A low-voltage disconnect protects a battery by shedding the load when the voltage falls too far, but the setpoint everyone focuses on is the one where it disconnects. The setpoint that actually decides how a site behaves after a scare is the other one: the recovery, or reconnect, threshold at which the load is switched back on. Set the gap between disconnect and reconnect wrong and a site either flickers on and off at dawn, hammering its equipment, or sits dead all day when it should have recovered. This page focuses on that reconnect setpoint and the hysteresis band between the two thresholds, and on how to tune them so an RTU comes back cleanly instead of chattering or stranding itself.
LVD Recovery Threshold in one line: The low-voltage disconnect recovery threshold, also called the reconnect setpoint, is the higher battery voltage at which a controller switches the load back on after a low-voltage disconnect has shed it. The gap between the lower disconnect voltage and this higher reconnect voltage is the hysteresis band, and it exists to stop the load from rapidly cycling on and off. Set the band too narrow and the RTU chatters on and off as the voltage bounces around the setpoint; set the reconnect too high and the site can stay stranded off-line all day even after the battery has partly recovered.
A low-voltage disconnect has two setpoints, not one. The disconnect threshold is the lower voltage at which the controller opens the load to save the battery from deep over-discharge, and the reconnect threshold is the higher voltage at which it closes the load again once the battery has recovered enough to carry it. The vertical gap between these two voltages is the hysteresis band, and it is the single most important tuning parameter for how a site behaves around a disconnect event. The disconnect setpoint gets all the attention because it is the protective action, but the reconnect setpoint and the band around it determine whether the recovery is graceful or disastrous.
Hysteresis exists to prevent chatter, which is the rapid on-off cycling that would otherwise happen at the disconnect voltage. The moment the load is shed, the battery voltage rebounds, because a battery under load reads lower than the same battery at rest, so removing the load immediately lifts the measured voltage. If the reconnect threshold were set at or just above the disconnect voltage, that rebound alone would trip a reconnect, which would reapply the load, drag the voltage back down, trip the disconnect again, and repeat endlessly. The hysteresis band separates the two setpoints far enough that the voltage has to genuinely recover, not just rebound from load removal, before the load comes back.
The size of the band therefore has to account for how much the battery voltage rebounds when the load is removed, which depends on the load size and the battery's internal resistance. A larger load or a weaker battery produces a bigger rebound, so it needs a wider hysteresis band to avoid false reconnects, while a light load on a healthy battery rebounds little and can tolerate a narrower band. Choosing the band is really about ensuring that a reconnect only happens when the battery has actually gained real charge from the array, and not merely because the voltage jumped up the instant the load was disconnected. That distinction is the whole point of the recovery threshold.
The two failure modes of a badly tuned band sit at opposite extremes, and both show up most clearly at dawn. If the hysteresis band is too narrow, the site chatters: as the morning sun begins to trickle in, the battery voltage rises just past the reconnect threshold, the load switches on, the load drags the voltage back below the disconnect threshold, the load switches off, the voltage rebounds past reconnect again, and the RTU flickers on and off repeatedly through the early morning. This rapid cycling is hard on the controller's relay, on the RTU and radio that keep power-cycling, and it fills the SCADA system with a burst of connect and disconnect events, all while the site does no useful work.
The opposite mistake is setting the reconnect threshold too high, which strands the site off-line long after it could have safely come back. If reconnect is set at a voltage the battery only reaches when nearly fully charged, then after an overnight disconnect the site stays dark through the morning and into the day, waiting for the array to push the battery all the way back up before it will reconnect, even though the battery recovered enough to carry the load hours earlier. On a short winter day the array may never reach that high reconnect voltage at all, so a site that dipped low overnight can remain disconnected for the entire day, reporting nothing, when a lower reconnect threshold would have brought it back at mid-morning.
Tuning the band is about threading between these two extremes. The reconnect threshold should be high enough above the disconnect voltage that a reconnect signals real recovered charge rather than a mere load-off rebound, avoiding chatter, but not so high that the site waits pointlessly for a near-full battery before it will resume, avoiding an all-day strand. Practical tuning often starts from the controller or battery manufacturer's recommended values, then adjusts the band based on the site's actual load and how the voltage behaves around the setpoints, widening it if the site chatters and lowering the reconnect if the site stays down longer than the battery's real recovery warrants.
Because the recovery threshold governs how a site comes back to life, its behavior is exactly the kind of thing remote monitoring is meant to catch, and the symptoms of a mistuned band are unmistakable in the data. A cloud SCADA platform such as Merobix logging connect and disconnect events alongside battery voltage will show a chattering site as a rapid burst of alternating disconnect and reconnect events clustered around dawn, a pattern no healthy site produces. Seeing that fingerprint tells an operator immediately that the hysteresis band is too narrow for the site's load, which is a setting change rather than a hardware failure and can often be corrected remotely.
The opposite fault, a reconnect set too high, appears as a different and quieter pattern: a site that disconnects overnight and then simply stays off-line and silent through the morning and into the day, reconnecting far later than the battery's recovery would suggest, or not at all on short days. Because a stranded site reports nothing while it is disconnected, it can masquerade as a comms outage or a dead site, so correlating the last-known battery voltage before the silence with the reconnect setpoint is what reveals that the site is not broken but is waiting behind an unreachably high reconnect threshold. That diagnosis turns an apparent site failure into a setpoint adjustment.
Trending this behavior across a fleet also lets an operator tune the band with evidence rather than guesswork. Watching how far the battery voltage rebounds when the load sheds, and how long each site actually takes to recover its charge in the morning, shows whether a given band is too narrow, too wide, or about right, and whether the reconnect threshold matches the real recovery the site achieves on its worst days. Because these are configuration values rather than physical faults, catching a mistuned band in the telemetry often means the fix is a remote setpoint change that spares a truck roll, which is precisely the kind of low-cost, high-value intervention that continuous monitoring of a remote site is supposed to enable.
The disconnect threshold is the lower battery voltage at which the controller sheds the load to protect the battery from over-discharge, and the reconnect threshold is the higher voltage at which it switches the load back on after the battery has recovered. The gap between them is the hysteresis band, which keeps the load from rapidly cycling on and off. Both setpoints matter, but the reconnect threshold and the band around it decide how cleanly a site comes back after a disconnect.
That chatter almost always means the hysteresis band is too narrow. As morning sun trickles in, the voltage rises just past the reconnect threshold, the load switches on and drags the voltage below the disconnect threshold, the load switches off and the voltage rebounds past reconnect again, and the cycle repeats. Widening the band, by lowering the disconnect or raising the reconnect enough that a reconnect signals real recovered charge rather than a mere load-off rebound, stops the flickering.
Yes, and it is a common tuning mistake. If reconnect is set at a voltage the battery only reaches when nearly full, a site that disconnected overnight waits through the morning and into the day for the array to push the battery all the way back up before it reconnects. On a short winter day the array may never reach that high threshold, so the site stays dark and silent for the whole day when a lower reconnect setpoint would have brought it back hours earlier.
Merobix reads your field devices into a cloud SCADA - the real thing behind these terms, live in days from any browser.