Warning: don't leave your Pandora plugged in for too long

[Grench]

There has been some exploding iPod Touches before, if any of it was true.

Sell millions of a device and anything that can happen will at least once and probably twice.


With only 1000 of these in the wild, having 2 (or more) displaying swollen batteries when under ideal conditions...

 

[Prometheus]

There has been some exploding iPod Touches before, if any of it was true.

That was true (and it affected a number of other products, and resulted in large recalls), however from what I can tell from this thread, it is not related to this issue. The issue you're recalling related to batteries that had been manufactured incorrectly.

 

[sinoth]

Just to let you know how other devices handle this, my Android phone will vibrate and bring up a huge CHARGING DONE, PLEASE REMOVE CHARGER alert. The Pandora could probably use something similar, cause not everyone is gonna comb over the manual looking for battery 'best practices'.

 

[EvilDragon]

It should not matter. Handheld radios, cell phones, laptop computers, smart phones - they all get left plugged in 24/7 in store displays. I haven't heard of one burning down a store yet. There are no public service announcements going off telling everyone to unplug all battery operated devices each night before we go to bed or risk fire and certain death.

THe batteries don't burn the house down. LiPolymer usually just get swollen.

If your device is -able- to supply additional full-charging power to an already fully charged battery (regardless of type - and regardless of it's remaining charge capacity), then it is dangerous. The fact that it was designed to do so???? Huh? I don't think blaming TI cuts it. Who designed the battery?


Most modern Li-Ion cells are built with an overcharge protection circuit built into them. Was the Pandora battery built without an overcharge protection circuit?

And so does the Pandora battery.


I had the same on my Motorola V600, one of my old HTC phones, etc.


And you can read in this thread that the same seemed to have happened to other guys here as well who kept mobile phones plugged in for a long time.


If you look at google, there seem to be a lot of guys having similar issues with batteries from different manufacturers.


Most of the times it seems to happen if the battery is faulty or it gets old (my V600 battery was getting old, same the HTC. After two years of usage, that happened).


That could be the problem here: Even if the the overcharging circuit works, it will stop charging. But as soon as the battery goes below the level, it will start charging again.


As the Pandora is using battery power when connected to the AC Adaptor, it is basically permanently charging and discharging.


If you do that for a few months straight, I can imagine the battery is becoming old WAY faster than usually. It eats charging cycles like hell...


If there was no overcharging protection, the battery would die after a few days / weeks, not after a few months.


I'm pretty sure 99% of most modern mobile phones would suffer the same issue if you keep them charging for months without ever removing them from the AC adaptor.

 

[Exophase]

So ED, are you or craig or whoever going to send me a new battery for free, do I have to pay, or am I just not getting anything either way? ;p

 

[EvilDragon]

So ED, are you or craig or whoever going to send me a new battery for free, do I have to pay, or am I just not getting anything either way? ;p

Well, you need to talk to your distributor. If that's Craig, then that's up to him. I handle my customers, he handles his.

 

[Exophase]

Well, you need to talk to your distributor. If that's Craig, then that's up to him. I handle my customers, he handles his.

I just figured you'd all have the same policy on this, regardless of the one whom I'd actually be in contact with.

 

[EvilDragon]

I just figured you'd all have the same policy on this, regardless of the one whom I'd actually be in contact with.

No, not really. We are two different stores and companies

 

[Grench]

And so does the Pandora battery.


If there was no overcharging protection, the battery would die after a few days / weeks, not after a few months.

Whew - good to hear that these protections were not given a short cut.

That could be the problem here: Even if the the overcharging circuit works, it will stop charging. But as soon as the battery goes below the level, it will start charging again.
As the Pandora is using battery power when connected to the AC Adaptor, it is basically permanently charging and discharging.


If you do that for a few months straight, I can imagine the battery is becoming old WAY faster than usually. It eats charging cycles like hell...

Would it make any difference between using a charger Vs a USB powered connection for maintenance charging?


i.e. if I were to leave a Pandora for a few weeks, would it be better off if left on the wall charger or if left on the lower amperage connection to a computer?


Batteries are wear out devices - and should be expected to be changed periodically in a long term usage device. However, the device should not be able to damage the battery.


Is there a solution where we can set a software switch to -stop- charging the battery? You mentioned a way to turn the unit on when only running on the AC power adapter. If started using that method, with a battery in the device, would it ignore the battery and not charge it?


Thank you for your assurances on the built in protection circuits and pressure switch. It would be interesting sometime to see a tear down or pre-assembly picture thread on the internal workings of the Pandora battery packs.

