The direct answer
If your off-grid battery is lower every morning than at dusk, something is drawing power while you sleep — and in most 12V systems the culprit is an inverter left switched on, idling away 6–18W for hours. The fix is meter-first: quantify the watt-hours lost overnight, then pull fuses one circuit at a time until the draw vanishes. With a clamp meter and an hour of patience, most drains are found the same evening.
Key takeaways
- Something is always on. Even 0.5A at 12V (~6W) burns ~72Wh across a 12-hour night.
- The #1 cause is the inverter left ON. Typical no-load (idle) draw at 12V is 0.5–1.5A (6–18W), and bigger inverters idle worse — check your datasheet for the exact figure.
- Meter first, guess never. A shunt-based monitor gives the real overnight Wh loss; a DC clamp meter finds the circuit responsible.
- Some drain is normal. Lead-acid self-discharges up to 3–5% per month, LiFePO4 typically under 1–3% per month — chemistry, not a fault.
- Fixes are cheap: a hard OFF switch or inverter remote, load-terminal discipline, and right-sizing the inverter.
Step 1: Quantify the overnight loss (meter first)
Before hunting a culprit, put a number on the loss. “The battery seems lower” is a feeling; “I lost 130Wh overnight” is a diagnosis. The meter-first approach uses a battery monitor with a shunt — a precision resistor in the main negative line that measures every amp in and out of the bank.
Overnight Wh lost = (bank Wh at dusk) − (bank Wh at dawn)
Most shunt monitors report cumulative amp-hours or watt-hours, so this is two readings. If yours shows only state of charge, convert: a 2,560Wh bank dropping from 100% to 95% lost 2,560 × 0.05 = 128Wh.
Worked example: 2,560Wh at dusk, 2,430Wh at dawn. Loss = 2,560 − 2,430 = 130Wh. Over a 12-hour night that averages 130 ÷ 12 ≈ 10.8W ≈ 0.9A at 12V — the target the fuse-pull test must explain.
Sanity-check self-discharge first. A 2,560Wh LiFePO4 bank at under 1–3%/month loses roughly 2,560 × 0.02 ÷ 30 ≈ 1.7Wh per night; lead-acid at 3–5%/month loses roughly 3.4Wh per night. Both are noise. Losing 100Wh+ means a load, not chemistry. (For how bank Wh is calculated, see battery capacity for solar systems.)
| Tool | What it tells you | Best for |
|---|---|---|
| Shunt battery monitor | Cumulative Wh/Ah in and out | Quantifying the overnight loss |
| DC clamp meter | Instant amps, no contact | Baseline and fuse-pull testing on big cables |
| Multimeter (amps, in series) | Instant amps on small wires | Small accessory circuits, under its 10A limit |
No shunt monitor? A multimeter across the posts reads voltage, not Wh — fine for before/after comparisons, but it won’t give the clean watt-hour number the isolation test wants.
Step 2: The isolation procedure (find the circuit)
One safety rule first: never connect a multimeter in amps mode directly across the battery posts. In amps mode a meter is nearly a short circuit — across a battery it blows the meter fuse instantly and can arc. The 10A jack is a small-wire limit, not a battery-cable limit.
Two safe ways to measure:
- DC clamp meter (preferred). Set to DC amps and clamp around a single conductor — the main positive cable, or the wire inside a fuse holder. No contact, safe on big cables. (Clamp both conductors together and the fields cancel: zero.)
- Multimeter in series (careful). With everything off, pull the main fuse and insert the meter in amps mode in series — current flows through the meter, not across it. Stay under 10A; reserve this for small accessory circuits.
The fuse-pull procedure:
- Switch everything OFF — inverter, loads, chargers, anything with a light. Wait a few minutes for standby modes to settle.
- Measure the baseline on the main battery cable. Write it down.
- Pull one fuse at a time, watching the reading after each pull.
- When the reading drops, you’ve found the circuit. Re-insert the others to confirm nothing else contributes.
- Trace that circuit to the device, switching items off one by one until the draw disappears.
Small baseline (0.05–0.1A) with no single culprit? The drain is several standby devices added together. Reading stays high with every fuse pulled? The leak is upstream of the fuse block — the inverter’s direct connection, the charge controller, or corroded terminals.
Step 3: The usual suspects
Once the fuse-pull test points at a circuit, the culprit is usually in this table. Draws are typical no-load (idle) figures at 12V — planning ranges, not specs; confirm against your device’s datasheet.
