High-Level CMOS Driver Circuit

This invention relates to semiconductor devices, and more particularly to driver circuits of the type used in CMOS VLSI semiconductor memory devices or the like.

Semiconductor memory devices of the dynamic read/write type are constructed in N-channel technology as shown in U.S. Pat. No. 4,239,993, issued to McAlexander, White and Rao, assigned to Texas Instruments. As set forth in such Patent, some of the clocks and control signals are preferably boosted to voltage levels above the supply. For example, the row lines are boosted so that a full one level can be written into the storage capacitors. Thus, the row decoder which activates the row lines must produce an output exceeding the power supply voltage. This is accomplished in NMOS circuitry by bootstrapping, but is not as easily achieved in CMOS technology.

Accordingly, one of the problems encountered in CMOS memory designs is that of getting the gate of N-channel transfer and load devices above the supply voltage Vcc in order to eliminate threshold losses on the signal level. Due to forward biasing the drain junction of the P-channel load device, the output of an inverter cannot be booted above Vcc as with NMOS drivers. One solution to this problem in CMOS is to provide a Vcc+ voltage, i.e., a voltage supply higher than the Vcc level supplied to the chip. Generated on chip from the Vcc supply, this voltage is sufficiently greater than Vcc that any NMOS device will be driven into the triode region, eliminating the Vt loss problem. The difficulty of this approach is generating sufficient current to handle the current requirements of any CMOS inverters connected to this Vcc+ supply. If the load to be driven has large capacitance and must be driven with fast transitions, this requires wide load transistors in the driver to meet the transition requirement. The Vcc+ thus has not only the capacitor charging current requirement but it must also be able to meet the large switching current in the inverter driver stage, when both the N and P channel devices are on.

It is therefore the principal object of this invention to provide improved driver circuits for semiconductor dynamic memory devices or the like, particularly using CMOS circuitry. Another object is to provide an improved driver circuit which produces an output exceeding the power supply voltage, but does not exhibit a high current drain upon an on-chip high voltage supply circuit.

In accordance with one embodiment of the invention, a CMOS driver circuit for producing a high-level output voltage, exceeding the voltage supply uses an inverter and an output stage. The inverter circuit is supplied from the usual voltage supply for the chip, and this inverter is followed by the output stage which is supplied from an on-chip generator producing a higher level voltage. The current requirement of the on-chip generator is kept to a minimum by charging the output capacitance mainly from the voltage supply, and using the output stage only for the latter part of a charging cycle.

The novel features believed characteristic of the invention are set forth in the appended claims. The invention itself, however, as well as other features and advantages thereof, will be best understood by reference to the detailed description which follows, wherein:

FIG. 1 is an electrical schematic diagram of a driver circuit according to the invention;

FIG. 2 is a timing diagram showing voltage for various nodes, and current from the high voltage supply, as a function for time, in the circuit of FIG. 1;

FIG. 3 is an electrical schematic diagram of a driver circuit according to another embodiment of the invention; and

FIG. 4 is a timing diagram showing voltage for various nodes, and current from the high voltage supply, as a function of time, in the circuit of FIG. 3.

Referring to FIG. 1, a CMOS driver circuit according to the invention receives a low-level input "t" at an input node 10 and generates a high-level output "T" at an output terminal 11. This circuit uses the standard Vcc voltage on line 12 as supplied to the chip, usually +5v dc, and also uses a Vcc+ supply on line 13. The Vcc+ supply is at least one Vt above Vcc. A CMOS inverter including a P-channel transistor 14 and an N-channel transistor 15 receives the input 10 on the transistor gates, and the output 16 of this inverter is connected through an N-channel transfer device 17 to a node 18. The inverter is supplied from the Vcc voltage 12, and the gate of the transfer device 17 is connected to this supply. The output circuit for producing the high-level output T employs P-channel transistors 20 and 21 each having its source connected to the Vcc+ supply 13. The transistor 20 couples the node 18 (which is the same as the output node 11) to Vcc+ when turned on by a zero voltage on its gate node 22; this gate node 22 is coupled to Vcc+ through the source-to-drain path of the other P-channel transistor 21, and coupled to the input node 10 via an N-channel transistor 23 which has its gate connected to the node 18. The transistor 21 has its gate connected to output terminal 11.

