Metering is where a charger stops being an electrical product and becomes a commercial instrument. Once a session is billed, the reading has to be defensible to a customer, and potentially to a regulator.
The decisions that determine whether it is defensible are made during hardware design and cannot be corrected in software afterwards.
Establish the regulatory character first
Whether the transaction is a sale of electricity, a service charge, or an amenity provided at no cost changes what the metering must satisfy. Markets treat these differently, and the requirement for a regulated sale is materially stricter.
This is a commercial and legal determination rather than an engineering one, and it must be settled before component selection because it constrains the meter, its placement and its documentation.
Accuracy class is not a specification detail
Meters are classified by how much error they may exhibit across their operating range. A meter adequate for indicative reporting may be inadequate for billing, and the difference is not something firmware can compensate for after the fact.
Choosing the class late, or choosing a lower class to save cost, is the most common way a charger becomes unsuitable for the commercial model it was built for.
Where the measurement point sits
Measuring upstream of the contactor and cable includes losses the driver did not receive. Measuring at the output includes less. The difference is small per session and becomes material across a year on a busy site, and more importantly it becomes an argument during a dispute.
The defensible position is a documented measurement point that both parties understand, rather than one that happens to be wherever the meter fitted mechanically.
What makes a reading defensible
- A meter of the required accuracy class for the transaction type.
- A measurement point that is documented and consistent.
- Calibration where the class or market requires it, with records.
- Start and stop readings recorded against the session, more than a computed total.
- Timestamps that agree with the platform, so time-of-use tariffs can be applied.
- Tamper evidence appropriate to the deployment.
The firmware coupling
The meter communicates with the controller over some interface, and firmware reads, scales and reports those values. A meter substitution that is electrically equivalent but uses a different protocol, register layout or scaling factor will produce wrong numbers rather than obvious failures.
This makes the meter one of the clearest examples of a firmware-coupled component, and one of the worst places to make a component substitution on cost grounds alone.
Reporting is a separate problem from measuring
A perfectly accurate meter is useless if the values reported to the backend carry the wrong unit, the wrong context, or arrive without the start and stop boundary readings that make a session total verifiable.
Metering design therefore extends past the hardware into what the charger reports and how, which is where a considerable share of billing disputes actually originate.
Design decisions that are expensive to reverse
Each of these is inexpensive to decide at architecture stage and expensive to change once a product is in production, which is the general shape of most metering regret.
- Accuracy class, which determines the component and sometimes the enclosure layout.
- Measurement point, which is mechanical as much as electrical.
- Whether per-connector metering exists, which matters for any multi-connector unit.
- Whether calibration values are stored per unit, and how they survive a firmware update.