The marginal cost of running a charging network is field dispatches. Everything else — the platform, the payments, the support inbox — scales cheaply. A van does not. So the operational question that matters most is not "how do we fix this charger" but "can we fix it from here".
The answer improves dramatically when triage is a fixed sequence rather than an individual’s judgement. Seven rungs, always in this order.
The ladder
- 01Read the connector state. Available, Occupied, Faulted, Unavailable. A unit that is Unavailable was usually set that way by a person, and the fix is a conversation rather than a repair.
- 02Read the fault code, including the vendor field. Classify it: self-clearing, remotely recoverable, or requires a person. If it is the third, stop here and dispatch.
- 03Read the last hundred OCPP messages. What was the unit doing when it failed? A StartTransaction that never got a response looks nothing like a MeterValues stream that stopped mid-session.
- 04Check connectivity. Has the unit been sending heartbeats? A charger that is silent is a network fault, and network faults are fixed differently from charger faults.
- 05Change availability — set inoperative, then operative. This clears more state than people expect and interrupts nothing if the unit is idle.
- 06Soft reset. Reboots the application without cycling power. Recovers most transient states.
- 07Hard reset. Full power cycle. Last resort remotely, because a unit that does not come back has just become a definite dispatch instead of a probable one.
The three faults that leave the ladder early
Ground faults and residual-current trips exit at rung two and dispatch. So does any report of physical damage or a burning smell, whatever the protocol says the unit’s state is. And a unit that has been silent for more than a few hours exits at rung four to a connectivity check, because resetting a charger you cannot reach is not a thing you can do.
Make the visit cheap when it happens
A dispatch that follows a completed ladder is a different job from one that follows a phone call. The technician leaves knowing the fault code, the message history, what has already been tried and which part is most likely — which usually means they carry the right spare and finish in one visit.
Record the outcome against the ticket: what the fault turned out to be, what fixed it, whether a spare was used. Six months of that data tells you which rungs are earning their place and which faults deserve their own shortcut.