Low Power Logic Array Device

The present invention relates to integrated circuits of the class known as programmable logic array devices, and in particular to such logic array devices with circuit elements for reducing power consumption of the devices.

Various forms of logic array devices have been in use for years. These devices have architectures based on various combinations of gate arrays, such as programmable-AND/programmable-OR, programmable-AND/fixed-OR, NOR/NOR, etc. While these devices have the flexibility to provide a large number of different custom logic functions, their power consumption can be quite high. This is largely because each logic gate within a gate array continually consumes as much as 20 milliWatts of power. The heat generated by the power consumption combined with the limited heat-sinking capability of the devices and their packages creates a practical gate count limit for such logic devices.

Early forms of logic array device used mask-programmable array elements based on bipolar transistor technologies. Later devices were introduced which used fusible links as array elements, allowing the devices to be programmed after manufacturing. Prior U.S. patents describing these early logic array devices include U.S. Pat. Nos. 3,541,543; 3,566,153; 3,702,985; 3,816,725; 3,818,452; and 3,849,638. More recently, programmable logic devices based on CMOS technology have been introduced. Two of these devices are described by Hartmann et al. in U.S. Pat. Nos. 4,609,986 and 4,617,479. These devices use EPROM or EEPROM transistors as the logic array elements. For a given gate count, these CMOS devices operate at lower power levels, because the circuitry surrounding the logic arrays generally does not consume power. For this reason, the introduction of these CMOS-based logic devices has helped to raise the gate count limit. However, the logic gate arrays themselves are still continually drawing current and therefore consuming power.

Another way to increase the permissible gate count for logic array devices is to simply reduce the current drawn by each array gate. However, this approach decreases the speed performance of the device and makes it unsuitable for applications that require high operating speed.

A circuit operation common to many logic array devices is the transmission gating of logic signals from an input terminal to an output terminal. U.S. Pat. No. 4,713,792 to Hartmann et al. describes one implementation of this operation. Available implementations of transmission gating have so far had limitations of either actively consumed power or less than optimum switching speed.

It is an object of the present invention to provide a logic array device with means for reducing the power consumption without substantially affecting the speed performance or logic function of the device.

The above object has been met with a logic array device which includes input switch detection circuitry which detects logic level transitions on inputs to a device's logic arrays. When and only when such a transition occurs are the gates of the logic arrays enabled by an array enable signal generating circuit, because only then will a transition be needed on outputs of the arrays. After such time has passed as is necessary for any such output transitions to occur, the logic array gates are disabled, and their power consumption is essentially eliminated. The logic outputs are stored in a latch while the arrays are disabled. Thus, the total power consumed by the logic device is greatly reduced. Because each gate receives full power when it is enabled, switching speed is not adversely affected.

FIG. 1 is a block diagram of a reduced power logic array device of the present invention showing its various functional components.

FIG.2 is a schematic diagram of an input driver circuit in the device of FIG. 1.

FIG. 3 is a schematic diagram of an input switch detector circuit in the device of FIG. 1.

FIG. 4 is a schematic diagram illustrating interconnection of several detector circuits of FIG. 3.

FIG. 5 is a schematic diagram of an AND array enable signal operating circuit in the device of FIG. 1.

FIG. 6 is a block diagram illustrating a physical layout of detector circuits of FIG. 3 or 4 and an enable signal generating circuit of FIG. 5.

FIG. 7A is a schematic diagram of an AND/OR logic array configuration used in the device of FIG. 1.

FIG. 7B is a schematic diagram of an AND/NOR/Inverter logic array configuration equivalent to the configuration in FIG. 7A.

FIG. 8 is a schematic diagram of an AND array gate in FIGS. 7A and 7B.

FIG. 9 is a schematic diagram of a combination NOR/Inverter logic gate in FIG. 7B.

FIG. 10 is a schematic diagram of an OR array enable signal generating circuit in the device of FIG. 1.

