You put in a new battery — or jump-started the old one — and three days later it's flat again. A battery is a bucket: it holds a charge, it doesn't make one. If the bucket is new and it still empties overnight, something is drinking from it while the car sits. That's parasitic drain, and this is the complete walkthrough for finding it: the classic fuse-pull test, and the millivolt voltage-drop method for modern cars where fuse-pulling ruins the test.
Parasitic-drain testing takes patience. Don't start it for a problem that isn't a drain:
If the battery is young, the alternator charges fine, and nothing obvious is left on — and the car is fine when driven daily but dead after sitting a few days — you've got a drain. Time to measure it.
Your car never truly draws zero. With the ignition off, the clock, radio presets, alarm and immobilizer, keyless-entry receiver, and the ECU's keep-alive memory all sip a little power. Published guides agree on what healthy looks like:
For scale: a 50 mA draw would take a healthy 60 amp-hour battery roughly 50 days to fully discharge in theory — in practice a good car can sit 2–4 weeks without trouble. So if yours is dead in three days, the draw is big, and the meter will show it.
There are two ways to find the drain. One is the classic method every guide teaches; the other is the one modern cars demand. Picking wrong doesn't just fail — on a modern car it can reset the very thing you're hunting. Choose before you touch anything:
Pre-mid-2000s, a handful of modules, no network drama — or a car you know has no CAN-bus sleep cycle to worry about.
→ Method A: the fuse-pull hunt below.
Pulling a fuse wakes nothing important, and the classic test is quick and decisive.
BMW, Audi, newer Ford and GM, basically anything 2010s or newer with modules that talk to each other.
→ Method B: the millivolt voltage-drop test below.
Pulling a fuse wakes the module on that circuit — it chatters on the network, other modules wake, the sleep timer restarts, and your reading is now meaningless. Some guides even warn it can reset the ECU and hide the problem you came to find.
Not sure which you have? Use Method B. It works on any car, never disconnects anything, and resets nothing. The only trade: it needs a meter with a decent millivolt range.
Both methods need the same prep: the car must be fully asleep before you measure. Turn everything off — lights, HVAC, infotainment — unplug chargers and the OBD dongle, disconnect any underhood light. Close all doors and the hood; if you need access while it's asleep, fool the latches with a screwdriver so the car thinks everything's shut. Then wait 20–30 minutes (some cars need up to 45–60). Do not open a door or push a button during the wait — every poke restarts the clock.
What you need: a digital multimeter with a dedicated 10A DC current port — the $25 kind works — plus a fuse puller. Gloves and eye protection, and take off your watch and rings: you're working around a live battery.
Move the red lead to the high-amperage port (labeled 10A or 20A) and turn the dial to DC amps. Start on the high setting, not milliamps — if the drain is big and you're on the mA range, you'll blow the meter's internal fuse.
Loosen the negative (−) terminal clamp and lift it off the post. You don't need to touch the positive side at all.
One probe on the negative battery post, one on the disconnected cable clamp. All of the car's current now flows through your meter, in series. Alligator clips on the probes make this hands-free — recommended.
Under 0.050 A (50 mA) after the sleep wait: no drain — your problem is the battery, the charging system, or a connection, not a phantom draw. Over 0.100 A (100 mA): confirmed drain, start the hunt.
Find your fuse boxes — owner's manual or the lid diagram; many cars have one under the hood and one in the cabin. Pull a fuse, watch the meter. No change? Put it back in the same spot and move to the next. Reading drops dramatically — say from 0.200 A down to 0.030 A — that's the circuit. It's pure process of elimination; there may be more than one offender, so keep going after the first hit.
One honest caveat about this method: disconnecting the battery cable can reset the car's computers — sometimes that's a minor annoyance (radio presets), sometimes it clears the very intermittent fault you're chasing. That's exactly why modern CAN-bus cars get Method B instead.
The trick: you don't measure the drain at the battery at all. A fuse is a tiny resistor, and any current flowing through a resistor drops a tiny voltage across it. Measure that drop and you know current is flowing — fuse by fuse, with the battery connected and the car asleep.
Do the same sleep prep as Method A (everything off, latches fooled, 20–30 minute wait). The whole point here is that you never break the battery connection, so nothing resets and nothing wakes up.
Every blade fuse has two tiny metal test points on its top, exposed through the fuse housing. Touch one probe to each test point. Don't push hard — you're just making contact.
0.0 mV = the circuit is dead. Good. Any positive reading — even 0.5 mV — means current is moving through that fuse while the car is supposed to be asleep. Check every fuse in every box: engine bay and cabin. The drain could be more than one circuit.
The mV reading alone tells you the circuit is live; to know whether it's the killer, convert it to current. Published blade-fuse voltage-drop charts map a millivolt reading to amps based on the fuse's rating (its color/amp number) — search "blade fuse voltage drop chart" for your fuse type. A rough sense: about 0.5 mV across a 10-amp fuse works out to roughly 65 mA — already over the 50 mA healthy ceiling by itself. Any circuit converting to a meaningful share of that 50 mA budget is a suspect.
The fuse tells you the circuit, not the component. One more round of narrowing:
Across the published guides and owner forums, the same culprits come up again and again — check these against whatever circuit the meter points at:
The little switch that should kill the light when the lid closes fails, and the bulb burns all night. The phone-camera trick above catches it in seconds.
Aftermarket radios and amps are wired by humans, sometimes badly — a constant-power wire where a switched one belongs keeps the unit sipping overnight.
Add-on security and remote-start systems are a classic drain source, especially on used cars where the install is a mystery.
Some 12V sockets stay live with the key out. A dash cam or phone charger plugged in there is a small but constant draw.
The OBD-II port is always powered, so anything plugged into it — a dongle, a tracker, a monitor — keeps drawing with the car off. A typical dongle draws 20–50 mA by itself: not a killer alone, but stacked on other draws it matters.
Covered above — drains overnight, invisible to the fuse hunt. Test by checking for the drain, then disconnecting the alternator's wiring and watching the reading.
The procedures above come from published guides, not from my head: