Operation Manual

Reference Information B-5
8299APPB.DOC TI-82, Appendix B, English Bob Fedorisko Revised: 02/09/01 9:46 AM Printed:
02/09/01 12:44 PM Page 5 of 12
Accuracy Information
To maximize accuracy, the TI
.
82 carries more digits internally than it displays.
Computational Accuracy
Values in memory are stored using up to 14 digits with a 2-digit exponent.
¦
You can store a value in the
WINDOW
variables using up to 10 digits (12
digits for
Xscl
,
Yscl
,
Tstep
, and
q
step
).
¦
When a value is displayed, the displayed value is rounded as specified
by the
MODE
setting (Chapter 1), with a maximum of 10 digits and a 2-
digit exponent.
¦
RegEQ
displays up to 14 digits.
Graphing Accuracy
Xmin
is the center of the leftmost pixel,
Xmax
is the center of the next to
the right-most pixel. (The right-most pixel is reserved for the busy
indicator.)
@
X
is the distance between the centers of two adjacent pixels.
¦
@
X
is calculated as (
Xmax
N
Xmin
)
à
94
.
¦
If
@
X
is entered from the Home screen or a program, then
Xmax
is
calculated as
Xmin+
@
X
…
94
.
Ymin
is the center of the next to the bottom pixel,
Ymax
is the center of the
top pixel.
@
Y
is the distance between the centers of two adjacent pixels.
¦
@
Y
is calculated as (
Ymax
N
Ymin
)
à
62
.
¦
If
@
Y
is entered from the Home screen or a program, then
Ymax
is
calculated as
Ymin+
@
Y
…
62
.
Cursor coordinates are displayed as eight characters (which may include a
negative sign, decimal point, and exponent). The values of
X
and
Y
are
updated with a maximum of eight-digit accuracy.
minimum
,
maximum
, and
intersect
on the
CALCULATE
menu are
calculated with a tolerance of 1
E
L
5;
dy/dx
and
‰
f(x)dx
use a tolerance of
1
E
L
3. Therefore, the result displayed may not be accurate to all eight
displayed digits. (In general, for most functions, there are at least
5 accurate digits.) The tolerance can be specified for the command-line
functions
fMin(
,
fMax(
, and
fnInt(
on the
MATH MATH
menu.