DropTesting.pdf

Drop Testing Reliability of Solder Interconnections

Dr. John Pan

Micah Denecour, David Guerrero, Jen Van Donk

EXECUTIVE SUMMARY

For our Honors Research project, we decided to investigate the drop test

reliability of lead free solder joints. The reason for looking into this is that recently, a law

was passed in the European Union (EU) that prohibits the use of leaded solders for

consumer applications. Since such electronic devices are prone to being dropped, it is

critical to understand how lead free solder joints withstand forces similar to those

incurred during their use.

In July 2006, the EU passed the Restriction of Harmful Substances directive

prohibiting the use of leaded solders in consumer electronics. The reason for this move

away from traditional leaded solder is the harmful effects of lead. Consequently, the

industry has been forced to look for alternative solders, and tin silver copper is the current

favorite. For this reason, our research tests the reliability of a tin silver copper alloy used

as a solder.

To perform our research, we are adhereing to the industry standard set by the

Joint Electron Device Engineering Council to perform drop tests. This standard

designates the board design as well as the drop table set-up. The standard calls for 15 ball

grid array integrated circuits to be placed on one side of a board. Then, the board is to be

bolted to a drop table and subjected to an acceleration of 1500 G’s for a 2 millisecond

pulse duration. To collect data, our group has created a software and hardware package to

measure the voltage drop across each chip. In this way, we can tell if a solder joint has

failed on any one chip.

So far, we have completed drop testing of 2 of the 50 circuit boards we have

obtained. The reason for this is that we had mechanical difficulties with our drop table. In

the summer to come, we plan on continuing our research at Henkel, where we will have

access to a sufficient drop table. In the fall, we will begin to analyze the data that is

collected during the summer.

INTRODUCTION

A recent directive in the European Union has banned leaded solder for use in

consumer electronics like cell phones, cameras, and other small handheld devices.

Recently companies have transitioned to a tin silver copper alloy and though some

thermal cycling testing has shown results for reliability in high ranging temperatures of

the new alternative there has not been significant research on drop test reliability. This

project tests the drop reliability of the solder joints, analyzes the data for trends in failure,

and makes recommendations for how to increase the solder lifetime. This paper will

discuss the background behind our research, the theory of solder joint reliability, the

methods we used to test the joint reliability, and the conclusions we reached after

experimentation.

BACKGROUND

In July 2006, the EU enacted the directive known as the Restriction of Harmful

Substances directive which prohibits the use of leaded solders in consumer electronics.

The directive does not apply to applications where high reliability is necessary, such as

military, aerospace, or scientific applications. The reason for this is that the reliability of

lead free solders is not completely understood and if the newer solders do not work as

well it could potentially be dangerous.

The motivation behind moving toward lead free solders is the negative effects of

lead. For example, in the United States, leaded paints are banned due to their harmful

nature. In electronics, lead is potentially harmful not during the use of the device, but

when it is discarded. Due to the fast turnover in the electronics industry, consumers

purchase new devices at an increasing rate. After the device’s life is up, it is generally put

into landfills. From there, lead can seep into water supplies and contaminate animals and

humans alike. Lead has been shown to cause blood and brain diseases if it is ingested.

The front runner to replace leaded solders is tin silver copper. However, due to the

fact that lead has always been used, the properties of tin silver copper have not been

thoroughly investigated. Consequently, much research is currently being performed to see

if this new alloy can replace leaded solders in both consumer and high reliability

applications.

To perform our research, we will be following the standards set by the Joint

Electron Device Engineering Council (JEDEC). This organization determines everything

from the circuit board composition to the number of components to be placed on the

board. For example, there must be 15 daisy chained integrated circuits spaced equally on

the board. Also, the board must experience a peak acceleration of 1500 G’s on each drop

for a duration of 2 milliseconds.

THEORY

There has been much study recently on the failure modes, and the causes of

failures in board level electronic assemblies. Much study has gone into thermal cycling

of board level assemblies in order to characterize the reliability of electronics in harsh

environments where temperature is highly

variable. Recently with the rise in sale of

portable and hand held electronic devices, the

drop reliability is a problem many people are

investigating. Research by Lim, et al, from the

institute of microelectronics looked at the

Figure. Bending of the circuit board is the

primary reason for failure of solder balls

strains of a cell phone printed circuit board (PCB) and of a PDA PCB. They found that

bending occurs in the PCB board in both longitudinal and transverse axis. Other studies

found that this bending causes tensile strains on the solder joints that connect the

electronic packages to the PCB. Due to the scaling down of solder joints along with

electronic packages, these joints are experiencing higher amounts of strain which lead to

brittle fracture, or fatigue fracture. Some research in the fracture of lead free solder joints

has been conducted by a team at Flextronics.

Taking a closer look at the solder joint it can be seen that along its length its

properties vary. Solder joints have three zones where along the length of the joint three

different mechanical properties exist. There are two intermetallic zones which are very

hard, but brittle, and the center zone which is the material of the solder joint. The

intermetallic regions occur due to the fact that the solder balls used in creating solder

joints need to be heated. This heating allows for the solder material to interact with the

material from the solder pads creating regions that have elements from both materials

creating the intermetallic zones.

