IPC-TM-650 EN 2022 试验方法-- - 第467页

IPC-TM-650 Page 10 of 1 1 Number 2.5.5.5.1 Revision Subject Stripline Test for Complex Relative Permittivity of Circuit Board Materials to 14 GHz Date 3/98 7-2.1 .1 For the three point measurement described in 6.5.1 , th…

100%1 / 824
z
IPC-TM-650
Page 9 of 11
Number
2.5.5.5.1
Revision
Subject
Stripline
Test
for
Complex
Relative
Permittivity
of
Circuit
Board
Materials
to
14
GHz
Date
3/98
6.4.1
The
computer
sets
the
sweeper
to
a
selected
carrier
wave
frequency
without
an
AM
or
FM
audio
signal
and
to
a
desired
output
power
level,
such
as
10
dBm.
6.4.2
The
same
frequency
is
given
to
the
synchronizer
with
instructions
to
lock the
frequency
of
the
sweeper
to
the
speci¬
fied
value.
6.4.3
The
computer
checks
the
synchronizer
for
status
until
the
status
value
indicates
the
frequency
is
locked.
6.4.4
The
power
meter
reading
is
obtained
by
the
computer.
Since
it
takes
a
finite
amount
of
time
for
the
power
sensor
to
stabilize,
either
a
delay
is
used
or
the
reading
may
be
taken
repeatedly
until
consecutive
readings
meet
a
given
require¬
ment
for
stability.
6.5
Use
of
the
Network
Analyzer
for
Measurement
of
the
Specimen
An
automated
network
analyzer
may
be
used
either
by
operating
the
front
panel
controls
manually
or
under
computer
control
with
suitable
specialized
software.
The
fixture
with
the
specimen
is
connected
by
test
cables
and
adapters
as
a
device
under
test.
Set
up
the
instrument
so
the
Cartesian
screen
display
shows
the
S21
parameter,
the
transmission/incident
power
ratio,
in
negative
dB
vertical
scale
units
versus
frequency
on
the
horizontal
scale.
Select
the
start
and
stop
frequency
range
to
sweep
across
the
resonance
peak
and
at
least
3
dB
below
the
peak.
Adjust
the
start
and
stop
frequency
values
as
narrowly
as
possible,
but
still
include
the
resonant
peak
and
the
portions
of
the
response
curve
on
both
sides
of
it
that
extend
3
dB
downward.
6.5.1
The
first
option
is
to
get
the
three
points
(fr,
f
〕
and
f2)
as
described
in
6.3
or
6.4.
Determine
the
resonant
dBr
and
frequency
fr
values
for
the
highest
point
(maximum)
on
the
response
curve.
With
manual
operation,
instrument
program
features
may
be
available
to
do
this
very
quickly.
On
the
response
curve
to
the
left
and
right
of
fr,
locate
the
%
,
dB〕
and
f2,
dB2
points
as
near
as
possible
to
3
dB
below
dBr.
These
may
then
be
used
in
the
calculations
shown
in
7.2.
6.5.2
A
second
option
requires
a
computer
external
to
the
instrument.
Collect
from
the
network
analyzer
all
of
the
f,
dB
data
points
represented
by
the
response
curve
between
dB〕
and
f2,
dB2
and
apply
non-linear
regression
analysis
tech¬
niques
to
determine
statistically
values
for
Q,oaded)
fr
and
dBr
that
best
fit
the
f”
dB,
paired
data
points
to
the
formula.
dBj
=
dB「
-
10
loge(10)
loge
(1+4
Qloaded2
(f"
fr
-
1)2)
[1]
where
10
loge(1
0)
is
the
constant
for
converting
from
loge
to
dB.
This
formula
may
be
derived
from
formula
5
with
the
rea¬
sonable
assumption
that
fr
-
j
equals
f2
-
fr.
The
statistically
derived
values
for
fr
and
Q
would
then
be
used
in
formulas
2
of
section
7.1,
formula
3
of
section
7.2,
and
formula
6
of
sec¬
tion
7.3
respectively.
This
has
been
found
to
fit
the
collected
data
points
very
well
at
all
regions
across
the
entire
&
to
f2
range.
