Integrated circuit comprising an output transistor with a controlled fall time

Abstract

An integrated circuit electrically is supplied with a voltage and includes an output MOS transistor having a gate driven by an output of a logic circuit and a circuit for biasing the gate of the output MOS transistor. The circuit for biasing the gate is provided for lowering a gate-source bias voltage of the output MOS transistor in a conductive state in relation to the gate-source bias voltage that would otherwise be provided by the output of the logic circuit. The present invention is particularly applicable to output stages for I2C buses.

Description

FIELD OF THE INVENTION

[0001] The present invention relates to the field of electronic circuits, and, more particularly, to the control of the electrical characteristics of an output transistor in an integrated circuit.

BACKGROUND OF THE INVENTION

[0002] The present invention relates to output transistors, such as MOS transistors, connected to a data bus and presenting a stray capacitance. In particular, the invention relates to controlling a working point in a low state and control time for passing to the low state of such output transistors.

[0003] To aid in understanding, FIG. 1 shows schematically a conventional output stage 10 of an integrated circuit 20 (implemented with CMOS technology) connected to the input of a data bus 30 of the I2C type. The output stage 10 includes an inverting gate 11, the output of which drives the gate of an output transistor TOUT. The output transistor TOUT may be an NMOS transistor, for example. The inverting gate 11 includes a PMOS transistor T1 and a NMOS transistor T0 connected by their drains. The source of the PMOS transistor T1 receives the supply voltage Vcc of the integrated circuit 20, and the source of the NMOS transistor T0 is connected to ground (GND).

[0004] The inverting gate 11 receives as an input (i.e., the gates of the transistors T0, T1 receive) Binary data DTX to be emitted on bus 30, and delivers inverted data {overscore (DT)}X on the gate G of the output transistor TOUT. A bit at 1 at the output of inverting gate 11 corresponds to the application of a voltage Vcc to the gate of the transistor TOUT (i.e., the transistor T1 is conductive) and a bit at 0 corresponds to the connection of the gate of transistor TOUT to ground (i.e., the transistor T0 is conductive).

[0005] The transistor TOUT is connected by its drain D to the data transmission line 31 of the bus 30, and its source S is connected to ground. The bus 30, as seen from its input, presents a stray capacitance CBUS between transmission line 31 and ground. When the transistor TOUT is in the OFF state, the line 31 is maintained by default at 1 by a bias or pull-up resistor RBUS receiving a bias voltage V1. In practice, the voltage V1 may be equal to Vcc and is, for example, about 5V.

[0006] When the transistor TOUT receives a bit at 1 on its gate (i.e., the voltage Vcc), the transistor TOUT becomes conductive, the capacitance CBUS discharges and the output of stage 10 passes to 0. The discharge time is designated TFALL. An output 0 corresponds to a drain-source voltage VDS and a drain-source current IDS having stable values designated respectively by VOL and IOL and defining the working point in the low state of the transistor TOUT.

[0007] For historical reasons relating to the use of bipolar transistors in input or output stages of integrated circuits, the industrial specifications of I2C buses require a rather low voltage VOL on the order of 0.2 to 0.4 V, and a rather high current IOL on the order of 0.7 to 3 mA. With MOS technology, such a working point may be achieved only with an NMOS transistor having a gate width W much larger than the gate length L, i.e., a ratio W/L clearly larger than 1.

[0008] With the saturation current ISAT of a MOS transistor being proportional to the ratio W/L, the transistor TOUT is thus traversed by a significant current during its switching, the capacitance CBUS discharges almost instantaneously, and the time for passing to 0 TFALL is very short. However, the fact that the time TFALL is very short generates an undesirable electronic noise in transmission line 31. It is thus desirable that the time TFALL be lengthened to at least about 20 ns.

[0009] However, in an output stage 10 as described above, obtaining a rather long time TFALL is incompatible with obtaining a low voltage VOL and a rather high current IOL. As a matter of fact, the discharge of the capacitance CBUS is basically provided by the transistor current in the saturated state, or the current ISAT. Thus, if the ratio W/L of the transistor is decreased to limit the discharge the current ISAT of the capacitance CBUS, the voltage VOL is increased at the same time as the current IOL is decreased.

