Zero Export Device for Solar: How It Works, Wiring and How to Choose

A zero export device solar installation starts with the connection agreement. It sets what limit applies, where it is measured, and how the system must act when control fails. Buying a meter or setting an inverter to zero does not answer those questions on its own.

This guide helps installers and project buyers choose the control system. It also helps you work out the effect on energy. It covers device selection. For the inverter settings, see zero export solar inverter configuration. For sizing and design rules, see export limitation in commercial solar design. Wiring, protection settings, live tests, and energization are for qualified professionals only. They should work from the approved design and the exact equipment manuals.

What is a zero export device?

A zero export device measures net power at the agreed grid connection point. Net power is what the site takes from the grid minus what it sends back. The device then adjusts inverter output so that no active power is exported. Active power is the real power that does work. Curtailment means cutting back PV output.

A zero export device usually has three parts. They are a compatible meter, a controller or inverter control function, and communications. The site stays grid-connected and can still import. Export limitation gives no backup power. It does not replace anti-islanding protection, which stops your system from powering a dead grid.

The controller is often built into an inverter or plant manager. It is not always a standalone box. A current transformer (CT) is only a measurement part. It clamps around a wire to sense current. Signed real power comes from the meter’s voltage and current readings. It also needs correct phase mapping and a documented sign convention. Signed means the reading shows the direction of flow.

A utility billing meter is not automatically a control meter. Confirm the measurement interface, update rate, and supported equipment. Also confirm that the network operator accepts the arrangement. GoodWe’s ES manual gives an example. It says other meter brands cannot take over from its Smart Meter. The reason is that the communication protocol is built into the inverter and meter (GoodWe ES manual).

How does the control loop work?

The controller compares measured grid power with its target, then adjusts inverter output. If building demand falls, PV output has to fall too. If demand rises, output can rise within the available sun and the equipment limits. A battery or controllable load can absorb surplus. It needs capacity and permission to do so.

The basic loop is:

1. Measure signed active power at the approved connection point.
2. Compare it with the export limit or control target.
3. Send a power command to every inverter or storage system that takes part.
4. Measure the grid power that results and keep adjusting.
5. Enter the specified fallback state if measurement or control is lost.

The measurement point has to include the loads and generators on that side of the boundary. A sensor on an inverter feeder alone cannot measure whole-site net export. The design needs a clear place for each of these. They are existing PV, battery discharge, generators, and multiple connection points.

Does zero export mean no export at all?

No. A zero export setting is a control target. Sudden load loss can cause a brief deviation. So can measurement delay, communication delay, and inverter response. The connection agreement defines the allowed behavior and timing. It also defines the measurement and any extra disconnection protection.

Two vendor documents show the size and cause of that deviation:

• Victron’s ESS configuration manual calls its system feed-in limit a target. It says the target can be exceeded momentarily. This can happen after events such as a large load disconnecting or a sudden rise in solar production. It also sets a grid setpoint, 50 W by default, to prevent feeding back power when the regulation overshoots a little.
• GoodWe’s ES manual says there can be a deviation of up to 100 W when exporting to the grid. That holds even with the output power limit set to 0 W. A separate FAQ answer puts it at around 50 to 100 W.

Three-phase measurement needs care. Net export summed across phases and zero export on each phase are different limits. A site can import on one phase while exporting on another. SMA’s Data Manager M manual describes phase-exact zero feed-in. It monitors grid feed-in on all three grid conductors. It reduces the total feed-in power as soon as feed-in occurs on one of them. SMA says this is mandatory in some countries and may lead to more power being drawn from the grid. Huawei’s SmartAssistant settings offer the same choice. It is a Limitation mode, Total power or Single phase power. Single phase power works only on a three-phase four-wire system. Check which basis your operator requires and which the equipment supports.

When is export limitation required?

It is required when the connection conditions or local rules say so. Other customers choose it for contract or cost reasons. Having no net-metering credit does not by itself create a legal zero-export duty. Zero export also does not exempt a grid-connected generator from the usual rules. These cover application, registration, inspection, and protection.

Get the requirement in writing for the actual site, because country-level lists go out of date. See the grid export limitation rules by country for a starting point. Then confirm with your network operator. Record:

• the network operator and the jurisdiction (the area whose rules apply);
• the accepted connection point and the current approval;
• installed DC capacity, inverter AC capacity, and permitted export (each on its own);
• total-site or per-phase measurement, and how existing generation is treated;
• fixed, timed, flexible, or dynamic limits and the required control interface;
• permitted deviation, response criteria, fallback, protection, and witness tests;
• limits on settings changes, firmware updates, and later equipment additions.

