DFM Review for PCBA: What We Check and Why It Pays
Design for manufacturability is often treated as a formality that happens between design and production. In practice it is the step where the cost of a product is decided, because the choices that determine yield, assembly time and board cost are made in the design rather than on the line. A DFM review that finds a marginal footprint before the stencil is cut costs an afternoon. The same issue found at the line costs a re-spin, a new stencil and a delivery date.
What the review covers
- Panel utilisation and board spacing: how the board is arrayed on the fabrication panel, which is one of the largest single levers on bare board cost.
- Footprint and pad geometry: whether the land pattern suits the process, including the solder mask and paste aperture relationship that decides how much solder the joint receives.
- Placement feasibility: fine-pitch and high-mass components checked against equipment accuracy, nozzle availability and the reflow profile for the thermal mass of the board. The placement and soldering process these are checked against is described under SMT assembly.
- Stencil and paste design: aperture reduction for fine pitch, and the paste volume the process can actually print consistently.
- Component availability and lifecycle: long-lead, single-source and end-of-life parts identified against the BOM before the quotation is fixed.
- Stack-up and impedance: layer arrangement, copper weights and controlled-impedance requirements checked against what can be fabricated.
- Assembly and test access: whether the test points, fiducials and tooling holes the process needs are present on the design.
Where the money is
Real savings on a PCBA programme come from design and planning rather than from pressure on the hourly rate. Improving panel utilisation reduces the bare board cost. Replacing an obsolete or single-source part with a pin-to-pin alternative removes both a premium and a supply risk. Standardising the surface finish across a product family reduces setup. Consolidating quantity into fewer builds lowers the per-board share of setup. And letting the assembler buy the material turnkey removes the margin stacking of a multi-vendor supply chain, as set out under component sourcing.
What we need from you
The Gerber files including drill data and board outline, a BOM with manufacturer part numbers, the assembly drawing showing orientation and polarity, and any stack-up, impedance or coating requirement the product has. If the design has already been produced elsewhere and had yield issues, sending that history lets the review target the known problem rather than starting from zero. The full file set is listed on the quotation requirements page.
Turning findings into a decision
DFM findings are returned as specific observations with a stated effect on cost, yield or schedule, so the customer can decide which to apply. Not every finding has to be actioned: a recommendation that adds a design change for a small yield gain may not be worth making on a mature product, and we say so rather than issuing a generic checklist.
The findings we return most often
Most of what a review turns up falls into a small number of categories. Each is listed with what it typically costs if it is not addressed.
| Finding | What it looks like | Effect if left |
|---|---|---|
| Panel utilisation | Board outline that leaves unusable panel area, or spacing below what the router needs | Higher bare board cost on every order, not just the first |
| Paste aperture against land pattern | Aperture ratio outside what the printer holds consistently, usually on fine pitch | Solder bridging or insufficient fillet, found only at AOI |
| Thermal mass imbalance | Large copper planes or heavy components next to fine-pitch parts on the same side | Cold joints on the small features or overheating on the large ones |
| Missing process features | No fiducials, no tooling holes, no test points, or a panel break-off that leaves burrs | Hand alignment, slower assembly, no access for test |
| Component lifecycle risk | End-of-life, not-recommended-for-new-design or single-source parts in the BOM | A build that cannot be repeated, at a premium that grows each year |
| Impedance and stack-up gaps | Controlled-impedance targets or a stack-up with no fabrication equivalent stated | Boards that meet the drawing and miss the electrical requirement |
Who runs the review and how long it takes
The review is carried out by our own process engineers rather than passed to the fabrication house or the customer's design team, because the findings that matter most cross the boundary between board and assembly: a stack-up decision changes the reflow profile, and a component substitution changes the paste volume. A standard review is returned with the quotation, normally within one to two working days of a complete file set, and it is carried out against the process the boards will actually run on at the Shenzhen factory. Where the review is a joint exercise between the customer's design team and ours, we join a call rather than exchanging annotated drawings over several rounds, because most findings are resolved in one conversation and cost several days if conducted by correspondence.