SUMMARY OF THE INVENTION

[0010] An object of the present invention is to provide a method for lengthening the time TFALL for passing to zero of an output transistor of the type described above, without modifying its working point VOL, IOL in the low state.

[0011] Another object of the present invention is to provide such an output stage including a controller for controlling the fall time of the output transistor when it is connected to a capacitive data bus.

[0012] These objects are achieved by a method of lengthening the time for passing to zero a signal on a terminal of an output transistor of the MOS type connected to a data transmission line presenting a predetermined electric capacitance while keeping a substantially identical working point in the low state of the transistor and where the gate of the transistor is driven by a logic circuit present in an integrated circuit receiving a predetermined supply voltage. The method includes lowering the gate-source bias voltage of the transistor in the conductive state in relation to the gate-source bias that would otherwise appear at the output of the logic circuit. The transistor may be designed with an increased width over length ratio of the gate of the transistor for keeping the initial working point in the low state.

[0013] Moreover, the method may include connecting the gate of the transistor to a gate bias circuit. The gate bias circuit may be arranged to reduce the voltage in the high state delivered by the logic circuit on the gate of the transistor. Specifically, the gate bias circuit may include a first MOS transistor substantially identical to the output transistor and arranged as a diode, and a native transistor arranged as a diode and arranged in series with the first MOS transistor. Furthermore, a gate-source bias voltage lower than 2 V may be applied to the output transistor for turning it on.

[0014] The present invention also relates to an integrated circuit electrically supplied with a predetermined voltage and including an output MOS transistor, a logic circuit providing an output for driving a gate of the output MOS transistor, and a circuit for biasing the gate of output MOS transistor. Accordingly, the gate-source bias voltage of the transistor is lowered in the conductive state in relation to the gate-source bias voltage that would otherwise be present at the output of the logic circuit.

[0015] More specifically, the gate-source bias voltage of the transistor in the conductive state may be lower than 2 V. The bias circuit may be arranged to reduce the voltage in the high state delivered by the logic circuit. Also, the bias circuit may include at least a first MOS transistor arranged as a diode and substantially identical to the output transistor. The bias circuit may also include a native transistor arranged as a diode and in series with the first MOS transistor.

[0016] Furthermore, the output transistor may have a width over length ratio of the gate that gives to it a working point in the low state in accordance with the specifications of I2C buses and a fall time of the output signal at least equal to 20 nanoseconds. In practice, the integrated circuit may take the form of an EEPROM memory, for example, and the output stage may be arranged for delivering data read in the memory.

BRIEF DESCRIPTION OF THE DRAWINGS

[0017] These objects, characteristics and advantages of the present invention will be described with more detail in the following description of the method according to the invention and an embodiment of an output stage according to the invention, given by way of non-limitative example, in conjunction with the accompanying drawings, in which:

[0018] FIG. 1, previously described, is a schematic diagram of a conventional series output stage of an integrated circuit according to the prior art;

[0019] FIG. 2 is a graph illustrating current/voltage curves for a conventional output transistor according to the prior art and an output transistor according to the present invention;

[0020] FIG. 3 is a schematic diagram illustrating an output stage according to the present invention;

[0021] FIG. 4 is a schematic diagram of an exemplary embodiment of a bias circuit according to the present invention; and

[0022] FIG. 5 is a schematic block diagram of an EEPROM memory illustrating an application of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0023] As noted above, the achievement of a rather long time TFALL in the output stage 10 represented in FIG. 1 is in practice incompatible with the achievement of a working point including a rather low voltage VOL and a rather high current IOL. This paradox will be better understood in relation to FIG. 2, which shows the current/voltage graph IDS=f(VDS) of the output transistor TOUT. When the gate of transistor TOUT receives voltage VCC delivered by the transistor T1 of the inverting gate 11 (FIG. 1), the transistor TOUT progressively turns on. That is, the transistor TOUT passes from the saturated state to the linear state and its drain-source voltage VDS slides along graph C1 from its initial value, which is the bias voltage V1 of bus 31, to the value VOL. During the first part of the discharge period of capacitance CBUS, the transistor is in the saturated state because the voltage VDS is above the differential gate voltage VGS−VT. The current IDS, conventionally designated ISAT, is constant and follows the conventional equation:

IDS=ISAT=(KP/2)(W/L)(VGS−VT)2  (1)