 

[WizardStan]

i.e. if I were to leave a Pandora for a few weeks, would it be better off if left on the wall charger or if left on the lower amperage connection to a computer?

If leaving it for a few weeks, it is better to remove the battery and attempt to run entirely on AC. It's an experimental feature in that it doesn't work for everybody, but it should work for almost everybody.


I would advice against leaving it on USB charging. Some of the things I've read in the spec suggest it doesn't auto throttle the amps in the same way it does on AC, meaning it might actually be overcharging where the AC would have forced the amps quite low.

Is there a solution where we can set a software switch to -stop- charging the battery? You mentioned a way to turn the unit on when only running on the AC power adapter. If started using that method, with a battery in the device, would it ignore the battery and not charge it?

Not yet. It should be possible but I'm still experimenting with the driver, and then it's up to Notaz and MWeston (anyone I'm forgetting?) to decide if my experiments are in the right direction or if they'll actually cause more problems.

 

[EvilDragon]

LiPo batteries usually don't lose charge a lot.


So if you don't use your Pandora for a few weeks, remove the battery (that's what's recommended for any other battery device as well).


If you want to use it 24/7 on AC, remove the battery before plugging it to AC and that's it.


It just doesn't make sense leaving the battery in the Pandora charging 24/7 for weeks...

 

[Exophase]

It just doesn't make sense leaving the battery in the Pandora charging 24/7 for weeks...

It doesn't make sense that the Pandora doesn't stop charging the battery either, yet that's what it does. If the battery really does hold charge well on its own then why shouldn't the Pandora stop charging when it reaches 100%, then not start again until it's 80% or something? But of course that's what everyone else is saying. Probably the big problem is that it doesn't look like the power system is good at switching between power from the battery and power from the AC adapter. We have these big disclaimers about not inserting the battery when the unit is powered from the adapter and I guess you can't switch from off the battery either. Is this really a limitation of the hardware or is it something that can be coded?


Given that running it with the battery out is an unofficial, "experimental" feature you can't really criticize people for not having done it. It's also inconvenient to remove the battery - even if you have it plugged in for weeks that doesn't mean you'll know when you want to pick it up and unplug it. I've had small periods of doing this for whatever reason. Having the battery out also makes it less comfortable to hold since it's missing all that weight in the back. Obviously being able to leave it charging indefinitely with the battery in but you guys are acting like we're morons or crazy for wanting it.

 

[levi]

Thinking about it, ED's description above might explain it - if the Pandora will resume charging once the voltage drops below a high threshold, then it's constantly keeping the battery near 95-100%, and by doing so, keeping it fairly warm. Batteries don't like being stored in a warm place, and seem to last longer if they're allowed to discharge to a low percentage (I use 10%) every so often.


It's a shame the Pandora can't hot-swap batteries like a laptop, but I guess that shows off its mobile phone heritage - I don't know any phones that can hot-swap batteries. In fact, I don't think I know any that can run without a battery, like most Pandoras can.


Maybe it would be best if the Pandora had a mode with two charging thresholds - a 'stop' one where it currently is, and a 'start' one somewhere low, like 10%. You could put it into this mode when you know you are going to leave it for a while, and don't need it to have a full battery when you grab it to go out. Then it will charge up, stop, use the battery for a dozen or so hours until it's flat (and allow the charging circuits to cool down), before charging up again.

 

[WizardStan]

Thinking about it, ED's description above might explain it - if the Pandora will resume charging once the voltage drops below a high threshold, then it's constantly keeping the battery near 95-100%

As I mentioned earlier, when the charging chip does detect the battery is "full" (which it does for some, not others, and not always) it'll discharge to about 80% (sometimes even a little lower) before turning back on.


The problem definitely seems (to me) to be that the charge circuit kept dumping too many amps into the battery long after it became full. The TWP chip should be handling this properly all on it's own, but if it isn't we can fall back on software control, or a combination of automatic and software control.

 

[MWeston]

The problem definitely seems (to me) to be that the charge circuit kept dumping too many amps into the battery long after it became full. The TWP chip should be handling this properly all on it's own, but if it isn't we can fall back on software control, or a combination of automatic and software control.

It has been my experience that this is not true. Batteries decide how much current they will pull from your power supply based on their internal resistance. At least this is how it looks to me by doing the readings of current and voltage. It is up to the charger to limit the maximum current for safety, the same way hooking an LED directly to a 9V battery explodes the top off of it. The series resistor is the charger in this example, loosely speaking.


