Bare HDI Board Shelf Life: Who Declares Solderability at Handover

Surface finish is a specification. Solderability is a state.

Almost every HDI purchase order names a surface finish. Very few name a date.

That asymmetry is the whole problem. A finish type is a specification: it is either on the board or it is not, and it does not change after the board ships. Solderability is something else — a condition of the outermost few atomic layers of that finish, which interacts with air, humidity, handling and time from the moment the board leaves the line. It degrades. It degrades at a rate that depends on which finish was applied, how the board was packed, where it sat, and how often someone opened the bag to count parts.

So the PO documents a permanent attribute and stays silent on a perishable one. When boards solder cleanly, nobody notices. When they do not, both sides discover simultaneously that no document in the transaction says how long the boards were supposed to stay solderable, when that period began, or who was supposed to be watching it.

The dispute that follows is unresolvable by design, not by bad faith. The fabricator's evidence ends at outgoing inspection: the boards were solderable when they were packed. The assembler's evidence begins at the reflow oven: the boards did not wet properly. Between those two points lies an undocumented interval — shipping, customs, a warehouse shelf, a kitting bench, possibly a bag that was opened three weeks ago for a first-article build — and neither party holds a record of it.

Nothing about this is specific to HDI boards in principle. It is sharper on HDI in practice, for a reason worth stating plainly: HDI boards carry a higher density of small pads, filled and capped vias, and fine-pitch features where a marginal surface has less margin to absorb. The same degradation that produces a cosmetic complaint on a coarse board can produce an open joint on a fine-pitch array. HDI boards are also more expensive per piece, which means the question of who pays for the rework is a larger question.

The window is built from four dates and two conditions

A shelf life claim is only provable if six things are recorded. Most purchase orders record none of them.

The four dates:

1. Manufacture or finish completion date — when the surface finish was applied. This is the earliest possible start of degradation, but it is not necessarily the start of the agreed window.
2. Packing date — when the boards were sealed into their delivered packaging. For vacuum-sealed, desiccant-bagged shipments, this is normally the date the window should run from, because the sealed environment is what the shelf life claim is about.
3. Receipt date — when the buyer took physical possession. Needed to separate transit exposure from storage exposure.
4. Opening date — when the seal was first broken, and by whom. This is the single most frequently missing record in the entire chain, and the one that most often decides the dispute.

The two conditions:

1. Storage environment — what the boards were kept in, on both sides of the handover. A sealed bag in a climate-controlled stockroom and a sealed bag on a shelf above a loading door are not the same storage.
2. Packaging state over time — whether the original seal was intact, broken and resealed, or left open. A bag that was opened and resealed is not in its original state, and whatever shelf life was claimed for the sealed condition no longer describes it.

Drop any one of these and the window stops being provable. Drop the opening date and the window becomes unprovable in the most expensive way possible, because the boards may well have been fine on arrival and compromised afterwards — with no way to show it.

Note what this list does not include: a number of days. That is deliberate. The duration itself is a matter for the specific finish, the specific packaging and the written agreement between the two parties, and reasonable figures circulate in industry practice and in published standards. What this article argues is narrower and more durable: whatever duration you agree, it is only enforceable if these six items are recorded and attached to the shipment.

Packaging determines which date the clock starts on

The packaging is not protective wrapping around the deliverable. On a shelf life question, the packaging *is* part of the deliverable, because it defines the environment the claim applies to.

Three elements carry meaning, and each should be named in the RFQ rather than assumed:

The barrier bag. A sealed moisture-barrier bag establishes a defined internal environment. A shelf life figure quoted for sealed storage describes the boards inside that bag, not boards in general. If the boards arrive in ordinary anti-static bags, shrink wrap, or an unsealed box, any sealed-storage figure in the quote has no object to apply to — and nobody notices this until much later.

The desiccant. Desiccant inside the bag is what keeps the internal environment from drifting toward ambient. Its presence, and the fact that it was fresh at packing, is part of what makes a sealed-condition claim meaningful. Its absence is not a cosmetic difference.

The humidity indicator card. This is the only element of the package that reports the actual state of the contents rather than the intent of the packer. A card that reads within its acceptable range at opening is affirmative evidence that the barrier held through transit. A card reading beyond its range is affirmative evidence that it did not — which is exactly the kind of fact the dispute above lacked. It is also the cheapest evidence in the entire chain, and it is routinely thrown away without being read or photographed.

From these three elements follows the practical rule about which date the clock starts on. If the shipment is sealed, desiccated and indicator-carded, the window should run from the packing date, because that is when the environment being claimed came into existence. If the shipment is not sealed, the window cannot honestly run from packing, because the boards were in ambient conditions from the start — and in that case the agreement needs to say what period applies to ambient-stored boards, which is a different and shorter conversation.

Say which of these two situations you are buying in the RFQ. Suppliers can meet either requirement; they cannot guess which one you assumed.

After the seal is broken, the question changes

This is where responsibility actually moves, and it moves in a way that surprises buyers.

