Solder-paste printing cell with the stencil clamped in frame and the squeegee resting at its edge as boards feed in

ENGINEERING NOTE · SOLDER PASTE PRINTING

Most SMT defects start at the printer

Industry-published estimates put 60–70% of SMT assembly defects at the printing step; placement and reflow usually take the blame. Reflow can pull an offset part back into place, but it cannot add paste that was never deposited. Here are the aperture criteria, four defect mechanisms, the SPI loop and the first two knobs to turn.

The process engineering team · 11 September 2026

THE GOVERNING NUMBERS

Area ratio decides it

Two ratios decide whether a stencil aperture will hand its paste to the pad, and both come out of the same geometry. IPC-7525B is the stencil design standard we check against; the values below are its criteria, written as the limits we design apertures to.

Stencil design criteria and what each one governs
Criterion Value What it governs
Area ratioopening area ÷ aperture wall area ≥ 0.66 Paste release. Below this limit the deposit is pulled apart as the stencil separates: part of it stays on the aperture wall and the pad is left short.
Aspect ratioaperture width ÷ stencil thickness ≥ 1.5 Release on narrow apertures. A 0.20 mm wide opening needs a stencil no thicker than roughly 0.13 mm to stay inside the limit.
Area ratio below about 0.50 print yield unpredictable Outside the range where a stable print window exists. Settings that work on one panel stop working on the next, so we re-cut the stencil instead of chasing the printer.
Stencil thicknessdeposited volume 0.120 mm → 0.100 mm Thinning the stencil raises the area ratio of every aperture on it and takes paste volume out of all of them at the same time. The two effects never move in the same direction.
Stencil constructionfine pitch and micro chips laser-cut stainless, nano-coated 01005 chips and 0.4 mm pitch devices are printed with laser-cut stainless stencils carrying a nano coating, which lowers friction between the paste and the aperture wall.

0.4 mm-pitch QFN example published in the industry: 0.25 × 0.60 mm pad on a 0.120 mm stencil gives an area ratio of 0.735; dropping the stencil to 0.100 mm raises it to about 0.882 but removes roughly 17% of the deposited paste volume. Industry illustration, not a specification.

FOUR DEFECTS AND WHAT CAUSES THEM

Name the defect before you change a setting

Each of these four is produced at the printer and only fixed in place by reflow. They are separated here by mechanism, by where they appear first, and by the one adjustment that addresses the cause rather than the symptom.

Defect 01 · release

Insufficient paste

Mechanism: the aperture holds the paste instead of handing it over. Area ratio is the usual cause — wall area is large against the opening, so paste touching the wall stays behind as the stencil lifts — and fast separation makes it worse. It appears first on the smallest apertures: 01005 sites, 0.4 mm pitch pads and the pattern around a QFN thermal pad. First response: read the printed volume on SPI before touching the placement program, then check the area ratio of the aperture that came out low.

Defect 02 · volume

Bridging

Mechanism: deposited volume is larger than the space between two adjacent pads, so the two deposits meet — at printing, or later, when the paste slumps before reflow. Apertures cut too wide, a solder-mask web too thin to hold the paste back, or a stencil too thick for the pitch all produce it. It appears on 0.4 mm pitch leads, fine-pitch connectors and anywhere the aperture was sized from the pad rather than from the gap. First response: reduce aperture width, not stencil thickness, and inspect the mask web between the pads.

Defect 03 · rheology

Slumping

Mechanism: the print is correct on the stencil and loses its shape afterwards. Viscosity falls when the paste is warm, when it is past its work life on the stencil, or when the room is humid; a slow squeegee stroke shears the paste and warms it further. It shows as deposits that have spread or rounded off between printing and reflow — on SPI, area up and height down with volume still inside the window. First response: check paste temperature and the work-life clock before changing any machine parameter.

Defect 04 · transfer

Solder balling

Mechanism: paste ends up where it was never intended to be. It is pressed onto the solder mask through a gasketing gap, pushed under the stencil when the board is not supported flat, or left on the stencil underside and carried to the next board. Those fragments ball up in reflow and sit beside the joint rather than in it. It clusters at board edges, at mask steps and around large apertures. First response: check board support and the underside wipe interval before adjusting the print program.

WHAT SPI ACTUALLY MEASURES

Close the loop back to the printer

A print you cannot measure is a print you can only argue about. Our lines run an automatic printer with 3D SPI behind it, and the inspection is wired back into the printer instead of into a report that is read the following week.

