I'm with Coleslaw, except it may not be just Android but have something to do with the way the PC handles charging devices over USB.

I have a Galaxy Reverb phone and until I shut down certain features like mobile data, I was getting pretty high battery usage. I noticed that if I left the phone plugged in for a couple of days, when I unplugged it, the battery would be dead even though it had remained continuously attached.

So I decided instead to use a wall charger that I knew would deliver an constant current and the problem was solved.

edit: It just occurred to me that this might have something to do with the way USB ports handle external hard drives. IIRC there is a feature in Windows (or maybe it's the bios) that determines if external drives can be put to sleep.

My guess is that the OS/bios treats the phone as an external drive and once it stops drawing power for a certain period of time, it 'puts the device to sleep.'

So what is probably happening is the device is being fully charged but the USB port stops providing current after that happens. Why the Android OS is reporting 100% charge is probably a bug in the OS though so you might want to upgrade to the latest version compatible with your phones to see if that helps. The Reverb is a fairly new phone so I don't think it will fix the charge-then-discharge problem but it might fix the battery indicator.

Thanks for the replies, everyone. I would like to stay with the computer for charging because it has a more efficient power supply than the little wall warts which come with the phones. I checked the device manager in Win7 and verified that the phones are staying in D0 state (full power, active), so the standard 500mA from a full powered USB just can't keep up. Oh well; I can just limit the compute speed. I just wanted to check with the pros to see if there was a simple software adjustment I could do to keep the phone honest before handicapping it. Too bad this is a common Android issue...I imagine this scenario is quite common out in the wild as phones get bigger and take on more powerful roles in computing.

PS -- I'll try setting BOINC to compute above 90% rather than 100% and see if the charge meter remains accurate as long as the charge controller doesn't get parked in the "battery full" position like with computing only at 100%.

Twilyth -- Yes, those are actually the style which came with the phones. I just figured be less efficient than the 80+ gold power supply I use. I acknowledge that efficiency at such a low current is almost entirely semantics, and in all honesty, I'm just trying to keep all the crunching on one system.

An update for anyone trying to juggle all these different power supply/consumption/management variables, I have found this issue can be resolved by simply setting the minimum battery percentage to a value other than 100%. The charge meter and the associated compute suspend works normally. Edit: unless the charger gets ahead and fully charges the battery followed by the charger falling behind...same thing, then.

Well....I have had mine set to suspend BOINC if it is below 90% for months, so your results are not what I witnessed. My phone will charge to 100% and then when I remove the phone a few seconds later it will drop to like 5%. This is of course after running all night long attached to the laptop. And like I said, I also use my phone to tether the internet to the laptop when this happens. I don't charge my phone on the PC normally, so it never happens when plugged into the wall.