Public EV Charging Stations: A Site-Host & CPO Planning Guide

Public charging · site-host and CPO guide

Public EV Charging Stations: Plan the Service, Not Just the Hardware

Design the complete public service—from discoverability and payment to session recovery—before selecting charger power.

Technical review: MarvinInternational · destination-specificHG Power DC portfolio: 40–480 kW
Public-Ready ChainFind → Arrive → Connect → Pay → Charge → Recover
Four ledgersDemand · power · transaction · responsibility
Opening proofTest the unfamiliar-driver and failure paths
18 RFQ inputsMake site and supplier scopes comparable

Public EV charging stations work only when an unfamiliar driver can find the site, reach the correct bay, connect a compatible vehicle, understand the price, start and pay for a session, receive the expected energy, and get help when something fails. For a site host or charge point operator, charger power is only one input. Demand, ownership, public access, payment, software, accessibility, maintenance and recovery must be specified together.

The public-station decision in 60 seconds

Decide first Evidence to collect Do not buy yet if…
Service job Who will charge, why they stop, dwell time, energy need and arrival peaks “Public charging” is the only demand assumption
Ownership model Owner for site work, EVSE, CSMS, payment, customer support, maintenance and data Every supplier assumes another party owns recovery
Charging architecture AC/DC mix, ports, site limit, allocation and vehicle envelope Selection begins with the highest advertised kW
Public journey Discoverability, access hours, bay geometry, payment, pricing, receipt and help path The test plan ends at “charger online”
Destination package Connector, input, metrology, accessibility, cyber, conformity and local approvals One regional datasheet is being treated as global proof

The Joint Office Public EV Charging Infrastructure Playbook organizes deployment across planning, engagement, siting, funding, policy, regulation, procurement and revenue. That breadth is the right starting point: a public station is an infrastructure and operating program, not a cabinet order.

Decision model 01 · Public-Ready Chain

One failed link can stop the public session

01Findaccurate public data

02Arriveroute · bay · access

03Connectinlet · reach · handling

04Payprice · auth · receipt

05Chargevehicle · EVSE · CSMS

06Recoversupport · diagnose · restore

Public-ready means an unfamiliar driver can complete the chain—and the operator can recover it when a link fails.

What makes an EV charging station public-ready?

Use the Public-Ready Chain:

Find → Arrive → Connect → Pay → Charge → Recover

Every link requires a promise, an owner and a test.

Link Public promise Procurement question Opening test
Find Accurate location, access hours, connector, power and availability data Who publishes and corrects static and live data? Search as an unfamiliar driver and compare listing with the site
Arrive Clear entrance, wayfinding, usable bay and safe access Does the layout support actual vehicle movement and local accessibility rules? Drive each intended vehicle path, including an accessible route
Connect Connector and cable reach the inlet without obstructing users or traffic Which vehicles, inlet locations, cables and parking orientations were checked? Connect representative vehicles in forward and reverse orientations
Pay Price and method are clear before commitment; authentication works Which ad-hoc/member methods, tariff rules, receipts and refunds are included? Complete each payment path and reconcile the transaction
Charge A valid request becomes a stable energy-delivery session within agreed limits What is tested across EV, charger, CSMS and site-power conditions? Witness normal, limited-power, simultaneous and interrupted sessions
Recover A failed session creates an actionable support and repair path Who monitors, answers, resets, dispatches, stocks parts and restores status data? Inject faults and time the escalation, diagnosis and status restoration

A site may be physically open but not operationally public-ready. A broken map pin, unreachable cable, failed payment reader or unassigned help call can stop the same session as effectively as a power-stage fault.

Choose the public charging job before choosing AC or DC

“Public” describes access; it does not define one charger type. Start with why the driver stops.

Service job Typical behaviour to verify Usually evaluate Main failure if misread
Destination charging Driver is already parking for a meaningful dwell period Networked AC Level 2, sometimes a small DC option High-power hardware adds site burden without improving the visit
Opportunity charging Retail, hospitality or civic visit has variable but bounded dwell AC, DC or a managed mix Too few usable ports or a tariff that conflicts with the host's goal
En-route charging Charging is the primary reason for stopping; queue tolerance is low DC fast charging with clear expansion and recovery plan Nameplate power is high but site/vehicle/session throughput is weak
Urban hub Drivers may lack home charging; demand and vehicle mix vary Multiple AC/DC services, strong payment and support design One interface or parking geometry excludes part of the public
Public/fleet crossover Reserved fleet windows and public access share assets Controlled access, scheduling and dynamic allocation Public drivers arrive while capacity is contractually reserved elsewhere

For longer-dwell AC design, use the commercial Level 2 EV charger guide. For DC electrical, layout, FAT and SAT interfaces, use the DC fast charger station design guide. This page owns the public service that sits across either architecture.

