Comparator design using a 4-bit adder

7 replies · 14K views · last reply
ichabodgrant
Messages
48
Reaction score
0

Homework Statement


Hi, it's me again.
Now I am going to design a 4-bit magnitude comparator using just ONE 4-bit adder and infinitely large number of gates (AND, OR, NOT, NAND, NOR, XOR, XNOR) for signed numbers (negative binary).

Block diagram of a 4-bit adder with inputs a3 to a0 and b3 to b0, carry-in cin and carry-out cout, outputs s3 to s0, and an output L on the right.


Homework Equations


A > B => A3barB3 + A2barB2x3 + A1barB1x3x2 + A0barB0x3x2x1

similar for A < B and A = B.

The Attempt at a Solution


Circuit schematic of a 4-bit adder with inputs a3 to a0 and b3 to b0, cin, and outputs s3 to s0 and cout; b inputs pass through XOR gates with cin, and s3 feeds an inverter driving output L.
[/B]
This is my attempt. Frankly, this is my first time encountering problems related to comparator design because this is never said to be in the syllabus of the course (introductory course). I find this in a past paper... seemingly indicating that the instructor wants us to divide and conquer it within the 3 hours of exam.

I do this all by resources on the internet. So there may be a lot of mistakes.
 
I think you are on the right track, but I have some comments:
  • Where is the Cin coming from? You are only comparing two 4-bit numbers!
  • Are the numbers supposed to be in two's complement?
And - there are only 16 combinations possible. I suggest making a truth table before continuing.
 
A truth table looks like this? because I have never drawn a truth table for such a design...so I searched on google...
 
Attachments
  • Truth table of the 7485 4-bit magnitude comparator, with inputs A3,B3 to A0,B0, cascading inputs A>B, A<B, A=B, and outputs A>B, A<B, A=B; X marks don't-care values.
    Table_de_verite_du_CI_7485.gif
    32.9 KB · Views: 13,642
I have a question about the XOR gates. Are those XOR gates for doing 2's complement (i am thinking in this way)?
But I know that 2's complement is not just inverting 1 and 0, but also needs to add 1 after that. It seems that the above design cannot do that?
 
And you mentioned that there should not be Cin? Does this mean there are no cascading inputs?
 
ichabodgrant said:
A truth table looks like this? because I have never drawn a truth table for such a design...so I searched on google...
No. I meant the truth table for what you are trying to design. Since you are trying to compare two 4-bit 2's complement numbers, create a table with the A-numbers horizontally (-8 to +7) and the B-numbers vertically (-8 to +7). Now fill in the interior of the table with the result of the comparison you want (e. g. A>B). After you have done that, start thinking about how to implement it.