What ICT and FCT are
In-circuit test (ICT) checks an assembled board at component level. The board is pressed onto a bed-of-nails fixture whose spring-loaded pins contact test points, and the tester measures shorts, opens, resistor and capacitor values, diode orientation and similar properties. Many ICT setups also program firmware or configuration data into the board.
Functional test (FCT) powers the board, or the finished product, and checks its behavior: it boots, runs the right firmware, draws the expected current, reads its sensors, transmits on its radio, charges its battery and responds to its buttons and display. FCT finds faults ICT cannot see, such as a correctly soldered but defective component or a firmware problem. ICT finds faults FCT may miss or cannot locate, such as a wrong-value resistor that still lets the board pass a functional check today.
Where each test sits
| Test | Finds | Needs | Typical use |
|---|---|---|---|
| AOI | Visible placement and solder faults | An AOI program | After reflow, every board |
| ICT | Shorts, opens, wrong or missing parts | Test points and a dedicated fixture | Higher volumes, after board assembly |
| Flying probe | Similar faults to ICT, more slowly | Test points, no fixture | Prototypes, pilot builds, low volume |
| Board-level FCT | Board function | Fixture, test software, firmware | After ICT, or instead of it |
| End-of-line FCT | Finished product function | Test station and software | After final assembly |
Boundary scan (JTAG, IEEE 1149.1) can test connections to BGAs and other devices that cannot be probed physically, and is often combined with ICT or FCT. Products with radios, sensors or batteries usually add dedicated steps such as RF calibration and testing, sensor calibration and battery checks.
What a production test plan should define
- Coverage. Which faults each test is meant to catch, and which are left uncovered.
- Limits. Pass and fail limits for each measurement, who set them and on what basis.
- Fixtures and software. Who designs, builds, pays for and owns them; see NRE.
- Data. Results logged per serial number and retained, so failures can be traced later.
- Fixture checks. Verification at the start of each shift with a known-good unit, and ideally a known-bad one.
- Failure handling. What happens to a failed board: diagnosis, repair, retest rules and limits.
Some products also need a burn-in test between assembly and final functional test.
Testability starts in layout
Test coverage is decided largely when the board is laid out. Without accessible test points, ICT is impossible or expensive and flying probe becomes slow. Plan test points, programming headers and debug access with the assembler before layout is frozen; see DFT.
Questions to ask a supplier
- What test coverage do you propose for our board, and what does it miss?
- Who writes the FCT software, and who updates it after design changes?
- Are results stored per serial number, and for how long?
- How many retests are allowed before a board goes to repair?
- How do you check fixture health at the start of each shift?
- How is the firmware version verified on each unit?
Common mistakes
- Retest until pass. Re-seating a board until it passes hides intermittent faults. Limit retests and log every attempt.
- FCT that only checks power-on. A board that boots may still have a dead sensor or a weak radio.
- Limits widened to raise yield. This turns test failures into field failures. Find out why units fail before changing a limit, and record who approved each change and why. A first pass yield that rose only because limits widened is not an improvement.
- No test points. Discovered when fixture design starts, after layout is frozen.
- No data retention. Without logs per serial number, a field failure cannot be traced to its test results.
Where Zefinity helps
Fixtures and test software are built around the coverage you agree, so late changes to it are slow and costly. Through manufacturing engineering support, Zefinity works through test coverage with the factory, reviews fixtures and limits before the pilot run and reads the test data with you during ramp-up.
Frequently asked questions
Do I need both ICT and FCT?
Not always as separate stations, but you need both kinds of coverage: assembly faults and functional faults. At low volume, flying probe plus a thorough functional test is common. At higher volume, ICT can pay back its fixture cost through faster testing and clearer fault location. The right mix depends on the board, the volume and test-point access.
What is a flying probe test?
A fixtureless form of in-circuit testing in which a few movable probes contact test points one after another. It needs no bed-of-nails fixture, so it suits prototypes, pilot builds and low volumes, but it takes much longer per board than ICT. It still needs accessible test points.