PCB Material Selection

A PCB material choice is not complete when the laminate has been named. The copper, bonding system, solder mask, marking, and finished construction must also support the product. A useful selection record connects each material to a specific job and states what still needs verification before production.

The following is a hypothetical engineering exercise, not an icallin customer project or a tested reference design. It shows how a small sensor controller can produce a concrete material request instead of another generic list of “good PCB materials.”

Define the Board Before Opening the Catalog

Assume a rigid, four-layer controller measuring 60 × 40 mm. It sits in a dry indoor enclosure, receives 12 V from an external low-voltage supply, powers a microcontroller and sensor, and switches a small external load. The intended local ambient range is 0–60 °C. Assembly includes lead-free reflow and may require one repair cycle. The board has no dynamic bending requirement and no RF transmission line. These are example requirements, not material ratings.

The design is still missing several important inputs: actual component temperatures, the power trace geometry, connector loading, detailed stackup, and the final soldering profile. Mark them as open items rather than assuming that a high-Tg label resolves them.

Screen the Substrate Against This Particular Job

For this exercise, a rigid FR-4 construction is a reasonable starting candidate. A flex construction does not solve an identified mechanical requirement. A metal-core board does not solve an identified through-thickness heat-path problem. A specialty low-loss laminate is not justified merely by the existence of a digital clock. These are screening judgments based on the stated assumptions, not universal rankings.

As a real candidate family, Isola describes 370HR as an FR-4 laminate and prepreg system for multilayer boards and lead-free-compatible processing. That establishes a basis for discussion with the fabricator; it does not approve our hypothetical board. The proposed construction, assembly exposure, and relevant qualification evidence still need review.

The selection record should therefore say “370HR construction proposed; fabricator confirmation pending,” not “370HR guarantees reliability.” If the fabricator suggests another material, request the properties and process evidence needed to evaluate the change.

Give Each Remaining Material a Separate Responsibility

Copper forms the conductive network. Its finished thickness, width, plating, and local restrictions must support the electrical design. Bonding materials hold the construction together and establish dielectric spacing after processing. Solder mask defines protected areas and pad openings. Legend markings identify assembly and service information; they are not a conductive layer.

The example's decisions can be captured in one working table:

Item	Preliminary decision for this example	Evidence still required
Laminate and prepreg	Proposed rigid FR-4 construction; qualification pending	Approved stackup and assembly exposure review
Copper	Specify finished copper by layer	Current, voltage-drop, etching, and plating checks
Bonding system	Use the fabricator's qualified system for the selected construction	Pressed dielectric thickness and compatibility
Solder mask	Qualified mask compatible with the chosen finish and assembly	Opening geometry, adhesion, curing, and process approval
Legend	Polarity, connector identity, and revision remain readable	Assembly drawing review and pad clearance

This is a proposed decision record, not a fabrication recipe. In particular, the table intentionally does not prescribe a press temperature, mask cure cycle, or universal minimum feature size.

Figure 1: A material system is ready for release only when its functional requirements and manufacturing checks agree.

Avoid the Solder-Mask Color Shortcut

Mask selection is a chemistry-and-process decision, not a color preference alone. A technical data sheet may separate adhesion, solder resistance, solvent resistance, insulation, and thermal-shock results. The Taiyo PSR-4000 AM03TS data sheet, for example, presents these as separate properties and tests. A favorable result in one row is not evidence that every other requirement is satisfied.

For our controller, the designer should supply pad geometry, connector keepouts, and assembly needs. The fabricator should confirm a qualified mask system and the applicable process. A marketing phrase such as “high-temperature coating” is not a replacement for that exchange.

Use a Change Request to Test the Selection

Now change one assumption: the sensor must move through a hinge several thousand times. The original rigid-board material decision no longer answers the interconnect requirement. A sensible architecture may keep the controller rigid and introduce a separately designed flex interconnect. That flex needs its own conductor, adhesive, coverlay, and bend qualification; simply thinning the existing rigid board is not a complete response.

Alternatively, suppose the enclosure traps more heat than expected. First identify the hot component and available cooling path. The right change could be layout, copper distribution, mounting, airflow, or power dissipation rather than an automatic switch of substrate family. The material decision should follow the observed constraint.

These two changes illustrate why recording the starting assumptions is useful. The team can see exactly which decision became invalid and avoid redesigning unrelated parts of the stackup.

FAQ

Should a fabrication note specify only the laminate brand?

No. Identify the approved construction and the requirements that matter, including finished thickness, copper by layer, and relevant assembly or electrical constraints. A brand name without a product or acceptance basis leaves significant ambiguity.

Can the board house substitute any material with the same Tg?

Not automatically. A substitute must meet the approved criteria and construction requirements. Depending on the design, that review can include dielectric data, expansion behavior, bonding compatibility, and processing evidence in addition to Tg.

Release the Decisions, Not Just the Shopping List

The deliverable from material selection should be an agreed construction plus a short record of why it fits the product. Keep unresolved items visible and define who closes them. That makes a modest sensor board easier to manufacture and a later design change easier to evaluate.