[0024] where KP is a constant (representing mobility of the carriers), VT is the threshold voltage of the transistor, and VGS is the gate voltage. The gate voltage VGS being equal to Vcc, equation (1) may also be written:

IDS=iSAT=(KP/2)(W/L)(Vcc−VT)2  (2)

[0025] During the second part of the discharge period, the voltage VDS becomes lower than the differential gate voltage Vcc−VT. Transistor TOUT then works in the linear mode, and the drain-source current IDS follows the conventional equation:

IDS=KP(W/L)(Vcc−VT)VDS  (3)

[0026] When the capacitance CBUS is completely discharged, the transistor presents the working point POL represented in FIG. 2, defined by the specific values IOL and VOL of its drain-source current IDS and its drain-source voltage VDS. At the working point POL, equation (3) may be written as:

IOL=KP(W/L)(Vcc−VT)VOL  (4)

[0027] At the working point POL, the voltage at the terminals of resistor RBUS and the voltage at the terminals of transistor TOUT are bound with the following equation:

V1=RBUS*IOL+VOL  (5)

[0028] Thus, combining equations (4) and (5):

V1=RBUSIOL+IOL/[KP(W/L)(Vcc−VT)]  (6)

[0029] It may be deduced therefrom that the value of IOL is:

IOL=V1/[RBUS+1/[KP(W/L)(Vcc−VT)]]  (7)

[0030] Combining equations (5) and (7), it may also be deduced the value of VOL is:

VOL=V1/[1+(RBUSKP(W/L)(Vcc−VT)]  (8)

[0031] In equations (7) and (8), parameters V1, Vcc, RBUS are imposed by the environment of the output transistor TOUT. For lengthening the time TFALL, a solution could includes decreasing the ratio W/L of the transistor. This would allow the limitation of the discharge current ISAT of capacitance CBUS during the saturation period (equation 2), and the limitation of the current IDS during the linear period (equation 3). However, equations (7) and (8) also show that a decrease of the parameter W/L would lead to a clear increase of voltage VOL and a small decrease of current IOL. The working point POL would thus slide towards a working point POL′ on a graph C2′ represented in FIG. 2, which does not satisfy the industrial specifications.

[0032] Furthermore, the working point POL provided by the industrial specifications imposes the ratio W/L of the transistor, and thus imposes the saturation current ISAT. Providing a voltage VOL and a current IOL in accordance with the specifications is contradictory with providing a longer discharge time TFALL which would allow the decrease of the switching noise.

[0033] The present invention provides for a decrease of the saturation current ISAT of the output transistor TOUT without substantially offsetting the working point POL. Furthermore, the bias voltage VGS of the source of transistor TOUT is decreased while the ratio W/L is increased (instead of decreasing it). Before describing an example of implementation of this method, the theory of the method will be first explained.

[0034] The current IOL1 and the saturation current ISAT1 of the conventional transistor TOUT are given by the following equations (9) and (10) below, in which the characteristics previously indicated by IOL, VOL, W/L, ISAT are now designated IOL1, VOL1, W1/L1, ISAT1:

IOL1=KP(W1/L1)(Vcc−VT)VOL1  (9)

ISAT1=(KP/2)(W1/L1)(Vcc−VT)2  (10)

[0035] Furthermore, the following equations (11) and (12) relate respectively to the current IOL2 and the saturation current ISAT2 of a transistor according to the invention, having its gate biased by a voltage V2 different from Vcc, and having its drain still biased by the voltage V1 of bus 30:

IOL2=KP(W2/L2)(V2−VT)VOL2  (11)

ISAT2=(KP/2)(W2/L2)(V2−VT)2  (12)

[0036] According to the invention, it is desired that:

IOL2=IOL1=IOL=Constant  (13)

VOL2=VOL1=VOL=Constant  (14)

[0037] while satisfying the following equation:

ISAT2=N*ISAT1  (15)

[0038] This is because it is desired to lengthen the discharge time TFALL, where N is a number lower than 1 (ranging between 0 and 1). Combining equations (9) (11) (13) and (14), the following equation results:

KP(W1/L1)(Vcc−VT)VOL=KP(W2/L2)(V2−VT)VOL  (16)

that is:

V2−VT=(W1/L1)(L2/W2)(Vcc−VT)  (17)

