Calculation validation
An engine that agrees with itself proves nothing. Every calculator here is compared against numbers somebody else published — a worked example in the standard, an annex table, a reference case — and the comparison ships inside the tool as a panel you can open and run. This page is the summary of those panels: 142 comparisons across 10 calculators, none failing at the last check.
What each engine is checked against
| Calculator | Reference the engine is compared against | Checks | Failing | Status |
|---|---|---|---|---|
| #003 Substation grounding grid | Worked example of IEEE Std 80 Annex B: surface derating factor Cs = 0,743, touch voltage limit 838 V, step voltage limit 2 696 V, decrement factor 1,026 published worked example | 19 | 0 | PASS |
| #012 Short circuit, LV up to 1 kV | Worked 400 V example of IEC TR 60909-4:2000 clause 3: I″k3 = 34,62 / 34,12 / 6,95 kA at F1, F2, F3, peak current 81,36 kA by method b and 70,85 kA by method c, factor m = 0,197, Joule integral 83,6 (kA)²·s, plus the voltage factors of Table 1 published worked example | 18 | 0 | PASS |
| #010 Transformer sizing | Impedance bands of IEC 60076-5 Table 1 and the standard rating series; short-circuit contribution with cmax per IEC 60909-0 published table | 18 | 0 | PASS |
| #004 Cable ampacity | Conductor resistances of IEC 60228 Tables 2 and 3, thermal properties of IEC 60287-1-1 Tables 2 and 3 and IEC 60287-2-1 Table 1, skin-effect behaviour with cross-section published tables | 14 | 0 | PASS |
| #005 Reactive power compensation | kvar required to move 100 kW from cos φ 0,80 to 0,95, capacitor current at 400 V, and the overcurrent duty of IEC 60831-1 clause 21 standard formulas and clause limits | 14 | 0 | PASS |
| #009 Motor sizing | Efficiency limits of IEC 60034-30-1 Tables 5 and 7 (IE3, 50 Hz, 90 kW, 4 poles = 95,2 %; IE2, 50 Hz, 200 kW and above, 8 poles = 93,5 %) and the site ambient of IEC 60034-1 Table 11 published tables | 13 | 0 | PASS |
| #008 LV cable sizing | Example of IEC 60364-5-52 Annex E.52.2: a 39 A four-core PVC copper cable clipped to a wall selects 6 mm², the tabulated rating of that size is 41 A (Table B.52.10), and the harmonic factor moves the selection to 10 mm² published worked example | 12 | 0 | PASS |
| #002 Short circuit, MV 1 to 33 kV | Impedance-correction identities of IEC 60909-0 Formulas (11a) to (11c) and the KT relation, and the voltage factors of Table 1 clause identities, not a worked example | 9 | 0 | PASS |
| #013 Lightning risk | Cases of IEC 62305-2 Annex E: country house R1 = 2,51×10⁻⁵, office building with five zones R1 = 9,65×10⁻⁵, hospital with four zones R1 = 7,00×10⁻⁴ published worked examples | 6 | 0 | PASS |
| #001 Voltage drop | Formula of IEC 60364-5-52 Annex G with the resistivity in normal service (1,25 times the 20 °C value, 22,5 mΩ·mm²/m for copper): 400 V three-phase, 100 A at cos φ 0,8, copper 35 mm², 100 m gives 2,43 %, against the limits of Table G.52.1 published formula and limits | 1 | 0 | PASS |
| #011 PV string calculator | Module and inverter limits of IEC 62548 and IEC 60364-7-712 applied to the entered data. No comparison panel is built into this tool yet — it is on the list, and until it is there this row says so rather than implying a check that does not exist. panel not yet implemented | — | — | — |
The counts above were read from the tools themselves. Each panel runs in your browser when you open the calculator, so you are never taking this page's word for it: open the tool, scroll to its validation section and the same comparison runs in front of you.
Why the reference has to be someone else's number
There are two ways to test a calculation engine, and only one of them is worth anything. The weak way is to check the code against the formula it was written from — which confirms the typing, not the engineering. The strong way is to take a case that a standards committee published together with its answer, feed the inputs in and see whether the same answer comes back.
That is why the table above names an annex and a number rather than saying “tested”. Where a standard publishes no worked example — which is common — the next best reference is the tabulated value the standard prints, and the row says that instead. Where neither exists, the row says that too: #002 is checked against clause identities, and that is a weaker claim than a worked example, so it is written as a weaker claim.
Two published inconsistencies, recorded rather than smoothed over
In the low-voltage short-circuit reference case, IEC TR 60909-4:2000 disagrees with itself in two places. The tool records both instead of quietly matching whichever number is convenient:
- the positive-sequence impedance at fault location F1 is printed as 1,881 + j6,746 mΩ in clause 3.4.1 and as 1,881 + j6,764 mΩ in clause 3.5.1 — a transposition worth 0,2 % in the resulting current;
- Table 4a prints |Z(0)| = 6,421 mΩ at F1, while its own components 2,140 + j6,009 mΩ give 6,378 mΩ — a 0,7 % difference.
Both are noted in the tool's validation panel with the tolerance each check uses. An engine that reported a perfect match on those two rows would be hiding something.
What this is not
These comparisons are a statement about arithmetic and method, not an approval. IECCalc is not certified, endorsed or approved by IEC, IEEE or any other body, and you will not find the words “IEC approved” anywhere on this site. Validation means the engine reproduces published values; the responsibility for the design that uses those values stays with the engineer, as section 2 of the terms spells out.
If one of these comparisons is wrong, or you have a published case that a tool fails, the forum is the place — corrections from readers are already in several engines.