Once the seal is broken, the supplier's sealed-condition claim has been consumed. It described boards in a sealed bag, and the boards are no longer in a sealed bag. Whatever happens from that point is governed by how the opening party handled them — not by the original shelf life figure.

That creates three distinct situations, which need to be distinguished before they occur rather than after:

Opened by the buyer, for the intended build. Normal and expected. What matters is that the opening date is recorded, the indicator card was read, and the boards went into assembly within whatever post-opening period the parties agreed. A post-opening period is a separate figure from the sealed-storage period and must be stated separately; buyers often assume the long sealed figure continues to apply after opening, which it does not.

Opened by the buyer, then resealed for later use. This is the case that produces the worst disputes, because the boards look like sealed-condition boards while no longer being sealed-condition boards. Resealing is a legitimate practice, but it transfers the claim: the resealing party is now the party asserting the internal environment, and the record needs to show who resealed, when, with what desiccant, and after how much ambient exposure. If that record does not exist, the resealed bag is evidence of nothing.

Opened by a third party. A contract assembler, a freight forwarder doing an inspection, a customs examination. The boards have been through an environment neither the buyer nor the supplier controlled or documented. Decide in advance who is responsible for capturing that event, because by default nobody is.

The underlying principle is simple enough to write into a PO in one sentence: the party that opens the package assumes responsibility for the condition of the contents from that moment, and must record the event for the responsibility to be traceable at all.

Three routes for stale boards, and what each one needs agreed first

When boards are past the agreed window — or when nobody can show whether they are — there are only three outcomes. Each requires a prior agreement to be usable. None of them can be improvised at the moment of discovery, which is when they are usually attempted.

Route 1: Verify before use. Test the solderability of a sample and decide on the result. This is the most defensible route and the one most often skipped, because it requires having agreed in advance on three things: what test method applies, what sample size is representative for the lot, and who pays for the test. Note also that solderability testing consumes the samples. If the order quantity has no allowance for destructive sampling, this route is not available no matter how sensible it sounds — which is a reason to think about sampling allowance at RFQ time rather than at the point of dispute.

Route 2: Rework the surface. Some finishes can be stripped and reapplied; some cannot, and on HDI boards the decision is constrained by the structure, not only by the finish. Boards with filled and capped vias, fine-pitch arrays or tight annular rings have less tolerance for additional process exposure. Two things must be settled before this route is chosen: whether the fabricator considers the specific build reworkable at all, and whether a reworked board still carries the original acceptance criteria and the original warranty terms. A reworked board is not automatically the board you ordered, and silence on that point creates a second dispute on top of the first.

Route 3: Do not accept. Clean, but only available if the agreement defines what evidence triggers it and within what period after receipt the buyer may invoke it. A rejection raised months after delivery, with no opening record and no indicator card, is a negotiation rather than a contractual right.

The common thread is that all three routes depend on facts that had to be captured earlier. By the time a board fails to wet in the oven, the cheap evidence is already gone.

What to add to the RFQ, and who fills each field

The useful output of all of this is a short list of declaration fields. Split them by who is responsible for each, because that is what makes them actionable rather than decorative.

Buyer states in the RFQ:

• The surface finish required, and whether a specific solderability period is required with it.
• Which packaging form is required: sealed moisture-barrier bag with desiccant and indicator card, or standard packaging.
• The intended storage conditions at the buyer's site, stated honestly. A requirement written against conditions you cannot provide is worse than no requirement.
• The expected interval between receipt and assembly, including whether boards will be held for a staged or repeat build rather than consumed immediately.
• Whether destructive solderability sampling is permitted, and whether the order quantity includes an allowance for it.

Supplier declares with the shipment:

• Finish completion date and packing date.
• Packaging configuration as shipped: bag type, desiccant, indicator card present or not.
• The storage period the declaration applies to, and explicitly the condition it assumes — sealed or ambient.
• The standard or internal procedure the declaration is made against, named rather than implied.
• Lot identification that ties this declaration to the boards in this box, so the declaration remains attached to the goods rather than to the paperwork.

Both parties agree in writing:

• When the window starts: finish date, packing date, or receipt date. One of the three, named.
• What period applies after the seal is broken.
• Who may open, reseal and re-declare, and what record that requires.
• What happens to stale boards: which of the three routes applies by default, who pays, and what the decision timeline is.
• What evidence is required to invoke a claim, and the period after receipt within which it may be raised.

Nothing on this list requires a supplier to accept unusual liability, which is why it is normally straightforward to get agreed. It requires both sides to write down what they were each already assuming — and assuming differently.

Five checks to complete on the day the boards arrive

All five are free, take minutes, and are impossible to perform retroactively.

