PCB Flying Probe Checklist: What to Check

A flying probe tester moves a few slender probes across a bare or assembled board. It touches pads and test points one at a time, so it needs no fixture and no mask. That flexibility makes it a strong fit for prototypes, low volume runs and boards with frequent changes. The checklist below covers what a buyer or engineer should confirm before accepting test results from a supplier. Each item connects to a decision made earlier in the build.

Flying probe testing is not a single step. It is a program, a set of fixtures, a board design and a reporting routine that must agree with each other. If the design leaves no probe access, the best tester cannot help. If the program skips a net, a passing report means little. The checklist therefore starts with the data sent to the factory and ends with the records returned to you. That order keeps the review honest.

Design Data and Probe Access

Start with the Gerber files, the netlist and the bill of materials. The tester program is generated from this data, so any gap in the files becomes a gap in coverage. Check that every net you care about has a named test point. Confirm that pad sizes meet the probe’s minimum contact area. Confirm that no tall component blocks the path of a probe. These points are reviewed during a proper PCB assembly data review, before any tooling is built.

Access matters more than most buyers expect. Probes approach from above, and some machines also test from below. A test point hidden under a shield or a connector may be unreachable. Ask the factory to list the nets it cannot touch and explain why. If the list is long, consider adding test points in the next layout revision. A small change in copper can turn an untestable board into a fully covered one, and the cost is usually trivial.

Program Setup and Verification

The test program defines which nets are checked, what limits apply and how the machine reports a failure. Ask for the program revision that will run on your order. Confirm that it matches the current netlist, not an older version. A mismatch here is one of the most common causes of false failures and missed defects. The factory should show that the program was validated on a known good board before the production lot started.

Limits deserve a close look as well. Resistance and capacitance thresholds should reflect the design, not a default template. A loose limit passes boards that should fail. A tight limit rejects good boards and wastes time on rechecks. Ask how the limits were set and whether they were confirmed against measured values from a sample. That answer tells you whether the program was tuned or simply copied.

Fixture and Contact Checks

Although flying probe testing needs no dedicated fixture, it still depends on stable contact. Probe tips wear, and a worn tip gives noisy readings. The machine should have a schedule for tip inspection and replacement. Ask how often probes are cleaned and how contact force is verified. A simple continuity check on a reference coupon before each lot is a quick way to catch a bad probe before it affects real boards.

Board support is the other half of contact quality. A thin board can flex under probe pressure, which shifts the touch point and distorts readings. The tester needs adequate backing pins or a vacuum plate. Confirm that the support plan suits your board thickness and panel size. For thin or large boards, ask for a first article check that includes contact repeatability. This is a detail that rarely appears in quotations but often explains mysterious failures.

Coverage and Test Limits

Coverage is the share of nets and components the program actually verifies. Ask for a coverage report, not a verbal assurance. The report should list tested nets, untested nets and the reason for each gap. Components such as connectors, shields and large capacitors may be partly tested or skipped. Knowing this in advance prevents a false sense of security. It also helps you decide whether another test method should fill the gap.

Flying probe testing checks opens, shorts and component values on an assembled board. It does not prove that every function works under power. For many products, a functional check or a system test is still needed. A sound PCBA testing plan combines probe testing with functional coverage, so structural faults and performance faults are both caught. Ask the factory how the two methods divide the work on your board.

Handling of Failing Boards

Every test plan needs a clear rule for failures. A board that fails should be marked, logged and routed to a defined repair station. The repair record should show the failing net, the measured value and the action taken. Without that record, the same fault can repeat across a batch with no one noticing the pattern. Ask to see a sample repair log and confirm that it links back to the test result.

Rechecks matter too. A board that passes on the second attempt is not the same as a board that passed first time. Ask whether retests are counted and how they are reported. A high retest rate usually points to contact problems, marginal limits or handling damage. None of these is a board defect, but all of them cost time and can hide real issues. The factory should track retests as a process metric.

Records and Traceability

The test report is the product you keep after the boards ship. It should identify the board, the lot, the program revision and the date of test. It should show pass and fail counts and list any deviations. A report that only says passed is not enough for a buyer who needs evidence later. Strong quality management practice treats the test record as part of the batch file, not as an optional extra.

Traceability should reach back to materials as well. If a batch of components changes, the test data can reveal a shift in measured values. That link between material records and test results is what makes a field return investigation fast. Ask how long records are kept and in what form. Digital records that can be searched by lot number are far more useful than paper sheets stored in a folder.

Prototype and Low Volume Use

Flying probe testing earns its place on small lots. A fixture for in-circuit test costs money and time, and it must be rebuilt whenever the design changes. A probe program can be updated in hours. For a new design, that speed lets you test early samples and find layout problems before tooling is committed. Teams using rapid PCBA prototyping should ask for probe testing on the first batch, not only on later runs.

Early test data is valuable beyond pass or fail. It shows which nets are marginal, which pads are hard to contact and which components vary most. That feedback improves the next layout and the next program. Ask the factory to share a short summary of findings with the first article. A supplier that treats the first run as a learning step will usually deliver a smoother volume ramp than one that simply reports a number.

Supplier Questions Before Release

Before you approve a lot, ask five things. Which program revision will run? What coverage does it reach, and what is left untested? How are probe tips and board support controlled? How are failures and retests recorded? What does the final report contain? The answers should be specific and written. A supplier that cannot answer them clearly is probably not controlling the step, even if the machine is modern.

Use the answers to compare suppliers on equal terms. A low price with thin coverage is not a bargain, because defects reach the field and cost far more later. A clear program, a documented coverage report and a searchable test record are worth paying for. For buyers moving into volume, the same discipline supports SMT PCB assembly and keeps test data comparable across lots. The checklist is short, but it protects the whole build.