I'm trying to diagnose what is apparently a not-uncommon problem in Honda Accords (mine is a 2005 EX with 165K on it) in which the alternator decides to not charge the battery while driving, leading eventually to the car slowly dying as the battery (which is now basically trying to power the whole car) eventually fails. I've made some progress but I need more guidance on this.

The failure mode when starting with a freshly-charged battery is generally this. Period of time between one item and the next is usually several minutes:

1. Normal driving for some long period of time (especially when A/C is not in use, or in daylight with no lighting required).
2. Check Engine light comes on. (Ours set a bad Knock Sensor code, though this might really have been due to the falling voltage.) Normal driving continues.
3. Battery light comes on. Normal driving continues.
4. ABS light comes on. Failure imminent.
5. Luminescent dash gauges and indicators start flickering and failing; car coasts to a stop.

Our battery is in good shape and bounces back immediately, within minutes of connecting jumper cables or a charger, enough to start the car and stagger on home. I have also given it a long-term charge (a day or more), let it sit for over a week, and the car started with no loss of charge, so parasitic drain is not a factor.

The battery voltage is about 12 volts across the terminals but does not rise after the car is started. If left to idle, the voltage sinks steadily, and it will eventually die from a dead battery; it just takes longer to do so in the driveway than if driving around at the time.

I've read this PDF document on the Dual-Mode Charging System here:

https://opensiuc.lib.siu.edu/cgi/vie...text=auto_pres

...and that seems to fit what I"m seeing. I also found this 1993 procedure for testing the 4-pin connector at the alternator:

https://www.2carpros.com/questions/h...d-not-charging

...and although it's old, the procedure seems consistent with the 2005 alternator, and when following that test, I found that the White/Green connector did not show +12V, which is a problem.

I could only track that harness back as far as a second harness connector clipped to the exhaust manifold support, and the White/Green wire tested dead there too. This is starting to look like a bad Electric Load Detector, but I don't want to jump to conclusions (and I don't know where to go from here anyway). The alternator is a recent replacement, done before I heard anything about the ELD on Hondas, unfortunately. Any help would be welcome; thanks!

 

Install a Denso alternator.

 

Not sure how that would help...? I'm reading no voltage on the Control circuit (White/Green) going into the alternator at the 4-pin connector. It needs a high signal on the C input to tell the alternator go into high charging mode.

 

If the ELD is bad it will usually code and not affect the charging system.

Im assuming you tested/checked all relevant wiring,connections,fuses and the belt and gonna say you have a crappy alternator.

(W/G wire should have B+(12v) for high charging and B-(ground) for low charging.With the engine running, turn on the high beams and check for B+.( There is never a time when the ELD says NOT to charge, its either high or low) You can simply cut the W/G wire and it will eliminate the ELD(same as if it was faulty) and the alternator will go into full charging.

 

Okay, thanks for clarifying. I've checked every fuse under the hood and under the dash. The drive belt is tight and in good shape.

So in the test above, I'm basically trying to load up the main bus in the fusebox with enough current flow to trigger the high signal from the ELD; is that right? If I do see that signal, then I've got a bad alternator (with apparently the same flaw as the original alternator that I just took out).

Okay to disconnect the 4-pin plug at the alternator while the engine is running, or do I need to tap into the W/G wire some other way?

 

I assume you mean to test it?

Any time you test any electrical component on a vehicle,the test has to be done with the connector plugged in. (with the exception of resistance tests)

So either backprobe the wire at the connector or tap into it another way.

(You can disconnect the connector if it is easier to probe/pierce the wire that way,just be sure to replug the connector in for testing purposes)

 

Well, that was what I was asking, yes: whether I could just unplug the 4-pin connector at the alternator in order to test for +12V at the W/G pin, or whether unplugging the whole connector would just give me a misleading result due to the other three wires getting disconnected at the same time.

All righty; thanks. Will report back later...

 

Sorry for the slow reply; real world interrupted for a while...

Okay, back on this now. The car had sat for at least a couple of weeks after the last time the problem occurred (fortunately in the driveway), but the battery had not drained any further, showing about 10.5 volts before startup, and that was enough to start the engine right away. The alternator then brought the voltage up to 12-13 volts almost immediately. Backprobing the white/green ELD wire at the alternator plug (leaving it connected to the alternator at all times) showed about 7.2 volts.

Idling for a while showed no change in what looked to be healthy voltage. I then loaded it up with Max A/C, high-beam headlights, 4-way flashers and stereo. Voltage sunk again to around 10.5; no change in 7.2V reading on the ELD wire at the alternator. Car continued to idle and voltage stayed where it was. I then turned off all accessories again and voltage returned to 12-13 volts.

Finally I took a short test drive around the neighborhood, but the symptoms did not return, and testing when I got home showed healthy battery and 7.2V still showing on the ELD wire.

At this point, for lack of a better idea, I'm going to resume short trips around town again, and if/when I get the battery warning light again, I'll see what voltages I can measure, especially on the ELD.

 

10.5v with the engine running?
If so the alternator is bad. It should not sink lower than 12.5, even if the ELD has completely failed.