The manual is the one item on every pre-testing document list that reliably gets pushed back. The samples are packed, the slot on the schedule is booked, and the answer is: it is still with marketing, it is not final, can we send it later? Testing can usually start anyway - that is not the problem. The problem is that the manual is not just paperwork you owe the laboratory. It is a document the test report and the technical file will quote back at you, and the version question comes back months later, usually while a buyer is reviewing your documents.
This article explains what a laboratory actually reads in your manual, why the report references it, and what happens in practice when the manual you sent is not the manual that ends up in the box.
What the laboratory actually reads in your manual
Nobody at the laboratory reads your manual from cover to cover. Specific sections get used directly, and each one feeds something downstream:
- Model numbers and variants. The model list on the report is built from what you declare, and the manual is part of that declaration. When the manual, the samples and the enquiry each show different model numbers, resolving which product was actually tested becomes the project's first unplanned task.
- Ratings and power source. Rated input, the external adapter model and its output rating, battery type and capacity. The EMC and safety setups are configured from these - a different adapter rating means the setup was built for a different product than the one shipping.
- What is in the box. Every accessory - adapter, cable, cradle, remote - is part of the tested configuration and appears in the report's list of ancillary equipment. An accessory that never reached the laboratory is an accessory that was never tested.
- Intended use. Indoor or outdoor, consumer or professional, accessible to children. This decides which standard applies and, inside the standard, which clauses matter.
- Operating modes and how to enable them. EMC testing hunts for the worst case. If the manual describes a mode nobody mentioned in the enquiry - a second radio mode, a high-power mode - the scope can grow mid-project.
- Warnings and safety instructions. In hazard-based safety standards such as EN IEC 62368-1, user instructions are part of the safeguards the design relies on. A warning in the manual is sometimes the reason a given risk is considered controlled - which makes deleting it at print time a technical change, not a text edit.
- Symbols and their explanations. The symbol explanation page is what market surveillance compares against the markings on the product and the packaging.
- Languages. For consumer products placed on the EU market, the General Product Safety Regulation requires instructions and safety information in a language consumers in each member state can easily understand. An English-only manual is a real rejection reason in parts of the EU, not a formality.
Why the report quotes your manual back at you
A test report does not describe "a product" in the abstract. It describes one configuration: this model, this power supply, these accessories, this firmware version. Engineers write that description from the samples and the manual together, and the manual is cited as its source. This is the whole reason the version matters - the report does not update itself when your manual does.
Part of this is hard rule, not just laboratory habit. For equipment authorized under the FCC's Part 15 rules, specific statements are required in the user's manual - the compliance statement and, for digital devices, the interference information to the user. A manual missing those statements fails the paperwork review even when the hardware tested clean. On the EU side, the technical file behind the Declaration of Conformity references the manual as part of the product description, so a DoC signed over a file that cites a manual you never shipped is a DoC with a hole in it.
How the wrong version actually shows up
Four situations cause most manual-version problems. None of them involve anyone doing anything dishonest - they are all normal business decisions made without telling the laboratory.
The adapter was changed two weeks before shipment
The samples were tested with a 5 V, 2 A adapter. Before mass production, procurement finds a cheaper adapter from a different supplier - same output rating, different model - and the manual is updated to list the new one. Now the report, the manual and the physical product each describe a different adapter. A buyer's compliance check or a marketplace review finds the mismatch, and the question becomes whether the product was ever tested with the adapter it ships with. Sometimes the honest answer is no, and the affected tests have to run again on the new configuration.
An accessory was added at the last minute
A charging cable, a regional plug type or a carrying case gets added to the box after testing, because sales asked for it. The accessories list on the report came from the manual sent to the laboratory, so the new item appears on no document the laboratory ever saw. The fix ranges from a file correction to extending the report's configuration - cheap only if someone mentions it before the report is issued.
The model was renamed for a customer
An OEM project gets its own model number, or a customer requests a rebrand. The report still carries the original model; the listing and the printed manual carry the new one. Buyers and marketplaces check that the model on the report matches the model being sold - this mismatch is one of the most common rejection reasons, and the product itself was perfectly compliant.
The warnings were trimmed at print time
The final artwork is tight for space, so marketing shortens or removes warnings - a "do not cover during use" note, a battery warning, a temperature statement. If a warning was part of the safety reasoning in the technical file, removing it breaks that reasoning without anyone touching the product. This is also the change least likely to be reported to the laboratory, because it looks like editing.
Draft or final: what is acceptable at which stage
A draft manual is fine for starting. Mark it clearly as a draft, and expect one question back: which parts might still change? What cannot survive is silence between versions. The content has to be frozen before the report is issued, and the freeze list is short:
- Model numbers, covering every variant you intend to sell
- Ratings, and the adapter or battery model shipped in the box
- The complete accessories list
- Warnings and safety statements
- Languages, for markets with language requirements
Two habits prevent almost all of this. First, version the files - a filename with a version and date beats one called "manual_final_v3_NEW.pdf". Second, when the content does change, send the updated PDF to the laboratory the same week, even if the project is already finished. A correction while the project is open is routine; a correction discovered later by your buyer is a negotiation.
Getting the manual right before printing
The cheapest moment to fix a manual is before it goes to print - before the FCC statements, language coverage and rating lists are locked into artwork. Manual review is part of the document review we run at the start of every project: the testing and certification services page shows how that fits into a submission, and our laboratory in Shenzhen works with these documents on export projects daily. If you want the full pre-submission set rather than the manual alone, we covered the complete document list a laboratory works with separately - the manual is one row in that table, and this article is about why its version line matters more than most.
If your manual is still moving between versions, send the current draft together with the specification sheet before the samples ship.
We will flag what has to be frozen before the report is issued - model numbers, adapter and accessories, warnings, languages - and what can still change freely, so the manual does not quietly break the file two weeks before shipment.