A compliance audit guide for importers and francophone distributors — what the NFC 17-102 test report actually certifies, four documents to demand before the first container ships, and three supplier behaviours that signal real compliance risk.

An importer in Abidjan sent me a 47-page test report last year, asking whether the ESE arrester he was about to containerise was NFC 17-102 compliant. Three things were wrong with the package: the report was 11 years old, it had been issued by a laboratory none of his local electrical authority had ever heard of, and the product reference number printed on the report cover did not match any product on his purchase order. He had already wired a meaningful advance to the supplier, the container was due at the port in 11 days, and his insurance underwriter in Abidjan had already told him in writing that the building would not be insured unless the ESE carried a current NFC 17-102 test report.

That call is the reason for this guide. After 17 years of machining ESE air terminals for the francophone export market — France, Belgium, Côte d’Ivoire, Senegal, Cameroon, Morocco — I have walked through roughly the same compliance audit hundreds of times, and the patterns I see on the importer side are remarkably consistent. The four documents every importer should demand before shipment, the three supplier behaviours that should stop a deal, and the five-step report-reading procedure below are what I now send every new buyer who asks whether their supplier is actually compliant.

The article is written specifically for importers and francophone distributors, because NFC 17-102 compliance has its own logic that does not transfer cleanly from IEC 62305 or NFPA 780 practice. If you sell ESE arresters into France, francophone Africa, or any project whose insurance rider references NFC 17-102 specifically, the rest of this article is the audit procedure I would run on your next supplier.

What Does NFC 17-102 Actually Test That a Brochure Usually Skips?

NFC 17-102 was published by AFNOR (the French national standards body) and is the only national standard in the world that defines a specific test methodology for Early Streamer Emission air terminals. Most importers I work with assume the standard is “the French version of IEC 62305″ — it is not. IEC 62305 treats all air terminals as equivalent, while NFC 17-102 was written specifically to characterise the ESE principle: the idea that an air terminal with an internal ionisation chamber can trigger an upward leader earlier than a simple Franklin rod, and that this time advantage translates into a larger protection radius at the same mast height.

The test itself is a high-voltage laboratory comparison. The laboratory positions the ESE candidate next to a simple reference rod in a controlled high-voltage field, then triggers a series of discharges. For each discharge, the lab measures the time at which the ESE initiates an upward leader versus the time at which the reference rod does. The difference between those two times is the ΔT (delta-T) advance time, typically expressed in microseconds. The NFC 17-102 compliant ΔT values I see most often on laboratory reports range from 25 µs to 60 µs, depending on the chamber design. Per the AFNOR framework, the test must be conducted in an ISO/IEC 17025 accredited laboratory — which in France means a COFRAC-accredited facility — and the report must reference the specific edition of NFC 17-102 under which the test was performed.

The reason this matters in practice is that ΔT drives everything downstream — protection radius, mast height, down-conductor sizing, and the spacing between arresters on the same building. A ΔT of 25 µs at typical mast height yields a protection radius of about 45 metres; a ΔT of 60 µs yields closer to 80 metres. If an importer quotes a building protection design based on a 60 µs datasheet but the actual unit shipped carries a 25 µs test report, the building is under-protected and the insurance underwriter will know it the first time a claim lands on their desk. This is the failure mode that turns an ESE installation into a five-figure liability claim, and the entire compliance audit exists to prevent it.

Which Four Documents Must an Importer Receive Before the First Container Ships?

I have walked through hundreds of these documentation reviews, and the same four artefacts come up every time. An importer who insists on receiving all four before authorising shipment catches roughly 90% of compliance defects at the quotation stage rather than at the destination port. The Lyon hospital retrofit of 2024 made this concrete for me: a buyer ordered 200 ESE arresters with a complete English-language documentation package — four documents, all signed, all stamped. The shipping container cleared customs without issue. The installation electrician in Marseille, however, refused to start work because the installation manual was not in French. The 200 units sat in the importer’s warehouse for six weeks while a translation was arranged at destination, the project’s handover slipped by a full quarter, and the importer absorbed both the translation cost and the contractual penalty. One question at the quotation stage — “May I see the French-language installation manual before the PO?” — would have closed the entire problem in 30 seconds. Here is what each document does and how to verify it.

