EECS 270, Lecture notes 3: Minimization Page 1 of 11

Minimization of Combinational Circuits (6.2)

Notes:
• I will post a set of slides that does this very formally later. They are _really_ good. Look them over.
• The text does a pretty good job here (though not as good/formal as the slides).

One good question is “how do we minimize combinational circuits?” And the answer to that question is more complex
than you’d think. Part of the problem is it isn’t clear what it means to minimize. Do we mean to use the fewest number
of gates? If so, we could end up taking something like the canonical sum-of-products and making a really, really, long
path. That would greatly slow the clock.

We probably mean to reduce the number of gates as much as possible while keeping the delay bounded by some value.
In this class, we’ll just look at finding the minimal sum-of-products and product-of-sums forms for a given logic equation.
This is called two-level logic because the worst-case path through the circuit will consist of exactly two gate delays.

Motivational Example #1
Consider the logic statement F(a,b,c)=ab’c+abc+abc’. You’ll notice that we are using three 3-input AND gates and one
three input OR gate to implement this logic (let us assume that all literals are available, so a and a’ can both be used and
don’t require a NOT gate). What can you do to reduce the complexity of this circuit while keeping ourselves to two
levels of logic?

First notice that ab’c+abc can be combined (using the “combining rule” from the first lecture of the year). We can also

combine abc+abc’. So we get: __________________________ as a simplified version of this. As you’ve likely noticed, it
isn’t always easy to find pairs that can combine, and sometimes those pairs are well hidden inside of various terms.

Motivational Example #2
Consider the following: a’b’c+a’cd’+bcd. It turns out that equation can be reduced to a’c+bcd. How the heck can we
figure that out? One trick is to drop back to the canonical sum-of-products form. That gets us:

__________________________________________________________________________________

You might notice that all four the terms that have a’c in them are true. So we can combine those four into a’c. That
leaves abcd. It can be combined with a’bcd to make bcd. It turns out we can’t combine any more and this point and so
we’ve succeeded!

We got the right answer this way, but it wasn’t obvious. Heck, it wasn’t obvious that we are done. So the question is,
how do we make it obvious? How can we quickly see what terms combine and prove to ourselves that we are done?
What we’d like is an easy way to visualize exactly

What we are looking for

Say there are 3 variables: a, b and c. A given term (say abc) can be combined with three other terms (a’bc, ab’c and abc’
in this case). We want a way to see those connections quickly.
EECS 270, Lecture notes 3: Minimization Page 2 of 11

Karnaugh Maps
Consider ab’c+abc+abc’ from above. Let’s write the truth table in a rather odd way:

ab/c 00 01 11 10
0 0 0 1 0

1 0 0 1 1
Notice that box which represents abc is adjacent to the three different terms it can be combined with. In this case we

can quickly see how the combining rule can be applied. We end up with two product terms: _________ and _________

Now consider
ab/c 00 01 11 10
0 1 1 1 1
1 0 0 0 0
In this case, we can combine all four terms into a single term: ____________________

ab/c 00 01 11 10
0 1 0 0 1
1 0 0 0 0
And notice we “wrap around” the edges. So this combines to be: ____________________

Informal Algorithm
What we are going to do is circle groups of “1s” that are rectangles 1 where each side is a power of 2 in length. The first
K-map on this page had a 2x1 and a 1x2 rectangle, the second a 1x4 and the third a 1x2. We never circle a rectangle that
is part of a larger legal rectangle. We only circle enough rectangles so that every 1 is circled.

ab/c 00 01 11 10
0 0 0 1 1

1 0 1 1 0

Notice that in the K-map above there are three rectangles we could circle, but two of them cover all the “1s”.
Let’s practice a bit.

ab/c 00 01 11 10 ab/c 00 01 11 10 ab/c 00 01 11 10

0 1 1 1 1 0 1 0 0 1 0 1 1 1 1
1 0 0 1 0 1 0 1 1 1 1 1 1 0 1

__________________________ __________________________ __________________________

1
Recall a square is special case of a rectangle.
EECS 270, Lecture notes 3: Minimization Page 3 of 11
EECS 270, Lecture notes 3: Minimization Page 4 of 11

An interesting example
Circle all rectangles: Answer 1 Answer 2
ab/c 00 01 11 10 ab/c 00 01 11 10 ab/c 00 01 11 10
0 1 1 1 0 0 1 1 1 0 0 1 1 1 0

1 1 0 1 1 1 1 0 1 1 1 1 0 1 1

__________________________ __________________________ __________________________

Terminology (p308-310)
Notice that we are finding sum-of-products solutions.
• Recall that a minterm is a product term that includes all of the functions variables exactly once.
• The on-set of a function is the set of minterms that define when the function should evaluate to 1 (the minterms
that have a 1 in the truth table.)
o The off-set is the set of minterms that evaluate to zero.
• An implicant of a function is a product term that evaluates to 1 only in places that function evaluates to 1. (The
on-set of an implicant of a function is a subset of the on-set of the function.)

o Graphically, in a K-map an implicant is: _______________________________________

• An implicant covers those minterms that appears in its on-set.

o What is the on-set of the function F(a,b)=a? _______________________

o What minterms does that function cover? ________________________

• Removing a variable from a term is known as expanding the term. This is the same as expanding the size of a
circle on a K-map.

ab/c 00 01 11 10 ab/c 00 01 11 10
0 1 1 0 0 0 0 1 1 0
1 1 1 0 0 1 1 1 1 0

• Prime implicant: ____________________________________________________________

• Essential one 2: _____________________________________________________________

• Essential prime implicant: ____________________________________________________

ab/c 00 01 11 10
0

2
This term isn’t used by our text, they skip from prime implicant directly to essential prime implicant.
EECS 270, Lecture notes 3: Minimization Page 5 of 11
EECS 270, Lecture notes 3: Minimization Page 6 of 11

More Formal Algorithm

• Identify all prime implicants
• Identify all essential ones.
• Circle all essential prime implicants
• Cover the remaining minterms using a minimal number of remaining prime implicants.

