In-Process Inspection at Each PCB Fabrication Stage

The cheapest defect to fix is the one caught at the stage that created it. A short circuit found on the finished board means scrapping a panel that has already passed through drilling, plating, imaging, etching, solder mask and surface finish; the same short found after inner layer etching costs one panel core and a few minutes of rework.

That is the argument for in-process inspection. This article follows a multilayer board through the fabrication sequence and describes what is checked at each stage, what the check is looking for, and why some checks only make sense before the next process step covers the evidence.

Why In-Process Inspection Beats Final Inspection

Final inspection examines the finished article, by which point the causes of a defect are several process steps in the past and the evidence has been covered by the next layer of material. In-process inspection examines the board while the evidence is still visible and the process that produced it is still running, which makes it possible to correct the process rather than just sort the boards.

The distinction matters commercially as well as technically. A defect found at the inner layer stage affects one core; the same defect found at final test affects a finished multilayer panel whose value now includes every operation performed on it. The earlier the check, the smaller the quantity of material and labour that is written off with it.

Artwork and Film Verification

The first check is on the artwork, before any board exists. Every film is inspected visually after processing for broken traces, shorts, thin lines, registration between layers and physical damage to the film itself. Films are then protected in sleeves, because a scratch on a working film will be reproduced on every panel it is used for.

Registration is the item that deserves the most attention. Each layer’s film has to align with the others within the tolerance the design allows, and a film that is correct in isolation can still be wrong in combination. Where the artwork is generated digitally and imaged directly, the equivalent check is on the data and on the imaging parameters rather than on a physical film, but the purpose is identical.

Inner Layer Imaging and Etching

After the inner layers are coated with resist and exposed, the pattern is inspected for breaks, shorts, foreign material and registration errors before development and etching. Any of these defects that survives to the etching stage is etched into the copper, where it can no longer be corrected except by rework or scrap.

A second inspection follows the etching step, and this is the one that matters most on a multilayer board. The check looks for broken traces, shorts, discolouration, raised areas, foreign material, resist peeling, residual copper and scratches. The board is then inspected for these defects again after the copper has been built up, and the two inspections together decide whether the core is fit to be laminated into the stack.

Copper Plating and Hole Quality

After plating, the panel is checked for plating defects and for blocked holes. A hole that is partially blocked may still pass a continuity test while providing a connection with a fraction of the intended cross-section, which will fail later under thermal cycling. Plating thickness is checked on coupons that are processed with the panel, because the barrels themselves cannot be measured without destructive analysis.

Blocked holes are a particular risk on panels with a high aspect ratio, where the plating chemistry has to reach the middle of a deep, narrow barrel. The check is quick and the consequence of missing it is severe, which makes it a good example of an inspection step whose cost is trivially small compared with the failure it prevents.

Outer Layer Imaging and Etching

The outer layers repeat the inner layer sequence, with the same two inspections: after exposure, looking for breaks, shorts, foreign material, registration and damage; and after etching, looking for the same list plus discolouration, raised areas, peeling, residual copper and scratches. The difference is that the outer layers carry the pads and the fine geometry that will define assembly yield.

This is where a solder mask dam that is marginally too narrow, or an annular ring that is thin on one side, becomes visible. Both are within the fabricator’s process capability but outside the design intent, and catching them here allows the process to be adjusted before the remainder of the order is produced.

Surface Finish and Solder Mask

The surface finish is applied after the solder mask, and it is inspected for coverage, uniformity and contamination. An organic finish that has not formed evenly will show up later as a solderability problem, and a gold finish that is too thin over its nickel layer will fail at the interface under thermal stress. Neither is visible on a functional test.

Solder mask inspection covers registration to the pads, adhesion and the absence of voids over features that will be exposed. Mask that is lifted or blistered will fail during reflow, and mask that has crept onto a pad reduces the area available for the joint. Inspecting it before the finish is applied avoids finishing a board that will have to be remade.

Final Inspection and Packing

Boards produced as prototypes are usually tested with flying probes, and the prototype requirements for multilayer boards describe how that is specified. The design guidelines that keep a board manufacturable are the other half of the same agreement.

Final inspection on a bare board combines an electrical test with a visual check. The electrical test — flying probe for prototypes, a fixture for volume — confirms continuity and isolation on every net, and the visual check confirms the outline, the marking and the general condition of the board. Boards are then packed by part number so that a mixed shipment cannot contaminate a build.

The records from all of these stages are what make the process traceable. A defect found at final test can be traced back to the process step that produced it only if the earlier inspection results were recorded, which is the difference between a quality system and a set of inspections. For a buyer, the useful question is not whether inspections happen but what the fabricator does with the data.

Recording the Results and Reacting to Them

An inspection that produces a pass or fail with no record is only half a control. The useful output is a trend: the number of inner layer shorts per panel, the plating thickness distribution across the week, the registration offset measured at the outer layers. Those numbers show a process drifting before it starts producing scrap, which is what turns inspection from sorting into prevention.

The records also support the customer conversation. When a buyer asks how a defect was found and what was changed, a fabricator with data can answer precisely; one without data can only describe what is done in general terms. Mapping the inspection stages onto the characteristics used to judge PCB quality makes that conversation concrete, and it also tells the buyer which of the process limits their design is relying on.

FAQ

How many inspections does a multilayer board pass through? A typical sequence covers the artwork, the inner layer after exposure and after etching, the plated panel, the outer layers after exposure and after etching, the surface finish, and the finished bare board. Not all of them apply to a simple two-sided board.

What is checked on a plated panel? Plating defects and blocked or partially blocked holes. Plating thickness is verified on coupons processed with the panel rather than on the board itself, because the measurement is destructive.

Is final electrical test enough on its own? No. It confirms continuity and isolation, but it cannot see a marginal annular ring, a lifted solder mask or a finish that is too thin. Those are found at the stages that created them.