Skip to content
S.B. Electronics Printed circuit boards · since 2000, home page

Engineering guide · Design data & DFM

PCB DFM checklist: 12 things to verify before release

The twelve checks our CAM engineers run on every incoming order, from track and space to panels, so your data goes straight to production without a query.
Two-layer ESP32 flight controller with fine-pitch pads, vias, mounting holes and a dense silkscreen legend
FIG. 01Two-layer ESP32 flight controller with fine-pitch pads, vias, mounting holes and a dense silkscreen legend. Board design: ESP32 flight controller by Shazmeen Siddiqui, MIT licence; rendered from its Gerber files.
Written by
S.B. Electronics Engineering Team
Published
Updated
Reading time
7 min · 1,570 words
On this page · 9 sections
  1. 01What is DFM for PCBs?
  2. 02What minimum track width and spacing should you use?
  3. 03How big should vias, pads and annular rings be?
  4. 04Which solder-mask and silkscreen rules matter?
  5. 05What should you check for assembly and panelisation?
  6. 06Which problems does a design-rule check miss?
  7. 07Why do most DFM problems turn out to be data problems?
  8. 08How does DFM affect cost and lead time?
  9. 09What does the DFM review at S.B. Electronics cover?

Building this board?

Design for manufacturability (DFM) is the difference between a board that passes your CAD tool's design-rule check and a board that yields well in production. The rule check knows your rules; it does not know the physics of etching, plating, drilling and reflow. This checklist follows the review our CAM engineers run on every incoming order. Work through it before you release data and most engineering queries never happen.

What is DFM for PCBs?

DFM is the practice of keeping a design inside the process window of the factory that will build it — with margin. Every process has a capability limit, the finest feature it can make, and a comfortable zone where yield is high and cost is low. A design that sits at the limit everywhere can be built, but more slowly, at lower yield and at a higher price. A design that uses the limit only where it must, and standard rules everywhere else, is cheaper, faster and more reliable.

Table 1 · Our limits, and what we suggest designing to

FeatureS.B. Electronics capabilityOur guidance
Track / space6 mil / 6 mil standard; 4 mil / 4 mil fine-line6 / 6 mil or wider wherever routing allows
Minimum drill0.3 mm mechanical; 0.2 mm for fine-pitch designs0.3 mm vias wherever density allows
Annular ring0.15 mm minimum for drilled holes (design guideline)Pad diameter = finished drill + 0.3 mm or more
Outline tolerance±0.15 mm as standard; ±0.1 mm on requestKeep copper at least 0.3 mm inside routed edges
Board sizeMinimum 10 × 10 mm; maximum 500 × 600 mm (larger on request)Panelise small boards for assembly
Assembly board size50 × 50 mm to 510 × 460 mmAdd rails to small or irregular boards
Fig. 02DFM · the first measurements

ATrack, space, edge

6 / 6 mil (0.15 mm) tracks; copper at least 0.3 mm inside the routed edge.

BVia and annular ring

A 0.3 mm drill in a 0.6 mm pad leaves the 0.15 mm ring we design to.

CSolder-mask dam

0.5 mm pitch with 0.05 mm mask expansion leaves a 0.15 mm dam (0.1 mm minimum).

All dimensions in mm · drawn to scale within each panel

The five dimensions CAM measures first: track width, spacing, annular ring, solder-mask dam and copper-to-edge clearance.

What minimum track width and spacing should you use?

Use 6 mil / 6 mil (0.15 mm) as your default and drop to 4 mil / 4 mil (0.1 mm) only where a fine-pitch part forces it. Our process runs 6 mil / 6 mil standard; 4 mil / 4 mil fine-line. Copper weight changes the numbers: thicker copper etches with more undercut, so 2 oz (70 µm) copper needs roughly 8 mil / 8 mil.

