Neither method is better in absolute terms. Flying probe needs no fixture and adapts to a new revision through reprogramming, which suits prototypes and designs still in flux. ICT carries a one-time fixture investment and returns a cycle measured in seconds, which suits frozen designs in steady production. Three variables decide it: volume, revision frequency, and whether the board has test points.
What Is Flying Probe Testing?
Flying probe is a fixtureless electrical test method. Movable probes land in sequence on pads and component leads, taking each measurement point by point.
Probe paths are generated from design data, usually Gerber, ODB++, or a netlist, so a test program can be written within hours of receiving the files.
Configurations range from four probes on a single side to twenty probes testing both sides at once. More probes means a shorter cycle per board.
Some systems also handle boundary scan, LED polarity and color checks, and barcode reading. The equipment and process are described further in this overview of flying probe testing in PCB manufacturing.
What Is In-Circuit Testing (ICT)?
ICT powers the board through a custom bed-of-nails fixture. Spring-loaded probes are arranged to match the test point layout, and the board is pressed onto the fixture so every point makes contact at once.
Because nothing has to move between measurements, switching is electronic rather than mechanical, which is why the cycle is short.
The trade-off is that the fixture is built for one specific board, and the board must carry dedicated test pads. The process is outlined on the ICT testing service page.
What Defects Do Flying Probe and ICT Detect?
Coverage overlaps heavily. Both work at the electrical layer and detect:
- Opens and shorts
- Missing, wrong, and reversed components
- Resistance, capacitance, and inductance out of tolerance
- Basic conduction behavior of diodes and transistors
The shared blind spot is function. Electrical test confirms the board was built to the BOM (Bill of Materials) and netlist. It does not confirm the board performs as designed once powered, which is what functional test covers.
Surface defects sit outside both methods as well, and are handled by AOI (Automated Optical Inspection) and X-ray.
Flying Probe vs ICT: Side by Side
| Factor | Flying probe | ICT |
| Fixture | None required | Custom bed of nails, one-time cost |
| Preparation | Program written from design data in hours | Fixture build typically 2 to 4 weeks |
| Cycle per board | Measured in minutes | Typically under 30 seconds |
| Handling a revision | Reprogram, hardware unchanged | Fixture usually rebuilt |
| Test point dependency | Contacts pads and leads directly | Dedicated test points required |
| Access on dense boards | Probes reach fine-pitch locations | Limited by fixture probe spacing |
| Mechanical stress | Point contact, low stress | Full-board compression |
| Volume fit | Prototypes, low to mid volume | Steady production |
When to Use Flying Probe Testing
Four situations make flying probe the practical choice.
The design is still changing.
Each revision means another fixture, and fixture cost compounds with revision count. Flying probe only needs a program update.
Volume is too low to amortize a fixture.
On orders of tens to a few hundred boards, the fixture investment per board often exceeds the cost of the test itself.
Multiple board types run in parallel.
Building a fixture for each of five or six board types running at double or triple-digit quantities is rarely justifiable.
The board has no test points.
Flying probe contacts pads and component leads directly, which makes it the only viable electrical test for legacy products and boards built without design for test.
The cost of flying probe is speed. Sequential probing puts the cycle in the minutes range, and once volume rises, test becomes the line constraint and cycle cost overtakes fixture cost.
When to Use ICT Testing
Three conditions justify the fixture.
The design is frozen.
The fixture is built around the current test point layout and is usually scrapped by a revision, so a stable design is the precondition for spending on one.
The volume is steady and repeating.
Fixture cost is recovered through quantity, and a one-off order rarely recovers it.
Cycle time is the bottleneck.
When every other operation runs fast and only test lags, moving to ICT raises throughput across the whole line.
The cost of ICT is rigidity. Budget three items rather than one: the fixture, test program development, and program maintenance after component substitutions or minor revisions. Ask for them as separate lines so a quotation covering only fixture hardware is visible as such.
At What Volume Does ICT Pay Off?
Most published guidance offers a range along the lines of flying probe below a few hundred boards and ICT above a thousand. That range holds for boards with a moderate test point count, and on a real program the crossover is worth calculating.
