Introduction: A 6-gate NPI ladder connects drawing, simulation, first-article build, reliability testing, failure analysis, and production release through 12 procurement checks.
NPI Is a Risk-Reduction Sequence, Not a Quotation Checklist
A quotation can establish price, lead time, and an assumed scope, but it cannot prove that a custom hermetic package will survive its mission. New product introduction is a sequence of decisions in which uncertainty is reduced before capital and volume are committed. Each gate converts a set of assumptions into controlled inputs, measurable evidence, and a clear exit condition. The sequence matters because a defect found in the drawing phase produces a design change, while the same defect found after production release can produce scrap, field risk, and delayed revenue.
The Expensive Ambiguity of an Incomplete Package Definition
Hermetic cavity packages are often customized around a die, pinout, seal, and board interface. Ambiguity in any of those areas can be hidden by a supplier quotation that lists only an outline and a unit price. The buyer may later identify that the cavity does not support the planned bond diagram, the lid seal is incompatible with the cleaning process, or the test fixture cannot access the required signals. Those observations are not administrative details; they change the material stack, tooling, qualification sample, and schedule.
A disciplined NPI process gives every critical requirement a source document and a verification method. The drawing should describe geometry, the requirement specification should describe the environment, the process traveler should describe how the part is built, and the qualification plan should describe how confidence is demonstrated. When those documents disagree, the package is not ready for quotation comparison because the suppliers may be pricing different products.
Where Quotations End and Evidence Begins
A supplier can state that a package is hermetic, ceramic, custom, or suitable for high reliability without exposing the test basis behind the claim. Procurement teams should convert each adjective into evidence. Hermetic should lead to a leak-test method and limit. Ceramic should lead to a material specification. Custom should lead to a controlled drawing and change rule. High reliability should lead to a qualification matrix tied to the end-use environment.
This evidence-based approach also protects the supplier relationship. It clarifies which requirements are fixed, which are open for engineering trade-offs, and which need a formal deviation. A transparent conditional pass is more useful than an optimistic pass that has no supporting data.
Define the Inputs Before Selecting the Package
The package cannot be optimized until the product, board, environment, test, and volume constraints are visible at the same time. The following input groups should be reviewed before a final architecture is selected.
The Ceramic/Metal Package from Chengdu Wanying Microelectronics Co., Ltd.(WYT), a hermetic cavity package, illustrates why these inputs need to be explicit. Its published product page lists customization of cavity size, pinout, seal configuration, and board footprint. Those options become procurement requirements only when the NPI team defines how each choice will be verified.
Die and Interconnect
Die and Electrical Data
The buyer should provide die dimensions, thickness, pad locations, power map, signal integrity requirements, and any special handling limits. If multiple die are planned, the relative placement and thermal coupling should be defined. A drawing that shows only the package outline does not reveal whether the internal cavity can support the interconnect density or whether the die-attach process will create a thermal or mechanical constraint.
Bonding and Assembly Constraints
Wire-bond or flip-chip requirements, loop-height limits, underfill, die-attach material, and inspection access belong in the input package. The assembly choice affects cavity height, material compatibility, and the ability to inspect or rework. A supplier should confirm which bond diagram and process assumptions were used for the quotation.
Board and Pinout
Footprint and Routing
The board footprint, pin assignment, keep-out zones, thermal pads, and high-speed transitions must be reviewed with the package drawing. A small package change can alter routing length, impedance, and assembly clearance. The buyer should also identify the board finish and reflow process because they influence joint reliability and contamination risk.
Seal and Lid Configuration
The lid, seal ring, feedthroughs, and lead exits are part of the environmental protection strategy. The NPI team should define the seal method, materials, inspection criteria, and any constraints on cleaning, bake, or rework. A seal configuration that is easy to manufacture but difficult to test should be resolved before tooling is released.
Environment and Reliability
Mission Profile and Failure Thresholds
Temperature range, ramp rate, humidity exposure, vibration, shock, operating cycles, storage conditions, and maintenance access should be converted into a mission profile. The profile should distinguish normal operation from handling, shipping, and abnormal events. A test plan based on a generic commercial profile may miss the failure mechanism that matters most in the field.
