What this instrument
refuses.
A short-circuit study is defined as much by what it will not compute as by what it will. Here is the boundary, each refusal with its reason — the same §7.0 that prints on every issued document, given room to breathe.
The minimum is a compliance quantity
Maximum fault current sizes the equipment — breaking capacity, peak withstand, busbar bracing. The minimum fault current does a different job entirely: it is the evidence that a protective device still operates when the fault is at the far end of the run, the supply is at the bottom of its band, and the conductors are hot. In Australia that second number is the one an engineer has to be able to defend, and it is the one most quick calculations never produce. This instrument computes both, side by side, on every document — and the minimum arrives with its working shown: the declared c_min, the declared conductor temperature θ_e, and the per-element resistance multiplier the correction actually applied.
The correction is base-explicit — R(θ_e) = R_ref·[1 + α(θ_e − θ_base)] / [1 + α(θ_ref − θ_base)] — because the shorthand form assumes a base it never states and fails a simple round-trip test. Declare the base; the arithmetic stays honest. A minimum computed on cold-conductor resistance is not a minimum, and the difference lands exactly where it hurts: on whether a pickup is proven or presumed.
The voltage factor is yours to declare
Every fault-current tool applies a voltage factor. Almost none of them shows you which one, from which edition, on which basis. This instrument inverts that: the factors are required, empty, engineer-declared fields — labelled for IEC 60909-0:2016 Table 1, cited to the clause you type, recorded with your name and the date, and printed in §2.0 of the report. The engine holds no value to fall back to, so it computes nothing until you declare. That is not an inconvenience; it is the licensing posture and the professional posture in one move — the standard's values live in your licensed copy of the standard, and the document of record says whose declaration every number rests on.
The same discipline decides which voltage the method takes. The equivalent voltage source is c·U_n/√3 with U_n the nominal system voltage — 400 V on the ordinary Australian LV system — not the transformer's 433 V rated secondary. Both voltages are declared, both print, and neither stands in for the other; the 433 / 415 / 400 V note is the long version of why those numbers differ at all. Typed into the wrong slot, the rated voltage overstates every LV current by 8.25 %.
Meshed networks and near-to-generator faults
The network model is a single radial chain of series elements fed by parallel sources at one node. A mesh is not rejected by a validator — it is unrepresentable: there is no field in which to express one. That is deliberate. Applying a single uniform peak factor to a meshed network is wrong, not approximate, and the honest treatments — the equivalent-frequency method, per-branch superposition through the mesh — are a different instrument. Likewise near-to-generator faults: the factor families the method needs when a synchronous machine sits close to the fault are not implemented, so there is no generator source type at all. An engineer with a genset declares it as a stated current and X/R, and the document then says plainly that source decay is not modelled and the breaking and steady-state currents are not computed for that network.
A gap named on the face of the report is safer than an approximation the report cannot bound. Where the earth-fault return path is missing, undeclared, or blocked by a winding the model cannot pass through, the affected quantity is refused by name at the affected nodes — never substituted, never silently zeroed — and the adequacy badge at those nodes says the governing fault may not have been calculated.
Why this differs from ETAP, SKM or PowerFactory
A load-flow-mode short-circuit study starts from a solved operating point: actual tap positions, actual pre-fault voltages, shunts, load current superposed. IEC 60909-0's equivalent-voltage-source method deliberately does not — it replaces the network with a single source c·U_n/√3 at the fault location, represents everything else by its internal impedance, and neglects loads, shunt capacitance and tap position. The voltage factor c exists precisely to bound what the operating point would otherwise contribute. So the two families of tools answer differently because they answer different questions: one asks what the current is likely to be in a modelled operating state, the other what a design must cover across the band the standard defines. Neither number is wrong; a document that mixes them without saying so is. This instrument computes the 60909 quantity, names the method on the report, and prints its results as symmetrical RMS values with no decrement applied, each X/R labelled by the loop it belongs to — so what crosses into another study is a defined quantity, not a loose number.
The hand-off is a named document, not a loose figure
The shipped calculators that consume a fault level — cable sizing for breaking capacity and the adiabatic check, protection coordination for its maximum and minimum fault markers, earthing grid for grid current and decrement — each carry a fault source document field that prints on their own report. Type the issuing document number — AMP-FC-0001 Rev A — and the chain of custody is on paper: the study that produced the number is named on the document that consumed it. For the earthing grid the report block also labels each X/R by its loop, because the earth-fault decrement wants the Z₁+Z₂+Z₀ loop's ratio, not the positive-sequence one, and the two can differ by a factor that matters on a step-and-touch document.
The full exclusions list — line-to-neutral on MEN systems, I″k2E, K_T, tap position and tolerance, the AS/NZS 3000 loop-impedance distinction, and the rest — prints verbatim in §7.0 of every issued document. This page is the reasoning; the report is the record.