IPC-TM-650 EN 2022 试验方法.pdf - 第435页

• Sweep Frequency Generator Mainframe H.P. 8350B • RF Plug-In, 0.01 to 20 GHz H.P. 83592A. A plug-in of nar- rower frequency range (X-band) may be selected at consid- erable cost savings. 83545A 5.9 12.4 GHz. • Power Spl…

from

the base boards to the top of the 25.4 mm x 51 mm area

to which clamping pressure is applied. The minimum horizon-

tal dimension must be enough to extend at least 6.5 mm

beyond the center line of the vertical portion of the probe line

on either side. For the pattern card of Figure 4 and fixture of

Figure 12, these minimums are 38.1 mm x 68.6 mm. For the

smaller size, the clamp force in 6.1 or Table 1 is not changed

because the effective area over which the force is applied is

not reduced. The test fixture is designed to accommodate a

total specimen thickness of either 3.18 mm ± 0.22 mm or

2.54 mm ± 0.18 mm from an even number of layers.

NOTE: Testing of built-up specimens introduces error, which

can exceed 5% due to air gaps. Exact correlation factors and

techniques must be agreed upon or other methods of test

used. The 1 MHz method of IPC-TM-650, Method 2.5.5.3,

can be used with a correction factor based on tests of

samples of the nominal thickness of Table 1 using both tech-

niques.

With some material types not based on woven fabric rein-

forcement, it is possible to machine specimens to achieve the

nominal thickness for test.

4.0

Suggested Electronic Apparatus

The

principal com-

ponents required for the test setup consist of the test fixture

described in 5.0, a microwave signal source, an accurate

means of measuring the signal frequency, an accurate means

for detecting power level, and an accurate method of deter-

mining frequency values above and below the resonant fre-

quency at the half-power level for the test fixture loaded with

the specimen.

The microwave signal source must be capable of providing an

accurate signal. During the required time period and range of

frequency needed to make a permittivity and loss tangent

measurement, the source must provide a leveled power out-

put that falls within a 0.1 dB range. When the source is set for

a particular frequency, the output must be capable of remain-

ing within 5 MHz of the set value for the time required to make

a measurement.

The means for measuring frequency shall have a resolution of

5 MHz or less and an accuracy of 8 MHz or less. An error of

+8 MHz in measurement of a resonant frequency for a mate-

rial with nominal permittivity of 2.50 represents a -0.004 error

in permittivity.

The means for detecting the power level shall have a resolu-

tion of 0.1 dB or less and be capable of comparing power

levels withina3dBrange with an accuracy of 0.1 dB. An error

of 0.1 dB in estimating half power frequency points can result

in an error in the loss tangent of about 0.0001 for a material

with 2.5 permittivity. See 7.2, equation 5.

4.1

Manual Test Setup

The

method of determining the

half-power points depends partly on the type of signal source

used. If the power input to the test fixture is maintained con-

stant as the frequency is varied, then an SWR meter may be

used to determine the half-power points at the output of the

test fixture. This may be accomplished by using a leveled

sweep generator or by using a tunable klystron (at a consid-

erable savings) and manually adjusting the power input to the

test fixture to a prescribed level by use of a variable attenua-

tor. A typical equipment list is shown below. Equivalent makes

and models of equipment may be substituted where it can be

shown that equivalent results are obtained. For example, if a

leveling system is not used and the power output of the

source varies widely with frequency, a ratiometer may be sub-

stituted for the two SWR meters. If only permittivity is desired,

it is not necessary to level the input.

The following equipment, or equivalent, may be used.

• Sweep Frequency Generator H.P. 8350B or 8620C

• X-Band Frequency Plug in Unit H.P. 83545A or 86251A

• Frequency Meter H.P. X532B

• Crystal Detector (2) H.P. 423B (Neg)

• Matched Load Resistor for one Crystal Detector H.P.

11523A, opt. 001

• SWR Meter (2) H.P. 415E

• Directional Coupler HP 779D

• 10 dB Attenuator H.P. 8491B

• 8.9 kN Dillon Force Gauge, Compression Model X, Part

Number 381612301, with ± 1% full scale accuracy

• Vise or press that is able to exert controlled 4.45 kN force

on the test fixture and that opens at least 127 mm to accept

the force gauge and test fixture

• Semi-rigid Coaxial Cable and Connectors

• Waveguide to Coaxial Adapters (2) H.P. X281A

• The measuring equipment shall be connected as shown on

Figure 1.

4.2

A Test Setup for Computer Automation of Data

The

following

components or equivalent, properly interconnected,

can be used most effectively with a computer control program

for automated testing.