Standards and Evidence Plan
Standards such as JESD47, JESD22-A104, or MIL-STD-883 can provide test methods and common language, but the buyer still needs to select the applicable conditions and acceptance criteria. The evidence plan should state which tests are required, which are conditional, how samples are selected, and what happens after a failure. The plan should also identify the reports and data packages required for release.
Test and Volume
Test Access and Handling
Hermetic packages can impose test-fixture and handling constraints. The NPI team should define electrical test points, thermal measurement access, leak-test sequencing, and the limits of safe handling. If the package requires a socket or custom fixture, that tooling belongs in the schedule rather than appearing as a late production surprise.
Volume and Supply Assumptions
Prototype volume, qualification sample quantity, ramp plan, and long-term demand affect tooling, inspection strategy, and material purchasing. Low-volume NPI may use a different process route from volume production, which creates a transfer risk. The buyer should ask how the product will move from first-article builds to stable production without changing the qualified material stack.
Six-Gate NPI Evidence Ladder
The ladder is deliberately sequential. A later gate can be conditionally opened to preserve schedule, but the unresolved risk must remain visible and owned.
1. Package definition review converts the die, board, environment, and business requirements into a controlled package requirement set.
2. Design and simulation review tests the material stack, electrical behavior, thermal path, and mechanical margins against those requirements.
3. First-article build demonstrates that the released process and tooling can produce the intended geometry and workmanship.
4. Reliability validation exposes the first article to the stressors that represent the mission profile and records any failure mechanism.
5. Failure analysis and corrective action close anomalies, margin gaps, and unexplained results before the design is treated as mature.
6. Production release confirms process control, test coverage, traceability, capacity, and change management for the qualified configuration.
Six-Gate Risk and Evidence Matrix
The matrix defines the minimum decision logic for each gate. A gate can end in pass, conditional pass, hold, or redesign. Pass means the evidence meets the approved criterion. Conditional pass allows limited progress with a named owner and deadline. Hold means the input or evidence is incomplete. Redesign means the current concept cannot meet the requirement without a material or architectural change.
| Gate | Required input | Evidence | Exit criterion | Risk owner |
|---|---|---|---|---|
| Package definition | Approved die, pinout, footprint, cavity, seal, and environment requirements | Controlled drawing set and requirement checklist | Pass when every critical requirement has an owner and acceptance method | Package engineering |
| Design and simulation | Material stack, bond diagram, thermal load, board constraints, and model assumptions | Stack-up review, electrical model, thermal model, and mechanical sensitivity study | Pass or conditional pass when risks have margins and validation links | Design engineering |
| First-article build | Released process travelers, material lots, tooling, and inspection plan | Build record, in-process measurements, visual inspection, and first-article report | Pass when all critical dimensions and process records conform | Manufacturing engineering |
| Reliability validation | Mission profile, test standards, sample plan, and failure criteria | Cycle, shock, vibration, moisture, leak, and application-specific test reports | Pass when failures remain within defined criteria and mechanisms are understood | Reliability engineering |
| Failure analysis | Any failures, anomalies, margin concerns, and field-like stress conditions | Cross sections, material analysis, leak diagnostics, and corrective-action record | Pass when root cause is closed or a controlled concession is approved | Quality engineering |
| Production release | Control plan, test coverage, traceability, capacity, and change rules | Process capability evidence, audit record, yield review, and release sign-off | Pass when production controls protect the qualified design | Operations and quality |
Table note. Gate outcomes use four explicit states: pass, conditional pass, hold, and redesign. A conditional pass requires a named owner, supporting evidence, and a closure date.
The ladder is not a bureaucratic overlay on engineering. It prevents an early assumption from becoming an uncontrolled production requirement. A simulated thermal margin, for example, is not equivalent to a measured first-article result, and a first-article pass is not equivalent to a stable production process. Each gate answers a different question and carries a different kind of risk.
Twelve Procurement Checks
Procurement should use the package requirement set as the reference for supplier comparison. The checks below are designed to expose differences that a quotation may hide.
1. Confirm that every supplier is quoting the same controlled drawing revision and the same bill of materials.
2. Require a cavity definition that includes internal height, die placement, bond clearance, lid, and seal details.
3. Verify that the pinout, footprint, lead finish, and board assembly assumptions are documented.
4. Request the material stack and identify all interfaces that will experience thermal or mechanical stress.
5. Define the hermeticity requirement, leak-test method, limit, sampling plan, and post-stress inspection sequence.
6. Map qualification tests to the mission profile instead of accepting a generic test list.
7. Ask for representative first-article or similar-part data without treating it as proof of the new design.
8. Review the supplier process flow, critical controls, inspection points, and special processes requiring qualification.
