Binary multiplier using ternary code

Claim 2

A binary multiplier circuit for receiving in the form of logic electric signal a first binary number X and a second binary number Y, so as to supply in the form of logic electric signal the product of these two numbers, said multiplier circuit comprising a coder for converting n binary signals y j (j=0 to n-1) representing number Y into n pairs of binary signals (s j ,v j ), each pair representing one coefficient b j (j=0 to n=1), where b j can be 0, 1 or -1 and where one signal (s j ) of the pair represents the sign of b j and the other signal (v j ) of the pair represents the module of b j , where the product of two consecutive coefficients b j and b j-1 is zero, and where the coefficients b j are such that Y=b.sub.n-1 2.sup.n-1 +b.sub.n-2 2.sup.n-2 + . . . b.sub.1 2.sup.1 +b.sub.o, said coder comprising, for each integer value of indicia j between 1 and n, an EXCLUSIVE-OR gate receiving y i and y 1-1 , an AND gate receiving on a first input an output p j of the EXCLUSIVE-OR gate, and receiving on a second input the logical complement of v j-1 , said AND gate having an output supplying the binary signal v j , a further AND gate having a first input receiving signal v j and a second input receiving signal y j+1 , except that for j=n-1 the second input of said further AND gate receives y n-1 , said further AND gate having an output supplying signal s j , said coder having outputs for supplying binary signals s j and v j (j=1 to n-1) and further comprising two outputs for supplying respectively a signal v o which is equal to y o and a signal s o which is the logical AND of y o and y 1 .

Claim 3

A binary multiplier circuit according to claim 1 wherein said number X is the multiplicand and said number Y is the multiplier, the multiplier circuit comprising a routing circuit receiving binary signals representing number X and a binary adder stage, said routing circuit being controlled by the outputs s j , v j of the coder, and said adder receiving outputs from the routing circuit.

Claim 4

A binary multiplier circuit for receiving in the form of logic electric signal a first binary number X and a second binary number Y, so as to supply in the form of logic electric signal the product of these two numbers, said multiplier circuit comprising a coder for converting n binary signals y j (j=0 to n-1) representing number Y into n pairs of binary signals (s j ,v j ), each pair representing one coefficient b j (j=0 to n-1), where b j can be 0, 1 or -1 and where one signal (s j ) of the pair represents the sign of b j and the other signal (v j ) of the pair represents the module of b j , where the product of two consecutive coefficients b j and b j-1 is zero, and where the coefficients b j are such that Y=b.sub.n-1 2.sup.n-1 +b.sub.n-2 2.sup.n-2 + . . . b.sub.1 2.sup.1 +b.sub.o, said multiplier circuit further comprising a second coder for converting m binary signals x i representing number X into m pairs of binary sigals (r i ,u i ), each pair representing one coefficient a i (i=0 to m-1), where a i can be 0, 1 or -1 and where one signal r i of the pair represents the sign of a i and the other signal u i of the pair represents the module of a i , where the product of two consecutive coefficients a i and a i-1 is zero, and where the coefficient a i are such that X=a.sub.m-1 2.sup.m-1 +a.sub.m-2 2.sup.m-2 + . . . a.sub.1 2.sup.1 +a.sub.o said multiplier further comprising a routing circuit and a binary adder stage, said routing circuit receiving the outputs of said second coder and the routing circuit being controlled by the outputs of said first coder, and said adder stage receiving outputs from the routing circuit.

Claim 5

A binary multiplier circuit for receiving in the form of logic electric signal a first binary number X and a second binary number Y, so as to supply in the form of logic electric signal the product of these two numbers, said multiplier circuit comprising a coder for converting n binary signals y j (j=0 to n-1) representing number Y into n pairs of binary signals (s j ,v j ), each pair representing one coefficient b j (j=0 to n-1), where b j can be 0, 1 or -1 and where one signal (s j ) of the pair represents the sign of b j and the other signal (v j ) of the pair represents the module of b j , where the product of two consecutive coefficients b j and b j-1 is zero, and where the coefficients b j are such that Y=b.sub.n-1 2.sup.n-1 +b.sub.n-2 2.sup.n-2 + . . . b.sub.1 2.sup.1 +b.sub.o, said multiplier circuit further comprising a second coder for converting m binary signals x i representing number X into m pairs of binary sigals (r i ,u i ), each pair representing one coefficient a i (i=0 to m-1), where a i can be 0, 1 or -1 and where one signal r i of the pair represents the sign of a i and the other signal u i of the pair represents the module of a i , where the product of two consecutive coefficients a i and a i-1 is zero, and where the coefficient a i are such that X=a.sub.m-1 2.sup.m-1 +a.sub.m-2 2.sup.m-2 + . . . a.sub.1 2.sup.1 +a.sub.o said multiplier further comprising a routing circuit and a binary adder stage, said routing circuit receiving the outputs of said second coder and the routing circuit being controlled by the outputs of said first coder, and said adder stage receiving outputs from the routing circuit.

Claim 6

A binary multiplier circuit according to claim 3, wherein said second coder further supplies m binary logic signal r i * where r i * is the complement of r i if u i =1 and is zero if u i =0.

Claim 7

The multiplier as claimed in claim 4 wherein said routing circuit transmits to elementary adders of said adder stage having a first and a second pair of inputs the following signals depending on the four pairs of control signals: on the first pair of inputs (C i ,C' i ) ##EQU7## at the second pair of inputs (D i ,D' i ) ##EQU8## the signals of the type u i-k or r i-k being logic levels 0 if i-k is negative and r i * being the complement of r i if u i =1 or 0 if u i =0.

Claim 8

The multiplier as claimed in claim 5 wherein said elementary adders supply at their outputs output sum and carry-over pairs according to coding having the same significance as the coding of coefficient a i or b j into signals (r i ,u i ) or (s j ,v j ), and the "sum" outputs of an adder are each applied to an input of a respective OR gate, these gates further receiving the carry-over outputs from the elementary adder of immediately lower rank.

Claim 9

The multiplier as claimed in one of claims 1 to 6 further comprising a decoder receiving pairs of logic signals (t i w i ) each representing a coefficient c i with three values o,1 or -1, for converting the number Z represented by these pairs, Z=c m+n-1 2 m+n-2 + . . . c 1 2 1 +c o into a binary number represented by binary logic outputs z i with: Z=-z.sub.m+n-1 2.sup.m+n-1 +z.sub.m+n-2 2.sup.m+n-2 + . . . z.sub.1 2.sup.1 +z.sub.o

Claim 10

A multiplier as claimed in claim 7 wherein said decoder comprises an exclusive OR gate receiving the signals t i and w i and supplying a signal d i , an inverter receiving the signal t i and a binary adder having a carry-over input forced to the logic level 1 and pairs of inputs for signals to be added, each pair of inputs receiving the output of the exclusive OR gate and the output of the inverter.