PCB Design Problems That Block Production

Most boards that a fabricator or an assembly house rejects are not rejected for being electrically wrong. They are rejected because something in the pcb design cannot be built or assembled reliably, and the problems repeat across companies and projects. The list below covers the ones that appear most often.

Why Designs Get Rejected

The underlying cause is usually a gap in knowledge rather than carelessness. A designer who has never seen a placement machine may not know what a fiducial is for, and a company without a design standard produces boards that vary with whoever drew them.

The other cause is process. When no manufacturing engineer reviews the layout, the first person to examine it for manufacturability is the supplier, at which point the schedule is already running.

<img src="https://www.gopcba.com/wp-content/uploads/2024/09/Applications_of_Automobile_PCB.webp" alt="PCB panel with process edge, tooling holes and fiducial marks” />

Missing or Poor Process Edge

Assembly equipment needs somewhere to hold the board. Without a process edge along the sides, the conveyor cannot transport it and the machine cannot clamp it, so the first requirement is a rail of adequate width on two opposite sides.

The edge also has to be free of components and of anything that interferes with clamping. A rail that is present but covered in parts is functionally the same as no rail at all.

<img src="https://www.gopcba.com/wp-content/uploads/2023/05/2.jpg" alt="Assembly review of pad size and component spacing on a PCB layout” />

Tooling Holes and Location

Tooling holes position the board in the printer, the placement machine and the test fixture. They must be present, of a size the equipment can use, and placed where they will not be blocked by a component or lost when the panel is separated.

Their position relative to the circuit matters as much as their existence. A hole that is offset from the pattern it is supposed to locate introduces a registration error that appears as a placement offset across the whole board.

Fiducials and Machine Vision

A fiducial is a reference mark that the vision system uses to locate the board. Without them, or with marks that are too small or partly covered by solder mask, the machine falls back on mechanical location and the placement accuracy drops.

Good practice is a global fiducial set on the panel and local fiducials near fine pitch devices. The mark should be a defined shape with a clear area around it, and it must survive the mask process with adequate contrast.

Mounting Holes and Plating

A screw hole that is plated can be filled with solder during a wave process, leaving the screw unable to pass. Where that is a risk, the hole is left unplated, and the pad around it is designed with a spoke or flower pattern so that solder does not bridge it.

The annular ring on a mounting hole is often the largest on the board, and a pad that is too small for the screw head transfers the clamping load to the laminate, which can crack it.

Pad Size and Spacing Errors

Incorrect pad sizes produce defects that are blamed on the process. A pad that is too short gives a weak fillet, one that is too wide encourages bridging, and asymmetric pads on a two terminal part produce the tombstoning that appears as a standing component.

Spacing errors cause the same class of problem. Pads that are too far apart leave insufficient paste on the joint, and pads that are too close bridge during reflow. The pad design should come from the package specification rather than from a library that has never been checked.

Vias in Pads

A via placed in a solder pad draws solder down the barrel during reflow, leaving the joint short of material. Where a via must sit under a pad, it has to be filled and capped, as described for small BGA escape routing.

The same applies to a via placed close to a pad. Solder can wick toward the hole and away from the joint, so the clearance between a via and a pad should be adequate rather than minimal.

Test Points

Test points that are too small, or that are covered by a component after assembly, make in-circuit test impossible. A probe needs a defined area of exposed metal, and it needs to reach that area with the board in the fixture.

Test points should be on one side where possible, spaced according to the fixture capability and clear of tall components that would obstruct the probe travel.

Silkscreen and Solder Mask Over Pads

Print or mask that encroaches on a pad reduces the solderable area and is a common cause of intermittent defects. Reference designators that are missing, inverted or too small to read make assembly verification and repair harder than it needs to be.

The legend should be legible after assembly rather than before. A designator placed under the component it identifies is technically present and practically useless.

Component Spacing

Components placed too close together cannot be reworked and may not even be placeable. As a working figure, a minimum gap of about half a millimetre between chip parts on a reflow assembly is sensible, with more required for wave soldering.

Around a large area array package the requirement is greater, since the part and the rework tools both need room. Placing small parts within a few millimetres of such a device is a decision that will be regretted during the first rework, and the placement order should account for it.

Panelisation Problems

A panel that is poorly arranged produces interference between adjacent circuits, breakaway tabs that stress the board and v-score lines that pass through components. Assembling a panel that flexes also affects the placement accuracy on the outer edges.

Panel design should follow the assembly house’s requirements, including the rail width, the tab positions and the arrangement of the circuits. A panel built to a generic rule rather than to the equipment in use is a frequent cause of delay.

The Fix: DFM Review

Almost every item on this list is detectable before the data is released. A review against a written checklist, performed by someone who understands the assembly process, catches the majority of them in an hour.

That is the purpose of design for manufacture as a discipline: not to make the board cheaper, but to make it buildable. A design that passes the review reaches the line without questions, which is worth more than any individual optimisation.

Solder Mask Slivers and Dams

A solder mask web that is too narrow between two pads will lift or flake during assembly, leaving exposed copper where the design expected insulation. The minimum width the fabricator can hold should be respected, and the mask opening defined accordingly.

The same applies at the edges of a large pad. A mask dam that is marginal in width is a cosmetic detail on the drawing and a defect on the board, so the clearance between mask and copper should be generous enough to survive the process.

Checking a Design Before Release

A short checklist run before the data is sent catches most of the items in this article: edge rails, tooling and fiducials present, pads taken from verified footprints, vias clear of pads, test points accessible and the legend readable after assembly.

Then check the panel. The arrangement, the tabs and the rail width should meet the requirements of the assembly house that will run the boards, since a panel that suits one machine may not suit another. Once the data is sent, every one of these items becomes a question that costs a day.

FAQ

What is the most common design omission? Process features rather than circuit features: the edge rail, tooling holes and fiducials. They are invisible to the schematic and essential to the assembly.

Are these problems the designer’s fault? Usually not individually. They appear where there is no design standard and no manufacturing input during the layout, which is an organisational gap rather than a personal one.

How early should the review happen? Before the layout is finished. Placement decisions that break assembly rules are cheap to change on screen and expensive to change after tooling.