your post prior to this one suggests your battery is OK.
Here is the update: my multimeter does not have 10 AMP for parasitic draw test,
Even if you had one with 10A, there would be a good chance it would blow the fuse during the current surge when the car powers up after reconnecting the battery circuit. Measuring current on cars can be difficult. Many cool ways to do it though. My favorite is with a DC clamp meter. You clamp it around one of the battery cables and it measures current based on the induced magnetic field. There is no way to blow any fuse in that. For chasing a parasitic draw, best would be a pretty small unit, maxing out at around 100A, with a small clamp.
so I found on website a protocol to test alternator (below):
How to test the alternator
The easiest method to check the alternator is by measuring the voltage at the battery terminals when the engine is running.
- Start your Odyssey.
- Connect the black cable of the multimeter to negative terminal of the battery, and the red cable to the positive terminal.
- The voltage should be between 13.5 and 14.8 volts with the engine running. (My battery is 14.3 v)
- Switch on the accessories in your Odyssey, for example, headlights, AC fan at full speed, and interior lights.
- The voltage should remain around 14 volts. If it drops and remain below 13.5 volts, it indicates a problem with the current supply from the alternator. (My battery started with 14.03 v and slowly dropped to 12.08 v and seems to be stable there. When I turned off all the accessories, it came back to 14 v.)
- If the voltage drops below 12 volts with the engine running and the accessories turned on, it means the current supply from the alternator is completely cut off and all the electronics are solely powered by the battery.
The conclusion based on this test: the alternator is not completely bad, but intermittent bad. Is that right? Is the method reliable?
This definitely sounds like something is wrong with the alternator. If it is repeatable (i.e., you can repeat that careful test with identical results), I would not call it intermittent, but that is just terminology. The difference there is that if something is intermittent, I think of the failure as being GONE, not just undetected; in your case (if it is as you think), the fault (that voltage drops when loaded, after a while) is always present, just only detected under certain conditions. But it does sound like something an in-store bench test might miss.
Honda's ELD - Electric Load Detector - system is supposed to reduce alternator output in certain situations, when power is not needed and the purpose is to be more fuel efficient. But that would not make sense in this case unless the logic (it is PCM / ECU controlled, like most things) is broken, since when power demand increases, the alternator should not be throttling back.
On most cars there is a way to bypass the ECU's charging control - making the alternator charge at all times. Like back in the good old days, when if the engine was turning, the voltage regulator would be controlling the field current to hold system voltage steady at 14.4V. So reliable, and maybe slightly inefficient once in a long while.
If you're looking for a quick solution, I'd just replace or repair the alternator. If you really want to try harder on the diagnosis, I'd see if you can figure out how to hot-wire the charging control, and repeat your test to see if it holds or drops voltage. It could be that you just need to set a wire to ground or B+, or it could be that you need to set a pulse width modulated 5V signal.
And if you do take the alternator out, I'd just then start carefully taking it apart and observing. Oil? Debris? Loose screws? Loose 4P connector or B+ wire? Worn out brushes? Loose screws after the cover is off? Then there are a series of electrical tests you can do on it. I've posted a few times on this forum on alternator repairs, so if you search with my name, you should find some info.