Home - Power Supply Knowledges - Your LED Driver’s Certificate Says IEC 60950. That’s the Wrong Standard.
Your LED Driver’s Certificate Says IEC 60950. That’s the Wrong Standard.
A buyer in the UK emails us a PDF: a test report for an LED driver, four pages, issued by a lab in Shenzhen, with “IEC 60950-1” on the first page and a CE mark on the last. The product is being sold into a commercial lighting project and the electrical contractor has asked for the compliance file.
The report is genuine. The lab is real. The tests were actually performed. And the driver is, in the terms that matter for a European market surveillance authority, not certified against the standard that applies to it.
This is one of the most common compliance gaps in the LED supply chain, and unlike most compliance gaps it is not a case of fraud or a fake certificate. It is a case of the wrong standard being chosen years ago, filed once, and copied forward from product to product because nobody downstream knew to ask.
Here is what the standards actually cover, why the wrong one keeps appearing, and the five documents that settle the question in ten minutes.
Three standards, three scopes
The confusion exists because three different standards can apply to the same physical object depending on how it is sold and where it is installed.
| Standard | Scope | Applies to |
| IEC 61347-1 + IEC 61347-2-13 | Lamp controlgear — general requirements, and particular requirements for DC/AC supplied electronic controlgear for LED modules | The LED driver, as a component of a luminaire |
| IEC 60598-1 | Luminaires — general requirements and tests | The finished luminaire |
| IEC 60950-1 / IEC 62368-1 | Safety of information technology equipment / audio-video, IT and communication equipment | Power supplies for IT and AV equipment — not LED controlgear |
The relevant one for an LED driver is the first. IEC 61347-2-13 was written specifically to deal with the things that make an LED driver different from a generic power supply: the output is a current source or a regulated voltage feeding a semiconductor load, the load can be disconnected while energised, the output may or may not be SELV, and the driver may be built-in (inside the luminaire), independent (mounted separately, with its own enclosure) or integral (part of the luminaire housing with no separate enclosure).
Those distinctions matter because the test requirements change with them. An independent driver has to survive touch, moisture and mechanical tests that a built-in driver does not. A built-in driver can rely on the luminaire housing for some of its protection, which is exactly why you cannot take a built-in driver out of a luminaire and mount it on a wall without re-assessing it.
60950-1 was written for a different product category entirely. It has no concept of a constant current output driving an LED string, no requirements for output open-circuit voltage with the load removed, and no provisions for the thermal conditions inside a luminaire. Applying it to a driver is not an alternative route to compliance; it is a route to a different product.
Why 60950 keeps appearing
Four reasons, and none of them is sinister.
It used to be cheaper and easier. A lab that runs 60950 every day can turn a report around in two weeks. A 61347 report requires the lab to have the particular standard in scope and an engineer who has read it. Historically, fewer Chinese labs had that scope, so the market supplied what was available.
Many of these products started life as generic power supplies. A 24 V 5 A enclosed switch-mode supply is a legitimate ITE product with a legitimate 60950 report. When the same platform is repurposed as an LED driver — sometimes with no change other than the label — the report comes along with it. The hardware is identical; the intended application is not.
In some markets, nobody checks. A driver sold as a spare part, or into a project where the luminaire carries the certification, may never be examined at the component level. The gap is invisible until an audit, a customs inspection, or an insurance claim.
The 62368-1 transition confused everyone. 60950-1 and 60065 were replaced by IEC 62368-1, a hazard-based standard that merges IT and AV equipment. In the EU, EN 60950-1 was withdrawn on 20 December 2020. Suppliers who had 60950 reports had to do something, and many moved to 62368-1 rather than to 61347-2-13 — which is still the wrong direction for a driver. 62368-1 is a better standard than 60950, but it is not the controlgear standard.
What actually differs in the test
Four areas where the choice of standard changes the outcome, in rough order of how often they cause a project to fail.
Output under fault and no-load conditions. 61347-2-13 requires the output voltage to be assessed with the load disconnected, and limits it to SELV levels (60 V DC or 42.4 V peak) if the output is declared SELV, or requires additional protection if it is not. A driver designed as an ITE supply may have no defined behaviour with no load, or may climb to a voltage that is fine inside a PC and not fine on an accessible terminal block in a ceiling.
Thermal testing in the intended environment. Controlgear testing is done at the declared maximum ambient (tc point), with the driver in the enclosure it will actually be used in. ITE testing assumes an open, ventilated environment. This is the difference that produces field failures in sealed luminaires: the report says 40 °C ambient, the installation is at 55 °C inside a box, and nobody recalculated.
Spacing and insulation coordination. 61347 references IEC 60664 for clearance and creepage with pollution degree assumptions appropriate to lighting installations. The values are not always stricter than 60950 — in some places they are looser — but they are different, and a design that passes one by 0.2 mm may fail the other by 0.2 mm.
Marking and instructions. 61347-2-13 requires specific marking: rated output voltage and current, ta (maximum ambient), tc (case temperature reference point) and the independent/built-in/integral classification. An ITE supply marks input, output and a power rating. If the label has no tc point marked, the thermal design has never been declared, which means the life calculation in the datasheet is not anchored to anything.