  1. Track and space against copper weight. 6/6 mil is comfortable on 1 oz copper and 4/4 mil is our fine-line limit. On 2 oz copper, design 8/8 mil or wider — the etch undercuts more as the copper gets thicker.
  2. Current-carrying tracks. Size power tracks from IPC-2152 charts for the temperature rise you can accept, not by eye. Outer layers carry more current than inner layers at the same width, because they shed heat to air.
  3. Copper balance. Large empty areas opposite heavy pours make a board bow and twist in lamination and reflow. Fill empty areas with hatched or solid copper, and balance copper between layers that mirror each other in the stack.
  4. Acid traps and slivers. Acute angles between tracks trap etchant; thin copper slivers between pours can lift and short. Use 45° or 90° junctions, teardrops on small pads, and a minimum pour width.

How big should vias, pads and annular rings be?

Make every pad at least 0.3 mm larger than its finished drill. That leaves a 0.15 mm annular ring all round — our design guideline — which absorbs drill wander, layer registration and plating so the hole never breaks out of the pad. Keep the aspect ratio, board thickness divided by drill diameter, at or below 8:1.

  1. Annular ring. Via pad ≥ drill + 0.3 mm; component pads larger still. A 0.3 mm via gets a 0.6 mm pad. For Class 3, where no breakout is allowed, add margin.
  2. Aspect ratio. Plating struggles to reach the middle of a deep, narrow hole. On a 1.6 mm board, 0.2 mm (8:1) is the practical floor. Our drilling runs 0.3 mm mechanical; 0.2 mm for fine-pitch designs — prefer 0.3 mm wherever density allows.
  3. Via-in-pad. An open via in an SMT pad wicks solder away during reflow and leaves a starved joint. Move vias off pads with a short neck and tent or plug vias next to fine-pitch pads. If a BGA leaves no room, specify filled vias on the fab note and agree them with your fabricator before layout.

Which solder-mask and silkscreen rules matter?

Leave at least 0.1 mm (4 mil) of solder mask between adjacent pads so the mask dam survives development, and make each mask opening about 0.05 mm larger than its pad on every side to absorb registration. Keep legend ink off pads and exposed copper, and keep text at least 0.8 mm tall with 0.15 mm strokes so it prints legibly.

  1. Mask dams. Below about 0.1 mm the sliver of mask between fine-pitch pads lifts or never forms, and solder bridges become likely. If the pitch is too fine for a dam, use one gang opening across the row and say it is intentional.
  2. Mask expansion. Solder-mask registration is typically ±50 µm. An opening 0.05 mm larger than the pad on each side keeps mask off the pad; much larger openings expose neighbouring tracks.
  3. Silkscreen. Legend on a pad becomes a soldering defect. Clip silkscreen to the mask openings — most tools do it automatically — and place polarity marks where they stay visible after the part is fitted.

What should you check for assembly and panelisation?

Give the assembly line what its machines need: three global fiducials, clearance between parts and the board edge, and a panel when the board is small. Our assembly line takes boards from 50 × 50 mm to 510 × 460 mm; anything smaller is built in a panel with rails.

  1. Fiducials and edge clearance. Place three global fiducials — a 1 mm copper dot in a 2–3 mm mask-free clearance — asymmetrically, so the board cannot be loaded the wrong way round. Keep copper and components at least 0.3 mm (0.5 mm is better) from routed edges, and well back from V-score lines, which cut through copper.
  2. Panelisation. Agree the panel with your fabricator and assembler rather than guessing. V-score suits straight-edged rectangular boards; tab-routing with mouse bites suits irregular outlines and parts that overhang the edge. Add 5–10 mm rails with tooling holes and fiducials when a board is small or has parts near its edges.

Which problems does a design-rule check miss?

A design-rule check is only as good as its rule deck, and several manufacturing problems sit outside what most rule decks test. Look for these by eye, or with a dedicated DFM tool, before release:

  • Copper slivers and starved thermals — narrow necks between pours, or thermal-relief spokes so thin they etch away.
  • Hole-to-copper clearance on inner layers — a hole that passes close to a plane it must not touch risks a short after normal drill wander. Keep plane clearances around holes generous.
  • Mask slivers — strips of mask under 0.1 mm wide that the rule deck allows because the pads and tracks themselves are legal.
  • Silkscreen over vias and pads — legal in most rule decks, a defect on the finished board.
  • Floating copper islands — copper connected to nothing, which serves no purpose and can couple noise. Remove it or stitch it to ground.
  • Unequal thermal mass on small parts — an 0402 resistor with one pad on a plane and the other on a thin track can tombstone in reflow. Use thermal reliefs on plane connections.

