How can inspection data improve electronics manufacturing yield?

Every assembled board leaves a trail of information. Electronics inspection data equipment measures solder paste, cameras flag potential assembly faults, and electrical tests record whether the finished circuit behaves as expected. When a board needs attention, operators add another part of the picture.

The volume can be substantial. A 2023 study used production data from 15,387 PCBs, covering almost six million component pins in just nine days. Researchers linked solder-paste measurements with later optical inspection findings and operator assessments.

Electronics inspection data becomes useful when those records help explain what happened and guide the next manufacturing decision. That might mean correcting a printing issue, investigating a recurring test failure or reducing unnecessary manual reviews.

Why inspection should explain defects, not just reject boards

A failed inspection identifies something that needs checking. It does not, by itself, explain why the result occurred or what should change.

Automated optical inspection, usually shortened to AOI, uses cameras to assess visible features such as component position and solder joints. It helps flag potential faults consistently, but an inspection system can also flag acceptable variation as a problem.

In the same 2023 study, the authors reported that almost 20% of components labelled defective by AOI were subsequently found not to be defective under microscope inspection. The dataset covered one PCB design, so this illustrates the issue without establishing an industry-wide false-call rate.

These false calls take time to investigate and can make an inspection pass rate look worse than the underlying manufacturing quality. Genuine defects need a different response: their causes must be understood so that the next batch does not repeat them.

Recording the confirmed outcome makes that distinction possible. Was the board acceptable without intervention? Did it need repair? What was changed, and did it then pass testing? A generic “checked” entry leaves too much unexplained.

Those decisions should feed back into process reviews. Repeated false calls may justify checking the inspection programme against agreed acceptance criteria. Repeated genuine faults may point towards printing, placement, soldering or design changes. Any adjustment to inspection limits should be checked against known defects so that reducing unnecessary reviews does not weaken detection.

Connecting SPI, AOI, test logs and reflow electronics inspection data

Each stage answers a different question.

Solder-paste inspection, or SPI, measures the paste deposited before components are placed. Information about its volume, height and position can show whether printing is drifting or whether particular pads are receiving inconsistent deposits.

AOI after soldering shows the visible assembly result. Electrical and functional tests add evidence about circuit behaviour, including faults that cannot be confirmed from an image. Keeping the failed test step and measured value is more useful for investigation than retaining a pass/fail result alone.

Reflow data adds the heating conditions used to form the solder joints. Here, oven settings and a measured board profile provide different information. A populated board fitted with temperature sensors shows how representative locations actually heat and cool as they travel through the oven. AIM Solder recommends this approach and retaining profiles to support later troubleshooting.

The profile needs to be associated with the relevant product, oven setup and production run. It should also be clear whether it came from a representative profiling board or from monitoring individual assemblies.

To connect these records, engineers need a consistent way to identify the board. That could be a serial number or a panel ID combined with the board’s position within the panel. Product revision, production time and component reference then help narrow the investigation further.

Without those links, an engineer may know that several boards failed at a particular joint but struggle to find the corresponding paste measurements, assembly conditions or repair notes.

The information can also influence production directly. An example presented in iNEMI’s 2021 PCB assembly project used SPI measurements to adjust component placement according to the actual solder-paste position. It reported yield improvements of up to 2%, depending on printing quality. This was a specific application of connected equipment, with results dependent on the process conditions.

The same principle applies to an engineering review: a measurement becomes more valuable when it informs a controlled change. Teams can start by linking a recurring defect to the records they already collect, then investigate which part of the process needs attention.

Using defect patterns to improve first-pass yield over time

First-pass yield measures the proportion of units that pass a defined process first time, without repair, rework or retesting. A board that eventually passes after repair still represents a first-pass failure.

Define the process being measured and keep that definition consistent. An AOI pass rate describes one inspection stage; it does not establish that every board has passed all the checks needed for a finished product. Track false calls separately from confirmed manufacturing defects so that improvements in inspection efficiency and product quality remain understandable.

Start with a recurring problem that has a meaningful effect on production. For example, repeated insufficient-solder findings at the same component might lead engineers to compare SPI measurements from affected boards with measurements from boards that passed.

If the affected boards consistently received less paste, that gives the team a reason to investigate stencil openings, board support or the printing process. If the paste measurements look similar, placement, heating conditions and other causes remain worth examining. The pattern directs the investigation; it does not prove the cause on its own.

Trial a defined correction, record what changed and check the next comparable builds. Compare the same product revision and inspection criteria where possible, because a change in product mix can make an overall yield figure misleading.

Small improvements can have a useful effect. As a hypothetical example, increasing first-pass yield from 96% to 98% across 10,000 boards means 200 more boards completing the defined process successfully first time. The operational benefit will depend on how much investigation, repair or other handling those failures would otherwise have required.

Review first-pass yield alongside rework time, scrap, false calls and failures found at later tests. Together, these measures help show whether a change has improved the process or simply moved the problem elsewhere.

For TAD, connecting design, prototyping and manufacturing creates an opportunity to carry that learning into future builds. Recurring assembly findings can inform PCB layout decisions, while production experience can improve test planning and manufacturing documentation. Using electronics inspection data in this way helps turn individual checks into practical improvements that last beyond one batch.

Click here to get in touch or here to read more.

‘Engineering Design, Imagine what could exist’

Got a web design question or mobile application need? Our in-house design agency, Bluebrick Studios, has you covered. Check out their site to find out how they can help you achieve your mission.

FAQs

What is electronics inspection data?

Electronics inspection data includes measurements, images and classifications collected while checking PCB assemblies. Examples include solder-paste measurements and optical inspection results. Linking them with electrical test records, operator decisions and repair outcomes helps engineers understand recurring defects.

How does AOI improve electronics manufacturing yield?

AOI identifies visible assembly problems so they can be investigated and corrected. Its findings support yield improvements when confirmed defect patterns lead to changes in printing, placement, soldering or design. Recording operator decisions also helps identify repeated false calls.

What causes low first-pass yield in PCB assembly?

Causes can include inconsistent solder-paste printing, placement errors, unsuitable reflow conditions, component problems and designs that are difficult to assemble. Inspection false calls and test-fixture issues can also lower reported pass rates, so confirmed defects should be separated from measurement problems.