FIG. 11 is a timing diagram illustrating the relative switching times of AND and OR array enable signals SALE, BLPD and TPD with respect to input signal switch times.

FIGS. 12-12 are schematic diagrams of alternative circuits for generating and holding AND array gate outputs used for generating control signals in the device of FIG. 1.

FIG. 15 is a block diagram illustrating interconnections to a flip-flop storage element in the device of FIG. 1.

FIG. 16 is a schematic diagram of a low power, high speed, N:1 multiplexer in the device of FIG. 1.

FIG. 17 is a schematic diagram of an output driver in the device of FIG. 1.

FIG. 18 is a schematic diagram of an alternative embodiment of the input switch detector circuit in FIG. 3.

FIG. 19 is a block diagram of a control signal generating circuit using a one-time programmable element.

FIG. 20 is a schematic diagram of a NOR array gate for use in circuit in FIG. 21.

FIG. 21 is a schematic diagram of an alternative array gate with switch connectable elements.

FIG. 22 is a block diagram of an alternative embodiment of the flip-flop configuration of FIG. 15.

With reference to FIG. 1, a reduced power logic array device of the present invention has one or more arrays 31 of logic gates. Typically, the logic gates are AND gates, although NAND or NOR gates may also be used, and arrays 31 will be referred to hereafter as AND arrays 31. Input pins I.sub.0, I.sub.1, . . . , I.sub.n, input/output pins I/O and flip-flop outputs 33 all drive input buffers 35. Each input buffer 35 drives inputs 37 to an AND array 31. Each input buffer 35 also drives an input switch detector 39. The outputs 41 of the switch detectors 39 combine in an array enable signal generator 43 to generate signals that enable and disable the AND arrays 31 and OR gates 45. Each AND array 31 produces outputs on product term lines 47 which drive inputs to OR gates 45, flip-flops 49, flip-flop control signal generators 51 and output enable control signal generators 53. Multiplexers 55 receive as inputs both a combinatorial output from OR gates 45 along sum term lines 57 and a register output from flip-flops 49 along lines 59, and selectively pass one of these outputs to output drivers 61. Output drivers 61, when they are enabled by signals from output enable control signal generators 53, drive input/output pin I/O.

With reference to FIG. 2, a circuit implementation of input buffers 35 receives a signal INPUT from an input pin, an input/output pin or a flip-flop output. Signal INPUT goes through level shifting and gain inverters 63, producing a signal IT which is an input to an AND array. Signal outputs I and IB produced by input buffer 35 are used in the input switch detector 39 in FIG. 3. Signal IB is inverted compared to signals I and IT.

With reference to FIG. 3, an input switch detector 39 comprises a pair of pull-down transistors 65 and 67 whose gates are connected to respective pass transistors 71 and 77. Pass transistors 71 and 77, when on, transmit signals I and IB to the gates of pull-down transistors 65 and 67. The gates of pull-down transistors 65 and 67 are also connected to respective turn-off transistors 73 and 75. The gates of pass transistors 71 and 77, and of turn-off transistors 73 and 75 are toggled by a signal from a delay circuit 83. Transistors 71 and 75 are directly connected, while transistors 73 and 77 are connected indirectly through an inverter 84. Delay circuits 83 comprises an odd number of inverters 79-81 which invert and slow down the propagation of signal IT from input buffer 35.

