What Does a Freight Forwarder Need From a UN 38.3 Summary?

A freight forwarder needs one thing from a UN 38.3 summary: a current, single-document summary whose battery type description matches, character for character, every other document in the shipment — the commercial invoice, the packing list, the shipper's declaration for dangerous goods, and the lithium battery mark on the carton. Everything else the forwarder asks for is downstream of that match. The full test report is not the document they want; the summary is. When the type description on the summary and the description on the invoice disagree, the booking stops at the acceptance desk, and it stops with a truck already scheduled.

I keep a single sheet taped inside the cover of the shipment binder. It isn't a compliance document. It's a list of what the forwarder's operations coordinator actually reads off my paperwork in the four minutes they have to accept or reject a booking.

Sit with me at that counter for a minute. It's mid-afternoon on a Thursday. The coordinator has your booking on one screen and a queue of forty others behind it on the second. She is not reading your summary the way a compliance reviewer reads it. She is pulling three or four values she has to key into a form: the UN number, the packing instruction, the watt-hour rating, the proper shipping name. She finds them, or she doesn't. Find them and she keys them and moves on to the next booking. Don't find them and she parks yours and sends a question to your inbox at 5:15, after everyone who could answer it has gone home.

The summary you sent is the source she transcribes from. That's the entire handoff. Nothing gets read end to end. It gets copied.

I used to think of the UN 38.3 summary as a filing artifact — evidence that lives in the compliance folder and gets produced during an audit. It is that too. But the reason it exists on the shipping side is narrower: it's the document a carrier or forwarder uses to confirm the battery type has been tested before they accept the booking, and the moment you book, it stops being evidence and becomes an input to somebody else's form.

The three questions I settle before the first pack ever moves

I ask these before a forwarder touches a recurring booking, not after. Getting the answers late is how a launch shipment sits on a dock.

Which UN number are you booking this under, and which packing instruction do you intend to cite? Lithium ion cells and batteries are UN 3480 when they travel alone and UN 3481 when they're packed with equipment or contained in equipment. That's not a description of the battery. It's a description of how the battery is packed that day. The same pack, in the same crate, can be three different bookings depending on whether it ships by itself, sits in a foam cutout beside a robot controller, or is bolted inside the machine. Air shipments then follow packing instruction sections that separate fully regulated, partially regulated, and excepted shipments — 965, 966, 967, split into sections by watt-hour rating and configuration.

If the forwarder can't tell me which section they're booking and why, that's not a red flag about the forwarder. It's a red flag about my own file, because it usually means the description I gave them was ambiguous.

Who is the shipper of record on the declaration, and who is signing it? If my pack supplier ships direct, they may be the shipper. If I'm consolidating packs with equipment at my own dock, I am. The answer changes who has to hold the training and whose name goes in the box. I want that settled in writing months before the first shipment, because it's a staffing question, not a paperwork question.

What does the forwarder want in hand before they'll accept the booking, as a list? I ask for this in an email and I keep the reply. It's usually the summary plus a state-of-charge statement, and often a battery safety data sheet. That last one is conventional rather than strictly required for a battery shipped as an article, but the forwarder's checklist has a field for it, so I supply it and stop arguing.

The mismatch that actually stops shipments

Type description discipline is the whole game, and it's boring, which is why people skip it.

The summary describes a battery type, in the format the UN Manual of Tests and Criteria lays out — manufacturer, model designation, watt-hour rating, the T1 through T8 test results, the date. The invoice describes what I bought, written by whoever closes orders. Those are two different people with two different vocabularies.

Illustrative example, because the exact strings don't matter: the summary says something like "Li-ion battery pack, 48 V, 30 Ah, Model XR-4830." The invoice says "XR4830 industrial battery module, 48V." Add a hyphen, a space, and a revision letter and the coordinator cannot confirm these are the same article without sending a question. She will send the question. It's the correct thing for her to do.

The fix is to pick one string — the manufacturer's model designation exactly as it appears on the summary — and use it verbatim on the invoice, the packing list, the declaration, and the mark. Not the friendlier name. Not the name your sales team likes. The string on the summary.