In the operation of the circuit of FIG. 1, first assume that the input t at node 10 is initially high, as seen in FIG. 2. This forces node 16 to be low, as well as node 18. The low voltage on node 18 causes node 22 to go to Vcc+, via device 21. Since node 18 is low, transistor 23 is off and this allows node 22 to reach Vcc+. With node 22 at Vcc+, transistor 20 is off and there is no continuous current flow from Vcc+. When input t at node 10 goes low at time t1 of FIG. 2, the node 16 goes high, to Vcc. Node 18 follows node 16 until device 17 begins to turn off due to insufficient gate-to-source voltage (between Vcc and node 18). Device 23 is designed such that node 22 is discharged and turns device 20 on about the time that device 17 is turning off. Device 20 continues charging the output nodes 18 and 11 to Vcc+. There is a short period of time while the output is reaching Vcc+ that devices 21 and 23 conduct charge from Vcc+, so there is current flow Icc+ as seen in FIG. 2. As node 18 reaches a voltage within a P-channel Vt of Vcc+, device 21 turns off. At this time only charge to the load capacitor 25 is pulled from Vcc+ as node 18 charges to a full Vcc+ level.

At time t2 the input 10 goes back up to Vcc, and this turns on the transistor 15 and turns off transistor 14. The capacitance 25 discharges rapidly through node 18 and transistor 17. There is some transitory current Icc+ through P-channel transistor 20, but very little because the node 22 follows the node 10 high, since transistor 23 is on.

The circuit of FIG. 1 is designed to reduce the current requirements on the Vcc+ supply 13 by reducing both the capacitor 25 charging current and inverter switching current supplied by Vcc+. Essentially all of the switching current requirement is removed by connecting the inverter 14,15 to Vcc line 12 rather than to Vcc+ line 13. Also a significant portion of the load capacitance charging current 25 when t goes low is initially supplied by the Vcc supply 12 through transistors 14 and 17. The Vcc+ supply is required only to charge the load capacitor 25 from approximately 80% of Vcc to the final Vcc+ level. The transfer device 17 prevents the drain of device 14 from forward biasing when node 18 goes to Vcc+, and devices 20, 21 and 23 control the charging of the load capacitance 25 from Vcc+ during the final part of the charging.

Another embodiment of the circuit of the invention is illustrated in FIG. 3. The inverter including the transistors 14 and 15 is the same, and the node 16 is coupled through a transistor 17 to the node 18 (and output 11) to charge the output capacitance 25. But the gate of the transistor 17 is driven from a node 30 at the output of a CMOS inverter having transistors 31 and 32, instead of directly from Vcc. This inverter has the voltage from node 16 connected to the gates of the inverter transistors 31 and 32, so the gate of the transistor 17 is charged to Vcc-Vt when the input 10 is high (node 16 is low), with a Vt drop across a transistor 33; when the input 10 is low (node 16 is high), the gate of transistor 17 is discharged through transistors 32 and 33. On the high voltage side of the circuit the capacitor 25 is charged from the Vcc+ line 13 by P-channel transistor 20 as before, and the gate of transistor 20 at node 22 is held low at the beginning of a cycle by a transistor 23 connecting this node 22 to the input 10, or held high by a P-channel transistor 21 at the final part of a cycle, similar to the circuit in FIG. 1. But in the embodiment of FIG. 3 the gate of transistor 21 is connected to a node 35 instead of to the output 11. The node 35 is coupled to the inverter output 16 by a transistor 37 having its gate connected to the node 38 at the gate of transistor 17. The node 35 is coupled to the high voltage Vcc+ line 13 by another P-channel transistor 39 which has its gate connected to the node 22. The gate of the transistor 23 is connected to the node 35 (instead of to the node 18).

The modified circuit of FIG. 3 gives some improvement in the reduction of charge required from Vcc+ (the current Icc+ of FIG. 4), at the expense of additional circuitry and thus additional area on the chip.

The basic operation remains the same but the reduction of charging current Icc+ from Vcc+ is due to

(1) charging the load capacitor 25 to a full Vcc level from node 16 before charging to Vcc+ through device 20. This is accomplished through bootstrapping node 38 above Vcc and allowing node 18 to track node 16 all the way. The timing is designed such that nodes 38 and 30 turn device 17 off at the same time node 22 turns device 20 on.

(2) charging node 22 to a higher voltage before node 18 falls, thus reducing current through device 20 upon the falling edge of node 18.

(3) by connecting the gate of device 21 to node 35 instead of 18, current in device 21 is reduced; this reduction is the result of the fact that node 35 is less capacitively loaded than node 18 and thus is more quickly charged to Vcc+ than node 18 (and 11) when the input 10 goes low.

As seen in FIG. 4, the voltage waveform for node 22 has much sharper transitions than that of FIG. 2. Also, the node 35 makes the transition from one level to the other very rapidly, thus reducing the switching current.

While this invention has been described with reference an illustrative embodiments, this description is not intended to be construed in a limiting sense. Various modifications of the illustrative embodiments, as well as other embodiments of the invention, will be apparent to persons skilled in the art upon reference to this description. It is therefore contemplated that the appended claims will cover any such modifications or embodiments as fall within the true scope of the invention.