[0039] Combining equations (10) (12) and (15):

(W2/L2)(V2−VT)2=N(W1/L1)(Vcc−VT)2  (18)

[0040] Introducing equation (17) into equation (18):

(W2/L2)[(W1/L1)(L2/W2)(Vcc−VT)]2=N(W1/L1)(Vcc−VT)2  (19)

that is:

W2/L2=(1/N)(W1/L1)  (20)

[0041] Introducing equation (20) into equation (17) results in the following:

V2−VT=N(Vcc−VT)  (21)

[0042] Thus, according to the invention, decreasing the differential gate voltage Vcc−VT by the factor N to obtain a differential gate voltage V2−VT in response to equation (21), and increasing in parallel, by the inverse factor 1/N, the ratio W1/L1 of a conventional transistor to obtain a ratio W2/L2 in response to equation (20), allows the design of a transistor having a saturation current ISAT2 decreased by factor N in relation to the saturation current ISAT1 of a conventional transistor. This is the case even while keeping the working point POL (VOL, IOL) of the conventional transistor imposed by the industrial specifications.

[0043] The current/voltage graph C3 of such a transistor according to the present invention may be seen in FIG. 2. The graph C3 presents a saturation current ISAT2 much lower than current ISAT1 (i.e., the flat part of the graph) but keeps a linear part identical to the linear part of graph C1 when the drain-source voltage VDS ranges between VOL and V2−VT. Such a transistor limits the saturation current that provides a large part of the discharge of the capacitance CBUS. The increase of the time TFALL is equal to the ratio between the surface defined by graph C1 between the voltages VOL, V1 and the surface defined by graph C3 between the voltages VOL, V1.

[0044] An output stage 40 according to the present invention is shown in FIG. 3. The output stage 40 is similar to the stage 10 of FIG. 1 except that a bias circuit according to the invention, represented in the form of a block 41, is connected between the gate of the transistor TOUT and ground. When the transistor T1 of the inverting gate 11 is conductive, the circuit 41 regulates the gate of the transistor and imposes a voltage V2 in response to equation (21) above. N is chosen to obtain a time TFALL which does not generate an undesirable switching noise, and TFALL is preferably above 20 ns.

[0045] According to an advantageous embodiment illustrated in FIG. 4, the bias circuit 41 includes an element T3 and a transistor T4 arranged as diodes and connected in series. The transistor T4 has substantially the same structure and size as the transistor TOUT. In these conditions, the voltage V2 at the terminals of circuit 41, applied to the gate of transistor TOUT, is equal to:

V2=V3+VT  (22)

[0046] where VT is the threshold voltage of the diode transistor T4, and V3 is a voltage appearing at the terminals of element T3.

[0047] With this arrangement, the temperature variations of the threshold voltage VT of transistor TOUT are partly compensated by the voltage V2. The threshold voltage VT of the transistor T4 is identical and varies in the same way as the threshold voltage VT of the transistor TOUT. Thus, the differential gate voltage V2− VT of the transistor TOUT is given by:

V2−VT=(V3+VT)−VT=V3  (23)

[0048] and only depends on the voltage V3. The element T3 may be a reference element providing a stable voltage according to temperature or, more simply, a native NMOS transistor arranged as a diode, the threshold voltage VTNAT of which is lower that the one of an enhanced transistor.

[0049] In these conditions, for a voltage Vcc of about 5 V, a voltage VT of 0.7 V and a voltage VTNAT of 0.4 V, the voltage V2 imposed by the bias circuit 41 allows the reduction of about 74% of the differential gate voltage in relation to the value that it would have in the absence of circuit 41. A decrease in the same proportions of the saturation current ISAT is also provided. This results from the following calculations:

V3VTNAT+VT=N(Vcc−VT)  (24)

that is:

N=(VTNAT+VT)/(Vcc−VT)  (25)

that is:

N=0.26  (26)

[0050] It will be apparent to those skilled in the art that the present invention is likely to have various alternative embodiments and applications. For example, the present invention may also be applied when the bias resistor of the drain of transistor TOUT is a resistor arranged in the output stage 10 instead of being the resistor RBUS. The drain bias voltage will then be the voltage Vcc instead of the voltage V1 (although in practice V1 and Vcc are generally identical).