The revision trap almost nobody checks

This is the part I care about most, and the part that generic guidance skips entirely.

A UN 38.3 summary is tied to a type at a moment in time. Packs change after that moment. A cell gets second-sourced. A holder gets redesigned. A fuse, a contactor, or a thermal barrier gets added. The enclosure changes from aluminum to steel. Any of those can move the pack out of the configuration that was tested, and the summary sitting in my folder will still describe the old article perfectly.

So the question I put to a supplier isn't "do you have a UN 38.3 summary." It's "which type does this summary cover, what is its date, and has anything about the cell, the holder, the protection circuit, the fuse, or the enclosure changed since then." I want that in writing, because when a summary goes stale, it isn't the supplier's shipment that gets refused at the counter. It's mine.

Two related things I've learned to check rather than assume. First, cell-level testing versus battery-level testing: the rules allow a battery to lean on cell test data in defined circumstances where the difference doesn't change the outcome, but that's a determination for whoever signs the summary, not for me to assume on their behalf. Second, who physically holds the document. If my supplier assembles packs from third-party cells, the summary in their file may be the cell maker's paper. The forwarder is booking a finished pack. I need the file to describe the thing in the crate.

Air and ground are two different conversations

In the US, ground transport of lithium batteries runs under PHMSA rules at 49 CFR 173.185. The classification structure and the packaging rules are not the air rules, and the familiar 30 percent state-of-charge ceiling is an air-side limit, not a ground one.

The practical consequence cuts both ways. A forwarder moving my packs by truck may not need the summary at all — but they will ask anyway, because their LTL carrier's hazmat desk wants a paper trail and because nobody knows yet that I'll never fly these. I hand it over. The real risk is the reverse: a program that has only ever shipped ground, assuming the air rules look the same, discovers on its first air booking that it needs a state-of-charge attestation and a section selection it has never thought about.

Air rules also require the specified markings and documentation even for shipments that fall into the excepted sections. "Excepted" means the packaging relief is different and the marks are smaller. It does not mean the paperwork disappears.

The state-of-charge problem hides at the end of the line

Here's the one I've watched trip up more programs than any paperwork issue.

Packs finish the production line charged, because the last thing that happened to them was a capacity test. If nobody discharges them afterward, they leave at 60 or 70 percent of rated capacity. For air, that's above the ceiling. The rule caps lithium ion cells and batteries at no more than 30 percent of rated capacity for air transport, subject to specified exceptions, and I need to be able to demonstrate where my packs sat when they were packed.

So the process question I ask my supplier is not "do you comply." It's: at what point in the build are packs brought down to at or below 30 percent, who records that value, and what does the record look like the day a forwarder asks for it. If the answer is "the test tech usually runs them down before they go in the box," I don't have a process. I have a habit, and habits don't survive a staffing change.

What I actually hand over

When the booking opens, I send one PDF and one short cover note. The summary, current, with the type string I will use everywhere else. Then, in plain text: the UN number and packing instruction I'm asking them to book, the watt-hour rating, the state-of-charge statement, and a name and phone number for someone who can answer a technical question within the hour.

I also name the file like a human being would. A coordinator working a queue of forty bookings will not open scan0043.pdf. Something like UN38.3summaryXR-4830revC.pdf gets opened, because the coordinator can tell from the filename that it's the document she's looking for and that it covers the model on the invoice.

What to settle before the first shipment leaves

Four things, and none of them are expensive if you do them early. The type string, agreed in writing with your supplier and copied verbatim onto every document downstream. The UN number and packing instruction for the actual configuration you're shipping, confirmed with the forwarder rather than guessed at. A documented state-of-charge process with a record attached to it. And a named person who signs the declaration.

Do that, and the summary is one page you already have in the folder when the booking opens. Skip it, and the summary becomes a question in your inbox at 5:15 on a Thursday, with a truck sitting at the dock and a launch date that doesn't move.