The free IEC 62305-2 tools, and which edition each one computes

1. What actually changed at DEHN

The story is usually told as "DEHN took the free lightning risk tool away". That is not quite what happened, and the difference matters if you still have it installed.

DEHNsupport Toolbox — the free desktop suite that a large number of engineers learned risk assessment on — is still free and still available. What it is not is current. It implements the 2010 edition, and it stays there: the 2024 edition was not brought to it. Ed. 3 support went into DEHNrisk instead, a separate cloud product on a subscription of about €99 per year, with a discount in the first year for existing Toolbox users. DEHNrisk offers a long, fully functional trial with one deliberate limit — it computes but does not print, because the report is the thing being sold.

So nothing was taken from anyone. The free tool was simply left on an edition that has since been withdrawn, which is a quieter problem and an easier one to miss. Nothing in the software tells you that the ground moved.

2. "Still free" is not the same as "still current"

A 2010-edition engine is not broken. It computes the 2010 method correctly, and for reproducing a historic assessment that is exactly what you want. The problem is only that the method changed, in ways that move numbers rather than tidy them.

Under the current edition, risk is judged per zone rather than as one structure-wide total; the occupancy scaling that used to dilute a small zone's risk is gone; the density input is the ground strike-point density rather than flash density alone; and the loss values were reorganised. A server room inside a large building can now drive the whole verdict — under the older method its risk was averaged away. The same building, the same inputs, and a different answer.

We wrote those changes up in detail, in our own words, in what changed in the 2024 edition. The practical upshot for this article: if you are issuing an assessment today against a specification that names the current edition, a 2010-edition tool cannot get you there, no matter how carefully you drive it.

3. What is actually available

Everything we could verify, free and paid, with prices as published by each vendor when we last checked them on 2026-07-18. Prices move — confirm before you rely on one.

Worth naming plainly: DEHN's published worked examples are among the independent sources we reconcile our own engine against. They are a century-old lightning protection business and their material is good. This section is about which edition a given piece of software computes, not about who knows the subject.

4. Which one you want, by how often you do this

The honest determinant is volume, not features. Work out how many assessments you issue in a year before you compare anything else.

• More than about three a year — buy a subscription. At €99 a year, DEHNrisk costs less than paying per report anywhere, and that includes here. We are not going to pretend otherwise; a tool used monthly should be bought by the year.
• One to three a year — this is the case a subscription serves badly, because you pay for eleven idle months. Pay per report, or engage a consultant if you want a signature as well as a document.
• Enterprise, or you need NFPA 780 in the same tool — LRA Plus covers both standard worlds and is built for that. It is priced for organisations rather than individuals.
• You need a number on screen and are not issuing a document — the free tools are genuinely useful for a feasibility sketch. Just record which edition produced the number, because you will not remember in six months.
• You are reproducing or defending an older assessment — the frozen 2010-edition tools are the right answer, not the wrong one. Matching the edition the original was issued under is the whole point.

5. Three checks that work on any tool, including ours

Vendor claims about editions are not verifiable from the outside, and "supports 2024" is written on things that do not. These three checks take about two minutes together, need nothing but the tool's own output, and are worth running on every engine you rely on — this one included.

Check 1 — ask it for R4

The current edition deleted the economic-loss risk R4 and the cost–benefit annex that went with it. If a tool still offers you a fourth risk alongside R1, R2 and R3, or offers to compare the cost of protection against the cost of loss, it is implementing the 2010 method regardless of what the version notes say. This is the fastest of the three because it needs no arithmetic — just look at the results screen.

Check 2 — make it compute a collection area you already know

The collection area of an isolated structure extends outward from the walls by three times the building height, rounded at the corners — not by the height itself. Getting this wrong is a single-character error with a large consequence, and it is invisible unless you check it, because the number looks plausible either way.

So give the tool a rectangular building on flat open ground with nothing taller nearby: 200 m × 200 m × 35 m. The collection area should come out at roughly 158 600 m². A tool that extends by the height instead of three times the height returns about 71 800 m² — near enough 45% of the correct figure, and every event rate downstream is scaled by that same factor. Our own figure for this building is 158 636 m²; we print the collection area on every report precisely so this check is available to you.

Check 3 — see whether the verdict is per zone

Under the current edition each risk zone is compared against the tolerable risk on its own, and the structure needs protection if any zone exceeds its limit. A tool that gives you one structure-wide risk figure and no per-zone breakdown is not just presenting the result differently — it is computing the older method, in which a small high-consequence zone gets averaged into the building around it.

If any tool fails these, that is worth knowing before it is under your name on a document. If you run them here and something looks wrong, we would rather hear it from you than find out later — the validation page sets out what our engine was reconciled against, including what did not match first time.