Are Test Points Still Necessary When There are Vias on The PCB Circuit Board

There are vias all over the board. Why add test points before mass production? The problem is not whether the copper is connected, but whether the probe can stabilize on it, whether it can still be measured after the components are installed, and what method the factory will use for testing.

Can the probe be inserted into the via?

When drawing the PCB, test points are often regarded as one of the "optional" things. Many boards have already been very crowded in the wiring, and when the manufacturer requests to leave TP, the first reaction is: there are vias everywhere on the board, why not just find a copper-free area?

From a purely electrical connection perspective, this idea is not completely unreasonable. Vias, through-hole pads, and special circular pads, as long as they are connected to the network that needs to be measured, have the potential to be used as test points. But being able to conduct electricity does not mean it is suitable for the probe to touch.

Many vias have solder mask coverage, and when the probe is inserted, what is touched is green oil. Some vias, although not covered with oil, have small diameters and narrow ring widths, and if the spring probe lands slightly off-center, it is easy to slip open. Touching once with a multimeter is possible, but measuring hundreds of boards continuously with a fixture is a different requirement.

The exposed vias on the PCB that can be detected. Whether they can be used as test points depends on the window opening, size, and probe landing point.

Can they be touched before soldering, and still be there after?

When conducting bare board electrical testing, the board is not yet with components, and the probe can touch many component pads. When entering the PCBA testing stage, the problem arises: the pads have been occupied by resistors, chips, connectors, etc., and the points that could be touched may no longer be there.

For example, the pad under a QFN, when the board is bare, of course has copper, but after soldering, it is completely inaccessible from the outside. Another via may just be located under a shielding cover, a connector, or a large capacitor. They are still connected to the same network, but for the ICT probe, they no longer have accessibility.

Here, two words are easily confused: Bare board electrical testing mainly checks for open and short circuits in the circuit; ICT is board-level in-line testing after assembly, and the scope of attention will expand to welding and component-related issues. The contact position required in different testing stages may also be different.

The spring probes in the ICT fixture-clamping device need to align with the corresponding contact position on the board.

Why are test points always placed on the back of the board?

In mass production testing, spring needles fixtures are commonly used, also called "fixture beds". After the PCB is placed in the fixture, the probe presses towards each network from a fixed position. Whether the probe can be pressed correctly and stably directly affects the detection efficiency of the entire batch of boards.

If both sides of the board need to be probed, the fixture structure will become more complex, and the cost and debugging workload may increase. Therefore, many projects will prioritize concentrating the testable networks on the same side, usually the side with fewer components. It is not that the back is naturally better, but it needs to be consistent with the actual fixture scheme.

Test points should also leave room for the probe to move. If there are high components on the edge, the probe head may not be able to land. If several test points are crowded together, it may only be possible to use a finer and more expensive probe. Seeing a circle mark on the board does not mean the factory will definitely be able to probe it.

Which points on the board need to be left with probe space when there is not enough space?

When the space is tight, it is not necessary to cram large pads onto all networks. First, confirm with the test engineer whether to use flying probes, ICT, or functional testing, and confirm which networks really need fixed probe contact. Then discuss the size, spacing, and accessibility of the pads.

Power, ground, reset, critical enable, debugging communication interfaces, etc., networks are usually more worthy of considering the contact position for accessibility. Some via pads that are already suitable for probing or uncovered vias can also be reused when meeting process constraints, and do not necessarily need to place a new TP.

Finally, don't forget to check the manufacturing output. The test point coordinates, layer, and network name should match the fixture data. The points that can be marked as test points in the software do not necessarily mean that the actual size and placement are correct; this part still needs to be reviewed according to the conditions of the board factory and assembly factory.

There is another aspect that is easily overlooked: randomly adding a long test branch to a high-speed signal line may add additional load and reflections to the signal. Not every high-speed network is suitable for directly adding circular test pads; sometimes, special test structures, connectors, or boundary scan methods are needed.

The fact that the PCB can be powered on only indicates that it has the possibility of operation. During mass production, whether bad boards can be quickly identified depends on these seemingly insignificant test contact points.