Freeze who owns the station—and who operates the service

The U.S. DOT project-planning checklist makes ownership, operation and maintenance early decisions. The available arrangements vary by utility and jurisdiction, but the commercial principle travels well: asset ownership and operational responsibility are not the same question.

Model Host usually controls Third party may control Contract questions
Host-owned / self-operated Capital plan, tariff, branding, customer policy and asset decisions Software, payment processing or field service may still be outsourced Does the host possess 24/7 monitoring, support, reconciliation and repair capability?
Host-owned / CPO-operated Site and hardware ownership, commercial objectives CSMS, payment, driver service, monitoring and maintenance Who owns data, merchant account, refunds, spares, firmware and exit/migration?
CPO-owned / hosted site Land/access, amenities and site-interface obligations Hardware, network, tariff, service and maintenance What access, revenue, energy, lease, branding, performance and removal terms apply?
Hybrid / service agreement Responsibilities divided by package Responsibilities divided by package Is every interface assigned once, with no gaps or duplicated authority?

Create a responsibility schedule with at least these rows: utility coordination; premises wiring; charger supply; installation; commissioning; CSMS; SIM/connectivity; payment acquiring; tariff configuration; tax/receipt handling; public data publication; cybersecurity updates; remote support; field service; cleaning; collision/vandalism response; spare parts; warranty administration; KPI reporting; and end-of-contract data/configuration migration.

If a proposal says “turnkey,” replace the word with that schedule.

Decision model 02 · Four Capacity Ledgers

Hardware selection sits at the intersection

Demandsessions · kWh · dwell · peaks · queue
Powersite · station · port · vehicle ceilings
Transactionaccess · price · payment · receipt · refund
Responsibilitymonitor · support · repair · spares · data
OPENING READINESS = all four ledgers have an owner, evidence and acceptance test

Prove the concept with four capacity ledgers

1. Demand ledger

Record observed or defensible inputs by time interval:

  • eligible vehicles and connector/inlet mix;
  • arrivals, dwell windows and repeat behaviour;
  • starting/target state-of-charge assumptions where known;
  • energy requested per session rather than only vehicle battery size;
  • weekday/weekend and seasonal peaks;
  • acceptable wait, abandonment and blocked-bay conditions;
  • future demand scenario and the trigger for expansion.

A useful planning identity is:

Daily delivered-energy target = expected sessions × average delivered kWh per completed session

It is a scenario input, not a revenue forecast. A national-laboratory analysis of 3,705 US public stations found utilisation associated with contextual factors including local EV adoption, surrounding network and charging power. Do not copy another station's utilisation into a business case.

2. Power ledger

Keep four ceilings separate:

  1. available site input after other loads and the agreed operating limit;
  2. station or group limit after allocation/load-management rules;
  3. charger/port voltage-current envelope at the relevant operating point;
  4. vehicle acceptance under its state, temperature and charging curve.

The delivered power of a session cannot exceed the lowest active ceiling. A 240 kW label does not guarantee 240 kW to every vehicle, at every state of charge, or on two outputs simultaneously. Ask for the full output envelope and allocation logic, then test it with the target vehicle/CSMS configuration.

3. Transaction ledger

Map every monetary and non-monetary path:

  • free, paid, validated, member, employee/fleet or promotional access;
  • price basis and all visible components;
  • ad-hoc payment, app/RFID, roaming and automatic-authentication scope;
  • pre-authorisation, settlement, receipt, refund and dispute owner;
  • metering and tax evidence required in the destination;
  • idle/occupancy policy and how it is communicated;
  • offline or communications-loss behaviour.

Rules differ. As one jurisdiction example, EU Regulation 2023/1804 includes requirements on ad-hoc payment and price transparency for publicly accessible charging points within its scope. It is not a global hardware prescription; the destination review must identify what applies to the actual project and commissioning date.

4. Responsibility ledger

For each subsystem, record:

owner → monitor → first responder → remote resolver → field resolver → evidence → escalation time

This ledger converts “support included” into an operating design. It also exposes whether the proposed service depends on a manufacturer, local partner, CPO, payment provider or site employee who has not accepted the role.

Select AC, DC or a managed mix

Do not equate public with fast.

