The Saw Function
Aleksandar Ciric
To get the Saw function, we first define the function
that is linear, periodical, and aslope on one side. The letter
a defines whether the function is aslope to the left or to the
right. a is a function that is positive on even integers of x
and negative on odd.
This is the system on which the a(x) works. For
negative x, function is -1, and for positive, function is 1.
Now we put two functions together, with a(x) function,
just with different signs in front. The graph seams to look
like the wanted Saw graph. Still, this graph is not a
function and it needs to be cut on different periods for
function one and two.
This is the definition of wanted function, which is due
the integer position.
f (x) x x 1
a 1
a (odd int) 1
a (even int) 1
f (x)
x
x
f1 ( x ) x x 1
f 2 (x) x x
x x 1_ if _ int(x) _ is _ odd
f (x)
x x _ if _ int(x) _ is _ even
sin( x )
sin( x )
f (x)
Now, we need some periodic function that works similar
to the a function. The sine function is the most convenient
in this case.
The sine function's period is not convenient for the
wanted Saw function, since we want that it changes
direction on every integer, not on every By the
definition, we find that the new period should be two and
constant
T1
2
b
T1 2
T2
2
2
b
b
f (x)
Now to use this function that we got. The aslope look
right, period looks right, just one part of function needs
vertical shift by 1.
f (x)
sin( x )
sin( x )
sin( x )
sin( x )
x
x
sin( x )
sin( x )
To bring this function up we need first to bring the whole
function by .5, and then we use the other, a(x) function,
multiplied by . For the part of function that was positive,
the 1/2a cancels the vertical shift, and that part stays the
same. The negative part is shifted by the positive , and
the another positive multiplied by the a(x) function. The
function now looks right.
This Saw function has only one problem. At the every
even value of x, the function is undefined, and it should be
equal to 1. To correct this problem, the more sophisticated,
but simpler method should be used. This is the method of
inverse functions. It is known that when the inverse is the
function of the original function, or opposite, they are
equal to f(x)=x. To demonstrate this we can use the simple
function and its graph.
We need some periodic function, which will change the
positive and negative values of previously used function
a(x). The best is to use the sine and inverse sine functions.
Note that in this case it is not possible to use the inverse
sine function as the g(x), since its domain is limited.
f (x)
sin( x )
sin( x )
x 1 sin(x ) 1
x
sin( x )
sin( x )
2 sin( x ) 2
f ( x ) .5 x
x .5g ( x )
g( x ) 2x
f (g ( x )) .5( 2 x ) x
f ( x ) sin 1 ( x )
g ( x ) sin( x )
f (g ( x )) sin 1 (sin( x ))
The function now exists for every value of x, just it needs
to be vertically shifted, vertically shrank, and horizontally
shrank. Fist we look the amplitude of the present function.
Its value is To change it to 1, we multiply the function
with 1/Now to bring it up, we shift it vertically by .
It seams that only the horizontal shrink is left for the
function to look correctly. To do this, we find that the
period of the function is 2, and it should be 2. For
correction we multiply the x with 2.
It still does not look perfect. The function needs
horizontal shift. The first negative zero of the function is
when x=.25. Thus, we subtract of x inside the function.
The function looks right and the work is done.
1
sin 1 (sin( x ))
1
f ( x ) sin 1 (sin( x )) .5
f (x)
f (x)
f (x)
1
sin 1 (sin(2x )) .5
1
1
sin 1 (sin( 2( x ))) .5
