X-Ray Void Measurement: A Complete Inspection Guide

X-ray void measurement is the practice of imaging a solder joint from above, counting the dark areas inside it and reporting the result as a percentage of joint area. It is the only practical way to see inside a ball grid array joint without destroying it, and it is the check that most often decides whether a batch of area array assemblies is released.

This guide covers how void percentage is calculated, how to set the grayscale threshold, the void criteria that apply to different joint types, and the process changes that reduce voiding when the measurement fails.

<img src="https://www.gopcba.com/wp-content/uploads/2026/08/Low-Volume-PCB-Assembly-Manufacturer-in-china-1024×480-1.webp" alt="X-ray void measurement on a BGA solder joint array” />

What X-Ray Void Measurement Shows

The image shows density, not solder. Solder absorbs the beam strongly and voids absorb it weakly, so a void appears as a bright region inside a dark joint. Everything else in the field, including copper planes, thermal vias and the package substrate, also absorbs, which is why the interpretation depends on the surrounding structure.

A two dimensional image also integrates through the whole joint. A void near the top of a joint and a void near the bottom can produce the same shadow, so the reported area is a projection rather than a slice. Where the through-thickness position matters, a computed tomography scan is needed. An X-ray void figure that cannot be repeated at a different magnification is a statement about the imaging setup rather than about the joint.

How Void Percentage Is Calculated

Void percentage is the total void area inside the joint divided by the joint area, expressed as a percentage. The calculation sounds simple, and the difficulty is in the two boundaries: deciding where the joint ends and deciding what counts as a void rather than as an edge artefact.

The joint boundary should be drawn from a reference slice of a known good joint and applied to the batch. In an X-ray void report the software often auto-detects the boundary, and the result should then be checked on a few joints by hand, because a boundary that is a pixel or two wide on each side changes the percentage more than most process changes do.

Setting the Grayscale Threshold

The threshold is the brightness above which a pixel is counted as void. Set it too low and the joint edges, the flux residue and the substrate texture are counted as voids. Set it too high and small voids disappear. Either error moves the X-ray void figure by several percentage points, which is more than most process changes achieve.

The threshold should be fixed for a given joint type and image setup, then verified on a reference sample that contains a known feature. Our notes on paste and joint inspection describe how a reference sample is prepared and used.

BGA Solder Joint Void Criteria

Criteria are usually written per joint type rather than per assembly. A thermal pad under a power device may be allowed a large total void area with no single void above a set fraction, while a signal ball is held to a much tighter figure. The distinction exists because the two joints fail for different reasons.

Where a customer standard applies, it takes precedence and should be quoted in the inspection instruction with its revision number. Where no standard applies, the shop should set its own X-ray void limit from the void percentage at which thermal or mechanical testing showed a change, and record the basis in the inspection instruction.

Sample Size and Which Joints to Measure

Voiding is not uniform across a package. The joints at the centre of a large body see a different thermal history from those at the corners, and the voids follow that pattern. Measuring a handful of joints in one region gives a misleading picture of the assembly.

A practical sampling plan covers the corners, the centre and one whole row or column, on several packages from each batch. Where a batch is rejected, the failed joints should be mapped on a drawing of the package so that the pattern can be compared with the oven profile.

Measurement Error and Repeatability

An X-ray void figure should be reproducible: the same joint measured twice by two operators should give the same number. Where it does not, the threshold, the magnification or the joint boundary is not fixed. A repeatability check on ten joints, measured twice by each of two operators, is the cheapest way to find that problem.

Magnification is the most common cause of disagreement. A joint that is imaged at higher magnification shows more small voids, so the figure rises without any change in the joint. The magnification should be written into the instruction and held for the life of the product.

Reducing Void at the Process

Voiding responds to the reflow profile more than to any other single variable. A soak long enough to let the flux volatiles escape before the alloy seals the joint removes the largest voids, and a slower ramp into the peak reduces the ones that remain. Our reflow profile guidance covers the soak settings that matter most.

Paste choice comes next. A paste with a more active flux and a finer powder wets faster and traps less gas, and a stencil aperture that matches the pad keeps the deposit height even. Where the voids sit over thermal vias, the via structure and its plating should also be reviewed, as described in our notes on thermal via design.

Reporting and Records

An X-ray void report should carry the joint type, the criteria used, the magnification, the threshold, the sample size and the highest single void as well as the total. A single percentage figure without those details cannot be compared with a later batch or with a customer’s own measurement.

Where a batch is accepted at the limit, the report should say so explicitly, because a trend toward the limit is an early warning. The general acceptance approach for these inspections is described by IPC, and the shop’s own limits should be traceable to its own test data.

Verification and Records

The first article confirms that the setup matches the intent, and it is the cheapest point at which a wrong setup can still be corrected. The cost of verification is small compared with the cost of a field failure, and it is paid at a point where the product can still be corrected.

Documentation exists so that a person who was not present can reproduce the work and reach the same conclusion.

FAQ

Is a smaller X-ray void always better? Not always. What matters is whether the remaining joint area carries the load and the heat. A joint with a moderate void level that is well wetted usually outperforms one with fewer voids but poor wetting at the interface.

Can X-ray void be measured on a two dimensional image only? Yes, and that is the normal production method. Computed tomography is used for problem solving and for qualification, because it takes far longer per joint.

Does an X-ray void cause a field failure? A void that reduces the load-bearing area or blocks a thermal path can. Voids that sit inside a well-wetted joint with sufficient remaining area are usually benign.