In operation, pull-down transistors 65 and 67 are normally off, and signal LISD is held high by a pull-up transistor 69, seen in FIG. 4. If the input signal IT to the AND array is high, transistor 71 is off and turn-off transistor 73 is on, keeping transistor 65 off. Also, transistor 75 is off, pass transistor 77 is on, transmitting signal IB, which is low whenever signal IT is high, and thereby keeping transistor 67 off. Whenever the input IT switches from high to low, signal IB goes high, thereby turning on transistor 67 which pulls down signal LISD. After some time interval determined by the delay elements 79-81 in delay circuit 83, turn-off transistor 75 comes on and pass transistor 77 turns off, thereby causing pull-down transistor 67 to turn off and allowing signal LISD to return high. Pull-down transistor 65 remains off, since turn-off transistor 73 is now off and pass transistor 77 is now on, transmitting signal I which is low. Likewise, if the input IT switches from low to high, signal I goes high and transistor 65, now on, pulls down signal LISD. Again after a delay period, pass transistor 71 turns off and turn off transistor 73 comes on, thereby turning off transistor 65 and allowing signal LISD to return high. Therefore, the local input switch detection signal LISD pulses low whenever the input signal IT to the AND array switches, either from low to high or high to low.

As seen in FIG. 4, several input switch detection circuits 39 may be grouped together with a single pull-up transistor 69 for holding signal LISD high. Signal LISD will pulse low whenever an array input transition has occurred in any one of associated input buffers 35. Global input switch detection signal GISD, produced by a pull-down transistor 87 connected to the LISD signal line 89 via an inverter 91, will likewise pulse low. Any number of input buffers 35 and switch detectors 39 may be grouped together in the configuration 85 of FIG. 4. However, a practical limit is set by capacitive loading created on signal LISD. Any number of groupings of switch detector circuits shown in FIG. 4 may in turn be grouped together in a hierarchical manner to create an individual GISD signal, but the attendant capacitance again creates a practical limit. FIG. 5, for example, shows the groupings 85 in FIG. 4, each outputting a global detection signal GISD, being combined into larger groupings 85' to produce signals GISD.sub.1, . . . , GISD.sub.n.

With reference to FIG. 5, an AND array enable signal generator circuit 86 includes a NAND gate 87 with inputs from several input switch detection circuits 39 in FIG. 3, 85 in FIG. 4, or hierarchic combinations 85' of circuits 85 in FIG. 4. Normally, the signals GISD.sub.1, . . . , GISD.sub.n on these inputs are all high, and bit-line pull down signal BLPD on the output to NAND gate 87 is low. The circuit in FIG. 5 also has a second NAND gate 89 with an input connected to the output of NAND gate 87. A second input PROGRAMB is normally high, except when the AND array is being programmed. The output SALE from second NAND gate 89 is normally high.

When an input switch is detected, at least one of the input switch detection signals GISD.sub.1, . . . , GISD.sub.n pulses low. As a result, signal BLPD pulses high and signal SALE pulses low. As will be discussed later, this enables the AND gates in the array.

A signal ISDEB may be used to enable or disable the power reduction circuitry 86. Normally, signal ISDEB is low, and the inverted signal generated by inverters 91 and input into NAND gate 87 is high. In this case, the pull up elements in inverters 91 serve to hold signals GISD.sub.1 . . . GISD.sub.n normally high. When ISDEB is high, signal BLPD is always high, SALE is always low and the AND array is always enabled, whether or not an input switch detection is indicated. In other words, the switch detection circuits 85 are disabled when signal ISDEB is high. FIG. 6, shows one possible arrangement of two groups of input switch detection circuits 85 input into an array enable signal generating circuit 86 in a logic array device of the present invention.

With reference to FIG. 7A, an AND/OR logic array configuration includes an array 31 of AND gates 93, a group of which is shown, and one or more OR gates 45. Input signals IT.sub.0, . . . , IT.sub.n on input lines 37 from input buffers 35 in FIG. 2 drive AND gates 93. Signal BLPD produced by the AND array enable signal generator circuit 86 in FIG. 5 is also used as an input to AND gates 93. Signal SALE, when high, turns off AND gate 93's pull-up element, thereby eliminating any power consumption by the AND gate. The product term lines 47 output by AND gates 93 drive one or more OR gates 45. A signal TPD, produced by a circuit in FIG. 10 described below, enables the OR gates 45. Like the signal SALE, signal TPD pulses low when an input transition occurs. When TPD is high, the pull-down elements in the logic array's OR gates 45 are turn off, eliminating any power consumption by the OR gates. In FIG. 7B, an equivalent AND/NOR/INVERT logic array configuration has identical elements except that a NOR gate 95 and an inverter 97 replaces OR gate 45 in FIG. 7A. In this case, signal TPD enables a pull-up element in NOR gate 95.

