larger than VDD will be considered. For example, it is a case wherein voltage of the second input data Din 2 is 5V. In that case, the voltage of the signal na 1 is 0V, and so the N-channel transistor N 1 will be turned off.

Moreover, since the Vt 3 is larger than VDD, the P-channel transistor P 2 will be turned off. On the other hand, the P-channel transistor P 3 will be turned on. Therefore, the voltage Vx at the node X will agree with the Vt 3 , and the P-channel transistor P 6 will be turned off.

By the way, the parasitic diode D 1 is attached between the drain electrode of the P-channel transistor P 1 and the common bulk region 103 a . When the voltage Vt 3 at the IO terminal T 3 is larger than the power source voltage VDD, that the parasitic diode D 1 will be turned on. Providing that forward descending voltage of the parasitic diode D 1 is Vf, the voltage Vddf of the common bulk region 103 a can be expressed by a equation: Vddf=Vt 3 −Vf.

Next, because the voltage of the inverted enabling signal ENN is 0V, N-channel transistor N 2 will be turned off. Also, since the Vt 3 is larger than VDD, the P-channel transistor P 3 will be turned on, and the voltage Vt 3 will be supplied to the Node X.

In addition to that, since the voltage Vx at the node X will agree with the Vt 3 and Vddf=Vt 3 −Vf, the P-channel transistors P 8 and P 10 forming the nand circuit NAND will be turned off. Also, because the voltage of the inverted enabling signal ENN is 0V, the N-channel transistor N 4 will be turned off.

Also, since the P-channel transistor P 4 is turned on, ENN=0, na 1 =Vt 3 and Vddf=Vt 3 −Vf, so the drain voltage of the P-channel transistor P 5 will be Vt 3 . On the other hand, source voltage of the P-channel transistor P 5 is VDD, and so that transistor P 5 will be turned on, and some electricity will be flown through the transistor P 5 . Meanwhile, the gate voltage of the P-channel transistor P 1 will be Vt 3 . Since the voltage Vt 3 is higher than the voltage Vddf at the common bulk region 103 a , the P-channel transistor P 1 will be turned off.

Therefore, in that case, there is no path for the current leak except the P-channel transistor P 5 . When Vt 3 is higher than VDD, Vx equals to Vt 3 . The P-channel transistor P 2 reduces Vx to 0V when Vt 3 comes to 0V. Further, the P-channel transistor P 3 itself can perform the same thing as the P-channel transistor P 2 does; therefore, the P-channel transistor P 2 may be omitted.

Next, the operation of the buffer circuit 100 in the output mode will be explained. In the output mode, the enabling signal OE is at the L-level.

The voltage of the inverted enabling signal will be VDD, and so the N-channel transistor N 2 is turned on, and the voltage Vx at the node X will be 0V. In the output mode, the voltage Vt 3 at the 10 terminal T 3 will never be higher than the power source voltage VDD; therefore, the P-channel transistors P 2 and P 3 are turned off.

On the other hand, the gate electrode of the P-channel transistor P 6 is supplied the voltage Vx (=0V), so it will be tuned on. Therefore, the power source voltage VDD is supplied to the common bulk region 103 a , and its voltage Vddf will agree with VDD. Also, bulk voltage of the P-channel transistors P 7 to P 10 that form the nand circuit NAND will be VDD, and so the nand circuit NAND will operate as a general inverted “AND” circuit. More specifically, the P-channel transistors P 7 and P 8 are always kept turned on since Vx=0V, and the P-channel transistor P 9 is turned off since ENN=VDD. On the other hand the N-channel transistor N 4 is turned on. By that, the nand circuit NAND will be similar to an inverter circuit that series-connects the P-channel transistor P 7 and the N-channel transistor N 3 . Therefore, the output signal na 1 of the nand circuit NAND will be inversion of the first input data Din 1 .

Moreover, because the gate voltage of the P-channel transistor P 5 will be VDD, the P-channel transistor P 5 will be turned off. In addition, the P-channel transistor P 4 will be turned off.

Therefore, the buffer circuit 100 in the output mode will be similar to a circuit that consists of a series connection of a first inverted circuit (corresponding to the nand circuit NAND and the nor circuit NOR) and a second inverted circuit consisting of the P-channel transistor P 1 and the N-channel transistor N 1 . The buffer circuit 100 can output, from the IC terminal T 3 , an output data Dout that has the same polarity as the first input data Din 1 and is amplified. Also, in the output mode, there is no path for unnecessary current leak.

In the input mode, although a little current leak flows into the P-channel transistor P 5 when Vt 3 is higher than VDD, the P-channel transistor P 5 is positively turned on only when 0V<Vt 3 <VDD. The object is to turn off the P-channel transistor P 1 by biasing VDD to the gate electrode of the P-channel transistor P 1 . Therefore, the size of the P-channel transistor P 5 may be small.

