Test Point Contamination and Contact Resistance in ICT

Test point contamination raises the resistance between the probe and the pad, and the in-circuit tester reads that added resistance as part of the circuit it is measuring. A pad that should show 5 milliohms of contact resistance can show 500 milliohms once a film of flux residue and oxide has built up on its surface.

The result is not a measurement error that can be averaged out. It is a shift in every reading taken through that point, which turns good boards into failures and, more dangerously, hides marginal joints that would otherwise be caught.

Why Contact Resistance Matters

An in-circuit test measures small analog values: a 10 milliohm resistor, a diode drop, a continuity path through a connector. The resistance budget for the measurement includes the fixture wiring, the probe, the pad and the solder joint, and only the last of those is the product.

Where the contact contribution grows, the tester sees a higher reading and compares it against a limit that was set on a clean board. The failure is reported against the component rather than against the contact, and the diagnosis then follows the wrong path. Where the contact contribution is large, the tester also loses the ability to detect a genuinely marginal joint, because the marginal signal is buried in the noise.

Sources of Test Point Contamination

Contamination arrives from several directions: flux residue left by the soldering process, oxide that forms on exposed copper or on a thin finish, handling oils from fingers, and debris transferred from a dirty probe or a worn tip.

The geometry of the point matters as much as the chemistry. A small point with a solder mask ring around it traps material, and a point placed under a component or beside a tall body is harder to clean and easier to contaminate. Contamination can also be transferred by the fixture itself, since a dirty probe carries residue from one board to the next across an entire batch.

Flux Residue and No-Clean Chemistry

No-clean flux residue is designed to stay on the board, and it is not always an insulator. Its resistivity depends on the chemistry, the amount left after reflow and the humidity of the environment, and a thin, tacky film on a test point is the worst case for a probe.

Where the process uses a no-clean paste and the fixture sees high contact resistance, the choice is between cleaning the assembly, cleaning the probes more often, or moving the test points to areas that are naturally free of residue. A tacky residue is worse than a dry one, because it holds particles against the pad and prevents the tip from reaching bare metal.

Probe Selection and Tip Geometry

Probes are available with a range of tip shapes, and the shape decides whether the tip penetrates a film or rests on it. A crown or a serrated tip cuts through residue and oxide, while a flat tip simply presses the film against the pad.

Spring force matters too. A probe that is compressed beyond its rated travel loses force with age, and a probe that is compressed too little does not break the film. Tip selection is discussed in the notes on ICT probe selection. Probes also wear in the barrel that guides them, and a worn barrel lets the tip wander so that it lands beside the point rather than on it.

Fixture Cleaning and Maintenance

Cleaning the probes is a scheduled task with a defined method, not a reaction to a failure. The tips should be wiped or brushed with a solvent that dissolves flux without leaving a film, and the fixture should be checked for debris that can bridge adjacent probes.

The schedule should follow the paste chemistry and the board volume. A fixture running a no-clean process needs more frequent cleaning than one running a water-soluble paste, as covered in the notes on test fixture maintenance. Probes should be replaced on a cycle count rather than on appearance, since a tip can look acceptable while its spring has already lost a third of its force.

False Failures and Their Diagnosis

A false failure has a signature. It appears on one channel, it moves when the fixture is cleaned, and it often disappears when the board is tested a second time in a different fixture position.

The first diagnostic step is to test a known-good board in the same position. If the failure follows the fixture rather than the board, the problem is in the contact; if it follows the board, the problem is in the assembly and the test is doing its job. A cleaning method that leaves solvent on the tip is worse than no cleaning, because the solvent dissolves a fresh film of flux on the next board.

Measuring Contact Resistance

Contact resistance can be measured directly with a four-wire check on a reference coupon, or indirectly by testing a board with a shorting bar that connects pairs of test points. Both approaches give a number rather than an impression.

A practical limit for a clean fixture on a finished pad is under 50 milliohms, and a reading above 200 milliohms is a reason to stop and clean. Those numbers should be established on the fixture in question, because plating, finish and probe type all move them. A stable reading across ten cycles on the same board is the simplest proof that the contact is sound, and it takes less than a minute at the start of a shift.

Effects on Yield and Rework Decisions

A yield drop that appears suddenly and affects unrelated circuits is usually a fixture problem rather than a product problem. Boards pulled for rework on that evidence are reworked for no reason, and the rework itself introduces the risk that the test was meant to remove.

The safeguard is a rule that connects the yield to the fixture: when the failure rate rises above the normal band, the first action is a fixture check with a golden board, not a change to the test limits, as set out in the ICT notes. Where a golden board fails in a fixture that passed the same check earlier in the week, the fixture is the suspect until the contact resistance has been measured again.

Records and Verification Interval

Records should carry the fixture identification, the probe type and age, the cleaning dates and the contact resistance measured on the reference coupon. gopcb keeps those values beside the yield data so that a drop can be attributed correctly.

The verification interval should follow the number of board cycles rather than the calendar, since a fixture that runs 200 boards a day ages faster than one that runs 20. A simple counter on the fixture is enough to make that visible.

FAQ

What causes high contact resistance in ICT? Usually a film of flux residue or oxide on the test point, combined with a probe tip that cannot break through it or a spring that has lost force with age.

How often should test probes be cleaned? The interval is set by the paste chemistry and the board volume, and it is best defined by the point at which measured contact resistance begins to rise.

Can a test point be placed under a component? It can, but it is harder to clean and more likely to collect residue, so points should be kept clear of tall bodies wherever the layout allows.