In either case, when an input switch is detected, signal BLPD goes high and signal SALE goes low enabling AND gates 93. Signal TPD goes low, enabling the OR gates 45 or alternatively NOR gates 95. After some propagation delay time, the new logic array data will appear on the output 57 of OR gates 45 or of the NOR gate 95 combined with inverter 97. After an internally timed delay period, signal SALE will go high and signal BLPD will go low. This will shut off any AND gate power consumption and drive the AND gates' output 47 low. Likewise, signal TPD will go high, disabling OR gate 45's output pull-down element (or NOR gate 95's pull-up element). In this manner, OR gate 45 or NOR gate 95 is effectively turned off and the output signal 57 is held by a latch 99.

With reference to FIG. 8, an AND gate 93 is driven by outputs 37 from input buffers 35. The outputs 37 and their complements 101 control the gates of EPROM transistors 103. The drains of the transistors 103 are connected together and to a source of a sensing transistor 105 at a node 107. When the AND logic array including gate 93 is enabled, by signal SALE being low, the sensing node 107 will be high only if the gate electrodes 37 and 101 of all the unprogrammed EPROM transistors are low, and if signal BPLD is high. If the gate electrode of any unprogrammed EPROM transistor is high, that EPROM will conduct, and nodes 107 and output 47 will be low. Programmed EPROM transistors have a higher threshold, and therefore do not conduct.

With reference to FIG. 9, a NOR gate 95 is driven by the outputs 47 of gates 93 in AND array 31. A high output 47 on any AND gate 93 will turn on one of the transistors 108 and will pull node 109 low and make inverter 97's output 57 high. Conversely, if no AND outputs 47 are high while the NOR gate 95 is enabled, i.e. while signal TPD is low, transistor 111 will pull node 109 high and the inverter output 57 will go low.

One factor crucial to the correct operation of the gate power-up and power-down cycles is the relative timing of the OR enable signal TPD to that of the AND enable signals SALE and BLPD. When an input switch has occurred, the OR gates must not be enabled until the AND array gates have been enabled long enough to produce any high outputs. Otherwise, the OR gate enable element will cause an undesired low "glitch" on the output. For example, in FIG. 9, if any of the AND gate outputs 47 have a high output the signal at node 109 will be pulled-down and the output 57 should be high. However, if signal TPD enables the NOR gate 95 too soon, the pull-up transistor 111 will pull-up the signal node 109 until one of the AND gate outputs 47 turns high. Thus, an undesired low pulse will appear at the output 57. Conversely, the OR gates must disable before the AND gates disable to prevent the OR gates' enabling elements 111 from switching their outputs 57 low after the AND gates' outputs 47 have gone low.

With reference to FIG. 10, an OR gate enable signal generator circuit 113 meets these timing requirements. OR gate enable signal generator 113 is driven by input switch detection signals GISD.sub.1 and GISD.sub.2 from detection circuits 85 in FIG. 4 just like the AND array enable signal generator 86 in FIGS. 5 and 6. However, EPROM transistors 115-118 together with elements 121-128 guarantee that signal TPD will not go low until a delay period has occurred to first allow the AND array to be enabled. The gates of EPROM transistors 115 and 116 are controlled by respective signals GISD.sub.1 and GISD.sub.2. The drains of transistors 115 and 117 are connected together and to a source of a sensing transistor 123 at a node 135. Likewise, the drains of transistors 116 and 118 are connected together and to source of sensing transistor 124 at a node 136. Signals GISD.sub.1 and GISD.sub.2 are normally high, EPROMs 115 and 116 normally conduct and nodes 135 and 136 are normally low. Transistors 121 and 122 are off, transistors 137 and 138 are off, node 139 is high and signal TPD is high. When either or both of signals GISD.sub.1 or GISD.sub.2 pulse low, the corresponding EPROM transistor 115 or 116 or both turns off. Transistor 121 or 122 or both turns on, but the gate capacitance of additional EPROM transistors 117 and 118 causes a delay in node 135 or 136 or both going high so that the output of inverter 125 or 126 does not immediately switch from high to low. Eventually, the output of inverter 125 or 126 or both does switch from high to low. Either or both of transistors 137 and 138 turns on and node 139 goes low. Accordingly, signal TPD also goes low.

