EV charging and load-shedding are the two facts of South African life that collide hardest at the charge point. A charger only earns when it is energised, drawing clean power, and finishing sessions cleanly, and in a country that has lived through Stage 6 and the constant threat of more, none of those things can be taken for granted. The good news is that most of the disruption is manageable in software. This is how a well-run charging operation stays online across the stages, and where the limits of software actually sit.
Why load-shedding is harder on chargers than on lights
When the power cuts to an office, the lights go off and come back on. A charger is different. It is a high-current device negotiating a live conversation with a vehicle, sitting on a shared electrical connection that also feeds the rest of your site. Load-shedding stresses a charging operation in three distinct ways, and each one needs a different answer.
- The supply itself disappears. Sessions in progress are interrupted, and the charger and your management system can briefly lose contact with each other.
- The grid connection gets crowded. When power returns, everything switches back on at once. If your chargers ramp to full current alongside HVAC, lifts and refrigeration, you can trip the main breaker on your own site.
- State drifts out of sync. A session that was active when the power dropped may not be active any more, and unless your system reconciles that on reconnect, you are left with ghost sessions, wrong billing, and an occupied bay that is actually free.
Hardware and backup power solve the first problem. The second and third are squarely the job of your charge point management system.
Load management: don't let your own chargers trip your supply
The most common EV charging and load-shedding failure is not the outage itself, it is the recovery. Power returns, every device on the site powers up, and your chargers add tens of kilowatts on top of an already-loaded connection. The breaker trips, and now you have a self-inflicted outage on top of the scheduled one.
Load management solves this by treating your grid connection as a budget rather than an unlimited tap. The CPMS knows the ceiling of your supply, watches live demand, and throttles charging current so the total never exceeds what the connection can safely carry. If two cars are charging and the site is already busy, each gets a fair, reduced share rather than both pulling flat-out and tripping the main.
Throttling beats tripping
A driver getting 70% of full speed for ten minutes barely notices. A driver whose session dies because your breaker tripped notices immediately, and so does every other tenant on the site whose power just went with it. Smooth, dynamic throttling is almost always the better trade.
Soft-stop: why graceful matters when the power goes
When a session has to end, whether load-shedding hits, a kWh cap is reached, or load management needs to free up headroom, how it ends matters enormously. The wrong way is to simply open a contactor and break the circuit while the vehicle is still pulling high current. The right way is a soft-stop: the system ramps the delivered current down to near-zero first, then stops cleanly, exactly as it would at the end of a normal session.
Soft-stop protects three things at once. It protects the vehicle's onboard charger and battery management system from an abrupt cut. It protects the charging hardware's contactors from the wear and arcing of breaking under load. And it protects your grid connection from the electrical transient that a hard break can throw back up the line, the kind of disturbance that can nuisance-trip protection elsewhere on the site.
Hard breaks at high current are a hidden cost
Repeatedly opening a contactor while a vehicle is drawing tens of amps causes arcing that pits and degrades the contacts over time. Multiply that by every load-shedding slot, every day, across every charger, and you are quietly shortening the life of your hardware and risking a transient that trips your supply. A CPMS that always ramps current down before stopping turns a violent event into a routine one.
Scheduling around the Eskom stages and surfacing them to drivers
Load-shedding in South Africa is, mercifully, mostly predictable. The schedules are published, the Eskom stages are known, and a charging operation can plan around them instead of being surprised by them. There are two halves to getting this right: what you do behind the scenes, and what you tell the driver up front.
Behind the scenes, you can lean into the windows you have. If a backup battery or solar array is in the mix, the CPMS can favour charging when on-site generation is available and ease off when the site is running on stored power. Even without batteries, knowing a slot is coming lets the system avoid starting a long session minutes before a scheduled cut.
On the driver side, honesty is the whole game. A driver who arrives, scans, and discovers mid-session that power is about to drop feels cheated. A driver who sees the current Eskom stage and the next slot before they commit makes an informed choice, and trusts you next time. Surfacing the live stage at the point of decision turns a frustration into a feature.