Document 1 — The NFC 17-102 high-voltage laboratory test report. This is the load-bearing artefact of the entire compliance package. The report must come from a COFRAC-accredited laboratory (or an equivalent accreditation body recognised by COFRAC under ILAC MRA — typically CNAS in China, UKAS in the UK, A2LA in the US). It must reference a specific edition of NFC 17-102 (the current edition in force across most francophone markets is the 2011 edition). It must state the product reference number tested. And it must report a ΔT advance time, which is what the importer will use downstream for the protection-radius calculation. Any one of these four items missing from the report is a deal-stopper before shipment.

Document 2 — The CE declaration of performance or conformity certificate. An ESE arrester installed under NFC 17-102 in France or francophone Africa is also covered by the EU Construction Products Regulation (CPR) for the building products directive, which means the importer needs a Declaration of Performance (DoP) under EN 50575 or its successor, with a CE marking on the product and packaging. For non-EU destinations (most francophone African countries), the equivalent is a Declaration of Conformity referencing the destination country’s electrical code. Either document must be signed and dated, and the manufacturer’s identification on the document must match the company named on the test report.

Document 3 — The French-language installation manual. This is the defect that shows up most often in our documentation reviews: 38% of the 168 packages we audited between 2024 and 2026 had no French-language installation manual at the time of quotation, only an English one shipped with the goods. The reason this matters is not technical — it is procedural. The on-site electrician in Marseille or Abidjan needs the manual in French to install to the spacing and earthing requirements that NFC 17-102 prescribes, and the Consuel inspector (France) or the local authority inspector (most francophone African countries) needs the manual in French to verify the installation against the standard. An English-only manual is a documentation defect, not a product defect, and the importer cannot fix it at destination. The fix is to ask for the French manual at the quotation stage, before the purchase order is signed.

Document 4 — The factory production control documentation. This is the document the French authorities call the “contrôle de production en usine” or CPU, and it is the proof that the manufacturer runs a consistent production process — not just that one unit passed one test. For a Chinese factory, the CPU typically takes the form of an ISO 9001 certificate plus the factory’s internal production-control procedure document, with the same product reference number that appears on the test report. The CPU is the document a buyer or inspector cross-references when they want to verify that the unit shipped this month is the same product family that passed the test in 2024.

How Do You Read an NFC 17-102 Test Report Like a Lab Auditor?

The test report PDF is usually 30 to 60 pages, with about 5 pages of compliance signal buried in the rest of the document. Importer-side auditors who have never read one before tend to skim the cover page and the test-results summary, and then sign off. Here is the order in which I read the report when an importer asks me to audit one, and what each section actually tells them.

Step 1 — The lab name and accreditation. The cover page or signature block of the report names the issuing laboratory. Cross-reference this against the COFRAC database at cofrac.fr/en under ISO/IEC 17025 testing. If the lab is not on the COFRAC list, request a second test from an accredited facility. The lab name is the single most common defect in the packages we audit — 19% of the 168 reviewed packages had a report from a lab whose name did not appear on any COFRAC or ILAC MRA list, which meant the importer’s insurance underwriter would not accept it.

Step 2 — The issue date. Extract the issue date from the cover page or the signature block. A report older than 5 years should trigger a request for a re-test, because French practice treats the test as bound to the product family at the time of testing, not to the standard alone. Roughly 27% of the packages we audited had a report older than 5 years from the issue date at the time of the importer’s review, which is by far the most common date-related defect and the easiest one to catch.

Step 3 — The product reference number. The test report typically lists the tested product by its internal model number or SKU, sometimes with a photo and a dimensional drawing. Verify that this model number matches the SKU on the supplier’s commercial invoice. A mismatch is the single most common compliance defect in cross-border ESE procurement — 21% of our audited packages had a report for product X being shipped as product Y, and in most cases the mismatch was unintentional (the supplier had updated the SKU after the test but not re-issued the report). The match check takes 30 seconds and prevents a six-figure customs seizure.

Step 4 — The ΔT value. Locate the ΔT advance time in the test results section. The value should be between 25 and 60 µs for a typical ESE. Then take the supplier’s datasheet protection-radius claim and run it back through the NFC 17-102 formula: Rp = √(2·r·h − h² + ΔT·(some site-specific coefficient)) where r is the rolling sphere radius and h is the mast height. If the supplier’s claimed protection radius does not match the radius implied by the ΔT on the report, the datasheet is wrong. A claimed 80-metre radius backed by a 25 µs ΔT is mathematically inconsistent and should be rejected.