The presence of bypass transistors 131 and 132 insures that signal TPD returns high and disables the OR gate before the AND array disables. As soon as signals GISD.sub.1 and GISD.sub.2 return high, bypass transistors 131 and 132 turn on, immediately shutting off transistors 137 and 138. Nodes 139 and signal TPD return high. Though it takes time for the nodes 135 and 136 to fully return to low due to the capacitance of EPROMs 117 and 118, this delay has no effect on signal TPD, due to bypass transistors 131 and 132.

Returning to FIG. 9, note that as the logic arrays enable, a high output 47 from AND gates 93 is able to switch the OR gate 57 high without having to wait for the NOR gate 95 to be enabled by signal TPD. This minimizes signal propagation time through the logic arrays when an AND output is going high. Therefore, the propagation delay time required for producing a high OR gate output is limited only by the enabling of the AND gate and is not restricted by the AND/OR array enable timing sequence constraints described above.

In FIG. 11, a waveform timing diagram shows input signals switching at a time 141, either from high to low or low to high. This switch is detected by detectors 39 in FIGS. 3 and 4 and enable signals SALE, BLPD and TPD are generated by circuits 86 and 113 in FIGS. 5 and 10. AND array enable signals SALE and BLPD pulse low at a time 143 shortly after the switch. At some time 145 later, after the AND arrays have enabled, signal TPD goes low, enabling the OR gates. As soon as the OR gate output 57 in either FIG. 7A or 7B is stored in latch 99, the signal TPD goes high at time 147, disabling the OR gates. Finally, AND array enable signals SALE and BLPD go high at time 149, disabling the AND arrays. At another time 151, another input signal switch occurs and the power up/power down cycle repeats.

With reference to FIG. 12, a circuit 153 for generating and holding a logic array output signal 155 from a single AND gate output 157 is shown. If the output 157 of AND gate 93 goes high when the gate is enabled, node 158 on the ouptut of inverter 160 will go low, and thus, pull-up transistor 159 will pull node 161 high and the logic output node 155 will go high. If the AND gate's output 157 remains low when the gate is enabled, then when signal TPD also goes low, transistors 163 and 165 will pull node 161 low and the output signal 155 will go low. When the enable cycle is complete, signal TPD will go high, turning off transistor 165. Then when the AND gate 93 is disabled, AND output 157 will go (or remain) low and pull-up transistor 159 will turn (or stay) off. The logic output 155 remains in its high or low state, held valid by latch inverter 167. Inverters 160, 169 and 171 provide any propagation delay that is need. One feature of circuit 153 is that the only power consumed by the circuit is in the AND gate 93. No other portion of the circuit ever creates an active current path from the power supply to ground. Another feature of the circuit is that like the OR gate output 57 in FIG. 9, the output 155 can switch high without having to wait for signal TPD to go low.

In FIGS. 13 and 14, other embodiments of the circuit 153 in FIG. 12 perform the same function but have faster signal switching. However, the circuit 172 in FIG. 13 has the disadvantage of consuming power in transistors 173 and 175 when the AND gate output 177 is high. The circuit 179 in FIG. 14 has both fast signal switching and low power consumption but its output 181 cannot switch in either direction until signal TPD goes low. All provide latching of output 155 with inverters 167 and 169.