Volume
mm³ per aperture
Height
µm per deposit
Area
% of pad area
Offset
mm from pad centre

Volume is the primary accept criterion, because it is the paste that decides fillet size. Height catches over-print, stencil wear and paste pushed down into the aperture. Area shows spreading, missing paste and mask contamination. Offset separates a print error from a placement error before anybody edits the placement program.

Limits are set per project. No single industry limit survives a change of pitch, stencil thickness and paste chemistry at the same time, so volume, height and area limits are written into the inspection program for the board being built and agreed with the customer at first article. The loop then closes on the printer: a volume trend on one aperture becomes a change in squeegee pressure, print speed, separation speed or wipe interval on the next run, not a note in a quality file.

ONE PANEL, FOUR APERTURE DESIGNS

What the panel showed

One panel, four groups of identical boards, four aperture designs, one paste and one machine setting. Everything was printed in a single run so that the stencil was the only variable that changed.

Design 1 — aperture matched to the pad, 0.120 mm stencil
1:1 baseline
Design 2 — aperture width reduced to lift the area ratio
area-reduced
Design 3 — stencil thinned locally at the fine-pitch device
stepped
Design 4 — one large aperture divided into windows
windowed

WHAT SPI READ, IN DIRECTION

Where each design landed

Readings are given as direction and window status rather than as numbers lifted from one board, because the useful result of the panel is which way each design moves the deposit.

Design 1, volume on 0.4 mm pitch pads
below window
Design 1, bridging at the fine-pitch device
present
Design 2, volume
inside window
Design 2, spread across the panel
wider
Design 3, volume
inside window
Design 3, spread across the panel
narrowest
Design 4, volume on the thermal pad
inside window
Design 4, print time per panel
longer

The trade the panel exposes is the one in the criteria table: thickness buys area ratio and costs volume. Design 3 lifted the area ratio of the fine-pitch apertures by thinning the stencil there, and gave up paste volume across those same apertures — in the published example, about 17% of it. Design 2 kept the volume by taking area out of the aperture instead. Neither is a general answer: the design that holds the print window is chosen per pad geometry, and the stencil is the cheapest place on the line to fix it.

THE FIRST TWO KNOBS

Pressure and speed, in that order

When a print goes out of window, these six steps are the order we work in. The first two carry most of the effect; the rest only make sense once the first two are settled.

Squeegee pressure

Change pressure first, in small increments. Too little leaves paste on the stencil; too much scoops paste out of wide apertures and drives it under the stencil. SPI volume and the state of the stencil underside tell you which side of the window you are on.

Print speed

Change speed second. A slower stroke fills small apertures more completely; a faster stroke shortens the time the paste spends being sheared and limits the temperature rise it picks up from the blade.

Separation speed

Only after the first two. Slower, controlled separation gives the paste time to transfer off the aperture wall, which is exactly where a low area ratio aperture loses its volume.

Support and wiping

Rigid support under the board protects gasketing, and the dry and solvent wipe interval is set from the paste and the aperture count rather than from habit. This is the step that clears solder balling at board edges.

Re-measure

Every change is re-measured with SPI at the same panel positions. Without a before-and-after reading on the same apertures, a change is a preference rather than a correction.

Freeze the program

Freeze the parameters per product and write them into the print program, so the next run starts where this one finished instead of starting from the machine default.

One variable at a time. Two adjustments made in the same run cannot be attributed to either cause, and the next build then starts from a setting nobody can explain.

RELATED CAPABILITIES

SMT assembly

SMT assembly

Seven high-speed SMT lines placing 01005 (03015 metric) parts up to 100 × 90 mm, with 0.4 mm pitch QFP and BGA placement and laser-cut nano-coated stencils.
Testing & inspection

Testing and inspection

3D SPI reads volume, height, area and offset inside the print loop, AOI follows reflow, and X-Ray inspection is configured per project for hidden joints.
Design & DFM

Design and DFM guides

Process edge of 5 mm or more, fiducials, land patterns and stencil aperture requests — with a first-pass DFM report inside 24 hours on prototype boards.
Softly lit out-of-focus circuit board surface with faint copper traces and via pads

BRING US THE BOARD

Send the panel and the print parameters

Tell us the pitch, the stencil thickness and the defect you are seeing. We read the print data against the aperture table and come back with the geometry and parameter changes that fit your board.

Reply within one business day