Testing

The testing activities are aimed at qualifying the developed technologies as well as verifying the models developed. Successfully developed models can be employed then to predict the reliability of the technologies and to evaluate the reliability of flexible board design from a thermo-mechanical point of view.

I Reliability Test Plan

This part details the test activities associated with the development of the technology.

The test plan is done first for testing the flex board with surface mounted components.

Test Specimen Description:

• Multuilayer Flex with Surface Mount Components
• Surface Mount Components/Packages – 168, 144, 80, and 44 I/O
• Via Chains
• Flex Board
• 4 to 6 layers
• Board Size – 101 mm x 48 mm
• Six tooling holes – one each corner and one in the middle of the longer side
• Via chains
• Six tooling holes non-plated 2.2
• Total Number of boards – 30 boards (Maximum)

The following Assumptions are made in the test plan.

The base multilayer flex laminate material has been tested for the following properties and has been demonstrated to pass all the standard tests and requirements.

1. 1. Dielectric constant, 2. Dielectric strength, 3. Dissipation factor 4. Moisture absorption 5. Insulation resistance (horizontal and vertical), 6.Tensile strength, 7. Elongation, 8. Glass transition temperature, 9. Plating adhesion, 10. Chemical resistance, 11. Peel strength, 12. Tear Resistance, and any other requirements.
2. The test results on the bare flex laminate are readily available for any future failure analyses diagnostics.
3. The specific test vehicle has been assembled with requisite hardware with optimized assembly process and meets all the end-of-line (EOL) quality criteria.
4. The new technology is equivalent or better than the technology it replaces in terms of product quality and reliability.

The reliability testing is planned to be carried out alongside with the current FR-4 organic laminate technology to establish a comparative performance. Alternatively, the test results on rigid laminates are already available for comparison purposes.

The objectives of these tests are as follows.

1. Verify and validate the mechanical and thermal cycling modeling results in terms of the relative of strains experienced during the mechanical drop loading thermal cycling on the package-to- board interconnection integrity.
2. Assess the relative performance of the flex laminate assembly with conventional FR-4 assembly.
3. Determine the corrosion and electromigration propensity in a temperature, humidity, and bias environment.
4. Determine the failure location, failure mode and mechanisms, so that appropriate materials and/or design changes can be implemented to increase the performance.

A) Temperature Cycling Test (life time prediction

The thermal cycling test is used to assess the package to board interconnection strength, the integrity of different kind of via structures such as blind, .m-via, buried, and through vias, and –interfaces under cyclic strain. While thermal cycling test is not a simulation of use environment it, however, constitutes an accelerated life test. It can also be used to compare different packages assemblies and also enable weak links in the system in question.

Thus, one can determine a) the reliability of Plated Through Holes (PTH) and/or microvias, b) reliability of solder joint connection under creep and fatigue stress, c) weakest interfaces inside the component and board, d) reliability of different materials in substrate and package, and also, e) discern any assembly related issues.

Thermal cycling of the assemblies is to be performed in the –40° C to 125° C range with 3-5 minute ramp and 10-15min dwell at each of the extremities of temperature as shown in figure 2. The minimum thermal cycling requirement is 500 cycles without failure. A failure is defined as the occurrence of >200 ns resistance peak over 1000 Ohms, followed by 9 consecutive resistance peaks over 1000 Ohms. Observation slot is 2 sec.

The test vehicle should be populated with daisy chained packages, daisy chained via structures, resistor networks, etc. Recommended daisy chain components of a given type should be a minimum number of 30. A minimum of 10 via chains is required to be monitored during the test.

An air-to-air temperature cycling chamber, which can be programmed to provide the thermal profile indicated in figure 2, shall be used. An event detector with computer capable of in-situ monitoring and continuous measurement of daisy chain resistance is employed.

The following procedure is adopted in carrying out the thermal cycling test.

All the cables must be connected in through holes by soldering. One risk net is connected to each measurement channel. Each package is to have its own grounding cable. Special attention should be paid on grounding of test setup. The soldered test vehicles are connected in such a manner that there is no external mechanical stress imposed on the soldering joints. Flex circuit boards may require work board holders and should be considered. Samples need to be located inside the chamber so that they are not in any blind spot and also be stable to the airflow.

It is also important to position in the boards in a manner that boards do not obstruct the airflow to the adjacent boards.

The test is continued until 1000 cycles or at least 50 % of components or via chains have experienced failures.

It is desirable to have cut outs around each package so that a failed component can be removed for analysis immediately after the detection of failure, while the test can continue with good packages. Alternatively, since the carrier is flexible, the failed component is easily separated without effecting other components.

Failure analysis results, which show the first failure mechanism, the other failure mechanisms and distribution of failure mechanism are provided.

In addition, thermal cycling test results should include a PWB material property table and surface finish for completeness of information.