It
is
a
simplified
version
of
the
non-linear
regression
method
for
complex
S21
parameters
described
by
Vanzura4.
7.0
Calculations
7.1
Stripline
Permittivity
Use
special
care
to
assign
the
correct
n
value
for
each
resonance
measured.
At
resonance,
the
electrical
length
of
the
resonator
circuit
is
an
integral
number
of
half
wavelengths.
The
effective
stripline
permittivity,
耳，
can
be
calculated
from
the
frequency
of
maxi¬
mum
transmission
as
follows:
加二
[n
C
/
(2
fr(L
+
AL))]2
[2]
where
n
is
the
number
of
half
wavelengths
along
the
resonant
strip
of
length
L
in
mm,
AL
is
the
total
effective
increase
in
length
of
the
resonant
strip
due
to
the
fringing
field
at
the
ends
of
the
resonant
strip,
C
(the
speed
of
light)
is
2.9978
1011
mm/s,
and
fr
in
Hz
(or
cycles/s)
is
the
measured
resonant
(maximum
transmission)
frequency.
The
resonator
ends
coincide
with
the
end
edges
of
both
the
dielectric
and
the
ground
planes.
The
relative
fringing
field
at
the
ends
becomes
extremely
small.
It
has
been
the
practice
with
this
method
to
ignore
this
fringing
field
and
consider
the
AL
value
to
be
zero
in
the
calculation
of
stripline
permittivity.
7.2
Calculation
of
Effective
Dielectric
Loss
Tangent
tan
6
=
1/Qunloaded
-
1/Qc
where:
1/QC
is
the
loss
factor
of
the
conductor
1/Qunioaded
is
the
total
loss
factor
of
the
unloaded
resonator
due
only
to
the
dielectric,
copper,
and
copper-dielectric
inter¬
face,
and
does
not
include
loss
due
to
coupling
of
the
probes.
7.2.1
The
resonator
loss
factor
The
measurement
of
the
resonance
gives
a
value
for
the
loss
factor
of
the
resonator
with
loading
due
to
probe
coupling
(1/Q,oaded).
IPC-TM-650
Page 10 of 11
Number
2.5.5.5.1
Revision
Subject
Stripline
Test
for
Complex
Relative
Permittivity
of
Circuit
Board
Materials
to
14
GHz
Date
3/98
7-2.1
.1
For
the
three
point
measurement
described
in
6.5.1
,
the
calculation
is
1/Q|oaded
=
%
一
i)
/
*]
[4]
A
more
exact
calculation
can
be
used
that
does
not
require
that
the
values
of
and
f2
be
at
exactly
half
the
power
level
of
the
maximum
at
resonance.
This
is
especially
suited
for
auto¬
mated
testing.
The
formula
is
1/Qloaded
=
(
1
-(fl/fr))
(10A1//1°
-
1)
-°-5
.
((f2/fr)-1)(1OA2/1°-1)-°-5
where:
A1
is
the
positive
dB
difference
in
power
level
from
fr
to
and
A2
is
the
positive
dB
difference
in
power
level
from
fr
to
f2.
7.2
.
1.2
For
the
many
point
measurements
of
the
resonance
described
in
6.5.2,
the
non-linear
regression
to
fit
the
formula
1
derives
the
Q|Oaded
value.
7.2
.2
Correcting
the
Resonator
Loss
Factor
for
Load¬
ing
The
probe
gap
set
for
about
50
dB
insertion
loss
at
resonance
is
intended
to
make
Q|Oaded
approximately
equiva¬
lent
to
QUnioaded-
Nevertheless,
corrections
in
the
measured
total
loss
value,
1
/Q|Oaded
are
desireable.
With
the
assumption
that
the
S21
parameter
with
straight
through
connection
with¬
out
the
test
fixture
is
at
0
dB,
dBr,
the
insertion
loss
or
S21
parameter
in
dB
units
at
the
resonant
peak,
is
related
to
the
power
ratio
by
P2/P1
=
10(-dBr/1O)
where
the
dBr
value
at
resonance
is
taken
as
positive.