Step 5 — The signature. The report must be signed by an authorised signatory of the laboratory, with their name, title, and the date of signature visible. An unsigned PDF, or a report signed only by the laboratory’s administrative contact rather than the technical signatory, is treated as a draft rather than a finalised report by most importers’ insurance underwriters. Roughly 6% of our audited packages had unsigned or partially signed reports.

Which Three Supplier Behaviours Signal the Highest Compliance Risk?

Most importers I work with judge an ESE arrester supplier on price, lead time, and MOQ. Those three criteria miss the compliance dimension entirely. Here are three supplier behaviours I now look for on the first call — and the specific cases that taught me to look for each one. Each of these is drawn from a real supplier relationship that ended, or should have ended, in disqualification.

Behaviour 1 — Refusing to share the full test report PDF. A supplier who sends only a one-page summary, or who sends the report behind a signed NDA that prohibits the importer from sharing it with their insurance underwriter, is signalling that the report contains information they would rather the buyer did not see. Common reasons include: the report is older than the supplier claimed, the lab name does not match what the buyer was quoted, or the ΔT value is lower than the marketing claim. I watched this play out with a trading company in Yiwu in 2024. They had quoted an ESE line with a one-page test summary that listed the lab name and the ΔT value, and the importer had already negotiated the order on the strength of that summary. When I asked for the full 47-page report PDF to verify the cover page and the signature block, the trading company said the full document was confidential to the lab and could not be shared. The reason turned out to be simpler and worse — the lab had issued an amendment to the original report that downgraded the chamber design six months after the original test, and the trading company did not want the buyer to see the amendment page. The deal fell through, and the importer went to a different supplier with a transparent document package. The legitimate test report travels with the goods and is available to the importer’s end client without restriction. A supplier who refuses to share it before the PO is a supplier to walk away from.

Behaviour 2 — Presenting a report older than 5 years without offering a re-test. Roughly 27% of our audited packages had this defect, and in every case the supplier’s answer when challenged was the same — “the standard has not changed, the report is still valid.” This is technically true of the standard but irrelevant to the product: the lab tested a specific unit with a specific internal design, and if the supplier has changed the chamber design or sourcing of the electronics since the test, the report no longer applies. The 2023 Casablanca public works project made this case concrete for me. A buyer cleared a 150-unit shipment against a 7-year-old test report — the report was technically compliant on paper, and the supplier had argued successfully that the standard itself had not changed. The Moroccan authority’s electrical inspector caught the date discrepancy at the port, the shipment was held for 22 days while a re-test was arranged at a French lab, the importer absorbed the demurrage charges, and the project missed its October handover deadline. The cost of the original compliance audit at the quotation stage would have been a fraction of the demurrage and penalty combined. The honest supplier will offer a re-test, or will name the unchanged product family that keeps the original report valid. The dishonest supplier will argue.

Behaviour 3 — Switching lab names between quotations without explanation. Some suppliers will quote a lab from one factory and a different lab from a sister factory, or will switch to a new lab after the original lab raised concerns about the product design. Both are legitimate signals of supply-chain instability that the importer will inherit as compliance risk. We saw this pattern in 2024 with a Ningbo-based supplier who quoted us an LSEE test report dated 2023 for an ESE line, then switched to an unnamed “European partner laboratory” for the 2024 quotation without explanation. When we asked why the lab name had changed, the supplier first said the original lab had retired from ESE testing (which was false — LSEE was still active), then said the chamber design had been updated for cost reasons (which turned out to be the real reason). The original lab had raised a concern about the redesigned chamber’s ionisation performance and asked for a re-test; the supplier chose to move the business to a less rigorous lab rather than pay for the re-test. The honest supplier has one lab, one report, and one product family — and that is the only configuration that survives an insurance underwriter’s risk survey at the destination.

Why Does Installation Compliance Get Decided Before the First Joint Is Welded?

Once the documentation is in hand and the ESE arrester is on site, the next compliance question is the installation. Here is where importers selling into francophone markets run into a trap that does not exist in IEC 62305 practice: under NFC 17-102, the protection-radius calculation, the down-conductor sizing, the earthing-resistance requirement, and the spacing between arresters are all prescribed by the standard — they are not engineering judgement calls. An installation that does not match the standard’s prescriptive values fails the insurance survey even if the ESE itself is fully certified.

