Your UN 38.3 certificate may not cover the pack you are submitting for PPAP

Attach the certificates, but verify their scope before the packet leaves your desk. The rejection that comes back on a pack PPAP is almost never a missing document. It is a UN 38.3 test summary or a UL listing whose model list, cell model, or series-parallel configuration does not match the part number on the warrant. Before anything gets signed, put each certificate next to the PSW and reconcile four things line by line: the model designation, the cell inside the pack, the pack construction, and the electrical ratings. If any one of those four drifts, you do not have coverage, and no cover letter fixes it.

I lay the signed warrant flat on the left side of the desk and the supplier's document pack on the right, and I read them against each other. That is the entire method. It is not clever, but it catches what a checklist cannot, because a checklist only asks whether a certificate exists. It never asks whether the certificate is about the thing you are shipping.

The first thing that trips people up is the language itself. There is no such thing as a UN 38.3 certificate.

What UN 38.3 actually gives you

No accredited body issues a UN 38.3 certificate. What exists is a test report from a lab, and a test summary extracted from that report, structured the way the UN Manual of Tests and Criteria lays it out. The summary has a defined shape: cell model number, battery model number, a physical description, mass, watt-hour rating, the list of tests with results, the facility that ran them, and a reference back to the full report.

That structure is your first scope check, and it is a fast one. Cells get T1 through T6 and T8. Batteries get T1 through T5, T7 and T8. If the summary in front of you lists only the cell model and the results stop where cell testing stops, you are holding cell-level evidence. A pack-level PPAP needs pack-level evidence. They are not interchangeable, and a supplier who sends the cell summary is usually not being evasive — they genuinely think the cell approval carries up to the pack.

It does not. A pack adds a protection circuit, a housing, connectors, and a series-parallel arrangement that did not exist at the cell level. Those are exactly the things a pack-level test is designed to stress.

Stop calling it a certificate in your own documents

Small thing, real consequence. If your PPAP appendix index says "UN 38.3 Certificate" and what is behind the tab is a test summary, an auditor or a customer SQE who knows the difference immediately discounts the rest of your packet. Write "UN 38.3 Test Summary" and reference the report number. Precision here buys you credibility you will need later when something harder comes up.

UL 2054, UL 62133-2, and IEC 62133-2 are not the same document

These three get requested interchangeably in RFQs and they are not interchangeable in a submission.

UL 2054 evaluates the pack as a product — protection circuit, enclosure, abuse behavior at the pack level. UL 1642 is the cell-only lithium standard and gets confused with it constantly. UL 62133-2 is the US adoption of IEC 62133-2 and covers lithium-ion cells and batteries under a harmonized framework; the IEC 62133-2 CB test certificate is the same technical content routed through the IECEE CB Scheme, with national differences attached for the countries where the product will be registered.

So when the customer's requirement document says UL 2054 and the supplier sends a UL 62133-2 listing, you have a decision to make that is not yours alone. Some SQE functions will accept it as equivalent coverage for lithium-ion. Some will not, and will bounce it. Ask which one is contractual before you build the packet, not after. And note that the cell inside the pack may be covered under a completely different certificate held by a completely different company than the one who assembled the pack.

That last point deserves its own sentence. The certificate holder on a pack listing is sometimes the cell maker, sometimes the pack assembler, sometimes a distributor. Check that the factory named in your PPAP is the factory named on the listing, or that it appears under the authorization to mark. A listing that exists but does not authorize your factory to apply the mark is a listing you cannot use.

Reading the model list without guessing

A UL listing's model designation field is often a base number followed by a run of suffixes. That is not enough information on its own. Request two more documents from the certificate holder:

The model designation procedure, which is the holder's own written rule for how each suffix maps to a configuration. It is the difference between seeing "ABC-123-A" on the list and knowing that the -A means a different connector versus a different cell.

The construction or ratings table that sits behind the certificate, showing the cell model, cell count, series-parallel arrangement, and electrical ratings the listing was evaluated against. Compare that table to your PPAP part drawing, not to the supplier's cover email.

I do this comparison with a highlighter and a pencil, physically, because screen-switching lets me skim. The cell model line is where most of the drift shows up. A supplier swaps a cell for a second source, the pack part number does not change, the PPAP does not get revised, and suddenly the listing is describing a pack you are no longer building.

When the certificate covers a pack that is almost yours

This is the case that eats the most time, and it is not a paperwork problem — it is a real engineering question wearing paperwork clothes.

You have a listing for a pack with the same cell, the same protection circuit topology, and the same chemistry, but a different series count, a different housing, or a different connector. Everyone in the supply chain tells you it is basically the same and the certificate covers it. Sometimes that is true. Standards allow certain changes to stay within a tested construction, and a competent certifier will tell you where that line sits for a specific configuration.

Here is the part I would underline. That determination belongs to the certifier and the lab, not to your supplier's sales engineer and not to you. Your job is to obtain the answer in writing, not to reason your way to it. Ask the certificate holder for a statement that the exact configuration on the warrant — spelled out, with the series-parallel count and the cell model — is covered by the document being submitted. If they hesitate, that hesitation is your answer, and you should be pricing a delta qualification instead of arguing.

Where the evidence goes in the packet

PPAP is process evidence: PFMEA, control plan, MSA, capability, dimensional results. Certificate scope is design and type-approval evidence. Those two things should not be stapled together, and mixing them is how a reviewer loses confidence in the packet.

What I build now is a certification index page that sits at the front of the appendix, behind the PSW. One row per document. Document type. Certificate or summary holder. Document number. Revision. Issue date. Models covered. PPAP part number. A short mapping note explaining how the covered model relates to the part number on the warrant.

That mapping note is the whole page. When a reviewer two time zones away opens your packet at eleven at night, the note is what lets them close the loop in thirty seconds instead of emailing you. It is also the thing that saves you when the person who built the packet has left the program.

Watch the revision column. An older revision is not automatically invalid — but if a newer revision exists, find out what changed. Sometimes the change is an address update. Sometimes it is the addition of the exact model you are claiming. You need to know which.

Certification before approval, or after

It depends on the market and on the contract, and the two are not the same thing.

Where the safety mark is a condition of putting the product on the market, certification is a gate. No certificate, no PPAP approval, and no amount of interim paperwork changes that. Take the date seriously and work backward from it, because lab queues are the long pole and they are not something you can compress by escalating.

Where the certificate is contractual rather than a market-access requirement, an interim PPAP is a legitimate mechanism, and I have used it. It requires a dated commitment, a written containment plan, and the customer's explicit agreement recorded on the warrant. What it does not tolerate is a placeholder date that everyone knows is fiction. If the interim expires and the certificate still is not there, you have converted a manageable gap into a supplier quality escape, and that follows the part number for a long time.

What I actually check now, every time

I open the summary and the listing, and I read the model designation out loud against the part number on the warrant. Out loud, because reading silently lets my eye fill in the suffix that should be there. Then the cell model against the drawing. Then the series-parallel count. Then the ratings. Then I check the factory name on the listing against the factory name in the PPAP. Then I check the revision date against the last change the supplier told me about.

Six checks, maybe ten minutes. It is the cheapest insurance in the whole submission, and it is the one step that gets skipped because the certificate arrived as an attachment and looked official. Official-looking is the problem. A document can be genuine, current, and completely about a different pack than the one on your warrant.