IPC-TM-650

Number

2.5.5.5

Subject

Stripline

Test for Permittivity and Loss Tangent (Dielectric Constant

and Dissipation Factor) at X-Band

Date

3/98

Revision

age2of25

电子技术应用　　　　　　　ｗｗｗ．ＣｈｉｎａＡＥＴ．ｃｏｍ

•

Sweep Frequency Generator Mainframe H.P. 8350B

• RF Plug-In, 0.01 to 20 GHz H.P. 83592A. A plug-in of nar-

rower frequency range (X-band) may be selected at consid-

erable cost savings. 83545A 5.9 12.4 GHz.

• Power Splitter H.P. 11667A

• Automatic Frequency Counter H.P. 5343A

• Source Synchronizer H.P. 5344A. Obtained as an intercon-

nected assembly with the counter.

• Coaxial cables and adapters.

• 10 dB Attenuator H.P. 8491B

• 8.9 kN Dillon Force Gauge, Compression Model X, Part

Number 381612301, with ±1% full scale accuracy.

• Vise or press that is able to exert controlled 4.45 kN force

on the test fixture and that opens at least 127 mm to accept

the force gauge and test fixture.

• Programmable Power Meter H.P. 436A

• Power Sensor H.P. 8484A with 70 to 10 dBm range.

• IEEE 488 (GPIB) cables

• Controlling computer with GPIB interface.

The above equipment is connected as follows as illustrated in

Figure 2:

• RF connections. The power splitter connects directly to the

RF plug-in output. One output of the splitter connects by RF

cable to the counter input. The other output is connected by

RF cable to the attenuator which connects to one of the test

fixture probe lines.

• Control connections. Connections between counter and

synchronizer are provided as specified by the manufacturer.

The FM output from the synchronizer connects by BNC to

the FM input on the sweeper. GPIB cables connect in par-

allel to sweeper, synchronizer, power meter, and computer

interface.

• Other connections. The power sensor is connected to the

other probe of the fixture and its special cable connects into

the power meter.

• A synthesized CW generator can be used to replace the

sweeper, plug-in, power splitter, connector, and source syn-

chronizer for the simpler set-up shown in Figure 3.

4.3

Automated Network Analyzer for the Test Setup

The

instrumentation described in 4.1 or 4.2 may be replaced

with either a scalar or vector network analyzer with test cables

connected to the test fixture of 5.0 as the device under test

(DUT). Examples of automated network analyzers known to

be suitable include the Hewlett-Packard 8510 vector network

analyzer or the Wiltron Model 561 scalar network analyzer.

These or equivalent may be used.

Such instruments may be operated either manually or under

computer control with suitable programming to locate the

resonant frequency and the frequencies above and below

resonance where transmitted power is 3 dB below that at

resonance. Network analyzers have several advantages over

the instrumentation described in 4.1 and 4.2. Data collection

is rapid and may be continuously refreshed with averaging.

The log magnitude response curve for ratio of transmitted to

incident power (the S21 parameter) as dB versus frequency is

visible on a screen for easy verification of a valid resonance. A

large number of dB frequency data points near the resonance

are readily available for optional use of non-linear regression

analysis techniques to determine the frequency and Q values

with statistically better degrees of uncertainty than those

attainable by the three point (f

and

)

method in either

section 6.2 or 6.3.

5.0

Test Fixture

5.1 Recommended Fixture Design

improved test fix-

ture design is shown that facilitates changing test pattern

cards and lends itself to control of temperature. The test fix-

ture shall be constructed as shown in Figure 4 through Figure

14.

The resonator circuit shown in Figure 4 is an example of a test

pattern designed for a material with a permittivity of 2.20. For

other permittivity values, different pattern dimensions will be

required as outlined in Table 1. It shall be defined on one side

of a material of similar type to that being tested, a laminate

with dielectric thickness of 0.216 mm ± .018 mm. The clad-

ding thickness is normally specified as MF-150F designation 1

copper (nominal thickness of 0.036 mm but designation down

to Q (0.010 mm) may also be used. Designation Q is preferred

for high permittivity materials as covered in 4.2 and 9.7, Note.

The reverse side of the circuit board has all copper removed.

The copper foil shall be of IPC-MF-150, type 1, electro-

deposited, type 5, wrought, or type 7, wrought-annealed. The

type of copper foil and the treatment for adhesion will affect

the Q measurement. The 1/Qc values in Table 1 do not take

into account surface treatments or higher resistivity values for

the conductor that are encountered with the specified foil

types.

