Battery Charge Light

[chris_c]

http://chrisc.bedroomcoders.co.uk/?p=77 leave this running while in low power mode and the machine will power


off once charged, even when its connected to the mains - this is fine for the charger, it doesn't even get warm.


For those that have batteries that don't trip the charged status, WizardStan made a little app to temporarily


alter the threshold, I've tested both working together and it works fine...


End result is a fully charged Pandora in the morning (I leave charging over night usually) that hasn't


been baking its battery all night (if you have one of the few that don't trigger charged status)

Hm, how does this work? Is it possible to disable the AC Input via software to stop charging and shutdown?


If yes, that would be awesome - I didn't know that!


If possible, I'd love to get proper charging added to HF6. Are you subscribed to the developer mailing list?


I want to share my ideas there and get replies about possibilities, etc.

No it waits for the charge led to go out (lame i know) then disables AC - WizardStan helped with the i2c details after the twl data sheet near did my head in.


I'd of thought people would have been more excited that I had wake on alarm working but I guess no one see's the pim possibilities ....


There's a develeopers mailing list?

 

[tsh]

So we could combine this shutdown on full with an alarm wake-up to keep the battery topped up at 7am every morning?

 

[gruso]

I'd of thought people would have been more excited that I had wake on alarm working but I guess no one see's the pim possibilities ....

This is the first I'm reading of your work. Did you start any threads?

 

[WizardStan]

No it waits for the charge led to go out (lame i know) then disables AC - WizardStan helped with the i2c details after the twl data sheet near did my head in.

I've added "idle" status to the TWL driver when it is plugged in but not actively charging. I'll send you my diff tonight when I get home so you can build it into your kernel and test that as a more official way of getting charge status rather than relying on the LED.

There's a develeopers mailing list?

Right here


I didn't know about it until ED just mentioned as well. Signed up yesterday.

 

[NoeleyC]

My previous Pandora charging light never seemed to go out when I left it to charge overnight, so it would always be at 100% when I woke up in the morning. My new replacement one does not do that, and stops charging a little bit after reaching 100%, then the battery level decreases as normal (didn't realise it until the next day!). My Pandora shutting down automatically on full charge seems like a great idea to me (provided it was already on low power mode, otherwise that could be annoying)

 

[tonsofquestions]

My previous Pandora charging light never seemed to go out when I left it to charge overnight, so it would always be at 100% when I woke up in the morning. My new replacement one does not do that, and stops charging a little bit after reaching 100%, then the battery level decreases as normal (didn't realise it until the next day!). My Pandora shutting down automatically on full charge seems like a great idea to me (provided it was already on low power mode, otherwise that could be annoying)

My newly aquired Pandora (finally!) seems to be doing this as well. I plug it in at night before bed, the red light comes on, and then it charges to full and stops - red light off, power draining. The end result is that when I wake up in the morning, it's only 80-90% full, depending on how much I'd used it the day before. This doesn't seem normal - is there anything I can do to make it do otherwise? I've been putting it in low-power mode overnight, and I could try full-power, but I don't want it to be even lower!


For reference, as discussed in this thread, I have one of the PowerPax chargers.


I also find the battery-life while in low-power mode to be unimpressive. I go out for a day and use the Pandora very lightly (or not at all, as was the case yesterday) and after starting on a mostly full charge, it's below 50%. Seems about right from the Wiki, which suggests 24 hours of low-power battery, but with a battery as large as we have, that's not a lot compared to other mobile devices - a friend's phone only lost a few % after a day in sleep mode. Granted, I realize it's still a WIP. What kinds of things still need to be worked on for improvement? Anything a programmer with very small amounts of kernel experience can help with?


Anyway, insterested in #2, but really want to fix #1.

 

[tonsofquestions]

The end result is that when I wake up in the morning, it's only 80-90% full, depending on how much I'd used it the day before. This doesn't seem normal - is there anything I can do to make it do otherwise? I've been putting it in low-power mode overnight, and I could try full-power, but I don't want it to be even lower!

it is "normal", a 3% loss per hour in low mode.


Try avoiding it running Wizardstan's hack


when it reaches ~92% it will restart charging.

Yeah, I realize the 3%/hour in LP mode. But in the linked thread, they're talking about how dropping out of charging mode, when still plugged in, is strange, and that it should be switching to a trickle charge, instead. Mine stays plugged in all night. Or did something change and I missed mention of it in a thread somewhere?


Neat thread about the hack. Must've missed in my search. Taking a look now.

 

[WizardStan]

Yeah, I realize the 3%/hour in LP mode. But in the linked thread, they're talking about how dropping out of charging mode, when still plugged in, is strange, and that it should be switching to a trickle charge, instead. Mine stays plugged in all night. Or did something change and I missed mention of it in a thread somewhere?

