SMTAi Paper_Printing Small Aperture components Final Aug_2_2019 - 第3页

variations in shape. B ased on previous exper iments we determined that the 0.00 2” thick stencil had the best transfer efficiency. When speci fying a stencil thickness more often we take in consideration t he two ends o…

Printed Circuit Boards (PCB):

As Process Engineers, most PCB designs are created long

before we become involved. Design for manufacturing

considerations are most often out weighted by cost

requirements. We are often faced with having to design the

process to the board rather than the other way around where

adjustments and compromises are made to accommodate

shortfalls in the PCB design. This is where the Process

Engineer needs to take an active role to make sure that

success is designed into the PCB to insure repeatable

results. So far, applications have been focused on micro

components with little to no mixed technology on multi-up

panels. The problems with mixed technology, also referred

to as the broadband printing issue, is not the focus of this

paper though this would need to be addressed before

proceeding. Most applications to date have been using PCB

thicknesses of 0.030” (0.762mm) or less, so this should be

expected. Based on the PCB thickness and the amount of

routing, special carriers maybe needed to transfer the

product from machine to machine as well as stabilize it

during the reflow process. One of the key elements to

successful printing which the PCB plays a major role is the

capability to form a seal between the stencil and the PCB

commonly referred to as gasketing. Gasketing in turn plays

directly into the process to transfer the paste from the stencil

aperture to the PCB pad efficiently and repeatably.

Maintaining a consistent board thickness, by elimination of

variance in mask thicknesses from PCB to PCB, is key as

this determines the position of the PCB to the bottom stencil

surface. Mask encroachment on to the pad surface should be

eliminated as this will greatly affect the capability to seal

the aperture. Nomenclature and silk screens should be

avoided in the PCB design. Issues arise when silkscreen is

equal to or greater than the stencil thickness being applied

along with being located adjacent to the component. Logos

and identification information should be presented on the

back of the PCB or located in consideration to cause the

least effect on the critical components. The use of barcode

labels should be avoided as this is the source of many

printing issues with less challenging components. Barcode

labels should be applied after the print process is completed

if possible. One of the critical mistakes when designing

PCBs for micro component printing is to make sure the pad

is not significantly below the mask height. When designing

a board with a bare copper pad, the pad should be just below

or equal to the mask height. If the pad is below the mask,

this creates a gap that the paste must now overcome when it

is printed into the aperture. Because the paste must flow

between the aperture opening on the bottom of the stencil to

the pad surface, the paste will be unable to secure a proper

adhesion to pad surface. The result is erratic print results

that is represented in opens and insufficient volume or pad

coverage. In some cases, the PCBs with this issue become

non-manufacturable due to erratic results from the printer.

To insure the pad is positioned correctly to the mask height

and has a flat planar surface, an electroless nickel

immersion gold (ENIG) plating should be considered. ENIG

plating, consisting of an electroless nickel plating covered

with a thin layer of immersion gold, protects the nickel from

oxidation has shown to have the best results. In the board

design, the decision to use mask defined pads often present

issues. Accuracy of the pad locations as well as pads that are

sized larger than the specification have been issues when

using mask defined pads. Location of the fiducials,

especially when implementing a multi-up panel, is critical

for the machine vision systems to properly perform

alignment. The fiducials should be part of the artwork and

be present on the PCB image for best results. Avoid locating

the fiducials on the breakout panel as this tends to add to

any alignment error. Stretch and step and repeat errors

should be avoided as we are dealing with a recommended

pad size of 0.005” X 0.006” (0.127 x 0.1524mm) where as

much as a 0.001” (0.0254mm) error can have significant

consequences. Investment up front in the PCB design and

manufacturing will insure success where some of the

previously described issues are often difficult if not

impossible to overcome.

Squeegee Blades/Enclosed Heads:

A best practice is to separate a set of squeegee blades

specifically used for micro-component printing. This insures

that the blades being used are undamaged and not worn. The

squeegee requirements for micro-component printing is

simple, spring steel blades with a squared edge is all that is

required. A blade angle of 55 degrees is also recommended

where standard blade angles are set to 60 degrees. This

change in angle allows more surface area of the blade over

the aperture to promote an improved aperture fill. It also

improves sheering off the paste at this angle when the blade

passes over the aperture to prevent paste drag out and erratic

aperture fill. Blade length should match the PCB as closely

as possible with a maximum size range within 2 inches of

the PCB size in X direction. This will center the squeegee

pressure on the PCB as well as prevent long term damage to

the stencil. Inspect the blades every time prior to use for

cleanliness and for damage. It should be noted that enclosed

heads have been used for fine featured printing applications

with great success in the market today. The extrusion flow

from the pressurized chambers are compatible for repeatable

aperture fill for micro-component printing. Some Type 6

pastes have a limited stencil life where enclosed chambers

address this issue and minimizes paste waste. For this test

we focused on squeegee blades as this represents most of

the process applications on the market.

