Agilent Technologies 4395A - Page 421

Calibration
for
Network
Measurement
Modifying
Calibration
Kits
F
or
most
applications
,
use
the
default
cal
kit
models.
Modifying calibration
kits is
necessary
only
if
unusual
standards
are
used
or the
very highest
accuracy is
required. Unless
acal
kit
model
is
provided
with
the
calibration
devices used,
a solid
understanding of
error correction
and
the
system
error
model
are
essential to
making modications
. Read
all of
this section.
During
measurement
calibration,
the
analyzer
measures
actual,
well-dened
standards
and
mathematically
compares
the
results
with
ideal
\models" of
those standards
.The
dierences
are
separated
into
error
terms
which
are later
reduced during
error correction.
Most of
the
dierences
are
due
to
systematic
errors
- repeatable
errors introduced
by the
analyzer,
test
set,
and
cables
-
which
are
correctable.
However,
the dierence
between the
standard's
mathematical
model
and
its
actual
performance
has an
adverse aect;
it reduces
the system's
ability
to
remove
systematic
errors
,
and
thus
degrades error-corrected
accuracy.
Therefore,
in
addition
to
the
default cal
kit models
,a
\user
kit"
is
provided
that
can
be
modied
to
an
alternate
calibration
standards
model.
Several
situations
exist
that
may require
a user-dened
cal kit:
Y
ou
use
a
connector
interface
dierent
from
the
four
built-in
cal kits
.
(Examples:
SMA,
or
BNC.)
Y
ou
are
using
standards
(or
combinations
of
standards)
that
are
dierent
from
the
predened
cal
kits
.
(F
or
example
,
using
three
oset
SHORT
s
instead
of
an
OPEN,
SHORT
,
and LO
AD to
perform
a
1-port
calibration.)
Y
ou
want
to
improve
the
built-in
standard
models
for
predened
kits
. Remember
that the
more
closely
the
model
describes
the
actual
performance
of
the
standard,
the
better
the
calibration.
(Example:
The
7
mm
LO
AD
is
determined
to
be
50.4

instead
of 50.0
.)
Unused
standards
for
a
given
cal
type
can
be
eliminated
from
the
default
set,
to
eliminate
possible
confusion
during
calibration.
(Example:
A
certain
application
requires
calibrating
a
male
test
port.
The
standards
used
to
calibrate
a female
test
port
can
be
eliminated
from
the
set,
and
will
not
be
displayed
during
calibration.)
Glossary
This
section
provides
a
glossary
of
terms
related
to
calibration.
A
standard
is
a
specic
,
well-dened,
physical
device
used
to
determine
systematic
errors
.
A
standard
type
is
one
of
ve
basic
types
that
dene
the
form
or
structure
of the
model to
be
used
with
that
standard
(for
example
,
a
SHORT
or
a
LO
AD).
Standard
coecients
are
numerical
characteristics
of the
standards used
in the
model
selected.
A standard
class is a grouping of one or
more standards that determines which standards are
used in a particular calibration procedure
.
Dening the Standards
Standard denition is the process of mathematically modeling the electrical characteristics
(delay, attenuation, and impedance) of each calibration standard. These electrical
characteristics (coecients) can be mathematically derived from the physical dimensions and
material of each calibration standard, or from its actual measured response. The parameters of
the standards can be listed in Standards Denitions,Table A-2.
Basic Measurement Theory A-45