How to Diagnose a Solar Site That Browns Out Overnight

Back to Blog

First Checks

Pull the battery voltage trend and look at the shape of the night: when the decline starts, how steep it is, and at what hour the site actually drops. Then look at the day before: did the charge reach absorption and settle into float, or was the controller still in bulk at sunset? A bank that starts the night already low points to a harvest or capacity problem; a bank that starts full and dives points to load or a failing bank. Check the calendar too, because a site that only browns out in the darkest months may simply be operating past its design margin for days of autonomy.

Note whether the site drops at a consistent voltage. If it disappears each time the bus reaches the same value, you are seeing the low-voltage disconnect operating as configured, and the question becomes why the bank reaches that voltage so early, not why the site turns off.

Cause 1: The Bank Has Lost Capacity

The most common culprit on a site more than a few years old is a bank that still charges to a normal-looking voltage but holds far fewer amp-hours than its label claims. Aging and sulfation both present this way, and the trend signature is an overnight decline that grows steeper month over month at the same load, often with a late-night knee where the voltage lets go quickly.

The test is a capacity check: a controlled load test against the datasheet rating, or at minimum per-unit voltage readings under the real overnight load to catch a single collapsing unit dragging the string. The fix is replacement, or equalization where the manufacturer documents it for flooded banks; nursing a sulfated bank through another winter rarely ends well.

Cause 2: The Overnight Load Grew

Sites accumulate loads: a radio reconfigured to a higher power, a camera and its heater, an extra sensor, a port left always-on. Each addition moves the overnight energy draw away from the design that sized the bank. The test is measurement, not memory: clamp the battery output current after dark and compare it with the site's power budget, remembering that duty-cycled loads need observation over time rather than one glance.

The fix is either shedding and duty-cycling the growth or honestly resizing the system around it. Comparing the measured draw against the documented load profile makes the conversation concrete: the site is not misbehaving, it is doing exactly what the new load list demands.

Cause 3: The Day's Harvest Is Short

If the bank starts the night below full, the day is not refilling it. Soiling, snow, new vegetation and the lower winter sun path all cut harvest, and partial shading costs disproportionately more than the shadowed area suggests. A failing panel, a corroded connector, or a controller fault does the same. The trend test: charging current well below expectation in good sun, absorption never reached, float never seen.

The site test is a daytime visit in decent sun: measure array current into the controller, inspect for shade and soiling, and compare panel siblings against each other. Fixes range from a wash and vegetation cut to panel or controller replacement, and if the site is simply under-arrayed for winter, that is a design conversation about the worst month rather than a repair.

Cause 4: Voltage Drop and Measurement Illusions

Sometimes the bank is healthier than the readings suggest. Resistance between battery and load exaggerates sag under load pulses, so a radio transmit burst dips the measured voltage at the RTU far below the true battery voltage, low-voltage logic reacts, and the site browns out with energy still in the bank. Loose or corroded connections, undersized runs, and measuring at the controller instead of the battery all contribute.

The test is simultaneous measurement at the battery terminals and at the load during a heavy pulse; a large difference indicts the path between them. The fix is mechanical: clean, torque, shorten, upsize. It is worth ruling this out before condemning a battery bank, because it is the cheapest fix on this page.

When to Escalate

If the bank tests healthy, the loads match the budget, the harvest is reasonable, and the site still cannot make it through the worst nights, the design itself is short and no amount of maintenance will manufacture the missing amp-hours. That is the point to rework the sizing rather than continue nursing it, using the site's measured loads and the season that actually breaks it as the design case, per worst-month design thinking. Escalate immediately, rather than iteratively, if you find heat-damaged wiring or a swollen battery, because those are safety findings first and power findings second.

Frequently Asked Questions