IPC-4552.pdf - 第21页

IPC-4552 Proposal June 2001 -0 . 1 5 -0 . 1 0 -0 . 0 5 0.00 0.05 0.10 0.15 0.20 0.25 0 0.375 0.75 1.125 1.5 1.875 2.25 2.625 3 3.375 3.75 4.125 4.5 4.875 5.25 5.625 6 6.375 6.75 7.125 7.5 7.875 8.25 8.625 9 9.375 9.75 ti…

IPC-4552 Proposal June 2001

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.38

0.75

1.13

1.5

1.88

2.25

2.63

3.38

3.75

4.13

4.5

4.88

5.25

5.63

6.38

6.75

7.13

7.5

7.88

8.25

8.63

9.38

9.75

time in seconds

mN/mm

20 mins plating (4uins)

2 mins plating (1 uins)

Fig 9: Wetting balance data for Vendor D post 18 hours 85/85 conditioning

-0.15

-0.1

-0.05

0.05

0.1

0.15

0.2

0.25

0.375

0.75

1.125

1.5

1.875

2.25

2.625

3.375

3.75

4.125

4.5

4.875

5.25

5.625

6.375

6.75

7.125

7.5

7.875

8.25

8.625

9.375

9.75

time

mN/mm

Fig 10: Comparison of one microinch gold deposit tested at 8 months shelf life

Vs as received & 85/85

90 DAYS OLD

8 MONTHS OLD

POST 85/85

Page 18 of 10

Download From http://bbs.infoeach.com

IPC-4552 Proposal June 2001

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

0.25

0.375

0.75

1.125

1.5

1.875

2.25

2.625

3.375

3.75

4.125

4.5

4.875

5.25

5.625

6.375

6.75

7.125

7.5

7.875

8.25

8.625

9.375

9.75

tim e in seconds

mN/mm

18 hrs 85C/85% R.H.

575 days old

90 days old

APPENDIX 5 Fig 11: Comparison of one microinch Gold deposit after various storage times /

conditions

With all the data from the above testing the task of determining the deposit ranges began to be a simpler if not still a

very controversial task. Based on the performance of the one microinch gold deposit, it was felt that by doubling this

thickness as the minimum value would provide the end user with a significant amount of shelf life, there being very

little in the way of solderability performance improvements for the thicker deposits. Controversy with regard to what

2 microinch minimum actually meant raised the following points:

1) Was it the minimum average?

2) Was it the absolute minimum?

3) Would you accept a data point at 1.99 microinches?

Following some heated discussions the group proposed the following - the minimum gold thickness deposit is 2

microinches at minus four standard deviations from the mean value.

Based on all the testing to date and the input on plating times from the five chemical suppliers, a typical gold deposit

thickness range will be in the 3 to 5 microinches. A number of committee members looked to have an upper

specification for the gold thickness. We believe that it is not necessary as the reaction is self-limiting and based on

the XRF data, some of the gold systems may never be able to achieve an upper specification number. What is most

important is not an upper number but rather the performance of the deposit, which is governed by the minimum

thickness of the gold. This maybe a good point to mention that Gold embrittlement is not an issue with ENIG over

the ranges of deposit thickness measured in the XRF study.

Based on these studies and industry consensus, the thickness specification was set at:

5.1.1 Electroless Nickel Thickness:

The electroless nickel thickness shall be 3 to 6 microns (120 to 240 µin).

5.1.2

Immersion Gold Thickness:

The minimum immersion gold thickness shall be 0.05 microns (2 µin) at four sigma below the mean; the typical

range is 3 to 5 µin. Setup and measurement methodologies are crucial for accuracy (see appendix for recommended

Page 19 of 10

Download From http://bbs.infoeach.com

IPC-4552 Proposal June 2001

measurement techniques).

One of the byproducts of such an extensive series of tests was the identification of the variability that exists between

ENIG suppliers and that while the deposit thickness values maybe different the performance from a base

solderability standpoint is the same. It is clear that, all involved in ENIG – supplier, board house and OEM – need

to quantify their individual processes and have the necessary data to demonstrate that the process is known and

understood from a statistical / chemical and solderability viewpoint.

Contact Resistance Testing:

One of the functions that ENIG provides as a surface finish is that of a contact resistance surface that typically

interfaces with a “soft touch” membrane keypad. The membrane is usually carbon impregnated but other metal

dome switches are also used for this type of application, (the group is still requesting performance data on this type

of switch). While the group had a lot of anecdotal evidence of its use, cell phones, keypads etc, nobody had data for

this type of switch that was not company specific and therefore not shareable with the IPC community. In an attempt

to obtain shareable data, the committee decided to test some different patterns commonly in use in the industry.

After much searching we came upon S/G Industries, a supplier of membrane overlays to the automotive industry

among others. With their help, three different designs were fabricated:

1) Interlocking Square contacts

2) Interlocking Round contacts

3) Half Moon contacts

Note: At the time of writing this paper, test data from only the Interlocking square pattern is available.

To increase the data set it was decided to have two of the ENIG chemical suppliers plate these patterns. The testing

would be run until failure or until the resistance change would be considered non-useable for the switch application.

From the data from the first pattern testing at the time of writing, it would appear that we will stop test at 2 million

cycles with no negative, i.e. increase in resistance, being evident. The committee feels that 2 million cycles is more

than adequate to prove the use of ENIG as a suitable surface for this type of application.

The test protocol was as follows:

1) Each PWB/keypad assembly consists of five (5) pads.

2) Each pad represents a different carbon contact and trace size ranging from 2.0mm to 4.0 mm.

3) A total of six assembly sets will be tested

4) There are two ENIG suppliers, Vendor’s C&E

5) Initial testing consists of applying a 1000 gram load to the entire keypad (1000 grams / 5 switches per

keypad = 200 grams per switch).

6) Initial resistance measurements are taken from each switch location with this force applied without

passing current through the switch (mechanical durability test only)

7) The PCB/keypad assembly is then actuated in 50,000 cycle intervals with resistance measurements

taken at each interval. The forces applied during resistance measurements are identical to those used

during the initial resistance measurements (200 grams per switch).

8) Reference of positions measured: Position 1: 2.0 mm carbon contact, Position 2: 2.5 mm carbon

contact, Position 3: 3.0 mm carbon contact, Position 4: 3.5 mm carbon contact, Position 5: 4.0 mm

carbon contact.

The results from 1.85 million cycles for the interlocking square pattern are detailed below.

As can be seen from this summary data set the general trend in the resistance data is down, the direction of choice.

In discussions with S/G Industries regarding these phenomena, they explained that the carbon particles that are

impregnated into the pill tend to be compacted as the switch is compressed over time. This therefore reduces the

contact resistance, which is clearly demonstrated in the above graphs. A visual

Page 20 of 10

Download From http://bbs.infoeach.com