Workshop
Electronics
Tracking a 2012 Honda CR-V battery drain to Fuse 28
Taking a computer technician's troubleshooting habits under the hood: a car that needs to go to sleep, two meters, and a much smaller place to look for a battery drain.
A 2012 Honda CR-V kept ending up with a dead battery after sitting parked. As error messages go, a car that will not start leaves quite a bit to the imagination.
I am much more comfortable troubleshooting computers than cars. Give me a system with several things talking to each other, an intermittent problem, and a resource disappearing somewhere, and I have a reasonable idea how to start. Under the hood, I need a little more time with the map.
Still, this seemed like a question I could work with. Something might be using electricity while the car was parked. Could I measure it, then narrow down where it was going?
By the end of the session, I had a repeatable way to make most of the measured draw disappear. I did not have a confirmed failed component or a finished repair. I had made the mystery considerably smaller, which felt like progress worth keeping.
AI-generated illustration made for this post. It represents the troubleshooting setting, not the actual vehicle, meter connections, or a recorded reading.
Something is still using the battery
A dead battery gives you several possible directions: the battery itself, the charging system, or a load that keeps consuming power after the ignition is off. The symptom does not choose one for you.
I wanted to get a number for that last possibility. This is called a parasitic draw. Some electrical activity while parked is expected; the interesting part is how much remains once the car’s electronics have settled down.
I used a Fluke 289 in series at the negative battery connection to watch total key-off current. A second meter, a Klein Tools CL110, was used with its test leads in DC millivolts mode at the fuses. One meter could watch the whole system while the other helped look at individual branches.
One recorded Fluke reading was 0.0874 A, or 87.4 mA. At another point the car reached 0.0072 A, or 7.2 mA.
That difference was worth following. The car could get into a much lower-current state. Now I needed to understand what changed between the two.
Apparently, the car needs to go to sleep
The word off was doing a lot of work here.
I am used to computers continuing to do things after their screens go dark. The car had its own shutdown behavior to account for. Turning the key off did not mean every electronic module immediately stopped working, and opening a door or operating a latch could wake things again. Pico Automotive’s parasitic-drain testing guidance explains why disturbing the vehicle can restart that shutdown period.
There is a familiar troubleshooting catch: while trying to find out what is keeping a system awake, you can become the thing keeping it awake.
The ignition was off and the key removed. I kept the hood physically open for access and manually clicked its latch into the closed position to simulate a closed hood. Doors and the tailgate also needed to be treated as closed. Then I watched the current over time.
If you manually close the hood latch for a test, release it again before lowering the hood. Keep fingers clear of the mechanism. That is one bit of state I would prefer not to discover by closing the hood on it.
In one reproduced high-draw run, the reading remained near 90 mA after about 40 seconds, two minutes, and five minutes. Watching it persist was more useful than grabbing a number immediately after reconnecting the battery.
The MICU sleep-mode procedure reproduced from Honda service information describes a drop below 35 mA in less than 40 seconds under its specified conditions. MICU means multiplex integrated control unit, part of the body-electrical system. That is a particular service check, not a universal rule that the entire car must be asleep after 40 seconds. The applicable procedure and the vehicle’s configuration matter.
Looking for a branch without disturbing everything
The fuse boxes gave me a way to divide the problem into smaller pieces. Before pulling things out, though, there was another useful measurement to try.
A blade fuse has a small amount of resistance. Current flowing through it produces a small voltage drop that can be measured across the exposed test points on top. That lets you look for activity while leaving the circuit powered. Pulling and reinstalling fuses can itself change module state, which Pico also discusses in its explanation of the method.
I liked the idea of getting some information before changing the thing I was trying to observe. That part of the problem felt comfortably familiar.
During the high-current condition, the Klein displayed approximately 0.8 mV across under-hood Fuse 28.
There are two limits worth keeping with that observation. First, millivolts are voltage, not milliamps of current. Turning that reading into a current estimate requires the appropriate data for the fuse type and rating.
Second, looking back at the CL110’s published specifications, that is a very small reading for this meter. Its 200.0 mV DC range has 0.1 mV resolution and accuracy of ±(1% of the reading + 8 display digits). At 0.8 mV, that works out to about ±0.81 mV under the specified conditions. That bound does not tell me the meter’s actual error during the session, but it does make this a tentative clue. The stronger evidence came from watching the Fluke when I removed the fuse.
AI-generated illustration of the voltage-drop concept. The fuse holder and blank meter screen are illustrative; this is not a CR-V fuse-location diagram or evidence of the 0.8 mV reading.
For the fuse voltage test, the leads belong in the meter’s voltage inputs, with DC millivolts selected. Keep the tips on the intended test points and avoid bridging neighboring contacts. It is a different connection from measuring battery current in series.
Fuse 28, with an important qualification
The useful change came from Fuse 28 in the engine-compartment fuse box: 7.5 A, labeled INTR LT, for interior lights.
There is also a Fuse 28 under the dashboard. That one is the 15 A washer fuse. Same number, different box, different job. A fuse number needs its full address.