Like I said in the first post of that thread, "trickle charging a lipo is BAD". Originally it was thought it would be ok, there were older discussions about it. Trickle charging older li-ion batteries is typically OK, but apparently lipo are different enough from traditional lithium batteries that trickle charging causes problems. Stopping the charge is the right thing to do, give the battery a chance to "relax" a bit before starting to charge up again. I checked it out on my new Lenovo and it does very similar, stops charging when it reaches 100%. It only discharges to 98% before starting again, I think it's charge circuit has a much tighter voltage resolution than the Pandora's does, but we take what we can get.

 

[ZXDunny]

I checked it out on my new Lenovo and it does very similar, stops charging when it reaches 100%. It only discharges to 98% before starting again, I think it's charge circuit has a much tighter voltage resolution than the Pandora's does, but we take what we can get.

It's interesting that your ChargeHack1 only turns my LED off when it's been sitting at 100% for about an hour and a half. Why are we not using the same method of gauging battery charge levels as the battery indicator on the taskbar? At least it would go off at 100% and then come on again at 98% - as it is, it goes off much later than 100% and comes back on at about 90%.


D.

 

[benji_stein]

Like I said in the first post of that thread, "trickle charging a lipo is BAD". Originally it was thought it would be ok, there were older discussions about it. Trickle charging older li-ion batteries is typically OK, but apparently lipo are different enough from traditional lithium batteries that trickle charging causes problems. Stopping the charge is the right thing to do, give the battery a chance to "relax" a bit before starting to charge up again. I checked it out on my new Lenovo and it does very similar, stops charging when it reaches 100%. It only discharges to 98% before starting again, I think it's charge circuit has a much tighter voltage resolution than the Pandora's does, but we take what we can get.

Ahh. Good to know. Was this then changed in HF5? Because the other thread was still talking about trickling. I'll try your script tonight.


Out of curiosity, what model of Lenovo do you have? I have one that's a little over a year old, and it doesn't oscillate in charge - it stays at the particular value and cuts of the battery circuit entirely, running off of the wall, instead of the battery. Or, at least, that's what I was told it was doing. Then again, I have mine set to not charge until it drops below X%, and stop once it reaches Y%. Of course, I realize the Pandora can't do it that way, so yours is a nice solution.

 

[WizardStan]

It's interesting that your ChargeHack1 only turns my LED off when it's been sitting at 100% for about an hour and a half. Why are we not using the same method of gauging battery charge levels as the battery indicator on the taskbar? At least it would go off at 100% and then come on again at 98% - as it is, it goes off much later than 100% and comes back on at about 90%.

Because 100% doesn't mean it's fully charged, just that it's reached 4.2 volts. The charge chip fills the battery to 4.2 volts at the full 1amp (or 500ma if on USB) and then keeps dumping a little bit more energy into it at lower amperages to top it up completely. When the amount of energy it can put into the battery drops below a certain threshold, the chip detects that the battery is really full and stops the charging. It's kind of like if you fill a cup with water, you poor it from the tap really fast, but then as it gets to the top, you slow down the tap so it dribbles slower to top it up.


My hack isn't doing any logic itself, it's just tweaking the threshold that the charge chip already has hard wired into it. It can take more or less time to reach this "fully charged" state after reaching 100% depending on the usage. If your Pandora is actively in use with the LCD on, then I don't think it ever reaches that threshold: the amount being used out of the battery is greater than the threshold, so it just keeps topping it up. Close the lid and let it idle, it should only take about a half an hour. With the wifi on, LCD off, ssh'd in and working remotely, I've seen it take a little over an hour. If it takes you an hour and a half even with the screen closed and not in use, that just means I need to tweak the setting some more, or perhaps some kind of autocalibration is in order.

 

[WizardStan]

Ahh. Good to know. Was this then changed in HF5? Because the other thread was still talking about trickling. I'll try your script tonight.

No it's not? From the very first post it says trickle charging is bad and we shouldn't do it. Which posts are still talking about trickle charging?

Out of curiosity, what model of Lenovo do you have?

T410. I just looked it up and it uses li-ion though, not lipo, so I don't know whether it actually shouldn't be oscillating the power, or if it's using some different kind of battery technology that falls between the two or what.

 

[benji_stein]

Ahh. Good to know. Was this then changed in HF5? Because the other thread was still talking about trickling. I'll try your script tonight.

No it's not? From the very first post it says trickle charging is bad and we shouldn't do it. Which posts are still talking about trickle charging?

Sorry, I was referring to this thread, starting at about here


Where some people were turning off, and some were trickling.

T410. I just looked it up and it uses li-ion though, not lipo, so I don't know whether it actually shouldn't be oscillating the power, or if it's using some different kind of battery technology that falls between the two or what.

Ahh. A nice machine. It's possible you have some similar setting on. You could try checking under the advanced battery management settings.

 

[tonsofquestions]

Ahh. Good to know. Was this then changed in HF5? Because the other thread was still talking about trickling. I'll try your script tonight.

Yep, seems to have worked. 96% this morning. You rock, WizardStan!


Edit: Doh! Those previous couple of posts should have been as me.


That'll teach me to use a shared computer without first checking who was logged in!