| Suspect | Typical draw at 12V | Overnight cost (12h) | Notes |
|---|---|---|---|
| Inverter left ON (idle) | 0.5–1.5A (6–18W) | 72–216Wh | The #1 cause. Bigger inverters idle worse; some offer power-saver modes |
| Propane / CO detector | 0.1–0.3A | 1.2–3.6Wh | Safety device — don’t disable it; account for it |
| Charge controller load terminals | 0.02–0.1A | 0.24–1.2Wh | Controller electronics plus anything left on its load outputs |
| Standby electronics (per device) | 0.02–0.2A | 0.24–2.4Wh | Chargers, radio standby, gauges — small each, but they stack |
| Indicator LEDs | ~0.01–0.02A each | ~0.1–0.3Wh each | Trivial alone; a panel full of them adds up |
| Aging battery self-discharge | up to 3–5%/mo lead-acid; <1–3%/mo LiFePO4 | ~1–4Wh/night | Normal chemistry, not a fault |
| Phantom converters | often 0.02–0.1A | 0.24–1.2Wh | DC-DC converters or adapters that stay warm with nothing connected |
The inverter row dwarfs everything else: even the bottom of its range (0.5A) is five times the top of the detector range. The small rows matter by addition — five standby devices at 0.05A each total 0.25A, about 3Wh a night. If the drain survives every fuse pull, look at the battery itself: aging lead-acid can self-discharge faster than the healthy ranges, and old cells develop internal shorts. On lithium, the internal BMS is part of that picture (see what a solar battery management system does).
Step 4: Worked example — a 0.9A mystery draw
The system: a 200Ah 12V-class LiFePO4 bank. Nominal capacity = 200Ah × 12.8V (LiFePO4 nominal) = 2,560Wh. The symptom: dusk 2,560Wh, dawn ~2,430Wh.
Step A — quantify: overnight loss = 2,560 − 2,430 = 130Wh; average draw = 130 ÷ 12h = 10.8W ≈ 0.9A at 12V.
Step B — context: usable capacity at ~80% depth of discharge ≈ 2,560 × 0.8 = 2,048Wh, so 130 ÷ 2,048 = ~6.3% of usable capacity gone every night — call it roughly 5–6%. Over a cloudy week that’s 130 × 7 = 910Wh, nearly half the usable bank, lost to nothing you ever turned on.
Step C — fuse-pull test: everything off, the clamp meter reads 0.9A on the main positive cable. Fuse-block pulls change nothing — until the main inverter fuse comes out and the reading drops to ~0.02A (controller electronics only). The inverter circuit is the drain.
Step D — confirm: the inverter is a mid-size pure sine unit left ON “so it’s ready.” Its idle draw sits inside the typical 0.5–1.5A (6–18W) range for 12V pure sine inverters — 0.9A × 12V = 10.8W, right in the middle. Bigger inverters idle worse, which is why an oversized unit makes this problem bigger.
The verdict: ~11W × 12h = 130Wh/night ≈ 5–6% of usable capacity, all of it inverter idle draw. The fix costs nothing: switch the inverter off at night.
Step 5: The fixes
1. Hard OFF switch or inverter remote. A physical DC disconnect on the inverter’s battery cable — or its remote on/off panel — turns a 6–18W idle burn into zero. Make “inverter off at bedtime” a habit. Power-saving (“load sense”) modes can cut idle draw, but check the datasheet: some still draw a fraction of full idle and add a sensing delay to every load start.
2. Load-terminal discipline. Controller load terminals are for small, always-on loads — not a dumping ground for everything in the shed. Anything you don’t need at 2am belongs on a switched, fused circuit. Keep the detector on its own unswitched fuse.
3. Right-size the inverter. Idle draw scales with inverter size, so a 3,000W unit idling for a 300W load set wastes capacity every hour. Match the inverter to your real continuous loads, or add a small dedicated unit for the few AC devices you run at night. Choosing between types? See pure sine vs modified sine inverters.
4. Kill the small stuff. Unplug or switch standby electronics you don’t use. One 0.05A device is noise; five are a slow leak you’ll chase again next season.
5. Re-verify. Repeat the dusk/dawn readings for two or three nights. A healthy system loses single-digit watt-hours overnight. Still climbing? Run the fuse-pull test again — there’s a second draw you haven’t found.
Common mistakes
- Measuring amps across the battery posts. A meter in amps mode is nearly a short circuit; across a battery it blows the fuse instantly. Clamp big cables; series-insert on small circuits only.
- Clamping both conductors at once. The fields cancel and read zero. Clamp a single conductor.
- Blaming self-discharge for a big drop. Healthy self-discharge is a few watt-hours a night, not 100+. Big drops mean a load.
- Testing with devices “off” but plugged in. Standby electronics draw 0.02–0.2A each while appearing off. Kill the fuse, not the soft button.
- Pulling fuses with a load running. You’ll chase the wrong circuit. Establish the everything-off baseline first.
- Skipping the morning reading. Without dusk/dawn numbers you never know whether a fix worked. Two readings, two nights.
FAQ
How much overnight battery drain is normal?
Why does my inverter drain the battery when nothing is plugged in?
Can I use a multimeter to find the parasitic draw?
Do charge controllers draw power at night?
Should I disconnect my battery every night to stop the drain?
Next logical reads
Solar battery not charging? (troubleshooting checklist) Battery capacity calculator for solar systems Solar battery management systems (BMS) explained Pure sine vs modified sine inverter Solar battery maintenance guide How long will a 100Ah battery run?