Honda lists those assignments on printed pages 357–358 of the 2012 CR-V owner’s manual. The engine-compartment page and interior fuse page are also available as reproductions of that manual.
Here is what happened with the under-hood fuse:
| Condition | Total key-off battery current |
|---|---|
| High-draw state, under-hood Fuse 28 installed | Approximately 87–100 mA |
| Fuse 28 removed during the high-draw state | Approximately 7–7.2 mA |
| Fuse 28 reinstalled | Approximately 90 mA |
Pull the fuse, and the current dropped almost immediately. Put it back, and the higher reading returned. I could repeat the change during the session.
That was the part that got my attention. I could make a specific change and watch the system respond.
Using 90 mA and 7.2 mA gives a difference of 82.8 mA, or about 83 mA. The initial 87.4 mA observation was a separate reading; these were approximate states rather than one perfectly fixed number.
A smaller mystery is still progress
It would have been easy to turn that result into “found the bad part.” What I had actually found was a dependency: most of the extra draw disappeared when I removed this supply.
That strongly implicated the branch powered through Fuse 28, including any effect it had on the car’s electronic state. It did not identify an individual load or explain why it was active. The fuse itself could be doing its job perfectly well.
This is a distinction I recognize from troubleshooting software. Disabling a service can make a symptom disappear without telling you which part of the service is wrong, or whether it was responding to something elsewhere. It gives you a smaller area to investigate. That is valuable, provided you keep the conclusion the same size as the test.
The interior-light label gave the next search a starting point: dome and map lights, cargo lighting, vanity lights where fitted, and the switches or latch inputs involved in courtesy lighting. Aftermarket wiring or LED replacements would be worth checking if present; I had not established that either was involved. The exact vehicle wiring diagram needs to decide what is actually on the branch.
The next job would be to reproduce the high draw and isolate the relevant loads one at a time, letting the car settle after each change. Door and tailgate status would also be useful to inspect with a Honda-capable diagnostic tool. The interior-light cut relay and MICU-related controls remained possibilities if the simpler loads and inputs were eliminated. None of this was enough evidence to order a MICU.
The battery and charging system still deserved their own checks, too. Honda’s November 2014 battery-management product-update announcement describes a charging-related problem affecting certain 2012 CR-Vs and references service bulletin 14-071. I did not establish this car’s update status. Checking applicability and completion by VIN is a separate task; it does not make the measured key-off draw disappear.
Where I left the investigation
At the end of the session, under-hood Fuse 28 was removed and set aside. The interior/courtesy lights were no longer working. With the normal battery connection restored, the car appeared to start and operate normally otherwise. That was a limited observation, not a check of every affected function.
The plan was to leave the fuse out for several days and see whether the recurring dead-battery problem stopped. The available record does not confirm the outcome of that test. The failed downstream component and a lasting repair remain unconfirmed as well.
Removing the fuse was a temporary diagnostic isolation, not a permanent repair. An interior-light label does not guarantee that lamps are the only equipment affected. The vehicle-specific wiring and equipment matter, and being able to start the car does not establish that every disabled function is safe to leave off. A proper repair needs to restore the correct-rated fuse and normal circuit functions.
I was still looking at a car whose details I was learning. But I now had a repeatable measurement and a specific branch to investigate. The part of troubleshooting I was comfortable with had given me a useful way into the parts I was still figuring out.
The next question was much more manageable: what on that branch was still asking for power?
Practical notes before trying this
A car battery can supply far more current than a handheld meter can survive. These are the precautions that belong beside the measurements, even when the interesting part of the story is following the clues:
- Never connect an ammeter directly across the battery terminals. A series current measurement goes between the disconnected negative cable and its battery post, using the correct current input, range, leads, and intact specified meter fuse.
- Never crank or start the engine through the meter’s current circuit. Avoid operating high-current loads while it is connected. Restore the normal battery connection and remove the meter from that path before starting.
- Account for reconnection and wake-up surges before selecting an input or range. The recorded setup used a bypass connection during reconnection; that does not remove the need to stay within the meter’s ratings. Do not put an unknown surge through a sensitive milliamp input.
- Wear eye protection, remove jewelry, keep tools clear of the positive terminal and bodywork, and keep sparks and flames away from the battery. If you cannot safely manage the connection and surge current, have an automotive electrician perform the test.
The Fluke 287/289 user manual, “Measuring Current,” printed pages 37–40 explains the series connection, input selection, fuse checks, and the danger of putting current inputs across a powered circuit. Use the instructions for the actual meter in your hands.
Sources and illustrations
The session readings and sequence above come from the case as preserved in the later editorial brief. The two illustrations were generated with AI for this post at my request; neither is an original field photo, an exact wiring reference, or a measurement record.
The Honda manual supplies the fuse assignments, the reproduced service procedure supplies the bounded MICU sleep check, Pico explains the diagnostic method, and the Klein and Fluke manuals document the measurement limits and meter precautions. The Honda product-update announcement is separate charging-system background. None establishes which component failed in this car.