2.5 DFM Rules of Thumb for Non-Designers

Design for Manufacturing (DFM) is not about making a product "easy" to build; it is about obeying the physical laws of the factory floor. You do not need to be a layout engineer to identify high-risk features. A design can be electrically perfect in simulation yet functionally unbuildable at scale. DFM is the practice of respecting the limits of steel nozzles, liquid solder, and optical cameras. If you violate these limits, you trade yield for "compactness," a trade that rarely pays off.

Spacing: The "Courtyard" Rule

Every component needs a "Keep-Out" zone, often called a courtyard. Machines do not drop components from the sky; they lower them with a mechanical head.

• The Physics: If Component A is placed next to Component B, the nozzle placing B needs room to move without knocking A off its pads.
• If spacing is < 0.25 mm (10 mil) → Then the placement head will collide with adjacent parts, causing misalignment or chipping.
• If tall parts are next to short parts → Then the nozzle for the short part is blocked by the "shadow" of the tall part.
• Rule: Maintain a minimum 0.5 mm clearance between active components for standard assembly.

Fiducials: The Robot’s Eyes

A pick-and-place machine is a blind robot that counts steps. It needs a reference point to know exactly where the board is clamped. These reference points are called Fiducials.

• Definition: A distinct round copper pad (1 mm – 3 mm) with no solder mask covering it. It is not electrically connected to anything.
• If fiducials are missing → Then the machine cannot align itself to the board, resulting in placement drift across the panel.
• If solder mask covers the fiducial → Then the camera cannot see the contrast, and the board is rejected.
• Rule: Every board (and every panel frame) must have at least three global fiducials.

Polarity: The Ambiguity Trap

Ambiguous polarity is the leading cause of manual assembly error. Operators should never have to guess.

• The Problem: A silkscreen dot can be covered by the component body. A "+" sign can look like a "T".
• If the only polarity marker is under the chip → Then Inspection (QA) cannot verify it after assembly.
• Rule: Use unambiguous graphics. A thick line or a square on the silkscreen outside the component footprint is the standard.
• Pro-Tip: For LEDs, draw a diode symbol on the silkscreen. "Pin 1" is technical; a diode arrow is universal.

Stencil Apertures: The Paste Release

Solder paste is a sticky fluid. Pushing it through a hole in a steel stencil requires specific aspect ratios.

• The Physics: If the stencil hole is deep and narrow (like a straw), the paste sticks to the walls and won't release onto the pad.
• If the component lead is tiny (e.g., QFN/BGA) but the stencil is too thick → Then the paste clogs the stencil, and the pad gets zero solder.
• Rule: The area of the opening must be greater than the area of the aperture walls (Area Ratio > 0.66). Practically, this means you cannot use thick stencils for fine-pitch parts.

Test Access: The "Bed of Nails"

To test a board rapidly (ICT), you press it against a fixture of spring-loaded pogo pins. These pins need a target.

• Definition: A Test Point is an exposed copper pad (usually 1 mm circle) connected to a critical net (Power, Ground, Data).
• If there are no test points → Then technicians must probe legs manually by hand, which is slow, dangerous, and unreliable.
• If test points are covered by solder mask → Then the pogo pins cannot make electrical contact.
• Rule: Place test points on the bottom side of the board for every critical signal, spaced at least 2.54 mm apart if possible.

Final Checklist