Question Favors AC Level 2 Favors DC fast charging May favor a mix
Why is the vehicle parked? The destination is the main purpose Charging is the main purpose Multiple user groups share the site
Dwell Long enough for useful energy at lower power Short turnaround is part of the service promise Long-stay bays plus quick-turn bays
Grid/site capacity Constrained but many port-hours are useful A credible high-power utility/make-ready path exists Managed aggregate limit with differentiated services
Operating capability Light-touch access and longer recovery tolerance Strong monitoring, payment, support and field-service model CPO can manage distinct tariffs/SLAs
Expansion Add ports progressively Add dispensers/modules/site capacity by a frozen architecture Reserve civil/electrical space for both

High power is the wrong answer when the vehicle stays for hours, the site cannot sustain the aggregate load, the tariff cannot carry the fixed operating burden, or the operator cannot recover failed public sessions. Conversely, a few low-power ports may be the wrong answer where drivers stop specifically to replenish energy and leave.

Design for an unfamiliar driver

Find: publish accurate static and dynamic data

Static data includes coordinates, access hours, vehicle restrictions, connector, nominal power, payment methods, tariff information, amenities and contact details. Dynamic data can include availability and operational status. Decide which system is authoritative, who publishes updates, and how a repaired charger changes from “unavailable” back to “available.”

Do not treat map publication as marketing only. It is part of the service interface and should be tested at opening and after configuration changes.

Arrive: test the whole route, not just the bay

Check entrance visibility, circulation, turning path, queue storage, lighting, drainage, pedestrian conflict, impact protection, accessible route, bay markings and the route to amenities. Larger vehicles or trailers may require pull-through geometry rather than a standard parking stall.

For US projects, the U.S. Access Board technical assistance shows why accessible charging involves the route, ground space, operable parts, communication, cable handling and varied vehicle-inlet locations. Its recommendations and cited requirements must be interpreted with the responsible US designer; they are not universal dimensions for other countries. Every destination needs its own accessibility review.

Connect: prove cable reach with actual geometry

Vehicle inlets may be at the front, rear or either side. DC cables can be shorter and heavier. A plan view should show parking orientation, charger face, bollards, access aisle, connector holster and cable sweep. Test representative vehicles without allowing cable slack to become a trip, traffic or accessibility obstruction.

Pay: remove surprises before the session

An unfamiliar driver should be able to understand access eligibility, price, payment method and idle policy before starting. Test successful payment, declined payment, cancellation, receipt and refund—not only the happy path. If the station is free, specify authentication and misuse controls rather than assuming payment requirements disappear.

Charge: test the interfaces separately

The vehicle-to-charger protocol, charger-to-CSMS protocol, payment integration, roaming interface and field-service workflow are separate interfaces.

The Open Charge Alliance develops and tests the Open Charge Point Protocol. “OCPP supported” is incomplete. The RFQ should name version, required functions, security profile, certificate handling, offline behaviour, firmware/configuration management, data fields and target-CSMS test cases. OCPP alone does not prove that a payment terminal, roaming partner, Plug & Charge implementation or particular vehicle will work.

Recover: make failure observable and owned

The Joint Office charging user-experience program emphasizes payment processing, user interfaces, faults, diagnostics and data sharing. A recovery specification should answer:

  • How is a failed start distinguished from a vehicle, connector, charger, payment, network, CSMS or site-power issue?
  • What can be retried or reset remotely?
  • When is a port removed from public availability data?
  • Who speaks to the driver, authorises a refund and creates the incident record?
  • Which faults require a qualified site visit?
  • Which spares, tools, logs and firmware versions must be available?
  • What evidence closes the incident and restores public status?

Reliability: specify outcomes, not one attractive percentage

Uptime can be useful, but it does not describe every failed public experience. A port may be online while payment, connector reach, vehicle communication or energy delivery prevents a successful session. Define a KPI tree:

KPI Required definition Evidence source
Technical availability Exact “up,” “down,” exclusions and measurement window Charger/CSMS events plus outage records
Session-start success Denominator, retry treatment and excluded user/vehicle events Authorisation and transaction logs
Energy-delivery success Minimum evidence that a valid request delivered energy Meter/session record and termination reason
Payment success Method-specific attempts, declines, reversals and processor outages Terminal/gateway/merchant records
Data accuracy Static/dynamic field accuracy and update latency CSMS, API and public listing comparison
Mean time to acknowledge/restore Start/stop clocks, severity and excluded events Ticket, remote-action and field-service records

The US NEVI program's greater-than-97% uptime rule is a program-specific example with a defined calculation and exclusions, not a universal product claim. If a project adopts a percentage, copy the definition, not only the number.