Stencils:

For this experiment we used a 29” x 29” (736 x 736mm)

fine grain, laser -cut, Nano-coated, 0.002” (0.0508mm)

thick stencil. The aperture size is a square, 0.005” x 0.006”

(0.127 x 0.1524mm) that is one to one with no reductions or

variations in shape. Based on previous experiments we

determined that the 0.002” thick stencil had the best transfer

efficiency. When specifying a stencil thickness more often

we take in consideration the two ends of the spectrum for

paste requirements and find a compromise in-between. Most

applications so far using micro-component printing, have

had compatible component mixes where there was not a

significant difference in requirements. The frame size we

used was 29” x 29” (736 x 736mm), however, 23” x 23”

(584 x 584mm) stencil may be better suited based on

common average board size for 008004” (0201mm)

applications and stencil tensioning requirements outlined

below. It is recommended using a fine grain stainless steel

stencil that is laser cut. Electroform stencils have fallen out

of favor with reported issues such as variation on aperture

size and foil thickness and stretch being introduced to the

image. Recommended for this application is to use high

tensioned foils. Stencils have a range for tensioning

normally 28 – 40N/cm² (Newton/centimeter). Most stencil

tension falls into the lower 30-Newton range. Increasing the

tension into the upper 30-Newton range prevents stencil

drag. Stencil drag is when you are using a thin stencil foil

with a significant amount of aperture openings. The surface

tension of the paste that has now adhered to the PCB, pulls

at the stencil foil during release, resulting in lower paste

transfer efficiency. The higher tension results in a cleaner

more balanced release with no transfer issues. Nano-coating

is recommended as studies have proven it improves transfer

efficiency. Stencil manufactures have improved the

application methods for applying Nano-coating to stencils

that has improved its manufacturing life. However,

aggressive fluxes and repeated aggressive wipes will

eventually wear the coating off. Careful handling of thin

stencils should be emphasized as they are easily damage.

Special care should be used when storing and transferring to

and from the machine. Take care when handling blades over

the stencil in the machine as a dropped blade could ruin a

stencil quickly. Cleaning the stencil using ultrasonic

methods after printing is essential for continued stencil life.

Type 6 paste is difficult to clean and can become difficult if

not impossible to remove if not removed promptly after use.

Solder Paste:

The recommended solder paste for this aperture size is a

Type 6 powder size. The specification for Type 6 is a mesh

size of +635 mesh size with the ball size range of less than

20µ with an average of 10µ. Type 4 paste is the prevailing

powder size presently being used in the SMT market.

Significant improvements in powder size yields have eased

pricing for Type 4 and Type 5. However, Type 6 pricing has

remained constant where comparative pricing can be three

times the cost of the Type 4, they are presently using.

Squeegee speeds and release parameters are dictated by the

paste formulation and flux type. From the printing

prospective, Type 6 prints like any other paste, however

considerations of the requirements down steam need also be

considered. Matching the paste to the Pick and Place as well

as the requirement for using nitrogen during reflow must be

also considered when using Type 6 paste. As best practice

for a 0.005” x 0.006” (0.127 x 0.1524mm) aperture is a

Type 6 paste – experiments going forward need to be

performed to see if a hybrid Type 5.5 powder size or a Type

5 can be substituted for a Type 6.

PCB Support:

The consensus in printing is that tooling support is essential

to successful, repeatable print results. The aluminum tooling

plate is still the touchstone that all other forms of support

are tested against. Since most applications for micro-

component printing use PCBs 0.030” (0.762mm) or less, the

tooling in combination with vacuum assist to insure the

PCB is flat, level and supported will give the best results.

The plate should be designed so the PCB fits in a recessed

pocket with the PCB surface positioned above the tooling

surface. Support wings are also recommended to support the

squeegee outside of the print area to prevent long term

damage to the stencil. Recommended is a Venturi vacuum

system as the Hg (inches of mercury) produced by standard

vacuum systems may not be enough to flatten the PCB.

When implementing vacuum openings on the plate, take in

consideration the thickness of the PCB relative to the hole

size to prevent deflection or “dimpling” of the PCB surface.

Special attention needs to be focused on the leveling of the

bottom of the tooling plate fixture. This will be reflected on

how well the PCB gaskets to the stencil. Addressing how to

hold the PCB in place during the print process, vacuum is

the preferred method to insure a flat surface over top or side

clamping for PCBs thickness below 0.030” (0.508mm).

Wiping:

Wiping is the first defense against defects and can have both

a positive and negative impact on the process. Determining

the correct frequency, wipe sequence, compatible

chemistries and materials will impact repeatability and

eliminate potential defects. Micro sized apertures require

more frequent wiping where a simple experiment can

determine the starting point, however, over-wiping with

solvent can have the same negative effect as under-wiping.

The test involves printing a board and then drive the vision

camera under the stencil to inspect the apertures for any

paste squeeze out or clogged apertures. Note, that the

apertures will contain some paste that was not released

based on transfer efficiency and stencil quality. In most

cases this paste will be pushed out on the next print

sequence and will not require a wipe, please judge

accordingly. Continue this process of inspection until

defects are starting to form. Subtract 1 board from the total

and this can be your starting frequency. If a Solder Paste

Inspection machine is available, then based on results, this

can be used to determine the correct frequency of wipes.