B) Drop test for assembled FLEX

As flex assemblies are replacing the rigid assemblies in this effort, it is imperative that the flex assemblies meet the same mechanical reliability requirements. Hence mechanical drop test evaluations are to be made to the same standard. As the mechanics in each product are likely to be design specific and the use profile may vary considerably, it is well nigh impossible to correlate board level to product performance.Board level drop tests are used to evaluate package technologies, package-to-board interconnection integrity on flex, and determine the weakest links in the system, and also intercompare rigid and flex assemblies during early stages of technology development.

Board level drop test is used to evaluate flex assemblies - and m-via chain reliability against mechanical shock. The test vehicles can be populated with CSP components, LGAs, and other parts pertinent to the demonstrator vehicles.

The effect of a number of variables on mechanical shock can be valuated through this test. These include: weight distribution, relative strengths pf various interfaces, package/component locational effects, etc.

All the cables must be connected in through holes by soldering. Each package must have its own dedicated grounding cable. Special attention should be paid on grounding of test setup. One channel is designated for each package.

All the cable-soldering areas in single- and double-sided boards have to be protected from damage during drop by wrapping appropriately with a nonconductive tape.

A minimum of 40 components in five to six boards is to be subjected to the drop. 10 to 15 micro-via chains are to be monitored during the test.

A test board is to survive at least 10 drops without any failure. Resistance peak over 1500 Ohm and longer than 200 ns is counted as failure. Observation slot is 2 sec. When detecting the first failure of the tested structure, dropping has to be stopped immediately after the first failure.This has to be carried out with one test specimen in all the different test specimen cases.

The calibration with accelerometer for the jig should be about 1500 g +/- 20 % and length 1 ms +/- 30 % for shock pulse. Digital oscilloscope is used for pulse detection. The shock pulse must be checked always before a new test series. This helps in reference data analyzing.

Testing is continued to at least 30 drops. Then a failure distribution could be created. The test must be carried out in normal laboratory atmosphere.

Drop test results consist of the following

Failure plot in coordinates (x=number of cycles, y=cumulative failure) should be prepared. All the different test cases have to be presented separately. Lot size Number of failed test specimens and m-via chains should be indicated. Tested specimens codes, PWB material property table and surface, finish location of failures.

Failure analysis results, which show the first failure mechanism, all the detected failure mechanisms and the distribution of the different failure mechanisms should be reported.

A comparison of the test results with rigid boards similarly populated is to be included to judge the performance of flex assembly with rigid assembly under the same stress distributions.

C) Temperature, Humidity, Bias Testing

Purpose

The purpose of this test is to evaluate the performance of the assemblies in humid usage environments. The resistance of the flex printed wiring boards to surface insulation and inter-layer insulation resistance degradation due to corrosion and electromigration when exposed to high temperature and high humidity environments needs sets to be understood.

The resistance values between the comb pattern lines on the surface should not be less than 1 Meg Ohm. The conditions of the test are 85° C and 85% r.h. (non-condensing atmosphere) and should last at least 1000 hrs. The insulation resistance measurements are to be made every 100 hrs.

Equipment

A temperature and humidity chamber capable of attaining and maintaining 85° C and 85% relative humidity. Temperature is maintained within 0.1° C of the set temperature and humidity should be maintained within 2 % of the set value.

Insulation resistance tester capable of measuring 10E12 ohms or greater.

A minimum of 3 samples of each category is evaluated.

Procedure

The chamber is set to the appropriate temperature and humidity settings. The chamber is maintained at 85° C and 85 % relative humidity. The bias voltage is set 100 v.
It is important to ensure that there is no moisture condensation onto the test hardware at any time during the test.
Soldering to the through holes should make all connections to the board.
Resistance measurements of all the test patterns should be made initially before loading them into the chamber.
Test patterns should have individual grounding cables
Each test pattern should be individually connected to the measurement channel.
At the end of the test the hardware must be brought to the ambient temperature and humidity and resistance measurements should be repeated.

D) High Temperature/High Humidity Storage Test

Purpose

The purpose of this test is to assess the moisture absorption propensity of the flexible printed wiring board. Any absorbed moisture within the inner layers of flex might eventually lead to corrosion and electromigration phenomena leading to electrical shorts.

Equipment

1. A temperature and humidity chamber capable of attaining and maintaining 85° C and 85% relative humidity
2. A chemical balance capable weighing samples to nearest a tenth of milligram

Procedure

The multilayer flex assembly is dried for 4hrs at 120° C in a thermal chamber. The samples are cooled to the room temperature and weighed to the nearest milligram.