Then
the
correction
is
Qunloaded
-
^loaded
'
^P2/P?0.5]
or
Qunloaded
=
Qloaded
/
H
-
1
0
(由
/2。)]
[6]
As
can
be
seen
from
the
following
tabulation
at
high
degrees
of
insertion
loss
such
as
50
dB
errors
in
the
straight
through
connection
assumption
above
are
not
as
important
as
they
would
be
at
lower
values
such
as
20
or
15.
dB
60
50
40
30
20
15
10
5
QuQ
1.00
1.00
1.01
1.03
1.11
1.22
1.46
2.28
7.3
Calculation
of
1/QC
The
following
calculation
scheme
is
used
to
estimate
the
conductor
loss(5,6)
needed
for
formula
3:
1/QC=
%C/
(叫
同。5)
[7]
where:
ac
=
4
Rs
£r
Zo
Y
/
(3772
B)
=
attenuation
constant,
nepers/mm
Rs
=
0.00825
fr0-5
=
surface
resistivity
of
copper,
Ohms
Zo
二
377/(4
耳。
$
(Cf
+
(W/(B
-
T))))
二
characteristic
impedance
of
resonator,
Ohm
377
二
120
=
free
space
impedance,
Ohm
Cf
=
(2
X
loge(X+1)-(X-1)loge(X2-1))/
兀
Y
=
X+2
WX2/B
+
X2(1
+T/B)
loge
[(X
+
1)/(X-
1)]/k
X
=
1
/
(1
-
T
/
B)
£r
=
relative
permittivity
B
二
ground
plane
spacing,
mm
W
二
resonator
width,
mm
T
=
resonator
conductor
thickness,
mm
Proven
data
is
not
currently
available
for
correcting
this
calcu¬
lated
value
to
account
for
increased
conductor
loss
associ¬
ated
with
roughness
of
the
copper
foil
or
surface
treatments
for
adhesion.
When
smooth
rolled
copper
foil
is
used
in
Type
A
specimens
the
estimate
seems
quite
reliable
in
the
0.4
to
1
5
GHz
range
based
on
work
done
with
neat
(PTFE)
polytet¬
rafluoroethylene)
sheet
specimens(3).
8.0
Report
The
report
shall
contain
the
following:
8.1
The
type
of
specimen:
A,
B,
C,
or
D.
8.2
For
specimen
type
A,
if
not
copper
foil
type
W
(wrought),
grade
5
(as
rolled-wrought),
bond
enhancement
N
(none,
no
stain
proof),
or
for
specimen
types
B,
C,
or
D,
state
at
least:
metal,
type,
grade,
and
bond.
8.3
The
measured
length
of
the
resonator
and
specimen
dielectric.
8.4
The
measured
thickness
of
specimen
cards
or,
if
appli¬
cable,
of
stacks.
8.5
The
center
conductor
width.
8.6
The
center
conductor
total
thickness
(for
type
C,
this
is
twice
the
cladding
thickness).
IPC-TM-650
Page 11 of 11
Number
2.5.5.5.1
Revision
Subject
Stripline
Test
for
Complex
Relative
Permittivity
of
Circuit
Board
Materials
to
14
GHz
Date
3/98
8.7
The
temperature
of
the
test
fixture,
if
not
in
the
21
to
23℃
range.
8.8
Any
conditioning
prior
to
measurement.
8.9
The
orientation
of
the
resonator
with
respect
to
X
or
Y
axis
of
the
specimen.
8.10
For
each
resonance,
show
8.10.1
through
8.10.9.
8.10.1
The
node
number
n.
8.10.2
The
calculated
effective
stripline
permittivity.
8.10.3
The
calculated
effective
dielectric
loss
tangent.
8.10.4
The
resonant
frequency,
fr,
at
maximum
transmis¬
sion.
8.10.5
The
insertion
loss
at
resonance,
dBr,
at
maximum
transmission.
8.10.6
The
Q|Oaded.
(optional).
8.10.7
The
calculated
Qunioaded
(optional).
