A purchase order arrives with one line of technical instruction: Test to IEC 62368-1 and EN 55032, Class B. The product is an LED panel with an external driver. Both standards in that line are the wrong starting point - not because the buyer is careless, but because they copied the sentence from the previous product they sourced, a Bluetooth speaker. If you accept the line as written, you will pay for tests that do not apply and still arrive at the customs or marketplace check without the ones that do.
Standard selection is the least visible part of a testing project and the one that most often goes wrong before any measurement starts. It is not a lookup table. Six inputs decide it, and only two of them come from the laboratory.
The six inputs that decide the standard set
| Input | What it decides | Worked example |
|---|---|---|
| Market and legal route | Which rulebook the file has to satisfy | EU: harmonised standards cited in the Official Journal under the EMC, Low Voltage or Radio Equipment Directive. US: FCC Part 15. The two systems do not accept each other's documents. |
| Product function | Which product family standard applies | A luminaire is tested for emissions to CISPR 15, not CISPR 32, even if it contains a switching driver and a microcontroller. |
| Ports and interfaces | Which clauses inside that standard apply | An AC inlet brings conducted emissions plus harmonics and flicker. A sealed plastic enclosure with no cables brings radiated emissions and ESD, and nothing on the mains port. |
| Configuration and modes | Which operating states have to be measured | A device with Wi-Fi, Bluetooth and a 433 MHz remote has three radio configurations, not one, and the worst case has to be identified rather than assumed. |
| Edition in force at the assessment date | Which amendment is cited in the report | EN 55032:2015 with its amendments, not the 2012 edition that many older checklists still name. |
| The buyer's contract | Additional, narrower or older scope | A retailer's own specification, a marketplace template, or an OEM agreement that names a specific edition and a specific test house. |
The first input is legal, the second and third are technical, the fourth is engineering judgement, the fifth is administrative but decisive, and the sixth is commercial. A laboratory controls the third and fourth and advises on the rest. It does not get to invent the first, and it cannot ignore the sixth.
Product family standards sit above generic standards
Both the EMC and the safety systems are built as a hierarchy. A product family standard covers a defined group of products and takes precedence. A generic standard applies only when no family standard covers the product. Getting this order wrong is the most common reason a report passes but is useless.
| Product type | EMC standard | Safety standard |
|---|---|---|
| Household appliances, electric tools | CISPR 14-1 / EN 55014-1 | IEC 60335-1 with the relevant part 2 |
| Lighting equipment and luminaires | CISPR 15 / EN 55015 | IEC 60598-1; driver to IEC 61347-1 and the relevant part 2 |
| Information technology, audio/video, communication | CISPR 32 / EN 55032 | IEC 62368-1 |
| Industrial, scientific and medical equipment | CISPR 11 / EN 55011 | Depends on the equipment type; medical devices follow IEC 60601-1 |
| Stand-alone power supplies and adapters | Assessed with the end product, or as a separate product | IEC 61558-1 and the part 2 series, or IEC 62368-1 where it powers IT/AV equipment |
| Anything not covered above | Generic standards IEC 61000-6-3 (emission, residential) or 61000-6-4 (emission, industrial) | No generic safety fallback exists; the applicable product standard has to be identified |
Two consequences follow. First, limits differ between family standards, so a pass against the wrong one proves nothing - CISPR 14-1 and CISPR 32 do not share limits, and neither shares them with CISPR 15. Second, an EMC file assembled from generic standards is a fallback, not a shortcut: a buyer or a market surveillance officer who notices that a family standard existed will ask why it was not used.
Why the same product comes back with two different lists
This happens routinely, and it is not usually a mistake. Three mechanisms produce it.
- Classification. A kitchen appliance with a colour touchscreen and Wi-Fi can be argued as a household appliance or as multimedia equipment. Its primary function decides, not its components - and a lab that starts from the display may classify it differently from one that starts from the heating element.
- Visible ports. A USB connector used only for factory service, a detachable antenna connector, a DC barrel jack that will be fed by an adapter in the box: each changes the clause list. If the sample arrives without the adapter that ships with it, the lab tests the configuration it can see, and the report describes that configuration.
- Accreditation scope. A laboratory proposes what its accreditation covers. If its scope lists CISPR 32 but not CISPR 14-1, the quotation will lean toward the standard it can issue an accredited report against. That is a legitimate commercial behaviour and a reason to check the scope before comparing prices - the same check described in how to choose a testing laboratory.
When two quotations cite different standards, the useful question is not which laboratory is cheaper. It is which classification each one used, and which of them the buyer, the importer or the marketplace will accept. That comparison is what a scope-by-scope quote comparison is for.
The edition matters as much as the number
A standard number without an edition is an incomplete instruction. EN 55032 has been through a 2012 edition and a 2015 edition with subsequent amendments, and the two are not interchangeable documents. The same is true across the safety standards, where EN 60950-1 and EN 60065 were replaced by EN 62368-1; the presumption of conformity for the two older standards ended on 20 December 2020.
