At 3 a.m. yesterday, the production line suddenly triggered an alarm: AOI continuously intercepted 23 boards, and the tombstoning rate of 0201 resistors soared to 18%. When we pulled up the SPI data, we found that the Solder Paste thickness standard deviation across the entire board jumped directly from the normal 12 μm to 29 μm. The average thickness in the lower-left area was only 58 μm, while it reached as high as 92 μm in the upper-right area. This kind of regional uneven thickness cannot be fixed by minor parameter adjustments; we must immediately stop the line to inspect the physical condition.

We first used a feeler gauge to measure the stencil snap-off height and found that the deviation at the four corners reached 0.15 mm. Then we removed the stencil and measured it with a tension meter: the tension in the center area was only 32 N/cm², and the edge had fallen below the process baseline of 28 N/cm². This stencil had been in continuous use for 67 days and printed 142,000 times, far exceeding the mandatory replacement threshold of 30 days or 100,000 times. What was worse, the squeegee had just been replaced with a new blade that day, but the pressure sensor showed that the actual output was 12% lower than the set value. It turned out that there was a slight leak in the internal diaphragm of the pneumatic valve, causing attenuation in air pressure transmission.

During emergency handling, we did not touch any parameters. Instead, we replaced the original carrier with a thickened aluminum-based carrier with vacuum suction, quickly wiped the residual solder paste from the bottom of the stencil with alcohol wipes, and dried it with a nitrogen gun. Then we manually reprinted 5 boards in front of the SPI machine for thickness verification. We only released small-batch trial production after confirming that the thinnest point recovered to 71 μm. This is much faster than blindly increasing squeegee pressure or reducing release speed, because the latter would only exacerbate stencil deformation and solder paste shear failure.

The root-cause solution must be a combined approach. Stencil lifespan will no longer be controlled by days, but tied to the thickness CPK value from SPI per shift. When CPK is below 1.33 for two consecutive shifts, the scrapping process will be automatically triggered. Squeegee pressure must be measured and calibrated with a digital pressure gauge every 4 hours, and flatness detection patches with 0.02 mm precision will be installed on all squeegee mounting surfaces. Carriers will be fully phased out of ABS material, uniformly replaced with 6061-T6 aluminum alloy, and 4-point vacuum negative pressure monitoring points will be added to ensure that the deformation of the PCB during printing is controlled within ±5 μm.

The key to prevention lies in front-loading anomaly detection. The daily first article inspection will not only measure the average thickness, but also capture the standard deviation trend chart of 256 grid points across the entire board from SPI. Once the single-day fluctuation exceeds 15%, stencil tension retesting will be immediately initiated. All new stencils must complete three cycles of tension decay testing before being put into storage, and the difference between the initial tension and the residual tension after 24 hours must be recorded after each test. Any stencil with a difference exceeding 3 N/cm² will be rejected. Solder paste printing is never a problem of a single process; it is the combined result of the stencil, squeegee, carrier, solder paste rheological properties, and equipment dynamic response. A tiny drift in any link will be exponentially amplified at the moment of release.

Last week, the neighboring line switched to a new supplier's solder paste, with a nominal viscosity of 720 Pa·s but a measured batch-to-batch fluctuation of ±85 Pa·s. As a result, under the same parameters, the printing thickness range expanded from 18 μm to 33 μm. Therefore, we now require that all incoming solder paste must be accompanied by the rheological curve of the current batch, and must be left to stand in a constant temperature and humidity room for 24 hours before being opened for use, to avoid uneven internal stress release of the solder paste caused by temperature differences. These details seem trivial, but they are precisely the hidden drivers most often overlooked in SMT printing defects.