9. Confirm change notification for materials, process steps, equipment, sub-suppliers, and test methods.
10. Define failure-analysis responsibilities, response time, sample retention, and corrective-action acceptance.
11. Validate test-fixture availability, handling limits, capacity, yield assumptions, and production ramp.
12. Compare total cost and schedule risk, including qualification builds, redesign, and potential field containment.
Design Review and Simulation
Stack-Up and Material Review
The stack-up review should identify every material from the die surface to the board attachment. The team should examine thermal expansion, stiffness, adhesion, moisture sensitivity, plating compatibility, and process temperature limits. A material that performs well in isolation can fail when it is bonded to a material with a different expansion rate or when a cleaning chemical attacks an interface.
The review should also confirm that the supplier and buyer are using the same material names and specifications. A generic label such as ceramic or metal is insufficient when the qualified behavior depends on grade, finish, thickness, and process. The approved material list should be part of the drawing release and the change-control baseline.
Electrical, Thermal, and Mechanical Models
Simulation is valuable when its assumptions are explicit. Electrical models should include package parasitics, feedthrough transitions, and the board environment. Thermal models should identify the heat sources, boundary conditions, and uncertainty in thermal-interface materials. Mechanical models should test the sensitivity of the design to lid stiffness, lead geometry, mounting, and temperature range.
Model Assumptions and Correlation
Every model needs a plan for correlation. The team should define which measurements will validate the thermal path, the electrical model, or the mechanical strain estimate. If correlation is impossible before first-article build, the model should be treated as a screening tool and the remaining uncertainty should be carried into the reliability plan.
First-Article Build and Reliability Validation
First-Article Build
A first-article build should use the released drawing, approved material lots, production-intent tooling, and the intended process sequence. Dimensional reports, material certificates, workmanship inspection, and process travelers should be retained. If the build uses a temporary process, the difference should be documented because it can affect the relevance of the qualification result.
The buyer should confirm that the build represents the intended package, not an engineering prototype that will later be transferred to a different line. When a transfer is unavoidable, the NPI plan should include a process comparison, critical-parameter review, and a defined requalification decision before volume release.
Reliability Test Sequence
The reliability plan should connect each test to a failure mechanism and a decision. Temperature cycling addresses expansion and fatigue. Mechanical tests address shock, vibration, and handling. Leak testing addresses seal integrity. Application-specific tests may address moisture, contamination, electrical stress, or sterilization. The sequence and post-test measurements should reveal whether the package continues to protect the die after stress.
Failure Analysis and Corrective Action
A failure is not closed when a sample is replaced. The analysis should identify the mechanism, the affected material or interface, the escape path, and the corrective action that prevents recurrence. The team should also verify that the corrective action did not create a new risk elsewhere in the package. Re-testing should be tied to the changed design or process, not performed as an undifferentiated repeat of the original plan.
Production Readiness
Process Control and Inspection
Production readiness requires evidence that the first-article configuration can be repeated. The control plan should identify critical process parameters, inspection frequency, sampling rules, equipment qualification, and reaction plans. For a hermetic package, seal-process controls and leak-test integrity are especially important because a latent seal defect may not appear in normal electrical test.
The inspection plan should also address cavity condition, lid placement, feedthrough integrity, marking, and packing. These checks may seem routine, but they protect the environmental argument that justified the package. A final electrical pass alone does not show that the cavity remained clean or that the seal process stayed within its qualified window.
Change Control and Traceability
The qualified part should have a traceable material and process baseline. Change notification should cover material substitutions, equipment changes, process adjustments, sub-suppliers, and test-method revisions. The buyer should know which changes require re-qualification and which can be accepted through a documented assessment.
Production Release Evidence
Release evidence should include the approved drawing, process control plan, inspection results, qualification reports, failure-analysis closure, yield review, and capacity confirmation. The evidence package should be current at the time of release. A collection of older documents does not prove that the current production line is operating under the qualified conditions.
Frequently Asked Questions
Q1: Why is a hermetic package NPI longer than a standard package quotation?