[0051] Furthermore, although the bias circuit 41 is shown as connected to the gate of transistor TOUT in FIG. 3, it is also possible to connect this circuit in another point of the output stage. For example, the inverting gate 11 could be supplied with a regulated voltage lower than voltage Vcc and deliver, in the high state, a gate voltage VGS in response to equation 21 above.

[0052] Additionally, although the technical problem having led to the implementation of the present invention relates to buses of the I2C type, the present invention is not limited to such an application and may relate to other data transmission lines of the capacitive type.

[0053] An application of the present invention to electrically erasable and programmable memories, or EEPROM memories, provided with a serial output for a bus I2C is illustrated in FIG. 5. The EEPROM memory 50, represented very schematically in FIG. 5, includes a memory array 51 including memory cells 52 arranged in lines and columns, a line decoder 53, a column decoder 54, and a read circuit 55 connected to the column decoder. Read circuit 55 includes sense amplifiers 5A for delivering binary words Wi,j read in the memory array 51, e.g., bytes. The binary words Wl,j are applied to a register 56 with a serial output and parallel loading. The serial output of the register 56 is sent to an output terminal 57 of the memory 50 by an output stage 40 as described above.

Claims

1. Method of lengthening the time for passing to zero of a terminal of an output transistor (TOUT) of the MOS type connected to a data transmission line (31) presenting a predetermined electric capacitance (CBUS), while keeping a substantially identical working point (VOL, IOL) in the low state of the transistor (TOUT), the gate of the transistor being driven by a logic circuit (11) present in an integrated circuit (20) receiving a predetermined supply voltage (Vcc), the method being characterized in that it comprises a step of lowering the gate-source bias voltage (VGS) of the transistor in the conductive state in relation to the one (Vcc) which would naturally appear at the output of the logic circuit (11) if such a step was not provided, and a step of designing the transistor consisting in increasing the width over length ratio (W/L) of the gate of the transistor for keeping the initial working point in the low state (VOL, IOL) of the transistor.

2. Method according to

claim 1, comprising a step consisting in connecting the gate of the transistor to a gate bias circuit (41).

3. Method according to

claim 2, wherein the gate bias circuit is arranged to reduce the voltage in the high state (Vcc) delivered by the logic circuit (11) on the gate of the transistor.

4. Method according to

claim 3, characterized in that the gate bias circuit comprises at least, arranged in series, a first MOS transistor (T4) arranged in diode, substantially identical to the output transistor, and a native transistor (T3) also arranged in diode.

5. Method according to one of the

claims 1 to

4, wherein a gate-source bias voltage (VGS) lower than 2 Volt is applied to the output transistor for turning on the transistor.

6. Integrated circuit (20, 50) electrically supplied with a predetermined voltage (Vcc) and comprising an output MOS transistor (TOUT), the gate of which is driven by the output of a logic circuit (11), characterized in that it comprises a circuit (41) for biasing the gate of the transistor (TOUT), provided for lowering the gate-source bias voltage (VGS) of the transistor in the conductive state in relation to the one (Vcc) that would be fixed by the output of the logic circuit (11) in the absence of such a bias circuit (41).

7. Integrated circuit according to

claim 6, wherein the gate-source bias voltage (VGS) of the transistor in the conductive state is lower than 2 Volt.

8. Integrated circuit according to one of the claims 6 and 7, wherein the bias circuit (41) is arranged to reduce the voltage in the high state (Vcc) delivered by the logic circuit (11).

9. Integrated circuit according to

claim 8, characterized in that the bias circuit (41) comprises at least a first MOS transistor (T4) arranged in diode substantially identical to the output transistor.

10. Integrated circuit according to

claim 9, characterized in that the bias circuit comprises a native transistor (T3) arranged in diode, in series with the first diode transistor (T4).

11. Integrated circuit according to one of the

claims 6 to

10, characterized in that the output transistor (TOUT) presents a width (W) over length (L) ratio (W/L) of the gate that gives to it a working point in the low state (POL, VOL, IOL) in accordance with the specifications of I2C buses and a fall time (TFALL) of the output signal at least equal to 20 nanoseconds.

12. Integrated circuit according to one of the

claims 6 to

11, characterized in that it takes the form of an EEPROM memory (50), and in that the output stage (40) is arranged for delivering data (Wi,j) read in the memory.