Project-killer diagnostic table

Symptom Likely interface gap Verify Mitigation before opening
Station exists but drivers cannot find it Data publisher/coordinates/status owner missing Compare authoritative record with public apps/maps Assign data owner and update workflow
Repeated “charger unavailable” despite online status Payment, CSMS or vehicle fault hidden by uptime metric Correlate charger, CSMS, payment and session logs Add session-success and fault-code acceptance tests
Cables cross access route or miss some inlets Layout frozen before vehicle/inlet study Full-scale reach test with representative vehicles Reorient bay/charger or revise cable-management design
Queue blocks circulation Arrival peak and service time were averaged away Observe/simulate peak arrivals and queue storage Revise port mix, power allocation or circulation
Host and CPO dispute refunds/repairs Responsibility schedule incomplete Trace a failed-payment and hardware-fault scenario Assign merchant, help desk, dispatch and evidence owners
High nameplate kW produces weak service Site/vehicle/allocation ceiling not modelled Review time-series site limit and session profiles Resize architecture and state the real service envelope
Site cannot migrate providers Proprietary configuration/data/credentials not addressed Review export, certificate, SIM and decommission terms Add portability and exit deliverables to contract

Opening-day proof: test the public promise

Before public launch, witness and archive:

  1. correct coordinates, access hours, connector/power and status in the intended public channels;
  2. arrival, signage, circulation, lighting and accessible-route review by the responsible local party;
  3. cable-reach and connector handling with representative vehicle inlet positions;
  4. every authentication and payment path, including receipt, decline, cancellation and refund;
  5. single and simultaneous sessions under the agreed station limit;
  6. controlled communications loss, payment loss and charger-fault scenarios;
  7. help contact, incident creation, remote diagnostics, escalation and field-dispatch path;
  8. status withdrawal and restoration in public data;
  9. owner/operator training, spares, as-built information, software/configuration baseline and escalation contacts;
  10. an unresolved-items list with owner and deadline—never a verbal assumption that it will be fixed later.

Power-on and public opening are different milestones.

Cost and business case: price the obligations

The hardware quote is one part of the project. Build separate boundaries for:

  • utility/make-ready and premises electrical work;
  • civil works, bays, accessibility, protection, canopy/lighting and signage;
  • chargers, distribution/allocation and commissioning;
  • connectivity, CSMS, payment terminal/gateway, acquiring and roaming;
  • energy, demand-related utility charges and site lease;
  • customer support, monitoring, cleaning, preventive/corrective maintenance and spares;
  • insurance, permits, inspections, metrology, tax and conformity work;
  • data services, cybersecurity updates and end-of-contract migration;
  • contingency and expansion.

Use the commercial DC fast-charger cost guide to structure quote boundaries. Do not publish a payback period until utilisation, tariff, energy/demand costs, payment fees, uptime/session success, host spending and financing assumptions are all project-specific.

When should you not build a public charging station yet?

Pause when:

  • no measured mobility/dwell problem justifies the site;
  • access hours, lease rights or parking enforcement conflict with the public promise;
  • utility/site-power feasibility has no responsible owner or credible path;
  • local connector, payment, metrology, accessibility, data or conformity requirements are unknown;
  • there is no operator capable of monitoring, supporting and repairing the service;
  • the business case depends on copied utilisation or guaranteed charging-speed assumptions;
  • a private fleet, workplace/destination AC deployment or mobile charging alternative would solve the actual need with lower operating complexity.

Where HG Power fits

HG Power's commercial DC charging portfolio spans 40–480 kW. That range supports project discussions from lower-power commercial DC applications to higher-power multi-vehicle sites, but it is not one universal product specification.

For example, one supplied CCS1 manual documents named 120/180/240/360/480 kW models with its own input, output voltage/current, dual-connector, communications and OCPP 1.6J lines. Those values apply to that named family—not automatically to 40–80 kW products, CCS2 configurations, payment systems or another destination package.

For every quotation, freeze:

  • exact model and cabinet/dispenser architecture;
  • destination input and site limit;
  • output voltage/current envelope and allocation behaviour;
  • connector/cable configuration and reach;
  • CSMS/OCPP functions, payment integration and test scope;
  • environmental, protection, metering and cybersecurity requirements;
  • model-matched conformity documents;
  • FAT/SAT, training, spares, warranty and support boundaries.

For broader procurement sequencing, start with the commercial EV charger buyer's guide. For connector and vehicle-side compatibility, use the CCS charger buyer's guide. Review the DC EV charger portfolio, specifications and certifications as starting points; the current quotation and destination-specific document pack must control the final claim.

