吃瓜tv

What is PCBA quality control?

PCBA quality control is the coordinated checking and control of materials, assembly operations and product verification for a populated circuit board. It connects the released BOM and build instructions to the actual assembly, inspection results, test records and disposition. Its purpose is to make conformity demonstrable for the specified product, not simply to collect inspection photographs.

The PCB Assembly service defines the manufacturing scope. This page explains how to plan the quality controls around that scope: incoming parts, process setup, soldering, verification, nonconforming work and traceability. The wider quality assurance overview remains the place for supplier-system and credential discussions.

Begin by agreeing what will be built and what evidence will support acceptance. A general instruction such as 鈥渢est all boards鈥� is incomplete unless it defines the tests, limits, revision and record. Testing every unit is a statement about population; it does not automatically mean every potential defect or operating condition is covered.

Put the control at the point where it can act

Quality controls are most useful when they detect or prevent the relevant problem before more work hides it. Incoming checks address identity and material condition. Paste and placement controls act before or around soldering. Optical, X-ray and electrical methods answer different questions after selected process steps. The exact route depends on the assembly and the requirements.

Do not copy another product鈥檚 route simply because both use SMT. Hidden joints, through-hole connectors, coating, programming and product-specific tests change the control plan. Define who approves changes to the route and how the first affected serial or lot is identified.

Control the BOM, variant and supplied materials

The assembly release must identify the intended variant and exact approved purchasing identities. An item marked DNP must remain distinguishable from a missing fitted part. Customer-supplied components need the same clear identification and condition review as assembler-purchased material; ownership alone does not establish suitability.

Use the assembly BOM guide to prepare manufacturer names, exact part numbers and fitted references. Confirm how alternate approvals are recorded. A supplier should not turn a shortage into an unapproved electrical or mechanical substitution, even when the suggested part has a similar headline rating.

Record relevant handling requirements and packaging history. The provide examples of part-specific package, carrier and moisture-sensitivity information. Actual handling must follow the component and process requirements, rather than a universal baking instruction copied into every build.

Verify the first build against a controlled release

Setup review links the board revision, BOM variant, placement program, stencil and process instructions. Check polarity and orientation using the intended component and drawing conventions. Verify that the reviewed files are the files actually released to the line. An attractive rendering is not evidence that a feeder, program or physical part matches the release.

The first-piece inspection procedure handles the initial build and authorization gate in detail. Repeat the relevant setup review when changes invalidate the previous evidence. A component substitution, program edit or significant tooling change may need a targeted review even when the board part number remains unchanged.

Keep the reaction practical: identify the issue, affected work, responsible owner and authorization needed to continue. First-piece approval is not a lifetime certificate for all later assemblies. Continued process control and product checks remain necessary as materials, equipment condition and operating circumstances change.

Combine inspection and test without double counting coverage

SPI evaluates deposited paste characteristics before placement. AOI evaluates programmed visible features. X-ray can help examine selected hidden geometry. ICT and other electrical methods evaluate accessible circuit characteristics; FCT exercises defined operation. These methods overlap in some observations, but they are not interchangeable guarantees.

Choose the method from the requirement and physical access. A hidden interface cannot become visible because an optical program has more inspection points. A correct boot message does not establish every analog output or solder joint. The test-method comparison helps organize complementary coverage without repeating each specialist article here.

Treat a machine call, confirmed defect and escaped defect as different records. False-call review consumes time but is not itself a defective unit. A later-discovered issue may reveal a coverage or process problem; it should not be removed from the history just because the board eventually passes a retest.

Read first-pass yield without hiding repeat work

This calculator uses one identified cohort moving through three agreed checkpoints. Each successive count is the number that has passed every checkpoint so far on its first attempt, without repair or repeat processing. These nested counts let the tool distinguish cumulative first-pass yield from the final conforming count after the agreed disposition process.

Do not enter unrelated daily totals, sample-inspection counts or the number of times a machine ran. A board tested three times is still one unit. If the inspection route or cohort changes, separate the populations instead of forcing them into one apparently precise percentage. The result explains arithmetic; it does not establish a benchmark or authorize shipment.

Example: 90% first-pass and 97.5% final conformity

In an illustrative cohort of 200 assemblies, 190 pass the first checkpoint without repeat work. Of that first-pass group, 185 also pass the second checkpoint on the first attempt, and 180 pass the third. The cumulative first-pass yield is 180 divided by 200, or 90%. The checkpoint ratios use 200, 190 and 185 as their respective denominators.

If 195 of the original units are finally conforming after the agreed review and any authorized rework, final conformity is 97.5%. That number does not rewrite the first-pass result. Fifteen additional units have reached conformity beyond the untouched first-pass group; investigate the actual records before attributing all of them to one process or repair type.

Five units remain outside the final conforming count. They may be on hold, rejected or awaiting evidence; the yield calculation cannot decide which. Keep counts aligned with physical disposition and do not use 鈥測ield loss鈥� as a substitute for the actual status of each unit.

Separate workmanship criteria from functional requirements

IPC-A-610 addresses assembly acceptability, while J-STD-001 addresses requirements for soldered assemblies. Use the agreed document revisions, class and applicable customer requirements. The provide a scope reference; actual acceptance needs the applicable criteria and competent interpretation.

The IPC assembly standards guide helps turn a general standards request into a usable agreement. A workmanship class does not define a product鈥檚 firmware, voltage limits, communication behavior or complete functional test sequence. Those requirements need the product-specific verification plan.

Do not infer medical, automotive or aerospace conformity from an IPC reference alone. Additional product, customer and sector requirements must be identified for the project. Keep evidence requests specific so a supplier can confirm the applicable site, process scope and records instead of answering with a generic quality slogan.

Preserve failures, rework and retest history

When a check fails, separate the affected units and record the observed condition before attempting correction. Determine whether the cause could affect other boards from the same setup, material lot or program revision. Containment is about controlling the affected population, not only repairing the unit that exposed the problem.

Rework instructions should identify the authorized action and any limits or additional evidence required. Afterward, repeat the relevant inspection and tests; passing one repeated measurement may not address every consequence of the work. Preserve the original failure, the intervention and the subsequent result as related records.

For recurring problems, distinguish immediate containment from corrective action and verification of effectiveness. Replacing a component can repair a board without explaining why the wrong component was fitted. Customer communication should state the scope and proposed disposition clearly, especially when a deviation from the released requirement needs approval.

Make the quality evidence retrievable

Agree the traceability level before production: order or lot identity may be sufficient for one project, while another needs serial-level test and material relationships. The PCBA traceability guide explains the record structure and retention discussion. A barcode has value only if it resolves to the required evidence.

For functional operation, continue to the FCT execution guide for fixture, sequence, software version and per-unit records. Before shipment, use the final inspection and release guide to connect product status, quantity, documentation and packaging. This creates a handoff between process control and release instead of another disconnected checklist.

PCBA quality control questions