The networks drivers come back to aren't the ones that never lose power. They're the ones that are honest about it.
Session reconciliation: the part nobody sees until it breaks
Here is the quiet failure mode that costs operators the most. The power drops mid-session. The charger and the management system lose contact. When everything reconnects, the system still believes a session is active, but the car may have left, the meter has moved, and the real state is anyone's guess. Left alone, this produces ghost sessions: bays shown as occupied when they are free, sessions that never close, and billing that is either wrong or never happens.
Session reconciliation is the cure. When a charger reconnects after an outage, the CPMS reconciles its own record against the charger's reported state, closes out sessions that have genuinely ended, captures the final meter reading, and settles the correct amount. State converges back to the truth automatically, without an operator manually hunting through a dashboard for stuck sessions the morning after a Stage 6 night.
This depends entirely on a clean protocol relationship between charger and platform. Origami speaks OCPP 1.6J, the open standard the vast majority of South African hardware supports, and validates every message it receives, so a charger coming back online after an outage rejoins the conversation predictably rather than corrupting state. OCPP 2.0.1 exists in the wider market; we deliberately stay on robust, widely-supported 1.6J for now.
EV charging and load-shedding in practice
Load-shedding events and how a well-run CPMS responds
| Event | Without load management | With Origami |
|---|---|---|
| Power returns, whole site powers up | Chargers ramp to full current, main breaker trips | Charging current throttled to fit the connection's budget |
| Session must end mid-charge | Contactor opens under load, arcing, transient, unhappy car | Soft-stop ramps current down, then stops cleanly |
| Charger loses contact during an outage | Session stuck active, bay falsely occupied, billing wrong | Reconciled on reconnect, final meter captured, settled correctly |
| Driver arrives before a scheduled slot | Starts charging, power drops minutes later, trust lost | Live Eskom stage surfaced before they commit to a session |
| Multiple cars on a busy connection | Both pull flat-out, supply overloaded | Demand shared dynamically across active sessions |
Hardware, backup and the limits of software
Software can manage demand, end sessions gracefully, and keep state honest, but it cannot conjure electricity that is not there. Keeping chargers physically energised through an outage is a hardware and infrastructure question: a backup battery, a generator, a solar array with storage, or a grid connection robust enough to ride through the shorter slots. Those decisions depend on your site, your traffic and your budget, and they sit alongside the CPMS rather than inside it.
What the management layer guarantees is that whatever power you do have is used safely and accounted for accurately. It keeps you inside your connection's limits, stops sessions without damaging cars or contactors, surfaces the grid situation to drivers honestly, and reconciles every session back to the truth when the lights come back on. That is the difference between a charging operation that merely survives load-shedding and one that runs cleanly straight through it.
It is also the foundation of a charging business that actually earns. Reliable uptime, correct billing and driver trust are what turn a charger from a cost centre into income, the subject of our revenue playbook for site hosts. You can watch live telemetry, throttling and session handling on the live demo, and the operational detail lives in the docs.
POPIA-aware by design
Driver-facing flows are built around minimal data. A driver can scan a QR, pay by card, Apple Pay or Google Pay in Rand, and charge, with no app to download and no account to create. Hosts who offer free charging can use an emailed one-time code instead. Less data collected is less data to protect, which keeps you on the right side of POPIA without extra effort.
On cost, the maths is straightforward. Our software runs below the imported CPMS platforms most South African operators are quoted, and on energy we deliberately price a few cents under the big public networks on every kWh, with a revenue split agreed up front so hosts keep the majority of what their chargers earn. See the pricing for how the tiers work, then request access when you are ready to talk through your site.
Run a charging network that survives the stages
See how live telemetry, load management, soft-stop and session reconciliation keep your chargers online and your supply safe through load-shedding. Origami EV Connect is invite-only, so tell us about your site and we will be in touch.
Request access