Notice that there may be more than one

solution. Also notice that the last step is a bit
vague 

ab/c 00 01 11 10
0 1 1 1 0
1 1 0 0 1

4-variable
ab/cd 00 01 11 10
00 0 1 1 0
01 0 1 1 0
11 0 0 1 1
10 0 0 0 0

And some practice with these:

ab/cd 00 01 11 10 ab/cd 00 01 11 10 ab/cd 00 01 11 10

00 1 1 1 1 00 1 1 0 1 00 0 1 0 0
01 0 1 1 0 01 0 0 0 0 01 1 1 1 0
11 0 0 1 0 11 0 0 0 1 11 0 1 1 1
10 1 1 0 0 10 1 0 0 1 10 0 1 0 0

________________________ ________________________ ________________________

What’s left?
• Don’t cares
• 5+ variable
• Product-of-sums
• Programmable techniques
• Lots of practice.
• More context. Remember this is only for 2-level logic…
EECS 270, Lecture notes 3: Minimization Page 7 of 11

Don’t cares:
Note: I’m leaving the zeros blank to make things more readable!

F = ∑W , X ,Y , Z (1,5,7,11,15) + d (8,12,13,14)
F
WX
YZ 00 01 11 10
00 d d
01 1 1 d
11 1 1 1
10 d

• Use d cells to make prime implicants as large as possible.

• No PI should include only d’s
• Only 1-cells should be considered when finding the minimal cover set.

5-variable
EECS 270, Lecture notes 3: Minimization Page 8 of 11

Product of Sums:
Let F be: Then F’ is:

ab/cd 00 01 11 10 ab/cd 00 01 11 10
00 1 1 1 00 1
01 1

11 1
1
1
1 1 01
FF’
11 1 1
10 1 1 10 1 1

Find minimal SoP for F’.

Use deMorgans.

Programmable techniques
• Later in the semester, time allowing.

More context
Basically, just remember that this doesn’t find the “minimal” solution. If finds the minimal sum-of-products. It’s not
even clear how we would measure “minimal” over all. Least delay? Least number of gates? Least “gate inputs”? Recall
we did a “least delay” solution for GA1. This would have helped find a nice starting point, but wouldn’t have solved the
problem.

But this technique does let us find the minimal two-level solution (SoP or PoS). Which is pretty cool.

Questions:
1. Why isn’t this a great technique for a computer? Why is it good for people?

2. What might be easier for a computer?

ab/cd 00 01 11 10
00 1
3. Can you define all the terms we’ve seen?
• On-set, Off-set? 01 1 1 1
• Implicant, prime implicant? 11 1 1 1
• Essential one, essential prime implicant?
10 1
• Cover?

4. Why is the instance to the right tricky?

EECS 270, Lecture notes 3: Minimization Page 9 of 11

State minimization
One thing to be aware of is that the 2nd edition of our text somehow got worse at covering this material. I’ve posted the
relevant 6 pages from the 1st edition on the website, you should take a look at it.

Motivational Example
Consider the following two machines. Are they equivalent?
Inputs: x; Outputs: y
x’ x x’ x x’ x’ x x’

x x’ x x
S0, S1,
S0 S1 S2 S3
S2 S3
y=0 y=1 y=0 y=1 y=0 y=1

Say x were 0, 1, 1, 0, 0. What would be the output of the larger one? The smaller one? Are they always the same? How
can you be sure?

What makes states equivalent?

Two states are equivalent if
1. They assign the same values to outputs,
e.g. S0 and S2 both assign y to 0,S1 and S3 both assign y to 1
2. AND, for all possible sequences of inputs, the FSM outputs will be the same starting from either state
Let’s look at this example. Are any states equivalent?

Inputs: x; Outputs: y
x’
x’ x x’ x

x x’ x
S0 S1 S2 S3
y=0 y=1 y=1 y=1
Let’s use an implication table.
EECS 270, Lecture notes 3: Minimization Page 10 of 11

Inputs: x; Outputs: y
S1
x’ x x’ x x’

x x’ x
S2 S0 S1 S2 S3
y=0 y=1 y=0 y=1

S3

S0 S1 S2

Which states can’t be equivalent? Any with differing outputs! Cross those out.

Now we’ll follow the following algorithm. Do steps 2&3.

What does this look like after step 4?

EECS 270, Lecture notes 3: Minimization Page 11 of 11

Another Example Inputs: x; Outputs: y

x’

S1 x’ x x’ x

x x’ x
S0 S1 S2 S3
S2 y=0 y=1 y=1 y=1

S3 Now let’s do a final, less trivial, example:

S0 S1 S2

S Inputs: x; Outputs: y
x’ x’
x
S3 S0 S4
S
y=0 x’ y=0 x y=0
x’ x’
x x
S S2 S1
y=1 y=1

S S S S