With reference to FIG. 15, a flip-flop register 49 is typically a D-type flip-flop of the prior art. The output 57 from an OR gate 183 or equivalent NOR gate-inverter combination in accord with FIG. 9 drives a data input D and outputs 59 are produced. Control signal inputs AR, SP and CLK are driven by AND array programmable signals. The logic circuit 153 of FIG. 12, for example, may be used to generate a clock signal 155. Because of the structure of this circuit, the rising edge of the clock signal is not dependent upon the timing of the TPD signal, as explained above. This feature maximizes the operating speed of the logic array device when registered logic is used. The circuit 172 of FIG. 13 may be used to generate the flip-flop's synchronous preset control signal 178. Because there need be only a few preset signals in a logic array device, the additional power consumed by the circuit 172 of FIG. 13 is less of a disadvantage than the benefit gained from the circuit's increased operating speed. The flip-flop's asynchronous reset signal 47 is driven directly by the output of an AND gate 93. No latching element is needed because of the asynchronous nature of the reset signal.

With reference to FIG. 16, an N:1 multiplexer 55 is provided for the logic array device which has maximized switching speed and low power operation. The multiplexer inputs D.sub.1, D.sub.2, . . . , D.sub.n are driven by outputs 57 from one or more OR gates, outputs 59 from one or more flip-flops 49, and one or more other signals 185. The multiplexer outputs an output signal 187 selected from one of the inputs D.sub.1, D.sub.2, . . . , D.sub.n depending on which of the control signals C.sub.1, C.sub.2, . . . , C.sub.n is high. Each input signal drives the source of two parallel N-channel transistors 189 and 191. The larger transistor 191 is an enhancement type device with a threshold of 0.6 to 0.8 volts. This transistor provides the high speed signal transmission. The other transistor 189 has a threshold of 0.1 to 0.2 volts, and serves to pull the output node 187 up to a voltage level near that of the selected input signal.

Two methods are available for adjusting the threshold of transistors 189. One, a threshold-shifting implant may be formed during the device's manufacturing process. Alternatively, the threshold may be varied by choosing its channel length and width appropriately according to well-known techniques.

With reference to FIG. 17, an output driver 61 receives an enable signal OE generated by any of the logic generation circuits described above with reference to FIGS. 12-14 or form an input buffer, such as the one in FIG. 2, and also receives a data signal D to be output from either a logic OR gate, such as the one in FIG. 9, a flip-flop, such as the one in FIG. 15, or a multiplexer, such as the one in FIG. 16. Enable signal OE and data signal D drive inputs to NAND gates 193 and 195, where the data signal into NAND gate 195 is first inverted. The inverted outputs of NAND gates 193 and 195 drive gates of respective N-channel transistors 197 and 199. The output of NAND gate 193 also drives the gate of a P-channel transistor 201. The source of transistor 197, and the drains of transistors 199 and 201 are connected to each other and to an I/O pad at a node 203. The drain 207 of transistor 197 is connected via a resistor 205 to the device power supply. The source of transistor 199 is grounded, and the source of transistor 201 is connected to the device power supply.

When enable signal OE is low, then the NAND gates 193 and 195 output high regardless of the value of data signal D, and all transistors 197, 199 and 201 are off, and the driver is disabled. The output driver 61 is enabled when signal OE is high. When data signal D is low, transistors 197 and 201 are off and transistor 199 is on, pulling the output node 203 low. When data signal D is high, transistor 199 is turned off and transistors 199 and 201 are on. P-channel transistor 201 pulls the output node 203 up to the full voltage level of the device power supply. This reduces the power consumption in a system where the I/O pad is connected to inputs of other CMOS devices. N-channel transistor 199 pulls up the output node 203 more quickly than would be possible with P-channel transistor 201 alone, thereby maximizing the operating speed of the device.

