Siemens adds AI command search to its circuit-board manufacturing checks

Siemens’ Valor NPI 2610 update, announced on 26 October, adds natural-language command search, predictive suggestions and a Product Support Copilot to software used to assess printed circuit board manufacturability.[1]

The engineering problem sits between a circuit that works on paper and a board that a factory can produce. Valor’s established workflow compares design data with fabrication and assembly requirements. It can check electrical connections against manufacturing data, reconcile a bill of materials with the design, and assess whether approved alternative components physically fit. Supplier capabilities help determine which manufacturing rules apply.[2]

The manufacturing handoff has a physical shape too. Siemens’ description of Valor covers assembly panels with fiducials, tooling holes, breakaway tabs and scored separation features. These details show how geometry around a circuit matters to making it. The task extends beyond checking that the electrical connections match the design: the board also has to pass through fabrication and assembly.[2]

Three ways into the engineering tools

The underlying work still requires several kinds of engineering information. Siemens’ course for Valor NPI 2305 teaches the ODB++ board-data structure, component-library enrichment, fabrication analysis, assembly analysis and interpretation of results. These are connected activities: a useful answer about a manufacturing issue depends on understanding which part of the board model, which production step and which reported result are involved.[4]

Keeping a result attached to the right board

Xpedition engineering data management now connects verification to a particular PCB revision, retaining rules, results and reviewer feedback through check-in and approval. Classic Mode remains available during the interface transition.[1]

Consider a hypothetical late component substitution. An earlier analysis may describe the original part perfectly, yet answer the wrong question about the revised board. Keeping the design version beside the result lets a reviewer establish what was actually checked before deciding what needs repeating. The same logic applies when a rule changes: a different outcome can come from a different design, a different requirement, or both.

Valor’s existing product model already brings together information that can otherwise sit in separate drawings and documents, including parts data, surface finishes and the assembly panel. It also supports comparison between design revisions and changes made by a fabricator. That background explains why the revision link matters: the engineering object is a structured description of the product, with several kinds of information that must stay aligned.[2]

Turning standards into checks

The release also adds nine IPC standards profiles spanning performance classes and producibility levels.[1]

Performance class describes the finished product: basic function for Class 1; continued performance and extended life for Class 2; critical continued or on-demand operation for Class 3. Producibility levels A, B and C describe progressively more demanding manufacturing capabilities.[1]

The Global Electronics Association describes these profiles as a way to translate standards into engineering rules that software can apply. Its source requirements span board design, electrical and thermal considerations, fabrication and assembly. The profiles cover technologies including flexible boards, high-density interconnects, microvias, heavy copper and high layer counts.[3]

Two questions sit alongside one another here. A manufacturer’s custom rules describe what its processes can make. Industry standards connect the design with requirements for the intended product and its eventual acceptance. The association’s explanation treats these as complementary: a supplier-specific rule set may capture a factory’s capabilities without establishing the full link to the relevant standards.[3]

For example, the association links IPC-2221 design requirements with manufacturing and acceptance criteria such as IPC-J-STD-001 and IPC-A-610. The profiles supply values for automated checking; the underlying standards supply the context for when those values apply and how to interpret them.[3]

The attractive part of this approach is its proximity to a decision that has to be made. Assistance has a concrete purpose while a designer is choosing how to resolve an issue, with the relevant geometry, requirement and manufacturing context available. That gives AI a practical role in getting a product from its design to production.

Report an error or suggest a correction