Solder Ball Inspection Criteria for Assembled Boards
A solder ball is a small sphere of alloy sitting on or beside a joint where no joint should be. It might be a stray particle shaken from the paste, a bead expelled from under a component, or a fragment thrown clear by the reflow profile, and the difference matters because the causes are different. Inspection criteria exist to separate the harmless from the risky, and a criterion written without knowing which mechanism produced the ball is hard to apply the same way twice.
What a Solder Ball Is and Where It Comes From
Solder balls are typically between 0.05 mm and 0.5 mm across, and they sit on the solder mask, on a trace edge or beside a pad. They come from three sources: paste spatter thrown from the deposit as the volatiles escape, particles that were already separate in the paste, and beads expelled from beneath a component during reflow.
The source decides how the ball behaves. A ball sintered to the mask or trapped under a coating is fixed in place, while a loose ball on a bare surface can move during handling and vibration, which is what makes the location of a ball as important as its size. A ball that sits within a fraction of a millimetre of a pad is a different proposition from one in an empty area of mask, even when the two are the same size, because the first can bridge during a later process step while the second cannot.
Solder Balls Versus Solder Beading
Solder beading is a specific case in which a ball forms under the body of a chip component, held between the part and the mask by the surface tension of the flux. The bead is often larger than a spatter ball and sits close to the joint, and it appears when too much paste is printed or when the component presses the paste sideways.
The distinction matters for the acceptance limit, because a bead near a termination is closer to the electrical function than a stray ball in a wide keep out area. Both are counted as solder ball defects on many drawings, but a well written criterion treats them separately.
Why Balls Form During Reflow
During the ramp the flux boils and the paste releases solvent, and the escaping gas can carry small droplets of alloy with it. A droplet that lands where the flux has already been consumed does not wet the surface, so it solidifies as a ball instead of merging into a joint.
The alloy itself is not the problem, the absence of flux at the landing point is. That is why balls cluster near a via or a large pad where a lot of gas is produced, and why a drier paste or a slower ramp often reduces their number.
The Role of the Stencil and the Paste
The stencil decides how much paste is printed long before the first board is inspected. Squeegee pressure that is too high spreads excess paste across the mask, and a slack frame or polished aperture walls change the way the deposit releases from the opening.
Paste properties follow close behind. A paste that has stood on the stencil too long has lost solvent and prints with a dry, ragged edge that readily produces spatter, while a low viscosity paste tends to slump during preheat and let the flux carry alloy outward.
Moisture, Flux and the Profile
The profile decides when the volatiles leave. A ramp that is too fast for the flux chemistry drives the solvent off in one violent step, and a soak that is too short leaves moisture in the paste until the alloy melts, which is the point at which the gas has nowhere to go but sideways.
Board and component moisture contribute in the same way. A part that has absorbed water releases it at reflow, and the steam plume behaves exactly like the flux vapour, carrying solder away from the joint instead of leaving it in place.
Inspection Methods and Their Limits
Visual inspection under magnification finds the balls that sit on top of the surface and the ones beside a visible joint. It cannot see under a component or beneath a shield can, and its repeatability depends on the lighting and on the discipline of the operator. The inspection station should therefore define the magnification, the illumination and the time allowed per board, because a criterion that is applied with a different lamp on each shift is not a criterion at all.
Automated optical inspection is faster and more repeatable, but it has to be programmed with the same judgement the criterion expects, since a system trained to flag everything above a size threshold produces a long list of acceptable conditions. The trade offs are set out in the guide to automated optical inspection.
Writing an Acceptance Limit
A workable limit states a maximum ball size, a maximum number in a given area, and a distance from the nearest conductive feature. A single size figure without the other two produces arguments at the inspection station, because the same ball is harmless in one place and unacceptable in another.
The limit should also say whether balls must be attached or removable without residue. A loose ball on a finished assembly is a different risk from one firmly sintered to the mask, and the criterion needs to express which of the two is acceptable. Where a customer drawing is silent on the point, the shop should propose a limit in writing and have it agreed before production, because an unstated limit is normally resolved in the least favourable way after a rejection.
Cleanliness and Rework
Contamination changes the picture, because a ball in a residue rich area is far more likely to be trapped than one on a clean surface. Residue left under a component can hold a ball against the body, so cleaning is part of the inspection decision rather than a separate subject.
Rework is the last chance to remove what inspection found. A bead under a chip part is best cleared by lifting the part, cleaning the area and refitting it, while a loose ball on the surface can usually be brushed away once the assembly is clean enough for the ball to release.
Reducing the Rate on the Line
Reduction begins with the paste: a fresh jar, correct storage and a documented open time remove a large share of the cases. The stencil follows, with a clean underside, correct tension and a support system that keeps the board flat against the aperture.
The profile is the third lever, and slowing the preheat ramp is the usual first adjustment. Tracking the defect rate over time closes the loop, because a rate that climbs after a change of paste or a new stencil points straight at the cause. Reject boards are worth keeping for a short period, since a ball that has been photographed with its location and its process conditions is far more useful in a review than a verbal description of the same defect. The wider family of these defects is described in solder defects and board failures.
FAQ
Is every solder ball a defect? No. Many criteria accept small, attached balls away from conductive features. What matters is the size, the number, the distance from a conductor, and whether the ball could move during the life of the product.
Can balls be removed after assembly? Loose balls can be brushed or washed away once the residue around them has been cleaned. Beads trapped under a component usually require the part to be removed and refitted, which is why preventing them is cheaper than correcting them.
Why did the rate rise after a paste change? A new paste brings a different solvent system and a different rheology, so the old profile may now boil it too quickly. Re-qualifying the profile with the new paste is the first step, followed by a review of the stencil cleaning interval.
