In-circuit test (ICT) checks how a circuit board was built; functional test (FCT) checks whether it works. ICT probes individual components and nets to find manufacturing defects such as shorts, opens, missing parts and wrong values, and points to the exact faulty part. Functional test powers the assembly and exercises it as it will be used, proving it performs its job but often saying less about why it fails. Most production lines use both, together with flying probe and boundary scan where they fit better.
This comparison expands the test-methods section of our complete guide to automated test equipment (ATE). Below we compare the four main electrical test methods, show how to combine them for coverage, and how volume changes the right choice.
ICT vs functional test at a glance
| In-circuit test (ICT) | Functional test (FCT) | |
|---|---|---|
| Question answered | Was the board built correctly? | Does the board do its job? |
| Finds | Shorts, opens, missing, wrong or reversed parts, bad solder joints on probed nets | Firmware, timing, interaction and performance faults, plus many assembly faults |
| Power applied | Mostly unpowered; some powered tests | Fully powered and operating |
| Access needed | Test pad on nearly every net | Connectors and a few key test points |
| Fixture | Bed-of-nails, product-specific | Simpler functional fixture |
| Fault diagnosis | Precise: names the part and pin | Coarser: names the failing function |
| Test time | Typically under a minute | Seconds to many minutes |
| Program development | Largely generated from CAD and BOM data | Written from the product specification |
| Best for | Medium to high volume, accessible boards | Every product that must be proven to work |
The four electrical test methods compared
In-circuit test (ICT)
An ICT system uses a bed-of-nails fixture to contact every net, then measures each component in isolation: resistors, capacitors and inductors by value, diodes and transistors by junction behavior, and nets for shorts and opens. Guarding techniques electrically isolate the part being measured from parallel paths. Powered ICT adds checks of simple digital devices and on-board voltages. ICT is fast and very good at pinpointing faults, but its fixture is product-specific and it cannot reach nets without test pads.
Flying probe test
A flying probe tester performs similar measurements with a few robotic probes that move from point to point, so there is no fixture to build. It can reach small pads and vias that a fixture cannot. The trade-off is speed: tests take minutes rather than seconds, which suits prototypes, new product introduction and low-volume or high-mix production.
Boundary scan (JTAG, IEEE 1149.1)
Boundary scan uses test logic built into many digital ICs. Through a four- or five-wire JTAG port, the tester can drive and read each device pin, testing interconnects between JTAG-enabled parts, including BGA joints that no probe can reach. It needs only a connector, so it complements ICT and flying probe on dense boards. Its limits: it only covers nets connected to boundary-scan devices, and it does not measure analog values.
Functional test (FCT)
Functional test powers the board or product, loads firmware, applies inputs and checks outputs, communication and performance against the specification. It finds problems no structural test can, such as a timing marginality or a firmware configuration error, and it is the basis of most end-of-line testing. Because one failed function can have many causes, diagnosis is slower, so functional test works best after structural tests have removed most assembly defects.
Optical (AOI) and X-ray (AXI) inspection add visual coverage of solder joints and placement, which overlaps with ICT. See our machine vision guide for how vision inspection works.
Combining methods for test coverage
No single method reaches every defect. Test engineers plan a strategy so the methods overlap where defects are likely and cover each other’s gaps.
Worked example: combining ICT and boundary scan coverage
A board has 800 nets. The layout provides test pads on 720, so ICT has 720 ÷ 800 = 90% nodal access.
Of the 80 nets without pads, 60 connect between JTAG-enabled devices under BGAs. Boundary scan tests them, bringing structural coverage to (720 + 60) ÷ 800 = 97.5%.
The remaining 20 nets, such as an analog filter section, are covered by functional test measurements, and functional test also proves firmware and performance across the whole board.
How production volume changes the choice
Fixture cost, programming time and test time trade off differently as volume grows. A simple time comparison shows why flying probe and ICT suit different volumes.
Worked example: flying probe vs ICT test time
A board takes 3 minutes on a flying probe and 30 seconds on ICT.
At 5,000 boards a year, flying probe uses 5,000 × 3 = 15,000 minutes = 250 hours of tester time; ICT uses 5,000 × 0.5 = 2,500 minutes, about 42 hours.
Whether the 208 hours saved justify an ICT fixture depends on the fixture cost, tester availability and how often the design changes. At 500 boards a year the saving is only about 21 hours, and flying probe usually wins.
| Volume and mix | Typical strategy |
|---|---|
| Prototypes and new product introduction | Flying probe + boundary scan + bench or simple functional test |
| Low volume, high mix | Flying probe or boundary scan + functional test on a flexible tester |
| Medium to high volume | ICT (with boundary scan) + functional test + end-of-line test |
| High volume, dense boards with little access | AOI/AXI + boundary scan + functional test, with limited ICT |
Common mistakes
- Dropping test pads late in layout to save space, which forces a slower or less complete test strategy for the life of the product.
- Relying on functional test alone, so every assembly defect becomes a slow diagnosis at the end of the line.
- Running ICT without guarding, so parallel paths make good parts measure out of tolerance and limits get widened until they catch nothing.
- Not tracking which defects each stage catches and which escape to later stages or the field, so the strategy is never improved.
Need a functional or end-of-line tester for your boards?
Sciotex builds custom automated test systems for functional, end-of-line and production test, including fixtures, instruments and software, built and supported in the US since 1997.
Frequently asked questions
In-circuit test probes individual components and nets on a board to find manufacturing defects, usually with little or no power applied. Functional test powers the board and checks that it performs its intended function. ICT shows exactly what is wrong; functional test proves the board works.
Neither is better in general. ICT is much faster per board but needs a product-specific fixture, so it suits medium and high volumes. Flying probe needs no fixture and reaches smaller pads, so it suits prototypes and low-volume or high-mix production.
Boundary scan (JTAG, IEEE 1149.1) tests connections between digital devices that include boundary-scan logic, using only a JTAG connector. It is especially useful for BGA joints and dense boards where probes cannot reach.
Sometimes, for simple or low-volume boards. But without structural test, assembly defects are found later and diagnosed more slowly, which raises repair cost. Most volume production uses both.
Test coverage is the share of possible defects, nets or components that a test strategy can detect. It is often estimated per method and combined, for example ICT nodal access plus boundary scan plus functional test.
Related articles
- What Is Automated Test Equipment (ATE)? The Complete Guide
- Test Fixtures for Automated Testing: Bed of Nails, Pogo Pin and Functional Fixtures
- End-of-Line Testing: How EOL Test Systems Work and How to Specify One
- 7 Measurements to Automate with an ATE System (and How to Get Them Right)
- Automated Test Software: Test Executives, TestStand, LabVIEW and Python
- Hardware-in-the-Loop (HIL) Testing: How It Works and When to Use It