IPC-TM-650

Number

2.5.5.5

Subject

Stripline

Test for Permittivity and Loss Tangent (Dielectric Constant

and Dissipation Factor) at X-Band

Date

3/98

Revision

age3of25

电子技术应用　　　　　　　ｗｗｗ．ＣｈｉｎａＡＥＴ．ｃｏｍ

The

test pattern card shall have a permittivity equal to the

nominal value of the type being tested with a tolerance of ±

2.5% of the nominal value (measured by stacking sufficient

plies to the total thickness requirement of a specimen as

above. Use a photo resist and etching method capable of

reproducing circuit dimensions with ± 0.025 mm tolerance. All

copper shall be removed from the other side of the test pat-

tern card. See 9.7, Note, for special treatment of ceramic-

PTFE substrate types.

The pattern card of Figure 4 is 68.6 mm wide by 55.4 mm

high and is designed for the fixture hardware in Figure 5

through Figure 14. The length is cut so that when the pattern

card is clamped for the lap joint with the striplines on the base

card, the resonator is centered in the 51 mm high area above

the base plates of the fixture. For materials with permittivity

values higher than the nominal 2.50 shown in Table 1, please

see 5.2 for a discussion of recommended fixture modifica-

tions.

Probe line widths are based on ground plane spacing taken

as twice the nominal thickness of the two specimens plus

thickness of the pattern card and its 0.034 mm copper foil

pattern and computed as if the stripline were centered

between ground planes

(1,3)

Chamfer

values are based on published design curves

(2)

The

length of the four node resonator is given in Table 1.

Resonators of lower node values for the purpose of measur-

ing ∆L according to 6.1, will be proportionately shorter with

the probe lengths modified so that the gap is the same.

The values for conductor loss, 1/Q

in Table 1 are calculated

from known properties of copper, the test frequency, the cal-

culated characteristic impedance of the section of stripline

comprising the resonator, and its cross-sectional geometry

using published formulas

(1)

The values shown are usually

biased low giving a high bias to loss tangent results, because

conductor actually used may not have a smooth surface and

may include oxides, microvoids, or other sources of higher

resistivity.

5.2

Fixture Modifications for High Permittivity Materi-

als

Modification

of the fixture design of Figure 5 through Fig-

ure 14 and pattern card dimensions in Figure 4 are recom-

mended to overcome problems experienced with extraneous

transmissions and resonances at frequencies near the desired

resonant peak.

5.2.1 Replace

the coax-stripline launcher shown in Figure 7.

The part suggested has a tab width of 1.27 mm and may be

replaced with Omni-Spectra Part No. 2070-5029-02, or

equivalent, intended for 1.57 mm ground plane spacing and

with a tab width of 0.635 mm. A further acceptable alternative

is to redesign the base plates to accept another type of

coaxial fitting such as a flange mount jack, which can be

modified to provide a smooth, low-reflection transition from

3.0 mm semirigid cable with Z

50 Ohm, low permittivity

insulation into stripline with Z

50 Ohm, and high permittiv-

ity insulation in the fixture.

5.2.2

the stripline launcher in 5.2.1 is used, the edge at

the step to accommodate the launcher body on the base

plate should be machined with a slight undercut for an acute

included angle of about 80°. This, combined with a means to

press the launcher body axially against the edge, will assure a

well-defined ground connection from coax to stripline. A

poorly defined ground connection with ground current path

length varying or longer than that of the signal conductor has

been found to give rise to scattering, reflections, and reso-

nances in the open ended probe line that are evident as extra-

neous fixture transmissions that may distort the resonant peak

to be measured.

5.2.3

Omit

the conductor lap joints but keep the extended

base cards in the fixture assembly. See figures 13 and 14.

With high permittivity materials, the lap joint also gives rise to

unwanted scattering, reflections, and resonances in the open-

ended probe line, as discussed in 5.2.2. For this purpose the

resonator pattern card will have a longer vertical dimension to

extend down to the launcher pin replacing the spacer board

in Figure 13. It should still center the resonator in the clamp-

ing block area. The base dielectric boards will be etched free

of metal. The ground plane foils will also extend down to the

launcher.

The feature of extending the base dielectric boards upward

above the base plates is to be retained as a means to prevent

premature damage to the resonator pattern card with

repeated loading and unloading of the fixture. The base plate

with the deeper step will be on the side toward which the

resonator pattern faces to avoid straining the offset launcher

tab during assembly.

5.2.4

Scale

down the fixture dimensions to move remaining

probe line resonances away from the resonant frequency of

interest. For ε

10.5 material, the following dimensions were

found effective.

IPC-TM-650

Number

2.5.5.5

Subject

Stripline

Test for Permittivity and Loss Tangent (Dielectric Constant

and Dissipation Factor) at X-Band

Date

3/98

Revision

age4of25

电子技术应用　　　　　　　ｗｗｗ．ＣｈｉｎａＡＥＴ．ｃｏｍ