Advertisement
SECTION
2
OPERATI
NG
LI
M ITS
ACCUR ACY
The accuracy of the HP-35 depends on the operatIon
being performed. Elementary operations such as add, sub·
tract, multiply, divide, reciprocal. and square root have a
ma~imum
error of
±
one count m the tenth (least signifi·
ca
ll)
d
igit. Errol'S in these operations are caused
by
rounding answel'S to the tenth (least sig"l 'ficant) digit. An
e
~
ample
of round off error is seen when evaluating
(Vs)2.
Rounding
V5
to
10
signi
f
icant digits gives
2.236067977.
Squaring this number gives the
19
digit
product
4.999999997764872529.
Round ing the square
t
o
10
dIgIts gives
4.999999998.
If
the
ne
~
t
la rger
app
r
o
~
i
m~tlon
(2.2360671378)
i
~
~quorod,
the
.cou
l
t '
o
5.0000000 0223700
8484. Rou nd ing this
number t
o
10
sign if
ican
t d igits gIves
5.00
000
00
02. There simply
is
no
te
n
dIgit number whose square is
5.00 000000
0.
The accuracy
01
the remaining operations (trigonome tric,
logarithmic and exponential) depends upon the argument.
The answe
r
that is displayed will be the correct value for
an input argument with
i
n
iN
counts in the tenth least
si!Jliticant digi
t
of the origmal input argument.
Values
lor
N are:
OPERA
TI
ON
VALUE OF N
10gX.lnX,ande X
2
trigonometric
3
x y
4
for
X.and 7 fory
21
Advertisement