Design changes are yours to make
The review produces findings; the design belongs to the customer. Kilandu manufactures to a customer design and does not take over PCB layout or circuit design. Where a finding requires a layout edit, we state the requirement precisely enough that the change can be made in the customer's own tool, and where a review of the revised layout is needed before tooling is cut, we do that as part of the same exercise. This division is deliberate: it keeps design ownership where the intellectual property sits, and it keeps the DFM review honest, because a supplier that both reviews and redesigns has an incentive to find work.
When not to change a mature design
A design in volume production elsewhere is not automatically a design that should be changed. Where a product is shipping in volume with acceptable yield, applying a recommendation that changes a footprint carries a requalification cost that frequently exceeds the saving, particularly in regulated product categories where a change restarts an approval. In that situation the useful output of a review is narrower: which findings can be applied without touching the layout, such as a component substitution, a finish change or a panel change, and which should be carried forward to the next hardware revision. We separate the two lists rather than presenting one consolidated action list.
DFM before quotation, not after
The review sits before the quotation is accepted, not at the line, because that is the only point at which the findings are still cheap. Most DFM findings are inexpensive to apply while the design is being edited and expensive once tooling, stencils and material orders exist. That is why the review is included with the quotation rather than charged as a separate engineering service: a quotation issued against an unexamined design is an estimate, and the customer would be comparing an estimate to a quotation. The sequence that follows the review, from material plan to release, is described on the how to order PCBA page.
Before you send: what to check
Most of the round trips in a DFM review come from inputs rather than from findings, so a short self-check before the files go out removes most of them. None of the items below is a design change; they are the inputs the review works from.
- Gerber files include the drill data and the board outline, not only the copper layers.
- The BOM carries manufacturer part numbers and not descriptions.
- Quantities in the BOM are per board, not per order.
- The assembly drawing shows orientation, polarity and any keep-out area.
- A pick-and-place file is included, or it is understood that the positions will have to be derived from the Gerber.
- Parts that are already obsolete, single-source or on a long lead time are flagged.
- Test points, fiducials and tooling holes are present on the design.
- Panel break-off and board spacing are within what the router needs.
- Stack-up and any impedance target are stated.
- The acceptance class, Class 2 or Class 3, is decided.
- Coating or masking areas are defined on the drawing.
- Any known yield problem or field failure is described, rather than left for the review to rediscover.
Where one of these is missing, sending the design is still worthwhile; the review will say what it cannot assess without it. The full file set a quotation needs is listed under quotation requirements.
Send a design for review
Gerber files, a BOM with manufacturer part numbers and an assembly drawing are enough to start. Send them to alden@kilandu.com. If the design has a history of yield problems or a known field failure, say so and include what is known, and the review will target that rather than working from a blank sheet. The turnkey PCBA page describes what a build covers once the review is closed, and the quality assurance page covers how inspection records are held afterwards, and the workmanship criteria applied at inspection are set out under IPC-A-610.
What does the DFM review actually check?
Panel utilisation and board spacing, footprint and pad geometry against the process, placement feasibility for fine-pitch and high-mass components, solder paste aperture design for the stencil, component availability and lifecycle risk against the BOM, and stack-up or impedance issues that would affect yield. Findings come back with the quotation.
Is the DFM review chargeable?
The standard DFM review is included with the quotation rather than charged as a separate engineering service. Where the review identifies a change that requires redesign work, the effort involved is agreed with the customer before it is undertaken.
How much does a DFM change typically save?
The savings that recur are in yield and in schedule rather than in a one-off figure. Improving panel utilisation reduces bare board cost. Replacing an obsolete or single-source part with a pin-to-pin alternative removes a price premium and a supply risk. Standardising the surface finish across products reduces setup. Fixing a footprint that is marginal for the process removes a yield loss that would otherwise repeat on every build.
When should the DFM review happen?
Before the first quotation is accepted, not at the line. Most DFM findings are cheap to apply while the design is still being edited and expensive once tooling, stencils and a material order exist. If a board has already been produced elsewhere, sending that history along with the design helps the review target the right issues.