Under the EU system, a harmonised standard gives presumption of conformity only in the version cited in the Official Journal, and each citation carries a date on which the presumption for the superseded version ceases. That date, not the publication date of the standard, is what a compliance manager is checking when they reject your file. A report issued against a superseded edition is not fraudulent - it accurately records what was measured - but it no longer carries the presumption, and increasingly it is rejected at the first review.
This is also where a buyer's template can quietly date the whole project. A checklist written in 2018 and never revised will name editions that have since lost their presumption. Testing to it because the customer asked is a commercial decision; testing to it without telling them is a compliance problem waiting for the first audit.
Class A or Class B is a separate decision from the standard
EN 55032 classifies equipment by intended environment, not by price or by convenience. Class B applies to equipment intended for use in residential environments. Class A applies to everything else, and its instructions must carry the Class A warning that the equipment may cause radio interference in a residential installation. The limits differ, and Class A limits are the looser of the two.
The mismatch that shows up most often is a consumer product tested to Class A because the prototype only passed at those limits. The file is internally consistent and the label is legal, but the intended use is a living room, and the warning statement in the manual contradicts the marketing. Buyers who read manuals notice this.
Four situations that arrive every week
The list belongs to a previous product. An LED panel quoted against IEC 62368-1 and EN 55032 because the buyer's last project was a speaker. The correct set is IEC 60598-1 with IEC 61347 for the driver, and CISPR 15 for emissions. The quotation changes by more than the price: different standards mean different setups, different sample quantities and a different bench schedule.
Only one radio mode is listed. The buyer's technical file names EN 300 328 for the 2.4 GHz link, and nothing for the 433 MHz remote control that also ships. Under the Radio Equipment Directive the spectrum requirement in Article 3(2) applies to every transmitter in the product, and sub-GHz operation is assessed against a different standard (EN 300 220). Testing one mode and declaring the product compliant is the kind of gap that surfaces at a customs check or in a buyer's technical review, not at the bench.
The standard set was fixed before the adapter changed. A BOM frozen with a 12 V/1 A adapter, samples arriving with a 12 V/2 A unit from a second supplier. The adapter is a separate product with its own safety file, and the conducted emissions measurement was made on the original. Nothing about the test standard changed; the tested configuration did, and the report now describes hardware that is not shipping.
The checklist cites an edition that has been superseded. Common with safety files carried over from an earlier product generation, and with marketplace templates that were never updated. The fix is not to argue about it in the report. It is to agree in writing, before the bench is booked, whether the file is being built to the current edition or deliberately to the older one.
What a standard selection sheet should contain
Before the first cable is connected, you should be able to read a one-page sheet and check it yourself. Five lines per standard:
- The standard number and the full edition, including amendments.
- The route it serves - EMC, safety, radio spectrum, or a market-specific scheme - so it is clear which legal obligation it discharges.
- The ports and functions actually tested under it, and the operating modes selected as worst case.
- What it explicitly does not cover, so nobody assumes the file is complete when it is not.
- The acceptance date: the date by which the edition must still carry presumption, or the date the buyer's review happens.
That sheet is the document to approve before testing starts, and it is the same document the test plan is built from - the process described in what happens during the testing process. Approving a scope after the bench work is finished is not approval; it is a change order.
When the buyer insists on a standard that does not fit
Sometimes the customer knows their market better than the standard list suggests, and sometimes they simply need a document for a contract. Both cases have a clean answer: test to what they asked as a contractual test, and state in the report that it was performed at the client's request against the named standard. Keep the regulatory file on the standard that actually applies. What you should not do is substitute one for the other silently, because the two documents answer different questions and only one of them will satisfy a market surveillance authority.
The short answer
A laboratory does not freely choose a standard. The market decides the rulebook (EU harmonised standards cited in the Official Journal, or FCC Part 15 in the US), the product's primary function decides the product family standard (CISPR 15 for luminaires, CISPR 14-1 for household appliances, CISPR 32 for IT and AV equipment, IEC 60335-1 or IEC 62368-1 for safety), the ports decide which clauses inside that standard apply, and the assessment date decides which edition still carries presumption of conformity - for example, EN 60950-1 and EN 60065 lost their presumption on 20 December 2020 when EN 62368-1 became the reference for that equipment. The buyer's contract can add scope or freeze an older edition, but it cannot create a legal route that does not exist. If two quotations cite different standards, compare the classifications behind them before comparing the price.
Most of the scope reviews we run end with a shorter list than the customer expected: two standards removed because the product family does not match, one added because a radio mode was not declared, and one edition updated because the checklist predated the current citation. Our testing and certification programme is built around that sheet, because a correctly scoped project is cheaper than a correctly executed one that answers the wrong question.
Send the product description, the target markets and the standard list your buyer wrote - even if it looks wrong. We will come back with a standard selection sheet: each standard with its edition, the route it serves, the ports and modes it covers, what it excludes, and which items from your buyer's list drop out.