As said in the above, in the buffer circuit according to the first embodiment, there is no risk for problems such as latch up or the like because no unnecessary current flows although there would be a little current leak into the P-channel transistor P 5 when Vt 3 is high than VDD in the input mode. Moreover, comparing to the conventional buffer circuit shown FIG. 5, a chip area occupied by the buffer circuit 100 can be drastically reduced because the number of P-channel transistors and the N-channel transistors at the output step can be reduced.

Next, a second embodiment will be described. In the above-described buffer circuit 100 according to the first embodiment, a little current leak flows into the P-channel transistor P 5 when Vt 3 is higher than VDD in the input mode. In a buffer circuit 200 according to the second embodiment, consuming electric current can be reduced more than the first embodiment.

FIG. 4 shows a structure of the buffer circuit 200 according to the second embodiment. As shown in FIG. 4, the buffer circuit 200 has a P-channel transistor P 5 ′ of which gate length is longer than that of the P-channel transistor P 5 of the buffer circuit 100 . Also, the buffer circuit 200 has a delay circuit DL that delays an inverted enabling signal ENN and is configured between an inverter INV 1 and a gate electrode of a N-channel transistor N 2 . Other components are similar to those of the buffer circuit 100 according to the first embodiment.

The gate of the P-channel transistor P 5 ′ is long; therefore, on-resistance is higher than that of the P-channel transistor P 5 (FIG. 1 ). By that, in the output mode, when the voltage (Vt 3 —VDD) is impressed between a source electrode and a drain electrode of the transistor P 5 ′, electric current value flowing there is smaller than those of the transistor P 5 (FIG. 1 ). In a general integrated circuit, transistors are made up with the same unit; however, in this embodiment, the P-channel transistor P 5 ′ is made with other unit. For example, when a gate length of other P-channel transistors P 2 to P 12 is 10 micrometers, the gate length of the P-channel transistor P 5 ′ will be set at 100 micrometers. By doing so, when Vt 3 is higher than VDD in the input mode, it is possible to reduce electric current value flowing through the P-channel transistor P 5 ′ to about one tenth.

By the way, the gate electrode of the P-channel transistor P 1 holds stray capacitance because of wiring pattern or the like. By the stray capacitance, if the on-resistance of the P-channel transistor P 5 ′ is increased, time constant will increase. Providing that the signal na 1 is 0V in the output mode and the mode is changed to the input mode, the gate voltage of the P-channel transistor P 1 gradually increases from 0V to 3V by the large time constant. Therefore, there may be a risk for the P-channel transistor P 1 not to be turned off quickly. In this embodiment, the delay circuit DL is configured to supply the voltage VDD to the gate electrode of the P-channel transistor P 1 just after the change of the modes.

The delay circuit DL, as shown in FIG. 4, has a buffer B, inverters INV 2 , INV 3 , a P-channel transistor P 13 and an N-channel transistor N 7 . According to this delay circuit, because of transfer delay of each component, the inverted enabling signal ENN will be delayed and output as a delayed inverted enabling signal ENN′.

The N-channel transistor N 2 will be turned off slightly after the inverted enabling signal ENN is changed from H-level to L-level (i.e., when the mode is changed from the output mode to the input mode). As a result of that, the N-channel transistor N 2 has been turned on, and the voltage Vx at the Node X is kept at 0V; therefore, the P-channel transistors P 8 and P 10 shown in FIG. 2 will be turned on. Meanwhile, the gate voltage of the P-channel transistor P 9 has already been at the L-level, so the voltage of the output signal na 1 of the nand circuit NAND agrees with the power source voltage VDD.

Then, after a predetermined period (e.g., 15ns) elapsed, the N-channel transistor N 2 will be turned off. By that, the P-channel transistor P 5 will be turned on as similar to the first embodiment, and the gate electrode of the P-channel transistor P 1 will be biased at Vt 3 .

As said in the above, in the buffer circuit 200 according to the second embodiment, leak current can be reduced by increasing on-resistance by making the length of the gate of the P-channel transistor P 5 ′ long, and the P-channel transistor P 1 can be turned off certainly in the input mode by using the delay circuit DL.

In this specification, a first P-channel transistor in claim corresponds to the first P-channel transistor P 1 in the embodiments. A second P-channel transistor in claim corresponds to the third P-channel transistor P 3 in the embodiments. A third P-channel transistor in claim corresponds to the fourth P-channel transistor P 4 in the embodiments. A fourth P-channel transistor in claim corresponds to the fifth P-channel transistor P 5 in the embodiments. A fifth P-channel transistor in claim corresponds to the sixth P-channel transistor P 6 in the embodiments. A sixth P-channel transistor in claim corresponds to the second P-channel transistor P 2 in the embodiments.

The present invention has been described in connection with the preferred embodiments. The invention is not limited only to the above embodiments. It will be apparent to those skilled in the art that various modifications, improvements, combinations and the like can be made.