Loadable Ripple Counter

The present invention relates to loadable ripple counters.

Counters count sequential input pulses, providing a parallel output indicating the counted value. Ripple counters routinely employ flip-flops configured to alternate their outputs upon each clock pulse. An N-bit ripple counter employs N flip-flops connected in a chain so that each flip-flop output both clocks the next flip-flop and is used as a bit of the N-bit parallel output of the counter. A ripple counter having this configuration is described in Millman, Microelectronics, (McGraw Hill 1979), pp. 220-222.

Loadable N-bit ripple counters are configured to be loaded with an N-bit parallel input so that counting begins at the loaded number. An example of a common prior art loadable ripple counter is shown in FIG. 1 and described in detail below.

In general the invention features a loadable N-bit ripple counter having N bit subcircuits, each employing a flip-flop and a bit loading element. The flip-flop is configured to alternate its flip-flop output every time that it receives a clock pulse and has means to control its output to a known state when a flip-flop control signal is asserted. The bit loading element is inputted with the flip-flop output and a bit input of a multibit number being loaded into the counter and outputs a bit output of the counter. The bit output is controlled by the states of the flip-flop output and the bit input. Such a counter uses fewer gates than prior loadable ripple counters, saving space and expense.

In some preferred embodiments the bit loading element is a gate (most preferably an exclusive-NOR gate); and in some other preferred embodiments, the bit loading element is a multiplexer having two signal inputs connected to receive true and complement flip-flop outputs and a control input that determines whether the true output or the complement output is provided as the multiplexer output. In preferred embodiments the bit output, except for the most significant bit, serves as a clock for the next more significant bit subcircuit; the flip-flops are set to known states by clearing them; and the flip-flop output is connected to its D input to cause its output to alternate state. When implemented in gate arrays, fewer nets are needed than with prior ripple counters. In integrated circuits employing a plurality of ripple counters according to the invention, the savings in space may permit use of a smaller die size.

Other advantages and features of the invention will be apparent from the following description of preferred embodiments thereof and from the claims.

FIG. 1, labeled prior art, is a schematic of a known loadable three-bit ripple counter.

FIG. 2 is a schematic of a loadable three-bit ripple counter according to the invention.

FIG. 3 is a schematic of an alternative embodiment of a loadable three-bit ripple counter according to the invention.

In FIG. 1 is shown common prior art loadable ripple counter 20 that employs NAND gates 28-38 to control the set and clear functions of D-flip-flops 22, 24, 26 to load counter 20 with a three bit number indicated by the LD0H, LD1H, LD2H, LD0L, LD1L, and LD2L inputs. Counter 20 counts INC H pulses provided to flip-flop 22, the least significant bit flip-flop. The complement output of each flip-flop clocks the next flip-flop (except for flip-flop 26), is used as a bit of the three-bit counter output (CN0, CN1, CN2), and is provided to its D input (causing each flip flop 22, 24, 26 to alternate its output every clock period). A counter of this configuration has twelve gates and eight nets, i.e., connections between components, for each output bit when implemented in a gate array using a DEC15GA gate array library.

Referring to FIG. 2, ripple counter 40 utilizes clearable D-flip-flops 42, 44, and 46 and exclusive-NOR gates 48, 50, 52. Counter 40 counts pulses provided as the INC H input, provides a three bit parallel output (CNT0, CNT1, CNT2), and is loaded with a three-bit parallel input (LD0H, LD1H, LD2H). Flip-flops 42, 44, 46 are connected to be cleared by a flip-flop control signal, LOAD L. The complement output of each flip-flop 42, 44, 46 is connected to its corresponding D input, causing each flip-flop to alternate its output every clock pulse. The true output of flip-flop 42 is inputted to exclusive-NOR gate 48, which is also inputted with the least significant bit, LD0H, of the loaded number. The output of gate 48 is outputted as CNT0 and clocks flip-flop 44. The true output of flip-flop 44 is inputted to gate 50, which is also inputted with the second bit, LD1H, of the loaded number. The output of gate 50 is outputted as CNT1 and clocks flop-flop 46. The true output of flip-flop 46 is inputted to gate 52, which is also inputted with the third bit, LD2H, of the loaded number. The output of gate 52 is outputted as CNT2. A counter of this configuration utilizes eleven gates and five nets for each output bit when implemented in a gate array using a DEC15GA gate array library.

