Bare Board Electrical Test Methods and Net Coverage
A finished printed circuit board can look perfect and still be scrap. An open trace under a connector, a short between two nets hidden by the mask, or a plated hole that never made contact with an inner layer are all invisible to inspection. Bare board electrical test exists to find those faults before components are placed, when the cost of scrapping the board is at its lowest. This article explains what the test actually proves, how the limits are set and which method suits a given build.
What the Test Proves
The test compares the board against the net list it was manufactured from. Every pair of points that should be connected must show a low resistance, and every pair that should be isolated must show a high resistance. Passing the test means the artwork, the drilling and the plating produced the connectivity the design intended, which is the property that cannot be verified by looking.
It does not prove impedance, current capacity or long term reliability. Those depend on geometry, material and plating quality, and they are controlled by process data and coupons rather than by an electrical test. A board that passes electrical test can still fail a thermal cycle test, and the two are complementary rather than interchangeable.
Continuity Measurement
Continuity testing applies a small current through the net and measures the voltage drop, usually with a four wire method on critical paths so that probe and trace resistance do not contaminate the result. The pass limit is not zero resistance. A long thin trace has a real resistance, and a plated barrel contributes its own, so the limit is derived from the calculated resistance of the path plus a margin.
That calculation matters because a limit set too tightly fails good boards and a limit set too loosely passes marginal ones. Holes with poor plating are the classic marginal case: the connection exists but through a thin barrel, so the resistance is slightly high and the long term reliability is poor. Setting the continuity limit from the expected resistance of the longest path is what makes the test sensitive to that condition.
Isolation and Test Voltage
Isolation testing verifies that nets which should be separate are separate. The test voltage matters because a whisker, a conductive residue or a thin bridge may not conduct at low voltage but will break down at a higher one. Typical production test voltages are in the range of 100 to 250 volts for standard boards, with higher values used where the product demands them.
Higher voltage is not automatically better. It increases the risk of damaging a marginal board that would have worked, and on fine pitch designs with thin dielectric spacing it can create a failure rather than reveal one. The test voltage should follow the product requirement and the dielectric spacing, and it should be recorded with the test program so that the result can be interpreted.
Net List and Test Program Generation
The test program comes from the CAD data, not from a physical board. Generating it from the design ensures the test matches the intent, and it also captures the point locations from the same database the fabrication data came from. A program built by teaching points on a sample inherits any error in that sample and often misses nets that are not accessible on the surface.
Coverage should be reviewed before the first build. Nets that reach only an inner layer pad with no surface access cannot be probed directly, and they need either a test point added at design or acceptance of a known coverage gap. Coupon structures that document process capability, described in PCB test coupons, complement the electrical test but do not replace it.
Fixture Test Compared with Flying Probe
Fixture test uses a bed of probes matched to the board, so it measures every net in a few seconds and scales well to volume. The fixture is expensive and product specific, which means it only pays back when the quantity justifies it. For prototypes and low volume it is the wrong tool.
Flying probe test moves a small number of probes over the board under program control. It needs no fixture, so it can be used on a single board, and it reaches points that a fixture cannot. It is slower by orders of magnitude and it needs the board surface to be accessible, so it is the standard choice for prototypes and for boards that will never reach high volume. The trade offs are set out in flying probe testing.
Sampling, Records and Disposition
Not every board needs a full test. High volume production often runs electrical test on a sample basis once the process is proven, using coupons and process controls for the rest. That decision should be based on evidence of stability, on the cost of an escape and on the customer requirement, and it should be revisited when the design or the process changes.
Records are as important as the test itself. Logging the failure type, the location and the lot makes it possible to detect a drift in drilling or plating, and it turns individual failures into a pattern. Reviewing the failure data alongside the general criteria in judging PCB quality gives a complete picture of whether the board supplier is in control.
Handling, Marking and Traceability
Boards that pass the test still have to arrive at assembly in the condition the test verified. Handling between test and packaging is a real source of damage, particularly on thin boards and on panels with large cut outs, and a board that is dropped or flexed can crack a barrel that measured perfectly a minute earlier. Tested boards should be stacked in a way that avoids point loads, moved in rigid containers, and separated from untested boards so that the status is never ambiguous.
Marking closes the loop. A board that carries a lot code and a test status can be traced back to drilling, plating and test records, which is what makes it possible to contain a problem rather than to recall a whole shipment. Where the customer requires it, the test result can be tied to the individual board through a laser mark or a label, and the same data feeds the process control charts that show whether drilling and plating are stable.
Test Economics and Where to Apply It
Electrical test costs money per board and per program, so the decision should be made on the total cost of an escape. On a four layer board with wide traces, a single open circuit found at assembly is an inconvenience. On a fine pitch multilayer board with high value components already placed, the same open circuit can destroy an assembly that is worth many times the board itself.
The practical rule is to test where the loss is large relative to the test cost. That includes fine pitch designs, boards with many small vias, high layer counts and any product whose field failure cost is high. When volume is low, flying probe after fabrication is the usual answer, and when volume is high a fixture pays for itself quickly.
FAQ
Does a passed electrical test mean the board is good? It means connectivity matches the net list at the test conditions. It says nothing about impedance, plating thickness or long term reliability.
Why does the continuity limit allow some resistance? Every trace and barrel has resistance, so the limit must be above the expected value of the longest path. Tightening it below that produces false failures rather than better quality.
Can bare board test find a partially plated hole? Often yes, because the barrel resistance is higher than nominal. A cross section of a sample remains the definitive check on plating quality.
