PV Battery Discharges Too Fast vs SOC Stuck Decision Tree

Why this matters

A residential PV battery (Tesla Powerwall, Enphase IQ Battery, FranklinWH, SunPower Sunvault) that the customer says "doesn't last as long as it should" is presenting one of three distinct conditions: actual capacity loss, BMS state-of-charge (SOC) display fault, or load drawing from the battery that the customer did not predict. Each has a different remediation. Pulling a Powerwall under warranty for what turns out to be a 6 kW heat-pump dryer running on backup is an expensive misdiagnosis. The decision tree below isolates the cause in a defined order so the warranty claim is justified or the load conversation happens.

The decision flow at a glance:

  Battery discharges too fast or SOC stuck - why?
  |
  +-- 1. Runtime matches calc? ----------> LOAD
  |                                        MANAGEMENT,
  |                                        NOT FAULT
  |
  +-- 2. Short runtime, low load? -------> CAPACITY LOSS
  |                                        (WARRANTY)
  |
  +-- 3. SOC jumps 30% to 5%? -----------> BMS TRACKING
  |                                        FAULT: RECAL
  |
  +-- 4. Stuck at 100%, backup mode? ----> MODE CONFIG,
  |                                        NOT FAULT
  |
  +-- 5. Cold garage install? -----------> TEMPERATURE
  |                                        DERATE
  |
  +-- 6. 5-yr LFP, aggressive cycling? --> AGING: 80-85%
  |                                        NORMAL

Symptom presentation

Owner reports the battery drains in 2 to 4 hours when the documentation said 10 to 14 hours. SOC display drops fast and may "stick" at a value (e.g., reads 30 percent for hours then drops to 5 percent suddenly). Some installs report SOC reaching 0 percent while still delivering power; others report SOC stuck at 100 percent for hours after charging stopped.

Quick checks

1. Read the discharge log from the manufacturer's app or portal. Tesla Powerwall, Enphase Enlighten, FranklinWH, and SolarEdge BUI all show the actual kWh and kW discharge rate.
2. Compute average discharge rate. A 13.5 kWh Powerwall 2 reporting 4 hour runtime is averaging 3.4 kW; that maps to a baseline load plus an HVAC compressor running.
3. Check what loads are on the backup circuit. A whole-home backup with everything energized (HVAC, water heater, EV charger) drains a single-battery install fast. A partial-load backup with critical-load subpanel drains slower.
4. Read battery temperature data. Cold cells deliver less than rated capacity, so a garage install in a cold climate underperforms all winter and the owner reads that as a failing battery. Charging is the harder limit: lithium cells must not be charged below freezing without damage, which is why packs sited in cold spaces refuse to accept charge, self-heat first, or charge at a crawl. A battery that "will not take a charge" on a cold morning and behaves fine by afternoon is telling you about its location, not its health.
5. Check the battery age and cycle count. LFP batteries are rated 6,000 cycles to 70 percent retention but degrade faster under heat stress or full-discharge cycling.

Isolation tree

Step 1: Reconcile reported runtime with calculated runtime. Rated kWh divided by measured kW load gives expected runtime. If they match, the battery is performing to spec; the conversation is about load management. If they do not match, advance.

Step 2: If runtime is short and kW load is below rated capacity for the period, suspect actual capacity loss. Compare current full-charge kWh against the install commissioning value. Tesla Powerwall reports full-charge capacity; a value materially below the install spec is a warranty path.

Step 3: For SOC sticking: a BMS that displays 30 percent for hours then suddenly drops to 5 percent is showing a SOC-state-tracking fault, not a real capacity fault. The cell voltages no longer correlate with the BMS's calibration curve. Field action is usually a deep discharge and full recharge cycle to recalibrate the BMS. Tesla Powerwall does this automatically; Enphase requires a manual calibration on some firmware revisions.

Step 4: For SOC stuck at 100 percent after charging stops: the battery may not be configured for self-consumption or backup-only mode correctly. A Tesla Powerwall in Backup-Only mode does not discharge during normal use - it holds at 100 percent. The customer asked for a different behavior than what was programmed.

Step 5: For temperature derate: install a battery in a temperature-controlled space if possible. UL 9540 and NFPA 855 both speak to siting; many garage installs in cold climates derate winter performance significantly.

Step 6: For aging: a 5-year-old LFP battery cycled aggressively can be at 80 to 85 percent of original capacity. Manufacturer warranties typically guarantee 70 percent retention at 10 years; below that triggers warranty.

Confirming diagnosis

Capacity loss confirms when the commissioning full-charge value vs current full-charge value shows a meaningful drop. SOC tracking fault confirms when cell voltages do not match SOC display (read individual module voltages where available). Mode misconfiguration confirms when the operating mode in the app does not match the customer expectation. Temperature derate confirms by correlated battery temperature and capacity logs.

Remediation

For capacity loss above warranty threshold: open RMA with manufacturer.

For SOC tracking fault: follow the manufacturer's documented calibration procedure for that model and firmware, and only that. Most of them amount to letting the pack run down to its own lower cutoff and then charging fully so the BMS can re-anchor the ends of its curve, but the details differ per product and some units do it automatically with no field action at all.

Note the tension with the aging discussion above, and be deliberate about it: deep cycling is one of the things that ages a pack, so a calibration cycle is a one-time diagnostic tool, not a maintenance routine. Do it when the SOC is demonstrably lying. Do not put it on a schedule, and do not run it repeatedly hoping the number improves.

For mode misconfiguration: walk the customer through the available modes (Self-Consumption, Backup-Only, Time-of-Use). Set to the mode that matches the customer's goal.

For temperature derate: relocate to conditioned space if site allows. If not, set the customer's expectation - rated capacity is at 25 C cell temperature; cold reduces capacity for the duration.

For aging beyond warranty: discuss replacement timing. LFP battery costs are still falling; a replacement decision can wait if the customer's use case tolerates reduced capacity.

After any change, monitor performance for two weeks. Document commissioning capacity vs current capacity and the operating mode on the work order.

References

1. NEC 2023 Article 706 Energy Storage Systems and Article 480 Storage Batteries.
2. NFPA 855 Standard for the Installation of Stationary Energy Storage Systems.
3. UL 9540 Standard for Energy Storage Systems and Equipment.
4. UL 9540A Test Method for Evaluating Thermal Runaway Fire Propagation in Battery Energy Storage Systems.
5. Tesla Powerwall 2 Owner's Manual, Enphase IQ Battery 5P Installation Manual, and FranklinWH aPower Service Manual on capacity reporting, SOC calibration, and operating modes.