Same part number, new production process. Is the lot approved?
Illustrative image: Unsplash contributor on Pexels
Fictional engineering case. All case measurements and outcomes are constructed for this series; no company or reader is being quoted.
Last week’s answer: assess the startup trajectory
At a constant charging current, the ideal charge time is:
t = CΔV / I = 0.0047 × 48 / 1 = 0.2256 s
That agrees with the observed startup duration. With negligible load current, the MOSFET initially dissipates approximately 48 W. Its dissipation then falls as the capacitor voltage rises.
For an ideal constant 48 V source and a capacitor charging from zero to 48 V, the MOSFET dissipates approximately:
E = ½CV² = 0.5 × 0.0047 × 48² = 5.4144 J, approximately 5.4 J
The capacitor stores the same energy. This simplified result excludes other circuit losses, controller transitions and any load enabled during charging.
Lowering the current lengthens the ideal charging interval; it does not remove this energy. It changes the voltage-current-time trajectory, which may improve or worsen suitability for a particular device.
During that interval the MOSFET operates as a controlled pass element. Its fully enhanced on-resistance is not the relevant description of its instantaneous dissipation. The review needs the simultaneous drain voltage and current, pulse duration, initial temperature and restart conditions.
In the fictional follow-up, the team obtained manufacturer guidance for the measured pulse shape and mounting-base temperature. That review did not establish sufficient safe-operating-area margin for the substitute during startup. This was enough to reject the substitution for this application, but did not by itself prove the microscopic failure mechanism. Failure analysis would be needed before attributing the damaged parts specifically to local thermal instability. The original approval had never assessed that operating region.
The team assessed the complete power-versus-time waveform using the manufacturer’s guidance, rather than treating the ramp as a rectangular pulse with the same energy. A transient thermal calculation included the initial junction temperature and restart history, but was not treated as proof against local thermal instability. The 25°C single-pulse curves were not applied unchanged to a warm, potentially repetitive start.
The team qualified a device intended for the required linear operation and reviewed the controller’s current limit, startup timeout and retry behaviour together. Validation covered maximum input voltage, capacitance tolerance, initial temperature and the permitted restart sequence. A device surviving one cold start would not have closed the investigation.
This week’s challenge: the unchanged order code
A manufacturer receives a product change notification for an analogue interface IC. The supplier is adding an assembly-and-test site and changing the mould compound. The orderable part number, package outline and published electrical limits remain unchanged.
The notice lists the affected ordering codes and a projected first-shipment date. Its qualification summary reports successful tests on representative devices from the package family. The supporting document identifies the test vehicles, but does not clearly explain why they cover this exact device and compound combination.
The IC is used in an industrial controller that operates in a warm, humid environment. The product’s internal validation plan includes biased damp-heat testing and checks of low-level analogue accuracy. Its customer agreement requires assessment of relevant supplier process changes, but does not automatically require complete product requalification for every PCN.
Purchasing has six weeks of confirmed old-process stock. The supplier can provide new-process samples in two weeks. The distributor cannot yet confirm how old and new production will be distinguished on future deliveries.
Engineering has not identified a failure. Production wants a release decision before the next order is placed.
What evidence would justify release?
Which gaps are release blockers, which need clarification and which can be addressed through targeted application testing? How would you establish traceability and avoid mixing stock whose approval status differs?
Propose a proportionate acceptance plan. The task is to justify a decision from the available evidence, rather than assume that an unchanged part number guarantees approval or that every PCN requires a redesign.
The answer follows on 5 November. Share your approach or submit your own sourcing problem through the Troubleshooter page.