The calculation: (fixture cost + test program development) divided by (flying probe cost per board minus ICT cost per board) gives the quantity at which ICT becomes the cheaper option.
Fixtures commonly run USD 2,000 to 10,000 with a 2 to 4 week build, depending on test point count, board size, whether both sides need contact, and whether programming functions are integrated.
Three variables move the crossover significantly:
Test point count.
More points extend the flying probe cycle and pull the crossover earlier. On a simple board with few points, flying probe can stay economical into the thousands.
Revisions per year.
Two revisions a year means paying for three fixtures. This variable appears on no cost sheet and is frequently the deciding one.
Whether orders repeat.
A single order and a quarterly reorder amortize very differently. Run the calculation against annual volume rather than batch size.
What Test Points Does ICT Require?
Test point provision belongs in PCB layout. Adding points after the layout is fixed usually means a board revision.
Four items to settle at the design stage:
Test pad size and spacing.
Pads need enough contact area for a probe, with regular spacing between points. Exact values follow the fixture capability of the manufacturer building it.
Points concentrated on one side.
Double-sided access raises fixture complexity and cost noticeably.
Clearance for probe travel.
Points sitting against tall components may be unreachable.
Node coverage confirmed in advance.
Not every net can be reached. Request a coverage report before the fixture is designed, stating which nets are testable and which lack an access point.
Venture Electronics operates a dedicated PCB test fixture design function and develops test programs against board layout and access conditions, with coverage reports issued to the customer. For a project team, this means knowing what the fixture will actually reach before the fixture is paid for, rather than discovering an uncovered net afterward.
Can Flying Probe and ICT Work Together?
Yes, and most programs use them in sequence rather than choosing once. The common path is flying probe through prototyping, a fixture and ICT after design freeze, and a temporary return to flying probe for engineering samples or small revision builds during production.
Neither constitutes a complete test on its own. A typical production chain runs AOI for surface joints, X-ray for hidden joints under BGA (Ball Grid Array) and similar packages, electrical test for connectivity and parameters, and functional test for system behavior. The division of work is set out in this comparison of five PCBA testing methods.
Moving from flying probe to ICT calls for three things in place beforehand: fabrication data and netlist at the frozen revision, a written coverage target with acceptance criteria, and a plan for testing during the fixture build. Production continues shipping across those 2 to 4 weeks, usually on flying probe, which is why holding both capabilities at one supplier simplifies the transition.
Venture Electronics provides both flying probe and ICT in house, with no minimum order quantity on prototype and validation builds. Validation quantities and production quantities therefore stay with one supplier, and changing test methods does not require changing suppliers.
Set the Test Plan Alongside the Design
Test method selection feeds back into PCB layout, so it is worth aligning with the manufacturer during design rather than after boards exist. Share expected volume, revision frequency, and the reliability requirements of your industry, and let the test configuration be built around them.
Explore Venture Electronics' PCBA testing service and submit board type, test point count, and volume to have an electrical test approach assessed.
FAQ About Flying Probe and ICT Testing
Q1: Can flying probe fully replace ICT?
At prototype and low volume, yes. Once volume rises, the per-board cycle makes test the line constraint, so the cost of substitution shows up in lead time rather than in test pricing.
Q2: Does flying probe damage the board?
Not under normal parameters. Probes land on pads or leads with controlled force. Repeated testing at the same location can leave a light witness mark, so boards with cosmetic requirements are worth agreeing landing positions on in advance.
Q3: Can a legacy board without test points still use ICT?
Usually not, unless a revision adds them. These boards normally stay on flying probe, with boundary scan (JTAG) available to extend coverage on supported devices.
Q4: Can an ICT fixture move between suppliers?
In principle, provided wiring diagrams and working software transfer with it. Fixture hardware without the test program still requires development at the receiving site, which limits the actual saving.
Q5: Is a functional test still needed after the electrical test?
Yes. Electrical test confirms the board was built correctly, functional test confirms it operates as designed. The two answer different questions.
Q6: How should an RFQ be written to get comparable quotations?
State volume, board dimensions, test point count, whether double-sided access is needed, and whether programming is to be integrated. Without those, suppliers quote against different assumptions and the numbers cannot be compared.