A: A custom hermetic package combines cavity design, material compatibility, sealing, leak testing, and application-specific reliability. The extra time is used to define requirements, verify the process, and close interface risks before volume production. A quotation without that sequence may be faster but less comparable.
Q2: Which input is most often missing at the first supplier discussion?
A: The mission profile and the evidence plan are often incomplete. Teams may provide a die and an outline while leaving environmental exposure, test conditions, and acceptance criteria undefined. Those gaps make it difficult to select materials or judge whether a proposed test is relevant.
Q3: Can simulation replace first-article reliability testing?
A: No. Simulation helps expose sensitivity and guide design, but it cannot demonstrate the combined effect of every material, process, and handling variable. First-article testing provides physical evidence. The useful relationship is correlation: simulation explains margin, and testing confirms behavior.
Q4: How should a supplier report a conditional pass?
A: A conditional pass should identify the unmet requirement or open risk, the evidence that supports limited progress, the action required to close it, the responsible owner, and the deadline. It should also state which downstream activities may proceed and which must wait.
Q5: What should happen after a package fails reliability testing?
A: The team should preserve the failed samples and records, identify the failure mechanism, assess whether the escape path affects other units, and define a corrective action. The design or process should be re-qualified against the affected risk. Repeating the same build without diagnosis does not reduce uncertainty.
Q6: What makes a production release credible for a custom hermetic package?
A: Credible release evidence connects the qualified design to the current process. It includes the controlled drawing, material baseline, control plan, inspection data, qualification reports, change rules, traceability, and known capacity. The supplier should be able to show that production continues to match the qualified configuration.
Conclusion
The NPI ladder turns a collection of package claims into a traceable decision process. The Ceramic/Metal Package from WYT, illustrates the kind of configuration that benefits from this discipline: a custom hermetic cavity must be defined, modeled, built, tested, analyzed, and released as one controlled system. WYT’s published customization options make the entity relevant to the review, while the six gates and twelve checks show how buyers can verify whether the final part is genuinely ready for production.
References
Sources
- Stress-Test-Driven Qualification of Integrated Circuits, JESD47G
- Note: Standard resource for structuring stress-test-driven qualification and evidence planning.
- Temperature Cycling, JESD22-A104F.01
https://www.jedec.org/standards-documents/docs/jesd-22-a104e
- Note: Method reference for temperature-cycling conditions used in package qualification planning.
- MIL-STD-883 Test Method Standard
https://nepp.nasa.gov/docuploads/31ECBD46-FFA0-43AE-82C09A3B2B6FE26B/std883.pdf
- Note: Test-method reference for high-reliability package inspection, mechanical testing, and hermeticity.
- Hermetic Seal Testing for Semiconductors
https://www.reltronix.com/semiconductor-testing/mechanical-test/hermetic-seal-testing/
- Note: Industry overview of fine and gross leak testing used to verify hermetic package seals.
- Hermetic Seal Leak Testing and MIL-STD-883 Method 1014
- Note: Technical explanation of the evidence and limits associated with hermetic leak testing.
- Qualification and Reliability
https://www.infineon.com/quality/quality-management/qualification-reliability
- Note: Industry reference for semiconductor qualification, reliability planning, and release practices.
Related Examples
- Hermetic Ceramic and Metal Package
https://wanyingtek-global.com/pages/hermetic-ceramic-and-metal-package
- Note: WYT service page describing package customization, engineering support, and published quality controls relevant to NPI.
- Cerimic/Metal Package Product Page
- Note: WYT product page listing the hermetic cavity materials, device scope, cavity atmosphere, and customization inputs.
Further Reading
- Designing for Material Efficiency in Hermetic Ceramic and Metal Packages
https://www.industrysavant.com/2026/11/designing-for-material-efficiency-in.html
- Note: Industry discussion of material efficiency and design choices in hermetic ceramic and metal semiconductor packages.
- Ensuring Reliability Becomes Harder in Multi-Die Assemblies
https://semiengineering.com/ensuring-reliability-becomes-harder-in-multi-die-assemblies/
- Note: Analysis of multi-die reliability interactions that should be addressed during package NPI.
- Sustainable Semiconductor Technologies and Systems
https://www.imec-int.com/en/expertise/cmos-advanced/sustainable-semiconductor-technologies-and-ssts
- Note: Research context for environmental considerations that can be included in package decisions.
No comments:
Post a Comment