The samples are then set in a chamber set to 85° C and 85% relative humidity. The samples are removed for weighing at intervals of 1, 4, 8, 16, and 24 hrs, and then at intervals of 24 hrs (for a maximum of 96 hrs) or until the weights reached a steady state values. The samples taken out of the chamber at any interval is brought to ambient temperature, wiped with a dry cloth, and weighed to the nearest 0.1 milligram, and then returned to the chamber.

The rate of moisture uptake is calculated at each interval. The rate at 24hrs interval is to be used as the standard value.

E) Solder DIP Test

Purpose

This test designed to evaluate the effects of multiple component rework on the carrier such as delamination, blistering, via and plated through hole damage, etc.,

Samples of carrier materials are fluxed and dipped in solder pot set to 280° C for 10 seconds, taken out, and brought to the ambient temperature.

The samples are examined for surface damage, as well as with cross-sectioning for via integrity and internal damage.

The process is repeated with another set of samples where the solder dipping is done two and three times.

A minimum of 3X solder dip survival with no internal or external damage to the carrier needs to be accomplished.

II. Construction of a bending test apparatus (FhG-IZM)

For the investigation of the mechanical stability of test samples with embedded components, we constructed a bending test apparatus (s. fig. 8)

The test vehicle is pulled back and forth over an axis. The pull force is adjustable by a variable load. The bending radius is defined by the axis, which can easily exchanged. A control unit allows a setup of cycles and stops automatically after the run of bending cycles. The minimum length of test vehicle are some centimeters, the maximum width is 21cm.

III. Reliability testing of stacked vias assembled boards (Thales)

All tests have been performed with respect to IPC EIA J-STD-003A standard. Only connection reliability has been checked: the test has conducted using rigid substrates in place of flex ones, as handling and testing are much easier to perform. The stack-up was done using the soldering diffusion technique at 232°C (lamination pressure 30 bar, duration 2 hours).

Main differences between the three patterns are: the drilling diameters (500, 400 and 300 µm) and the via land diameter (900, 800 and 700 µm). The conductors' length remained the same equal to 300 µm between the vias.

The stack-up was done using the soldering diffusion technique at 232°C (lamination pressure 30 bar, duration 2 hours), we present the interconnection principle:

The daisy chain test vehicles electrically tested were submitted to a standard THALES Airborne Systems qualification test file:

• Zero-hour cross section,
• Solder float test (10 s dipping test in a 288°C melted SnPb bath),
• Reflow simulation (to simulate component assembly on substrate both sides),
• 300 Thermal shocks –55/+125°C (two chambers oven/ 30 min dwelling time),
• 300 Thermal cycles –55/+125°C (one chamber oven/ 2-3°C/min heating rate, 20 min dwelling time).

Zero-hour cross section

We present hereafter a typical cross section of the daisy chain test pattern used for the reliability assessment:

Whatever the joint created (SnAu or SnAg), the visual inspection gives the following result:

Solder float test

According the IPC EIA J-STD-003A test C standard, the solder float test is the most stringent and critical of all tests for plating adhesion. Solder float is the method most commonly used by PCB manufacturers to determine the reliability of a plated through hole.

We performed the following procedure:

• Before the test, samples were baked 4 hours at 150°C.
• Samples were then dipped into a SnPb solder bath at 288°± 5 °C for 10 seconds.
• After the test a measure of continuity was made. No variation was noticed.

The cross-sections after the test achieved both SnAu and SnAg assembled vehicles did not show any degradation of connections.

Reflow simulation

The objective of this test is to simulate a component to board soldering on PCB both sides:

• Before vapor phase samples are baked 2 hours at 120°C.
• Samples are exposed to the vapor phase at 220°C for 1 minute after a dwell time of 1 minute of preheating at 125°C.

This test has been done on one design for each circuit. After the test a measure of the continuity has been made and no variation has been noticed.

Cross-section made after the test did not show any degradation of connections.

Thermal assessment within the –55/+125°C temperature range

Two types of thermal solicitation were applied to daisy chain samples:

• Thermal shocks using a two chambers oven (joint "elasticity" involvement),
• Thermal cycling using one chamber oven (joint "plasticity" involvement).

Doing thermal shocks leads to expose samples to a series of high and low temperature excursions of 30 minutes into two chambers.

• Low temperature chamber is at –55°C and high temperature chamber is at 125°C.
• Transfer time between the chambers is less than 2 minutes.

Doing thermal cycling leads to expose samples to a series of high and low temperature excursions but at a much slower rate (2-3°C/min).

So samples were subjected respectively to those tests, and a measure of continuity of each daisy chain was done and no failure was noticed.

Conclusions

We have conducted typical qualification tests on SnAu and SnAg joints built using "standard" defined experimental conditions (i.e. assembly temperature 232°C/ lamination pressure 30 bar / lamination time 2 hours). Assembled test vehicles exhibit a comparable performance as standard plated through holes board do. A complete multilayer flex with very similar design as the one used for this work has been assembled.