What real project photographs can—and cannot—prove

Long row of EV charging positions in a supplied Haikou Railway Station project photograph

Supplied project photograph labelled Haikou Railway Station. It documents visible site layout only—not public operating status, power, utilisation, uptime or commercial results.
Multiple DC charging cabinets and parking bays in a supplied Shenzhen Shapu project photograph

Supplied Shenzhen Shapu project photograph showing charger-to-bay and cable geometry. It does not establish rating, session success, revenue or commissioning.

Image 1 — supplied Haikou Railway Station project photograph. The image shows a long outdoor row of charging positions and is useful for discussing circulation, bay count, cable route, lighting and expansion. It does not prove public operating status, exact power, utilisation, uptime or commercial results.

Image 2 — supplied Shenzhen Shapu project photograph. The image shows multiple DC cabinets, paired bays, wheel stops, bollards and visible cable geometry. It supports a physical-layout discussion, not a claim about rating, session success, revenue or commissioning.

Copyable RFQ input pack

Send these inputs before requesting a public-station configuration:

  1. destination country/city and intended opening date;
  2. site type, owner/tenant rights and public access hours;
  3. service job: destination, opportunity, en-route, urban hub or mixed use;
  4. target vehicle classes, connector/inlet requirements and representative models;
  5. arrival distribution, dwell, sessions/day and energy/session scenarios;
  6. acceptable queue/abandonment target and expansion scenario;
  7. existing electrical single-line, supply, transformer/service assets and site operating limit;
  8. required AC/DC port mix and simultaneous-session policy;
  9. parking/circulation plan, cable reach, impact protection and accessibility inputs;
  10. tariff, idle policy, free/member/public access rules;
  11. authentication and ad-hoc/payment methods;
  12. metrology, receipt, tax and payment-acquiring requirements;
  13. CSMS, OCPP version/functions/security and target integration;
  14. roaming/public-data/API requirements;
  15. monitoring, driver support and incident-severity model;
  16. field-service coverage, spares and restoration expectations;
  17. environmental, protection, local code and conformity-document requirements;
  18. FAT, SAT, opening-day test, training, as-built, warranty and handover requirements.

If some inputs are unknown, label them as open decisions. Do not replace them with the largest available charger.

Frequently asked questions

How do public EV charging stations work?

Site electrical infrastructure supplies an AC charger or the conversion stages inside a DC charger. The charger communicates with the vehicle, while a charging-management system may handle authorisation, status, tariffs and session records. Payment, roaming, public data and field service can involve separate providers. A complete public service coordinates all of them.

Are public EV charging stations free?

Some are free, some are paid, and some use validated, member or time-based access. The host/CPO must define the tariff, eligibility, idle policy, tax/metrology treatment and how those terms are shown before a session. Requirements vary by jurisdiction.

Do drivers need their own cable?

It depends on the destination and charger configuration. Some AC public points use sockets and require a driver cable; other AC equipment and DC fast chargers commonly use attached cables. State the connector and cable arrangement in public listings and on-site instructions.

How fast are public EV charging stations?

There is no single public-charging speed. Delivered power depends on AC/DC architecture, charger and site limits, output voltage/current, allocation across active ports, vehicle acceptance, state of charge, temperature and other conditions. Compare the expected delivered-energy service, not only nameplate kW.

How much does a public EV charging station cost?

The boundary can include power upgrades, civil work, bays/accessibility, hardware, software, payment, networking, commissioning, energy/demand charges, maintenance, support, spares and compliance work. A credible budget starts with site and operating inputs; a universal installed price is not defensible.

How do you start a public EV charging station business?

Begin with demand and a service model, confirm site/lease and utility feasibility, choose ownership/CPO responsibilities, model the AC/DC and port mix, identify destination rules, specify the full public journey, procure with acceptance evidence, then complete an opening-day and recovery test. Hardware selection belongs in the middle of that sequence.

About this guide

This guide is written for international commercial buyers. Product and project statements are limited to the supplied evidence; destination-specific engineering, installation, accessibility, payment, metrology and conformity decisions remain with the responsible local parties.

Technical review: Marvin

Primary sources

Get a configuration review

Send HG Power the 18-input RFQ pack above—especially destination, service job, vehicle/demand data, dwell, site power, port concurrency, connector, payment/CSMS, layout and support boundaries. We can review a project-specific equipment configuration and document list without pretending one 40–480 kW portfolio statement fits every site.

Configuration reviewTurn the 18-input pack into a comparable project brief.Contact HG Power →