1. Photograph the sealed package before opening it. Bag intact, label legible, in the condition it arrived. This single photograph resolves more disputes than any other record, because it is the only proof that the barrier was intact at handover.
2. Read and photograph the humidity indicator card at the moment of opening. Not afterwards. The card is a live reading of the package interior and it starts changing as soon as air reaches it.
3. Record who opened it, when, and why. A line in the receiving log is sufficient. This is the field whose absence decides disputes against the party that lacks it.
4. Check that the supplier's declaration is physically attached to the lot and that its lot identification matches the boards in front of you. A declaration that cannot be tied to specific boards protects nobody.
5. If the boards will not go into assembly immediately, decide the storage condition now and record the decision. Resealed with fresh desiccant, or held in controlled storage, or accepted as ambient-stored with a shorter usable period. Making this an explicit decision on day one is what keeps the question answerable in week six.

None of these five steps requires equipment, a standard, or a negotiation. They require someone at goods-in to have been told that they matter.

What this changes in practice

The reframing worth keeping is this: a surface finish is something you specify, while solderability is something you have to keep provable. The first lives in the PO. The second lives in a short chain of dates, conditions and signatures that has to be deliberately created, because no part of the standard purchase process creates it automatically.

Buyers who add half a dozen declaration fields to the RFQ and five checks at goods-in do not eliminate stale-board problems. They convert them from unresolvable disputes into ordinary commercial decisions with a known owner — which is the only outcome available, since the underlying chemistry does not negotiate.

If you are preparing an HDI order and want the declaration fields reviewed against your build and your storage reality before the RFQ goes out, send the fabrication data and the intended assembly timing to our engineering team through the HDI PCB page or the quote request form, and ask for the handover declaration items to be confirmed in writing with the quotation.

FAQ

Does a longer shelf life figure from a supplier mean better boards?

Not by itself. A figure only means something once you know what condition it applies to — sealed with desiccant, or ambient — and what evidence will accompany the shipment to show that condition was maintained. A shorter figure attached to a packing date, an indicator card and a lot-matched declaration is worth more than a longer figure attached to nothing.

Our boards are still in the original sealed bag. Does the clock stop?

Only if the bag is genuinely still in its original sealed state and the agreement runs the window from the packing date. The clock does not stop; sealed storage slows what the clock is measuring. If the bag was opened and resealed at any point, the original claim no longer describes the contents and the resealing party becomes the one asserting the internal condition.

Who pays when boards fail to solder and nobody recorded the opening date?

In practice this resolves by negotiation rather than by right, which is a bad position for both parties. The missing opening record is what removes the possibility of attribution — the degradation may have happened in transit or on your shelf, and neither side can demonstrate which. This is the specific reason the opening date belongs in the receiving log.

Can we just test the boards instead of tracking all this paperwork?

You can, and testing is the most defensible route when the window is uncertain. But solderability testing consumes samples, needs an agreed method and sample size, and needs someone to have agreed to pay for it. If the order quantity has no destructive-sampling allowance, the test is not available when you need it. Testing is a good backstop; it is a poor substitute for recording six fields.

Is this different for HDI boards than for conventional boards?

The mechanism is the same. The consequences are not. HDI boards concentrate small pads, filled and capped vias and fine-pitch features, so a marginal surface has less margin to absorb before it produces an actual open joint rather than a cosmetic complaint. Per-piece value is also higher, which makes the rework-ownership question larger. The practical conclusion is to treat the declaration fields as mandatory on HDI orders rather than optional.

Should the window run from the manufacturing date or the packing date?

Pick one and write it down — that matters more than which one you pick. For sealed, desiccated shipments the packing date is the more defensible choice, because the sealed environment is what the claim describes and it did not exist before packing. For unsealed shipments neither date supports a long window, and the agreement should say what shorter period applies instead.

Related PCB168 reading

• Moisture-sensitive device handling: what the PO must declare — the same logic applied to components rather than bare boards.
• PCBA packaging, ESD and transit protection in the PO — protection requirements for assembled product.
• HDI incoming acceptance: attributes you cannot verify on receipt — which characteristics need a supplier declaration because inspection cannot reach them.
• Shipment document package, CoC and lot traceability — how declarations stay attached to the goods.
• HDI PCB capabilities and manufacturing capabilities.

Sources and basis

• Published industry standards are referenced by name and scope only: the IPC/J-STD family addresses solderability test methods and requirements, and IPC-1601 addresses printed board handling and storage. Specific limits, durations and clause numbers are not reproduced here; consult the current revisions directly.
• No PCB168 storage periods, warehouse conditions, packaging specifications, certifications, equipment, yields, lead times or customer cases are claimed in this article. All durations and conditions are stated as matters for written agreement between buyer and supplier.
• Demand evidence: Google Search Console, pcb168.com, trailing 90 days, query-level impressions for the hdi pcb cluster.

Self-check

• Covers the handover declaration question only; does not restate HDI process, cost drivers, acceptance testing or incoming inspection scope covered by existing PCB168 articles.
• No numeric claims about PCB168 capability, storage, packaging or timing.
• Internal links verified present in out/en/ build output; locale-correct.
• FAQ answers are consistent with body sections and add no new numeric claims.
• Keyword usage is natural; no stacking of hdi pcb variants.

---