Why do most DFM problems turn out to be data problems?

Most of the issues we raise are not physics; they are data. A drill file exported before the last edit. An outline on the wrong layer. A fab note that says ENIG while the order says HASL. Re-export every manufacturing output after your final change, open the set in a Gerber viewer and compare it with the fab note. Our guide to preparing Gerber files has the full export checklist.

How does DFM affect cost and lead time?

Every feature at a capability limit narrows the process window. Fine lines need tighter etch control and closer inspection; small drills break more often and need slower feeds; tight tolerances need more measurement. None of that is a problem when the design needs it. When it does not, it adds cost and risk for no benefit. The cheapest reliable board uses standard rules everywhere they fit.

Table 2 · Design choices and their effect in production

Design choiceEffect in production
6/6 mil instead of 4/4 milWider etch window and higher yield
0.3 mm vias instead of 0.2 mmFaster drilling, easier plating, fewer broken drills
Balanced copper on every layerLess bow and twist through lamination and reflow
1.6 mm FR-4 with 1 oz copperStandard material, no special build
Rectangular board in a V-scored panelFast depanelling and better panel utilisation
Rigid one- or two-layer board within standard rulesPriced in seconds in the online instant quote

What does the DFM review at S.B. Electronics cover?

Every order's data is reviewed by a CAM engineer before production. The review checks your files against the rules above and our published capabilities, and flags anything that would cut yield or reliability — an unplated hole that should be plated, a mask sliver, a breakout risk, a missing outline. We raise one query with the proposed fix and wait for your approval before a panel is cut; we never change a design silently.

Written by

S.B. Electronics Engineering Team

The CAM, production and quality engineers at our factory in the Gandhinagar G.I.D.C. Electronics Estate — the people who review incoming data and build the boards. Published ; last reviewed and updated .

Ask the engineering team a question

Questions

Questions about DFM

Short answers to what engineers ask us most often about this subject.

What is the minimum track width and spacing S.B. Electronics can make?

Our process runs 6 mil / 6 mil standard; 4 mil / 4 mil fine-line. Design to 6/6 mil wherever routing allows and use 4/4 mil only where a fine-pitch part needs it. The board yields better and costs less, and the fine-line limit stays available for the few places that need it.

What is the smallest drill size for a PCB via?

Our drilling runs 0.3 mm mechanical; 0.2 mm for fine-pitch designs. Keep the aspect ratio at or below 8:1, so on a standard 1.6 mm board 0.2 mm is the practical floor. Use 0.3 mm vias where density allows: they drill faster, plate more reliably and leave more room for the annular ring.

Is DFM the same as a design rule check?

No. A design rule check tests your layout against the rules you entered in the CAD tool. DFM asks whether those rules, and the design, suit the factory's processes — etching, drilling, plating, mask and assembly — with margin. A design can pass DRC and still be hard to build if its rules sit at the limits.

Do I need to panelise my design for assembly?

Only if it is small or irregular. Our assembly line takes boards from 50 × 50 mm to 510 × 460 mm; smaller boards are built in a panel with rails, tooling holes and fiducials. Tell us whether you want V-score or tab-routing, or leave the panel design to our CAM engineers.

Should I add teardrops to pads and vias?

Yes, on small vias and narrow tracks. A teardrop widens the junction between a track and its pad, so normal drill wander cannot cut the track off at the pad, and the junction resists cracking under thermal and mechanical stress. It costs nothing in fabrication and most CAD tools add teardrops automatically.

Start a project

Send the data. We will check it before we build it.

Price one- and two-layer boards online in seconds, or send the full data pack — Gerbers, drill, fab note, BOM and centroid — and an engineer replies with a reviewed quote.