Resistor 205 limits the current through transistor 197 if output node 203 is at ground potential while node 209 is high. By putting the resistor 205 on drain 207 of transistor 197, the resistor has minimal effect on the operating speed of the device.

With reference to FIG. 18, the input switch detector 39 of FIG. 3 may be enhanced to further reduce the power consumption of a logic array device. The portion seen in FIG. 18 which is enclosed in dashed lines 39 is the same as the circuit in FIG. 3 except that delay elements 83 are replaced, outside the dashed line 39, by NOR gates 211 and 213 and inverters 215 and 217. When disable signal DISD is low, NOR gates 211 and 213 act as inverters. When disable signal DISD is high, transistors 65 and 67 will always be off and so detection signal LISD will not be pulled low even if the associated input signal IT switches. This is useful in situations where a particular device input or other logic array is not being used by the particular logic function programmed into the device.

With reference to FIG. 19, a one-time programmable MOS element 215 may be used to produce control signals, such as the input switch detector enable signal ISDEB in FIG. 5. One such configuration is disclosed in U.S. Pat. No. 4,713,792 to Hartmann et al. Typically, the programmable element 215 is a non-erasable floating gate MOS structure. By selectively programming these elements in a logic array device, a manufacturer can irreversibly alter the device to generate certain enabling control signals, so as to divide the device circuit components, emulate several different types of devices in a single device or replace circuitry within the device with a backup.

With reference to FIGS. 20 and 21, NOR logic gates 217 may be selectively grouped together into a single NOR gate. Selection control signals C.sub.1, C.sub.2, . . . , C.sub.n may be generated from either non-volatile programmable elements or from volatile storage elements. These control signals C.sub.1, . . . , C.sub.n, together with gate outputs 219 are input into NOR gates 221, the outputs of which connect to the gates of pass transistors 223. When a control signal C.sub.1, . . . , C.sub.n is low, the individual NOR output 219 corresponding to that control signal is selectively passed through the corresponding pass transistor 223 to node 225. When that control signal is low, the corresponding NOR output 219 is disconnected from node 225.

When a device's logic arrays 31 are in a disabled state, outputs 47 of the AND gates 93 are low. The gate of NOR gate 217's pull-up transistor 227 is at ground potential, so during this disablement period, the NOR gate output 219 will be held high by transistor 227. When an input level switch is detected, the AND array 31 is enabled. Any AND gate 93 with a high output 47 will pull down the NOR output 219. This, in turn, will turn on that NOR gate's transmission transistor 223 in FIG. 21 if the corresponding control signal C.sub.1, C.sub.2, . . . , C.sub.n is low, and node 225 will also be pulled low. If no AND gate output 47 in any of the connected NOR gates 217 is high, all transmission transistors 223 will be off and as the enable signal TPD goes low, node 225 will be pulled high by transistors 229. When the logic arrays subsequently disable, signal TPD goes high, turning off transistor 229. Transistors 231 and 241 hold node 225 in its latest high or low state. If any AND gate outputs 47 in FIG. 20 were high, they will go low, turning off all pass transistors 223 and eliminating any active power consumption in this AND/NOR configuration. Output signal nodes 233, 235 and 237 are available at all times to drive inputs of flip-flops 49 or multiplexers 55 shown in FIGS. 15 and 16.

With reference to FIG. 22, the AND/NOR circuit 217 of FIG. 20 may also be used to provide a data input signal D to a flip-flop storage element 239. When the logic arrays in FIG. 20 are in a disabled state, the output 219 is high, as explained above. Since the clock input 155 to the flip-flop is coming from the output of a logic gate 155 as in FIG. 12, the logic arrays will necessarily be enabled at times when the clock signal is switching. At such times, the logic output 219 goes low if any output of an AND gate in that array goes high. After a time delay period determined by inverter elements 160, 169 and 171 in FIG. 12, the clock signal 155 will go high, storing the desired data in flip-flop 239.