The transition that caught people out
Worth stating plainly because it is still causing confusion in 2026: EN 60950-1 was withdrawn in the EU on 20 December 2020. A report citing it is not evidence of conformity to a harmonised standard, because the standard is no longer harmonised. Reports citing it are historical documents. If a supplier’s compliance file still lists 60950 as the primary standard for a driver sold today, the file has not been updated in at least five years, which tells you something about how the product has been maintained.
In the US the picture is different but the conclusion is the same. UL 60950-1 is being superseded by UL 62368-1 with a defined transition, and neither is the right standard for LED equipment. UL 8750 is the standard for LED equipment, and for a driver sold as a standalone power supply, UL 1012, with UL 1310 if the output is Class 2.
The model list problem
Here is a detail that catches experienced buyers, not just newcomers.
A test certificate covers the models listed on it, not the factory and not the product family. A report might cover PS150-H1V12 and PS150-H1V24 — two output voltages of the same design — and nothing else. Order PS150-H1V48, which is the same board with a different transformer, and you have a product that is not covered, even though the label looks the same and the factory will happily ship it.
This is why the fifth document below matters more than the other four. It is also why a supplier saying “all our products are CE certified” is not answering the question: CE is a self-declaration by the manufacturer, and the declaration names the models it covers.
Five documents that settle it
Ask for these, in this order, and the answers will tell you whether the compliance file is real:
- The declaration of conformity (DoC), with the standards listed by number and year, and the model list. If the model you are buying is not on the list, stop.
- The test report’s first page, showing the standard, the report number, the issuing laboratory and the accreditation. The standard number is the whole question.
- The CB scheme report and certificate, if the product is sold across multiple countries. A CB report from a recognised NCB is the most portable evidence and it names the national deviations that were assessed.
- The tc and ta declarations, and where the tc point is on the housing. No tc point, no thermal claim.
- The model list appendix, cross-checked against your part number, including output voltage and any dimming option. Dimming variants are frequently not covered even when the base model is.
If a supplier can produce 1, 2 and 5 without a week of emails, they have a maintained file. If they send a CE certificate image with no standard number on it, they have a graphic.
The honest note about our own catalogue
This article would be hypocritical if it did not apply to us.
Several SANPU catalogue models — including parts of the indoor range built on the 60W IP20 indoor driver platform — are registered under GB 4943.1 / IEC 60950-1, because those models were originally developed and sold as general-purpose power supplies into markets where that registration was accepted, and the LED application came afterwards. That is the exact pattern described above, and we are not going to pretend otherwise.
What that means in practice:
- If your project is in the EU and the driver must be assessed to EN 61347-2-13, raise it at the RFQ stage. Do not order the catalogue model and then ask for the report; the assessment is a project, not a document request.
- If you need a CB report to IEC 61347-2-13 with the national deviations for your market, say so before the quotation, because the testing timeline runs in parallel with the production timeline and not after it.
- If your market accepts the existing registration, the catalogue model is the faster and cheaper route, and we will tell you when that is the case.
We would rather lose an order at the quotation stage than have a customer discover the standard mismatch during a market surveillance check, because at that point the cost is not the driver — it is the recall, the relabelling and the relationship.
The luminaire still needs 60598
One more thing that gets missed, sometimes by people who have just done the work of getting the driver right.
A 61347-2-13 compliant driver does not make the luminaire compliant. The luminaire needs IEC 60598-1, and the assessment includes the driver as a component: its tc point, its clearances to the housing, the wiring, the earthing, the thermal environment. A driver that is perfect on the bench becomes non-compliant when it is mounted 8 mm from an aluminium extrusion at 70 °C.
This is why the compliance conversation should happen at the luminaire design stage, not at the purchase order. The cheapest time to fix a spacing issue is on a CAD model. The most expensive time is after the tooling is cut.
FAQ
Is IEC 60950-1 still valid for LED drivers? No. It was written for information technology equipment, and EN 60950-1 was withdrawn in the EU in December 2020. The applicable standard for LED controlgear is IEC 61347-1 with IEC 61347-2-13.
What about IEC 62368-1 — isn’t that the replacement? 62368-1 replaced 60950-1 and 60065 for IT and AV equipment. It is not the controlgear standard. A driver tested to 62368-1 is not tested to the requirements that apply to a driver feeding an LED module.
Is a CE mark enough? CE is a manufacturer’s self-declaration. It is legally required and it is not evidence of testing. Ask for the DoC with the standards listed, and the test report behind it.
What is GB 4943.1 then? China’s national standard for IT equipment safety, historically aligned with IEC 60950-1 and, in its 2022 revision, aligned with IEC 62368-1. It is a legitimate standard for IT power supplies and not the LED controlgear standard.
Does the certificate cover all output voltages? Only the models listed on it. Different output voltages use different transformers and sometimes different secondary components. Check the model list appendix against your exact part number, including dimming options.
What is the ta and tc marking for? ta is the maximum ambient temperature the driver is rated for. tc is the case temperature reference point used for the thermal design and the life calculation. Without a marked tc point, the life figure in the datasheet has no defined basis.
Need a driver assessed to EN 61347-2-13 with a CB report for your market? Our OEM/ODM LED driver manufacturing line runs the standard assessment as part of the project, with the certification timeline planned alongside production rather than after it. Send the target market, the standard you need cited, and the volumes, and we will tell you what is achievable — including when the honest answer is to use the existing registration instead.