For project-specific research, start with the solar compliance resources. Then read the current operator documents directly.

Which control architecture should you specify?

Choose a documented, compatible system. It should cover every power source and how each fails. A brand name, a Modbus socket, or a smart meter does not guarantee that the parts work together. Modbus is a common data link between devices. Ask the supplier for exact models, firmware, and control functions. Ask for the measurement setup, any required certificates, and the acceptance test plan.

Give one party the job of owning the interfaces. Two controllers that send different setpoints can undo a good installation. A setpoint is a target value.

Zero export solutions by brand: meters, settings and CT direction

The table lists what each manufacturer’s own manual or support page says. It is not a ranking and not a compatibility guide. Models, firmware, and regional versions differ, so read the manual for your exact model.

CT placement and direction

A wiring diagram is the right way to show CT placement, and this page has none. In words: the CT or meter sits at the grid connection point. It goes on the supply side of every load and generator that counts toward the limit. Solis says the CTs go at the point of common coupling and not on a load branch circuit (Solis). The point of common coupling is where the site meets the grid.

CT direction depends on the manufacturer. Sungrow’s guide has the arrow pointing toward the house. Solis’s manual has P1 toward the grid. GoodWe’s troubleshooting section says to follow the direction sign on its CT (GoodWe). Follow the exact diagram for the model you install. Do not apply one brand’s rule to another. Solis also gives a check. With loads on, negative power on the EPM means the system is importing and the CT direction is correct. Positive power means the CT is reversed.

What happens when the meter or communications fail?

Specify, configure, and test the fallback for the approved system. Losing meter data does not always stop an inverter by default. The response depends on the equipment and connection conditions. It can be restricted output, standby, or protective disconnection. A qualified designer has to decide which one is accepted.

The manuals show how different the defaults are:

• Fronius’s optional Fail-safe reduces inverter output to 0% when the Smart Meter connection is lost. It puts the inverter in standby. Fronius advises against Wi-Fi between the meter and inverter when Fail-safe is on. Even short drops can shut the inverter down (Fronius).
• Huawei’s SmartAssistant sends a preset output limit to the inverter. This happens if it sees no meter data within the “Maximum protection time” (Huawei).
• Solis’s EPM stops the inverters if they lose communication with it while Fail Safe is set to its default “Run” (Solis).
• Victron’s Fronius AC PV zero-feed-in setup keeps generating by default when Modbus instructions stop. This changes only if Dynamic Power Reduction is configured (Victron).

The failure review should cover stale measurements and meter power loss. It should cover controller power loss and failed links to single inverters. It should also cover failed network equipment, startup, restart, and loss of a storage interface. Name who receives alarms and who restores operation. Also state whether settings survive a firmware update or a replacement.

Anti-islanding protection solves a separate problem. The problem is unintended energization of a disconnected network. Export limitation works while the grid is connected. Backup operation needs an approved isolation and switching setup. An export controller does not provide one.

How should you model curtailment and self-consumption?

Use time-matched site load and available PV power with clear measurement boundaries. Annual bills cannot show whether a factory’s weekend demand matches midday production. Interval data supports an energy estimate. But an annual energy simulation cannot show the fast control response needed at commissioning.

For a simple no-storage screen, PV used on site = min(available PV power, site load) x interval duration. The example below is illustrative. It assumes ideal zero export, no losses, and constant power within each one-hour interval. So it is arithmetic for explaining the method. It is not a generation forecast. In the table, kW is power and kWh is energy.

Available PV energy is 10 + 20 + 30 + 10 = 70 kWh. Of that, 36 kWh supplies the site and 34 kWh is curtailed. So the site uses 36 / 70 = 51.4% of the available PV energy. Site load is 6 + 8 + 12 + 14 = 40 kWh. So 36 / 40 = 90% of it is solar-supplied. These are different metrics. Calling delivered PV energy “100% self-consumed” in a zero-export model hides the curtailment. This happens when the energy that was never generated is left out of the denominator.

Use load-profile analysis to compare scenarios. Vary the season, operating day, and shutdown schedule. Vary the array size and expected losses too. Value the usable energy at the applicable tariff. A larger array can still pay even when it adds curtailment. So compare the extra cost with the useful energy instead of looking for one best size.

Zero export with a battery (BESS): can storage replace the controller?

A battery (BESS, a battery energy storage system) reduces curtailment. It needs charging power and free capacity at the right time. Some cases leave surplus that still has to be curtailed. These are a full battery, a charging limit, or a reserve requirement. They also include a temperature restriction or an unavailable battery. The controller stays. Something has to measure the grid connection point and command the inverters.

