The root cause of Solder Paste printing defects is often attributed to stencil or squeegee parameters, but our repeated verification on the production line found that when the sphericity of solder powder is lower than 90%, even if the stencil aperture accuracy is controlled within ±2μm, the squeegee pressure is stable at 4.5kg, and the speed is maintained at 40mm/s, local missing printing and edge burrs still occur frequently. This is because non-spherical solder powder cannot form uniform rolling when passing through the stencil apertures, and some sharp-angled particles get stuck at the micro-cracks on the aperture walls, causing the effective filling volume to fluctuate by more than 18%.

What is more troublesome is that solder powder with poor sphericity shows obvious lag during the demolding stage. Due to the tight interlocking between particles and uneven surface energy distribution, the solder paste is prone to tearing rather than overall peeling when detaching from the stencil apertures. We used a high-speed camera to record that the demolding time was extended by 0.15 seconds, which directly led to a 37% increase in the edge collapse rate of the solder paste. This is particularly obvious on 0.3mm pitch QFN devices, where the probability of bridging between adjacent pads increases to one board out of every ten.

In the actual production line, it was also observed that the Solder Paste Viscosity exhibits abnormal bimodal characteristics, that is, the apparent viscosity is low in the initial stage after stirring but increases sharply by 22% after 30 minutes. This stems from the unstable orientation arrangement of non-spherical solder powder in the Flux, which easily forms local cluster structures when standing still, resulting in smooth paste output at the front end but insufficient supply at the rear end when the squeegee pushes across the stencil. The final manifestation is that the front end of the printed pattern is full, while the tail end becomes thinner or even intermittent.

We once compared two batches of solder paste of the same model: batch A had a solder powder sphericity of 93.5%, while batch B only had 86.2%. Under the same SPI inspection conditions, the standard deviation of volume deviation for batch B was as high as 12.8%, nearly double the allowable upper limit of IPC-7527. Moreover, after continuous printing of 200 pieces, the residual amount at the bottom of the stencil in batch B was 0.7mg/cm² more than that in batch A, indicating that non-spherical particles aggravate the mechanical wear on the stencil aperture walls and accelerate aperture distortion.

During on-site debugging, if the hidden indicator of solder powder sphericity is ignored, simply adjusting the squeegee angle or demolding speed often yields little effect. This is because the fundamental contradiction lies in that the rheological behavior of the solder paste body has deviated from the design expectation. At this time, even if the printing speed is reduced from 45mm/s to 30mm/s, it cannot compensate for the forming out-of-control caused by particle-level motion instability, and may instead amplify the tailing phenomenon.