8.10.8
If
the
three
point
method
of
6.3, 6.4,
or
6.5.1
is
used,
report
the
frequency
and
dB
value
of
the
two
points
either
side
of
the
peak
(optional).
8.10.9
If
the
non-linear
regression
(NLR)
method
of
6.5.2
is
used,
report
the
number
of
data
points
used,
NLR
uncertainty
values
(for
fr,
Q|Oaded,
dBr)
and
the
standard
deviation
of
the
fit
in
dB
units
(optional).
9
.0
Notes
9.1
Permittivity
The
dielectric
of
a
stripline
circuit
affects
the
electrical
response
of
all
the
circuits
printed
on
it.
Velocity
of
propagation,
wavelength,
and
characteristic
impedance
all
vary
with
permittivity.
If
the
permittivity
varies
from
the
design
value,
the
performance
of
such
circuits
is
degraded.
Throughout
this
document,
the
term
u
permittivity"
refers
to
relative
permittivity
of
the
dielectric
material,
a
dimensionless
ratio
of
the
absolute
permittivity
of
the
material
to
that
of
a
9.2
Loss
Tangent
The
attenuation
and
Q
(figure
of
merit)
of
stripline
circuits
are
a
function
of
combined
copper
and
dielec¬
tric
loss.
An
excessively
high
loss
tangent
leads
to
loss
in
sig¬
nal
strength
and
to
degraded
performance
of
frequency
selec¬
tive
circuits
such
as
filters.
9.3
Dielectrics
Clad
with
Thick
Metal
on
One
Side
This
method
can
be
used
for
measurements
of
dielectric
sub¬
strates
with
thin
foil
on
one
side
and
thick
cladding
such
as
aluminum
sheet
on
the
other
by
using
the
Type
C
specimen
configuration.
In
some
cases,
with
very
thick
metal
cladding
it
may
be
necessary
to
use
a
modified
part
5.1
.2
(Figure
4)
with
a
reduced
thickness
dimension.
9.4
Anisotropic
Materials
For
anisotropic
materials,
test
methods
in
which
the
electric
field
is
not
imposed
on
the
dielectric
in
a
stripline
configuration
can
give
misleading
values
of
effective
stripline
permittivity
and
loss
tangent.
This
test
method
measures
an
effective
stripline
permittivity
when
the
specimen
configuration
is
close
to
that
of
the
application.
10
.0
References
1
.
Electrical
Performance
of
Microwave
Boards,
IEEE
Trans.
Components,
Packaging
&
Manufacturing
Technology,
Part
B,
vol.
18,
no.
7,
Traut,
G.
R,
Feb.
1995.
2.
The
Complex
Permittivity
of
RF
Circuit
Board
Materials
by
Resonances
of
a
Stripline
Section
in
the
0.2
to
15
GHZ
Range,
Traut,
G.
Robert,
Preprints
of
the
Measurement
Science
Conference
1997
January
23
&
24,
Pasadena
Convention
Center,
Pasadena,
CA
3.
Complex
Permittivity
Over
a
Wide
Frequency
Range
by
Adjustable
Air
Gap
Probing
a
Stripline
Resonator,
Traut,
G.
Robert,
Proceedings
of
the
Technical
Conference,
IPO
Printed
Circuits
Expo,
March
9-13,
1997,
San
Jose
Con¬
vention
Center,
San
Jose,
CA.
4.
The
NIST
60-Millimeter
Diameter
Cylindrical
Cavity
斤
eso-
nator:
Performance
Evaluation
for
Permittivity
Measure¬
ments,
Vanzura,
E.
J.,
Geyer,
R.
G.
and
Janezic,
M.D.,
NIST
Technical
Note
1354,
August
1993,
National
Insti¬
tute
of
Standards
and
Technology,
Boulder,
CO
80303-
3328.
5.
Characteristic
Impedance
of
the
Shielded-Strip
Transmis¬
sion
Line,
Cohn,
S.
B.,
IRE
Trans
MTT,
(July
1954):
pp.
52
-
57.
vacuum.
6.
Problems
而
Strip
Transmission
Lines,
Cohn,
S.
B.,
IRE
Transactions
MTT
3
(March
1955):
pp.
119-126.