Skip to the stories
Electronics & Semiconductors

Adapting PCB Reference Designs: What to Keep, Re-Specify and Validate

  • oliver-freeman
  • 912 reads

A proven PCB design can save time, but its assumptions, parts and layout need checking against your system before you route.

Adapting PCB Reference Designs: What to Keep, Re-Specify and Validate

A reference board can give a new PCB project a tested circuit and hard won layout decisions. Its value depends on carrying those decisions into the new design without carrying over assumptions that no longer fit. Differences in input range, load behavior, temperature, board stackup and component availability can turn a direct copy into an unstable or unbuildable product.

Treat the reference as engineering evidence to inspect and adapt. First check whether it was tested and whether its parts remain obtainable. Then sort each circuit block into three groups: keep, re-specify or remove. Validate changes before routing, and record what changed and why so another engineer can review the result.

Start by checking the design and its parts

Look at who produced the reference, when it was designed, and what it was optimized to achieve. A board designed for minimum cost may make different choices from one tuned for a small footprint or peak efficiency at a particular load. Those choices matter when deciding whether the design fits your system.

Check the bill of materials (BOM) before committing to its components. Parts in a design that is three or four years old may now be listed as end of life (EOL) or not recommended for new designs (NRND). The cited supply chain analysis reports that more than 620,000 electronic components were discontinued in 2025, and that 52% reached end of life without a formal Product Change Notification. Review how many distributors list each key part, and look for parts whose inventory repeatedly falls to zero. A component with limited supply can become a manufacturing constraint before the layout is finished.

Ask what evidence supports the design. Published test results and a validated layer stackup make it easier to judge how much of the design can transfer to your board. Without test data, you have less basis for treating the circuit as proven, even if its schematic looks plausible.

Recover the assumptions behind the schematic

A schematic captures connections, but not every condition under which the circuit is expected to work. Before copying blocks, identify the reference design's expected input voltage range, load profile and transient peaks, ambient temperature, airflow, duty cycle, EMC targets, isolation requirements and service life.

How to Apply Reference Designs in Projects Cover

Compare each assumption with your product. A power stage tested with open airflow may behave differently inside a sealed enclosure. A design built around one load point may need further checking if your load changes quickly or spends substantial time at a different operating point. The gap between those conditions is where seemingly small adaptations can affect stability, temperature or emissions.

Keep the parts that carry the design's hard won work

In a power circuit, the control topology, compensation network, feedback connections, protection and filtering often embody significant analysis and testing. Keep these intact when your operating conditions allow it. The same applies to sensitive layout patterns: examine the switch-node loop, sense-line routing, grounding and thermal-via arrangement before changing them.

A compact switch-node loop helps limit parasitic inductance, which in turn affects ringing and electromagnetic interference. Copying the schematic while changing these physical relationships can discard some of the reference design's most useful work. Preserve the layout intent where possible, then confirm that it still fits your board stackup and mechanical constraints.

Re-specify parts for the actual board

Some components may need different ratings or packages even when their circuit function remains the same. Multilayer ceramic capacitors (MLCCs) are a useful example: Class 2 ceramics can lose up to 80% of effective capacitance under DC bias near their maximum rated voltage. A capacitor selected for a 3.3V rail may therefore provide less capacitance when used on a 5V rail, potentially affecting regulator stability.

Recheck passive-component voltage, temperature and ripple ratings against your real operating environment. Confirm that package sizes and land patterns suit your assembly process, and select connectors that fit the product's mechanical interfaces. When a replacement exists in the same package, it can avoid moving the component and rerouting nearby traces. A footprint change may require both.

How to Apply Reference Designs in Projects The Pulse

Replace EOL and NRND parts with active alternatives, and review any non-cancellable, non-returnable (NCNR) components before ordering. If system requirements have changed, the microcontroller or regulator may also need to change. For every substitution, compare the datasheet ratings and package details; availability alone does not establish electrical or thermal suitability.

Changes to output capacitance or inductance can affect the regulator's compensation network. If either changes substantially, recalculate the relevant poles and zeros or run a loop-stability simulation. A network that was stable with the original component values cannot be assumed to remain stable after a significant substitution.

Remove evaluation features with care

Evaluation boards often include debug headers, test points, onboard programmers, indicator LEDs, unused peripheral circuits, selectable rails and margining circuitry. These features can take space and add cost when they serve no purpose in the finished product.

But appearance is a poor guide to function. A large input capacitor may look like test-bench convenience while actually suppressing input ringing. Removing it without checking its role can lead to poor transient behavior or instability. Identify what each part does before stripping it, especially in the power path.

Circuit blockTypical actionReason to check
Control loop and compensationKeepPreserves the established control behavior.
Switch-node and sense-line layoutKeepAffects interference and efficiency.
PassivesKeep, then reviewCheck voltage and temperature ratings for the new conditions.
ConnectorsRe-specifyMust match the enclosure and mechanical interfaces.
Obsolete or NRND ICsRe-specifySupports continued component availability.
Debug headers and jumpersUsually removeMay be evaluation features with no production role.
Margining circuitryUsually removeMay serve lab characterization rather than product operation.

Validate the changes before routing

Full-board simulation can help assess the design, but a focused set of checks is a practical minimum before routing. Check temperatures at the worst-case ambient condition and simulate load steps and input transients. Calculate the margins for every substituted part, verify footprints and land patterns, and run a loop-stability simulation if you changed the compensation network.

Thermal checks deserve particular attention when the new board has less copper area, different airflow or a hotter enclosure. Rapid temperature changes and mismatched thermal expansion among materials can contribute to microcracks; one cited manufacturing source attributes up to 15% of PCB failures in some production batches to such defects. That figure is specific to the reported batches, but it underlines why thermal and mechanical conditions should be considered together.

A layout copied across different stackups also needs scrutiny. Traces designed for a six-layer board can have different impedance on a four-layer board. Likewise, test-bench thermal results at 22°C may not represent a board operating at 50°C in a sealed plastic enclosure. Treat the reference conditions as test conditions, not as guarantees for your product.

Keep a record of every deviation

A deviation log makes the adapted design reviewable and gives the team a useful starting point if the board fails. Record the original part or value, the replacement, the reason for the change, the supporting evidence and who accepted the associated risk.

OriginalReplacementReasonEvidenceRisk accepted by
10µF 10V 0805 capacitor10µF 25V 1206 capacitorDC-bias loss and higher ambient-temperature deratingDatasheet thermal curveJ. Smith
LDO regulator ICDC-DC converterHigher efficiency targetBench test reportM. Davis

The log should cover layout as well as the BOM. Note changes to stackup, thermal vias, grounding and routing, since these can alter behavior even when the schematic remains the same. If the adapted board does not work, compare each recorded change with the reference, starting with substituted passives, compensation changes and layout differences. The record also gives a component manufacturer's applications team a clearer view of what changed if you need technical help.

Use part data to support the review

Part databases can speed up qualification by collecting lifecycle status, distributor stock, datasheets and CAD models in one place. For example, [Octopart](https://octopart.com/) lists BOM parts and links to datasheets and CAD resources, including schematic symbols, footprints and 3D models. Those details can support lifecycle checks, footprint review and the evidence recorded for substitutions. Availability and pricing can also help inform the deviation log, though each candidate still needs an engineering suitability check.

The practical test for a reference design is whether its proven parts still match the new product's electrical, thermal, mechanical and supply constraints. Preserve the circuit and layout choices that remain valid, verify every changed condition, and make the reasoning visible. That leaves the team with a design it can build, review and maintain.

Source: oliver-freeman · resources.altium.com

From the same set