During drop testing, two stages of

failure occur. The first is transitional failure.

This happens when the resistance of the solder

joint increases, but current can still flow

through the joint. A higher resistance can

occur in two way, the first is that the electrical

conduction is less due to a crack decreasing

the cross sectional area of the solder joint.

The second way is that any plastic

deformation in the solder joint will increase

its resistivity due to the increase in

dislocations within the bulk that can scatter

electrons. The second stage in failure is

complete failure. This occurs when the solder

joint has experienced complete fracture and

current can't flow due to a discontinuity within

the conduction path. When this occurs the two

pieces of the solder joint can actually continue to touch so the joint may continue to

function after complete failure, however, when complete failure occurs, most likely the

solder joint is discontinuous and no current will flow.

Figure . Shows solder balls used after the

soldering process has been done. Intermetallic

zones form at the boundary of the gold colored

solder pads and the solder ball.

METHODS

In order to test the reliability we had three key instruments; a drop table to create an

industry standard acceleration, a high speed data acquisition system to test the voltage

drop across each chip, and an accelerometer to ensure consistent acceleration.

The Table

Our key instrument was a large drop table. First we attached the board

to the top of the drop table using bolts in the four corners of the boards.

JEDEC standard specifies that the chips should face down toward the table

so that is how we mounted it. We then raised the table to allow a 50 in.

drop. This height created an average acceleration peak of about 600 Gs.

This was slightly lower than the standard 1500 Gs however it still allowed

us to see some trends of failure.

Initially we began with a brake enabled. This stopped the table after it

bounced once. Then we felt like we would better simulate a dropping

device if we allowed the table to bounce several times. Unfortunately we

could not get accurate readings on the acceleration of subsequent impacts

and therefore other researchers could not repeat the tests without the same

machine. This glitch would nullify all our data, and so we once again

enabled the break to allow only one impact.

B. The Data Acquisitions System

The data acquisition system consisted of two components. The

hardware was a system of resistors that connected to each chip. As the

board dropped the system measure the resistance across each resistor, and

therefore the voltage across each chip. A long cable that had leads for each

one of the chips on the boards as well as one ground lead connected the

board and resistor boxes. We soldered the cables onto the boards before

beginning each new test board.

The second part of the data acquisition system was software,

commonly referred to as DropGather, that compiled all the data collected.

DropGather read a couple thousand points of data over approximately 2

seconds and created graphs of the data. We could then look through the

graphs at each chip and find which ones had failure. The software also

allowed us to comment on which chips failed on a given drop and write it

directly into the file.

This data acquisition system was unique to our project because a

group of students developed it in the past few years specifically for this

project. Other research has measured the continuity by hand and though

this does show when the chips have completely failed, our method allows

us to see when even small transitional failure occurs. Furthermore, it is a

much more efficient method of testing because we do not have to stop

after each drop and measure each resistor. DropGather proved essential in

accurate and efficient data collection.

The Accelerometer

The third piece of equipment necessary for our tests was an

accelerometer. It helped find the correct acceleration during the initial

calibration. Later when we were conducting our data collection tests we

continued to measure the acceleration to ensure that the peak acceleration

remained consistent.

Our average acceleration turned out to be around 600 Gs, which is

only about 1/3 the 1500 G peak acceleration outlined in the JEDEC

standard. However we were able to still see some trends even with the

acceleration being significantly lower.

In future tests we would like to place the accelerometer on each of the

chips to discover the acceleration differences between each location on the

board. We believe the different accelerations could contribute to the

lifespan of each chip.

One difficulty we had with the software was that it did not plot

acceleration for more than 2 seconds, so we could not get accelerations for

the peaks if we allowed the machine to bounce, therefore we had to go

back to allowing it to bounce only once.

CONCLUSIONS

Through our initiation testing we found that:

The procedure developed for drop testing is valid and can be used on any drop

table.

The MTS drop table that we were using had about a maximum of 700 G

maximum acceleration with 2 ms pulse duration. This means that our tests were

not compliant with the JEDEC standard.

Chips located at positions where the printed circuit board has maximum bending

fail earlier than chips that are not located at these positions. These locations are

the center chips and the middle edge chips. This correlates with results from

many other researchers.

Recommendations:

After initial testing was completed:

Use strain gauges during calibration in order to get useful information about board

bending during the acceleration pulse.

A drop table capable of achieving the JEDEC standard acceleration pulse should

be used so that comparison of data with out researchers can be done.

Have a large sample size so that statistical analysis can be done to obtain useful

facts.

Use cross sectioning of the solder joints with optical microscopy and scanning

electron microscopy to find out about the failure modes occurring during our

tests.

BIBLIOGRAPHY

Callister, William, "Materials Science and Engineering: An Introduction", 6 th edition ,

page 622, 2003.

Haohuan, Lou, et al. Lifetime Assessment of Solder Joints of BGA Package in Board

Level Drop Test. In: Proc 6 th Intl Conf on Elec Pack Tech, 2005.

Plik z chomika:

Inne pliki z tego folderu:

Inne foldery tego chomika: