Brownouts and low grid voltage in Con Edison territory are shutting down solar and battery systems, causing lost production, battery failures, service calls, and added costs for customers and installers.
Solar installers are accustomed to troubleshooting equipment failures. A microinverter stops reporting. An inverter throws a grid fault. A battery refuses to reconnect. The customer calls the installer, and the installer is expected to fix it.
But increasingly, we are encountering another possibility: the solar equipment may be doing exactly what it was designed to do. The problem may be the voltage being delivered by the grid.
In parts of Con Edison territory, particularly during periods of heavy electrical demand, we have encountered customers whose voltage drops far enough that grid-interactive solar and battery equipment disconnects. What looks to the homeowner like a defective solar system can actually begin upstream of the solar installation.
The result is a surprisingly expensive problem. Customers lose solar production and sometimes backup capability. Installers make repeated service visits for equipment that may not be defective. Manufacturers get pulled into troubleshooting. And because the utility voltage often returns to normal before anyone arrives with a meter, identifying the underlying problem can be difficult.
One source of confusion is the voltage operating range printed on inverter specifications.
A typical residential service in Westchester may be nominally 120/240 volts, while Con Edison also supplies 120/208-volt service throughout its territory. Con Edison itself identifies both 120/208 V and 120/240 V as predominant secondary voltages in Westchester.
Some inverter documentation allows operation over a substantially wider range than the nominal service voltage. For example, an APsystems DS3-series datasheet lists a nominal 240-V output with a voltage range extending down to approximately 211 V. (APsystems)
That does not mean that 211 V should be considered normal 240-V utility service.
There is an important distinction between a manufacturer's protective operating envelope and the voltage customers should ordinarily receive from the utility. Con Edison's own distributed-generation specification states that steady-state system voltage is governed by applicable service-voltage limits established under ANSI C84.1 and by the New York Public Service Commission.
In other words, saying, "the inverter works down to 211 volts, so the grid is fine," misses the point.
An inverter is a piece of grid-protection equipment as much as it is a power converter. It continually watches the utility. When voltage or frequency becomes abnormal enough, it must protect the electrical system by reducing output or separating from the grid.
This is where seemingly brief brownouts can have an outsized effect on solar production.
New York's interconnection requirements specify voltage-response behavior consistent with IEEE 1547. When protective equipment disconnects generation because of abnormal conditions, the rules require the equipment to remain disconnected until acceptable utility voltage and frequency have recovered for the required period; the current requirements specify a minimum five-minute recovery period before reconnection in the circumstances described by the standard.
So consider a system experiencing intermittent low voltage.
The voltage drops. The inverter disconnects. Voltage recovers. The inverter waits before reconnecting. Then another voltage excursion occurs.
A customer can end up with repeated periods of zero solar production even though there is nothing mechanically wrong with the array.
From the customer's perspective, "my solar keeps shutting off."
From the inverter's perspective, it is complying with its safety requirements.
From the installer's perspective, it becomes a service call.
And from the utility's perspective, unless the voltage problem is captured at the correct time and location, everything may appear normal by the time it is investigated.
At Rivertown Solar, we have dealt with installations where production interruptions initially presented like solar-equipment problems but where low grid voltage became a significant part of the investigation.
One APsystems installation, for example, experienced intermittent production outages. The instinct in a situation like this is understandably to look at the micros, communications equipment, wiring or monitoring platform.
But when grid voltage is periodically falling outside the conditions under which the equipment can remain interconnected, replacing solar hardware does not solve the underlying problem.
This has consequences for both parties. The homeowner loses production while the installer spends hours diagnosing something originating outside the solar system.
And battery storage can make the problem considerably more complicated.
One of our more instructive cases involved an EG4 hybrid inverter and battery system serving a customer's critical-load panel.
The installation was intended to provide exactly the resilience customers expect from storage: keep selected household loads operating when utility power is unavailable.
But repeated low-voltage conditions created an unexpected operating problem.
When the utility voltage fell outside the inverter's acceptable grid conditions, the system separated from the grid and began supporting the critical loads from the battery.
That is normally exactly what it should do.
The problem was what happened next.
The low-voltage condition was not necessarily a clean, sustained blackout. Utility power existed, but its quality was poor enough that the inverter could not reliably remain connected to it. The battery therefore continued carrying the critical loads and gradually discharged.
Once the battery became depleted, the customer found himself in a particularly frustrating situation: utility power was technically present at the house, but the critical loads could not simply return to normal utility operation because the hybrid system kept attempting to qualify and reconnect to a grid condition it did not consider acceptable.
The energy-storage system had effectively become a choke point between the utility and the critical-load panel.
This was not simply a question of a battery "failing." EG4's 18kPV documentation specifically provides separate grid, off-grid and seamless-transfer behavior and requires the inverter to evaluate the grid before transferring loads between sources. The sophistication that makes modern hybrid inverters capable of seamlessly separating a home from an unhealthy grid also means that the inverter—not a simple piece of copper—is deciding whether that grid is acceptable.
Ultimately, we installed a manual bypass arrangement so that the critical-load panel could be supplied directly from utility power when necessary, rather than forcing every watt serving those loads to pass through the battery/inverter system.
That experience changed how we think about storage design.
A battery is supposed to eliminate a point of vulnerability. If the system architecture provides no practical way around the hybrid inverter when abnormal grid conditions persist, the resilience equipment itself can become a single point of failure.
Where appropriate and permitted by the equipment design, applicable codes and the authority having jurisdiction, we believe installers should seriously consider how a service or bypass path will work before the customer ever needs it.
The distributed-energy industry has spent enormous effort designing systems around blackouts.
A blackout is straightforward: the grid disappears. The inverter detects the outage, opens its grid connection and establishes a local source for the backed-up loads.
A brownout can be much messier.
The grid is still there.
Lights may still work. Appliances may still run. A homeowner may not even recognize that there is a utility problem.
But sophisticated power electronics are monitoring voltage continuously. Solar inverters and storage systems cannot simply ignore abnormal grid conditions because doing so could violate their interconnection requirements and potentially create safety or power-quality problems.
This creates an ironic situation: a customer's old refrigerator may continue operating while a $40,000 solar-and-storage system decides that the grid is unsafe to parallel with.
That does not necessarily mean the inverter is too sensitive.
It may mean that the inverter is the first device in the house sophisticated enough to tell us that something is wrong with the power.
At the moment, too often, the solar customer and installer do.
Suppose a homeowner loses several hours of generation over multiple sunny days. The financial loss from the electricity itself may initially appear modest.
But then add:
the installer's diagnostic visit;
remote monitoring and engineering time;
manufacturer technical-support calls;
repeat truck rolls because the condition cannot be reproduced;
battery cycling caused by repeated transfers;
customers losing confidence in their solar or storage systems; and
in the worst cases, critical loads losing power despite utility power being present.
Multiply that across a growing population of inverter-based DER and the economic cost becomes meaningful.
This is particularly important as New York electrifies buildings. The same customers installing solar are increasingly installing batteries, EV chargers, heat pumps and heat-pump water heaters.
Voltage quality that was merely annoying in an older house becomes much more consequential in a home full of power electronics.
There is also an information problem.
Solar installers now possess an enormous distributed network of voltage-monitoring devices attached directly to the grid. Microinverters, string inverters and hybrid battery systems continuously record conditions that historically would have required specialized utility monitoring equipment.
When dozens of devices begin reporting undervoltage events, that information should be valuable.
Instead, the homeowner often begins by calling the solar company.
The solar company then has to establish that the solar equipment is functioning properly, document the AC voltage, obtain inverter event logs and persuade everyone involved that an inverter grid fault does not necessarily mean an inverter failure.
There should be a much simpler process for DER contractors to submit credible voltage-event data to Con Edison and have the utility investigate the service, transformer or distribution circuit.
Con Edison's own engineering requirements recognize the importance of maintaining appropriate steady-state voltage and power quality on the distribution system. And New York's interconnection requirements explicitly contemplate DER disconnecting during abnormal voltage conditions and waiting for satisfactory conditions before reconnecting.
The industry therefore already recognizes the phenomenon technically.
What is missing is a better operational framework for dealing with its consequences.
Utilities, regulators, equipment manufacturers and installers should treat persistent undervoltage as a distributed-energy issue rather than simply a traditional power-quality complaint.
Utilities should make it easier to obtain interval voltage information and should have a clear escalation process when certified DER repeatedly records abnormal voltage.
Manufacturers should make grid-voltage events unmistakable in customer-facing monitoring portals so that a "grid undervoltage" event cannot easily be confused with an equipment failure.
Installers should capture line voltage and inverter fault history early in the diagnostic process rather than immediately assuming a defective inverter.
And battery installers should think carefully about failure modes and bypass architecture. Our EG4 experience demonstrated that a battery can successfully protect a customer from a short outage yet create a difficult situation during a prolonged, unstable-grid condition if the critical loads have no practical alternate path back to utility service.
Most importantly, customers should not be told that severely depressed voltage is acceptable merely because an inverter specification contains a low-voltage ride-through number.
There is a difference between what a grid-connected inverter can temporarily tolerate and what customers should reasonably expect their utility to deliver.
As rooftop solar and battery storage become a larger part of New York's electrical infrastructure, that distinction is going to matter more, not less.
Solar installers cannot regulate utility voltage. Equipment manufacturers cannot fix the transformer down the street. And homeowners should not have to pay repeated service charges to discover that their solar system is shutting itself down because it is correctly responding to an abnormal grid.
The grid and distributed-energy systems are now one interconnected system.
It is time we troubleshoot them that way.