Last-Time-Buy Planning — Protecting a BOM When a Component Reaches EOL

End-of-life planning is a procurement job, not just an engineering footnote #

Every active BOM eventually contains a part someone wants to stop making. The manufacturer issues a notice, the clock starts, and the people who feel it first are not in the design office — they are in procurement, trying to keep a line running after the last order window closes. A last-time-buy (LTB) decision is the single most common way a mature product survives an end-of-life (EOL) event without a redesign, and it is far more than "order a few extra reels."

Done well, an LTB turns a discontinuance notice into a planned, costed buffer that carries the product to its natural end of life. Done badly, it becomes a warehouse of dead stock — or, worse, a surprise line-down because the buffer was sized on a guess. This guide walks through the decisions that separate the two.

What a PCN or PDN actually tells you #

Two notices drive almost every LTB decision, and buyers confuse them constantly:

• A Product Change Notice (PCN) describes a change to a part that stays in

production — a new die revision, a different substrate, a lead-free plating switch. It may force re-qualification, but it is not a stop-ship.

• A Product Discontinuance Notice (PDN), sometimes called a discontinuation

or obsolescence notice, states the part is leaving the catalogue. This is the one that opens the last-time-buy window.

The PDN carries the dates that matter. Extract them deliberately rather than skimming the headline:

The trap is treating the LTB date as the deadline for *having parts in hand*. It is the deadline for *placing the order*. The parts may not arrive for another quarter, and any lot you buy then still has to clear incoming inspection, solderability and date-code checks.

Size the buffer before you open a purchase order #

The core LTB question is a quantity, not a yes/no. The number should come from the product's own demand signal, not from a round multiple of a reel:

``` required LTB quantity ≈ (remaining product life in months ÷ build interval) × monthly build volume × (1 + anticipated yield loss) × (1 + safety buffer) − existing qualified inventory ```

Three inputs decide whether that number is sane:

• Forecast horizon. Buy only as far as the product is genuinely supported.

A five-year buffer on a part whose platform dies in eighteen months is just capital tied up in a soon-to-be-obsolete component.

• Minimum order quantity (MOQ) and standard pack quantity (SPQ). The

economic order may force you above the pure math. Compare the forced surplus against the cost of a later bridge buy or a redesign.

• Shelf life, not just price. Long-tail buys of BGAs and QFNs sit in

storage for years, and moisture-sensitive parts degrade on the shelf. A ten-year stock of a BGA is only useful if its moisture-sensitivity level and date code keep it solderable when you finally need it.

This is where lifecycle stage matters. A mature, stable MCU such as the STM32F103C8T6 rarely vanishes overnight — but its clones and remarked stock do proliferate once the official channel thins, so an LTB here is as much about *source integrity* as quantity. A power stage like the TDA21590, tied to a fast-moving platform, may see its LTB driven by the platform's own sunset rather than the silicon. And a DDR4 device such as the K4A8G165WC-BCTD is now competing for fab capacity against DDR5 and HBM, so its commercial longevity is a portfolio decision at the vendor, not a quality issue — plan the bridge buy earlier than the notice alone suggests.

Qualify the replacement in parallel, not after #

The most expensive LTB is the one that runs out before the substitute is approved. The replacement named in the PDN is rarely a drop-in; it may need a PCB respin, a firmware change, or a full re-qualification. Start that work the week the notice lands, using the same disciplined process you would apply to any alternate part.

The trade-off to put in front of the business is concrete: spend engineering time now to qualify a replacement, or spend procurement budget now to bridge with last-time-buy stock and slip the redesign to the next revision. One of those is almost always cheaper than an unplanned line-down. The mistake is leaving the choice unmade until the buffer is gone.

The pitfalls that turn an LTB into a stockout #

A few patterns cause most failed last-time-buys:

• Underestimating qualification lead time. The replacement is "obvious" on

paper but needs twelve weeks of validation that nobody scheduled.

• Clone and remarked risk on long-tail buys. Large LTB orders of mature

parts attract the aftermarket. Verify traceability and lot consistency on the buffer itself, not just on routine replenishment.

• Ignoring date code and MSL on multi-year stock. A part bought today and

needed in five years must still be solderable then. Bake, dry-pack and record the date code at incoming, and re-check before it enters production.

• Assuming one notice covers the whole BOM. A PDN names specific orderable

codes. The -40°C grade or the tray-pack variant may have a different window than the reel you usually buy.

Where this connects to your sourcing workflow #

An LTB is rarely a single-part event. It usually arrives inside a broader supply signal — a vendor price action, an allocation, or a platform sunset that touches several lines at once. Tracking notices against the live BOM, rather than reacting to each email, is what keeps the buffer buys coordinated instead of frantic. When a discontinuance does hit a critical item, a clear availability and lot-verification path is the fastest way back to a known-good state.

Conclusion #

Last-time-buy planning answers one question: *how much of a discontinued part do we need to carry the product safely to its end, and what replaces it when that runs out?* Read the PDN for its LTB and last-ship dates, size the buffer from real demand and shelf-life limits rather than a reel multiple, qualify the migration part in parallel, and verify the buffer's date code and traceability before it sits in storage for years. Treat the LTB as a planned, costed decision — not a panic order placed on the deadline — and an EOL notice becomes a manageable event instead of a line-down.

Sources:

• Manufacturer PCN/PDN and last-time-buy notices published on vendor product pages (read the LTB and last-ship dates per affected orderable code).
• Internal LimChip EOL and lifecycle coverage, e.g. the second-source qualification guide and recent discontinuance notices in /news/.