Referring to FIG. 3, ripple counter 60 utilizes clearable D-flip-flops 62, 64, 66 and two-to-one multiplexers 68, 70, 72. Like counter 40, counter 60 counts pulses provided as the INC H input, provides a three-bit parallel output (CNT0, CNT1, CNT2), and is loaded with a three bit parallel input (LD0H, LD1H, LD2H). Flip-flops 62, 64, 66 are connected to be cleared by a flip-flop control signal, LOAD L. The true outputs of flip-flops 62, 64, and 66 are inputted to the A inputs of multiplexers 68, 70, and 22. The complement outputs of flip-flops 62, 64, 66 are inputted to their own D inputs (to provide alternating flip-flop outputs) and are inputted to the B inputs of multiplexers 68, 70, 72, respectively. The three-bit parallel inputs, LD0H, LD1H, and LD2H, of the loaded number are inputted to the control inputs, S, of multiplexers 68, 70, and 72, respectively. The output of multiplexer 68 is outputted as CNT0 and inputted as the clock for flip-flop 64. The output of multiplexer 70 is outputted as CNT1 and inputted as the clock for flip-flop 66. The output of multiplexer 72 is outputted as CNT2. A counter of this configuration utilizes ten gates and six nets for each output bit when implemented as a gate array using a DEC15GA gate array library.

Referring to FIG. 2, counter 40 is loaded with the loaded number indicated by the states of LD0H, LD1H, LD2H by asserting the flip-flop control signal, LOAD L. This clears the outputs of flip-flops 42, 44, and 46, in turn causing the outputs of gates 48, 50, 52, and thus the three bits of the counter output, CNT0, CNT1, CNT2, to be determined by the states of LD0H, LD1H, and LD2H. When LOAD L is deasserted, counter 40 counts INC H pulses. As each INC H pulse is received, the CNT0 bit of counter 40 transitions to the logical NOT of its current state. If this transition is from one to zero, then the next bit in the counter chain is clocked. This process "ripples" through to the most significant bit of counter 40. LD0H, LD1H, and LD2H must remain stable during the counting in order to allow the outputs of gates 48, 50, 52 to be affected by LD0H, LD1H, LD2H. Gates 48, 50, 52 thus act as bit-loading elements that each permit counter 40 to be loaded with a respective bit.

Referring to FIG. 3, counter 60 is loaded by asserting the flip-flop control signal, LOAD L, which clears the true outputs and sets the complement outputs of flip-flops 62, 64, and 66. The counter output, which is the output of multiplexers 68, 70, and 72, is now determined by the states of LD0H, LD1H, and LD2H, which are fed to the S multiplexer inputs. If the S input is high, multiplexers 68, 70, and 72 output the true input; if the S input is low, multiplexers 68, 70, 72 output the complement input. When the flip-flop control signal, LOAD L, is deasserted, counter 60 functions substantially the same as counter 40, the S inputs determining whether the true or complement outputs are passed through multiplexers 68, 70, 72. Multiplexers 68, 70, 72 thus act as bit loading elements that each permit counter 60 to be loaded with a respective bit.

Loadable ripple counters 40, 60 are advantageous in that they use fewer gates and nets than the prior art, thereby saving space. In integrated circuits employing a plurality of ripple counters according to the invention, the savings in space may permit use of a smaller die size. Ripple counters according to the invention also do not require settable flip-flops but can employ simpler, less-expensive flip-flops.

Other embodiments of the invention are within the scope of the following claims. E.g., the invention is applicable to both up counters, as in FIGS. 2 and 3, and to down counters.