Two manuals show how storage systems handle export:

• Victron ESS: a grid feed-in menu can turn AC and DC-coupled PV feed-in off completely. It can also cap the maximum feed-in power. Feed-in occurs only when surplus PV fully supplies the loads and the battery is charged or at its charge-current limit. A grid setpoint, 50 W by default, keeps a small import margin (Victron).
• GoodWe ES hybrid: the export power limit is set in the PV Master app with a compatible Smart Meter connected. The manual warns of a deviation of up to 100 W at a 0 W setting (GoodWe).

Include charging and discharge losses and degradation (wear over time) in the model. Include operating limits and replacement too. Compare the energy the battery recovers with its full extra cost. Do not count one stored kWh twice. It cannot be credited to both self-consumption and another service. The exception is when a dispatch model shows that both uses happen.

To test an export-limited design on your own load data, Book a Free Demo. SurgePV’s generation and financial workflow covers energy-yield and financial modeling. Use the demo to ask whether it covers your interval load, export-limit, and storage assumptions. Keep the supplier’s control design and the network operator’s proof of acceptance apart from the financial model.

What should commissioning and handover prove?

Commissioning is the final testing before handover. It should show how measurement, control, fallback, and protection behave in the approved test plan. A screenshot showing a zero setpoint is not enough. A billing meter showing no export during one observation is not enough either. Test evidence has to match what the operator and the equipment documents require. That covers the measurement resolution, timing, and operating conditions.

Qualified commissioning staff should record where each sensor sits. They should record the CT ratio and polarity and the phase-to-voltage mapping. They should also record the signed power checks and every controlled device. Add the configuration, firmware, and control priorities. The test plan should cover changing loads and high available PV. It should cover restricted or full storage. It should cover loss of communication, loss of measurement, restart, and recovery. It should also cover any separate protective disconnection that the connection conditions require. Do not disconnect live sensors or switch major loads to copy a generic internet test.

Hand over the approved diagrams and the model and firmware list. Hand over the settings record and the test results. Add the acceptance documents, access controls, alarm contacts, and operating instructions. State which changes need recommissioning or a new approval. The export-control scheme belongs on the single-line diagram. See how to integrate SLD generation into the design approval workflow.

How much does a zero export device cost?

Price the whole export-control arrangement, not the controller box. Get dated quotes for each of these. They are the compatible meter, CTs where needed, and the controller. They are also power supplies, communications, installation, and engineering. Add required protection, testing, operator fees, support, and maintenance. Solis’s EPM manual, for instance, lists CTs as an optional accessory. So a quote for the EPM alone leaves out a cost (Solis). A meter already installed for billing may still leave you needing dedicated control equipment.

Compare supplier quotes against one approved scope. This guide gives no installed price. Many things set the project cost. They include equipment architecture, retrofit work, site size, protection, and local acceptance rules.

Frequently asked questions

What is a zero export device for solar?

A zero export device measures net grid power. It tells supported inverters or batteries to hold active-power export at zero. This is at the approved connection point. It can be a separate controller or an inverter function used with a compatible meter. The site stays grid-connected, and anti-islanding protection still applies.

How does a zero export controller work?

The controller compares measured grid power with its target and adjusts inverter output. Surplus is curtailed (cut back), or absorbed by a battery or controllable load. A zero setpoint is a control target, not a guarantee. GoodWe’s ES manual says a 0 W limit can still show a deviation of up to 100 W. Victron says its feed-in limit can be exceeded momentarily. That can happen after a large load disconnects.

What is a zero export meter?

A zero export meter, with its CTs, measures signed power at the grid connection point. It feeds the controller. It has to be one that the inverter or controller supports. GoodWe states that other meter brands cannot replace its Smart Meter. Fronius, Sungrow and Solis each specify their own meter or CT arrangement.

Do I need a zero export device?

You need one when the grid connection approval requires zero or limited export. Some owners also choose it for contract or cost reasons. Having no net-metering payment does not by itself create a zero-export rule. An export limiter does not remove the grid-connection application.

How do you control export with a battery (BESS zero export)?

A battery absorbs surplus while it has charging power and free capacity. Victron’s ESS has a grid feed-in setting that turns feed-in off or caps it. It also has a grid setpoint (default 50 W) that keeps a small import margin. When the battery is full or charge-limited, the surplus still has to be curtailed.

Can a battery replace a zero export device?

No. A battery only takes surplus within its charging and capacity limits. A full battery or a charge limit still leaves surplus to curtail. So you still need measurement at the connection point, a control path to the inverters, and a defined fallback.