Separate the unknowns
Create a short uncertainty register before release. A new package escape, an unfamiliar material, an impedance target, and a stacked connection are different risks. Give each an owner, a proposed check, and a decision that the result will enable.
For example, a working processor may demonstrate basic connectivity while leaving the thermal endurance of its interconnects unresolved. Conversely, a good fabrication coupon does not establish software stability or the quality of the assembled power distribution network.
Preserve production relevance
Use the intended construction when the objective is to qualify that construction. A prototype built with more generous rules can help develop a circuit, but should be labeled accordingly. Record any departures in materials, thickness, panel arrangement, finish, via structure, or assembly sequence.
Cadence's manufacturing-file guidance treats the fabrication drawing as a place for material, panel, and special requirements. For a prototype, add a clear distinction between required features and experimental features so neither is changed casually during review.
Design measurements into the experiment
An illustrative first article for a compact sensor could include accessible power rails, representative impedance structures, and an agreed interconnect coupon. The point is not to add every possible test. It is to make the uncertain behavior observable without damaging the only functioning sample.
- Reserve accessible points for the measurements you actually need.
- Capture pre-assembly condition before thermal processing changes it.
- Keep sample identity linked to board and assembly records.
- Define acceptance and the action following a failed result in advance.
Close the loop before the next revision
Record findings against the original questions. Update source design files with accepted fabrication changes, and separate confirmed conclusions from unresolved observations. If a build works only after a manual modification, describe that modification precisely; otherwise the next prototype repeats the experiment without learning from it.
Adjacent-layer access, one connection at a time
Dielectric · geometry + material
Core or build-up dielectric
Illustrative only. Copper thickness, dielectric construction and via spans require a reviewed stackup.
Common questions
Is functional testing enough for an HDI prototype?
It is enough only for the functions it exercises under its test conditions. Manufacturing and lifetime questions may need separate structural or stress evidence.
Can I use a cheaper stack for firmware development?
Yes, if it meets that limited objective and is documented as a development substitute. Do not transfer its manufacturing conclusions to a different final stack.
References & further reading
Method & assumptions · General design guidance. Validate the proposed construction and acceptance criteria for your project.