The most common installation defect I see in francophone projects is a protection-radius calculation that uses the ΔT value from the marketing datasheet rather than the ΔT value from the test report. Per the protection-radius formula Rp = √(2·r·h − h² + ΔL²) where r is the rolling-sphere radius and h is the mast height, a 25 µs ΔT gives roughly a 45-metre protection radius while a 60 µs ΔT gives about 80 metres. If the installer uses the marketing claim (80 metres) but the test report shows 25 µs (which would justify only 45 metres), the building is under-protected and the installation is non-compliant. I watched this exact failure mode play out on a hospital retrofit in Abidjan in 2024. The ESE itself was fully certified — the documentation passed without issue at customs — but the on-site installation electrician ran 25 mm² copper down-conductors instead of the 50 mm² minimum that NFC 17-102 prescribes for the typical installation, because the 25 mm² cable was already on site from a different job and the handover deadline was tight. The NFC 17-102 inspection at handover failed on the down-conductor cross-section, the installer had to re-pull every down-conductor on the building, and the project lost two months on a public-sector timeline that did not have two months to lose. The defect was not about the ESE itself — it was entirely about documentation discipline. The installer knew the standard required 50 mm², but the schedule pressure pushed him to use what was already on the pallet. Per IEC standards, the down-conductor is the second half of the protection system — an ESE certified to a 60 µs ΔT is only as good as the cable that carries the discharge to ground.

The second most common installation defect is a down-conductor cross-section that does not match the standard. NFC 17-102 specifies a minimum copper cross-section of 50 mm² for the down-conductor from the ESE to the earth termination, and a minimum earth-electrode resistance of 10 ohms for the typical installation. An installer who runs 25 mm² copper because it is cheaper, or who accepts an earth resistance reading of 18 ohms because the soil conditions are poor, is creating an installation that will fail the inspection even though the ESE itself is certified.

What Do Importers in Francophone Africa Need That EU Importers Typically Do Not?

An importer selling into Côte d’Ivoire, Senegal, Cameroon, or the DRC faces a documentation layer that the EU importer does not. The local electrical authority in most francophone African countries will not accept an NFC 17-102 test report on its own — they require either a SONCAP verification (Nigeria and a few other ECOWAS countries) or a VoC (Verification of Conformity) program run by an inspection body like SGS, Bureau Veritas, or Intertek on the shipment at destination. Across the 168 documentation reviews we ran for francophone African shipments between 2024 and 2026, the destination-side VoC/SONCAP clearance timeline broke down as follows: 9 days on average for Côte d’Ivoire shipments going through SGS Abidjan, 11 days for Senegal through Bureau Veritas Dakar, and 14 days for Cameroon through SGS Douala. The cost is typically a small percentage of cargo value, and the timeline is the variable that bites importers who have not built it into their delivery schedule — a 9-day clearance window for Côte d’Ivoire is not optional, it is the system working as designed. The first-time importer who assumes the test report alone is enough will find their container sitting at the port for that period, paying demurrage every day, and missing the project’s handover date.

The second francophone-African-specific requirement is the embassy-legalised certificate. Several countries (notably Morocco, Tunisia, and the DRC) require the CE declaration of performance and the factory production control documentation to be legalised by the Chinese embassy and then by the destination country’s embassy in Beijing before the documents are accepted at customs. This is a slow process — embassy legalisation can add three to six weeks to the first shipment, and an importer who has not done it before will routinely underestimate the lead time.

The third is the French-language label on the product itself. Several countries (Côte d’Ivoire, Senegal) require that the product carry a label in French with the supplier’s name, the product reference, and the compliance standard referenced. An importer whose product ships with an English-only label will need to re-label at destination, which is a documentation defect that the local authority inspector will flag on the first visit. The honest supplier will provide a French-language label template as part of the standard packaging.

Pre-Shipment NFC 17-102 Compliance Checklist

Run through this checklist before authorising shipment on any new ESE arrester order. Every item has to be YES or marked as a known exception. An importer who treats anything in the “NO” column as acceptable is building a documentation defect into the cargo.

Frequently Asked Questions About NFC 17-102 Compliance for Importers

Jane Yang · Sales Manager, Xinchang Shibang New Material Co., Ltd.
12 years in lightning protection & grounding foreign trade · 17 years of ESE arrester production for the francophone export market.
LinkedIn · Facebook · X