The recommended sequence is a vacuum/vacuum/dry. The

combined vacuum strokes eliminate any paste pulled from

the aperture and left behind that was the result of the first

pass vacuum. Solvent should be used less frequently as the

purpose of this material is to address the flux that can build

up around the aperture opening. Recommended frequency

for a solvent wipe is every 4-6 wipe cycles. The

recommended solvent stroke sequence is a

solvent/vacuum/dry where a solvent application always

begins the sequence. Consult your paste manufacturer for

recommended solvents to ensure that the solvent used is

compatible with the paste flux. A quality paper should be

used as Type 6 paste can be difficult to clean, where

economy paper can have issues with retaining the solder

balls and contamination issues could result. [1]

The Printer:

The printer plays the major role in the success of printing

008004” (0201mm) components. It’s recommended prior to

printing micro apertures to make sure all preventive

maintenance and calibrations are up to date. The alignment

capability of the machine is vital for dealing with small

pads. Advancements in machine vision repeatability and

accuracy has kept pace with the introduction of micro-

components. However, if your machine was designed back

in the 1990’s, then it most likely will not have the accuracy

resolution to handle these component challenges. Testing

the machines vision alignment capability prior to developing

the process is recommended so that with the machine

verified, issues with alignment can be isolated and solved

more quickly. This can be done using a print verification

process, using embedded machine software that measures

paste deposits for accuracy and repeatability, the results will

determine if a vision calibration is warranted. Another key

calibration on the printer that is often overlooked is the table

leveling to the stencil rails. Since gasketing is paramount

when printing micro apertures, this calibration takes into

consideration the four corners of the table as it applies to the

stencil rails for proper seating between the PCB and stencil.

This calibration is overlooked as it was most likely done

when the machine was built by the manufacturer and never

addressed again after installation. One of the issues with

doing this calibration was the difficulty with the procedure

used. A feeler gage is employed to measure the distance

from the table to the bottom of the adjacent stencil rail. In

order to measure the four points, the gage is moved from

corner to corner repeatability to dial the distance to within

specifications. This process requires the machine to be down

significantly often taking hours to complete. A new tool

developed by MPM addresses this issue by adjusting all

four corners simultaneously. To date, specifications for table

to stencil leveling has been in the range of +/- 0.004”

(0.1016mm). However, studies have shown best results are

achieved when the specification is dropped to +/- 0.001”

(0.0254mm). To eliminate any tolerance issues between the

table and the tooling plate, the plate can be used as a

reference during this calibration. This specification can be

achieved using this calibration tool and has played a

contributing role in successfully printing micro apertures.

The time to complete this calibration has been reduced to

roughly 1 hour. Lastly, the printing machine should be

completely inspected for cleanliness and clean any paste

debris found. Root cause for many issues can be traced to

random paste deposits or residual paste that builds up over

time.

DESIGN OF EXPERIMENT:

Print Test for 008004 (0201mm)

Overview:

To demonstrate micro-component printing capability, with a

focus on the 008004” (0201mm) component using the new

SMTA miniaturization test vehicle. Using best practices

described above, examine the results to determine Cp, CpK,

Pp and PpK results. The goal is to achieve a process

capability, Cp, that is equal to or greater than 2.0 as well as

a CpK greater than 2.0 that demonstrates that the process in

within Six Sigma quality levels. The Pp and PpK numbers

will be examined to see how well the process is centered

with a goal of equal to or greater than 1.667. The test will

use the Edison platform to perform the printing using a

standard configuration. A description of the DOE and

details on the machine, materials and process used as well as

an examination of the results is as follows:

Design of Experiment:

The test consisted of printing a total of 24 PCBs with the

first 4 PCBs to be used as kneed boards to normalize the

process and get the solder paste to a working viscosity. A

wipe will be performed after each print to eliminate any

noise from the data. The remaining 20 PCBs will be

inspected by a Parmi SPI machine with the corresponding

data analyzed through a on board SPC package. Process

data for volume and height for the 008004” (0201mm)

components will then be gathered and displayed and studied

to determine the process capability of printing 008004”

(0201mm) components. All equipment used was recertified

prior to this test being performed.

Materials:

• Printed Circuit Board (PCB): The PCB used for

this experiment is the new SMTA miniaturization

test vehicle. (Figure 2) The board dimensions are

8.0” (203mm) in X and 5.5” (139.7mm) in Y with

a thickness of 0.062” (1.57mm). There are

approximately 400 pads per board with 200

positioned at 0 degrees and 200 pads positioned at

90 degrees. The 008004” (0201mm) pads are

0.005” x 0.006” (0.127 x 0.1524mm) with an air

gap between pads of 0.0047” (1.1938mm) and a

component pitch of 0.00126” (0.032mm)