The Pandora battery could maybe want 10A when it is dead (although protection chip in the battery pack will disconnect if current exceeds 3.25A +/-0.25A ever) but the charger is set to 1A for realistic charging. When the battery is full, it will only try to take a small amount of energy from the power source since the potential voltage difference is near zero. If you read back the fuel gauge when the battery has been full for hours and the AC is still plugged in and enabled, you will see the small trickle current going into the battery. I've seen it as low as 19mA and regularly at 40-80mA as it gets its last little top off level. When the voltage gets close to 4.2V, all the charger does is make sure you don't exceed its setting (1A overall) and that it doesn't over-volt the battery by regulating above 4.2V. That's all it can do. There is no physical disconnect in the TPS chip because the chip was designed like a phone (or as I would say a mobile gaming device) should be used.


If we want special charging modes that people would use to leave their device plugged in for months (which is still going to kill the battery faster, you can't get around that, even my Dell laptop eats batteries because I can't find a reason to unplug it often enough), that is software that needs to be written. In fact, this idea was talked about by the community long before the Pandora was ever released. It's just that some things stay at the discussion level indefinitely and other things are more important or fun.

 

[com64]

How about plugging the Pandora into one of those Christmas light timers? You could set it to turn on from 3:00-20:00, and shut off from 20:00-3:00, for example, so it isn't charging constantly. I have an analog timer that could turn something on and off at intervals of multiples of fifteen minutes.


Wait, do you really need it plugged in all the time? You could just shut it off. You could connect it to a power bar so you don't have to unplug it all the time. (I don't have one of those, so I have to unplug my SNES and NES all of the time; they share an outlet with my television. It gets really annoying... I also have to switch AV cables behind the TV... so annoying.)

 

[WizardStan]

It has been my experience that this is not true. Batteries decide how much current they will pull from your power supply based on their internal resistance. At least this is how it looks to me by doing the readings of current and voltage. It is up to the charger to limit the maximum current for safety, the same way hooking an LED directly to a 9V battery explodes the top off of it. The series resistor is the charger in this example, loosely speaking.

That may be true of lead acid or nimh batteries, but lithium will continue to absorb energy even after it is full, to the point that the chemistry of the batteries starts to change in ways I don't fully understand. That's why lithium batteries need to be charged so carefully: they won't simply reach 100% and then increase resistance to infinite; they reach 100%, and then start, I dunno, lithium plating the inside of the packaging because there's too many electrons going in or something. That's why lithium batteries puff when they're overcharged.


This is also why trickle charging an idle battery is very bad for lithium because they don't chemically stop the charging process the way older rechargeables do.


Trickle charging a live lithium battery is possible, if the charge uses a constant amp current exactly equivalent to the current being drawn from the battery, which is what I thought the TWL was doing. If it isn't, then it absolutely needs to be made to stop charging at some point, consistently, because trusting the battery to simply stop absorbing the energy it is being given will lead to disaster.

 

[Exophase]

So, anyone up for installing a GPIO controlled relay inside my Pandora to cut the battery off so I don't have to remove it? ;p

 

[Eridger]

That may be true of lead acid or nimh batteries, but lithium will continue to absorb energy even after it is full, to the point that the chemistry of the batteries starts to change in ways I don't fully understand. That's why lithium batteries need to be charged so carefully: they won't simply reach 100% and then increase resistance to infinite; they reach 100%, and then start, I dunno, lithium plating the inside of the packaging because there's too many electrons going in or something. That's why lithium batteries puff when they're overcharged.


This is also why trickle charging an idle battery is very bad for lithium because they don't chemically stop the charging process the way older rechargeables do.


Trickle charging a live lithium battery is possible, if the charge uses a constant amp current exactly equivalent to the current being drawn from the battery, which is what I thought the TWL was doing. If it isn't, then it absolutely needs to be made to stop charging at some point, consistently, because trusting the battery to simply stop absorbing the energy it is being given will lead to disaster.

So if this is the case, that the TWL isn't handling the battery properly, can something be done about this? Is there some sort of serious danger when using the Pandora 'normally'? I think that if this doesn't get fixed and the TWL is just pouring power into the battery, I'd imagine that even with typical usage the battery is being eaten up and would die A LOT sooner than it should, right?


From my basic knowledge of batteries, with Lithiums bloat, they can explode, which could seriously injure someone or cause damage to property. Even if this is an incredibly rare, 1-in-1000 chance of occurring, couldn't OPT be held responsible for not at least attempting to fix the problem?


That's assuming that there is a problem with the Pandora design, of course...

 

[WizardStan]

So if this is the case, that the TWL isn't handling the battery properly, can something be done about this? Is there some sort of serious danger when using the Pandora 'normally'? I think that if this doesn't get fixed and the TWL is just pouring power into the battery, I'd imagine that even with typical usage the battery is being eaten up and would die A LOT sooner than it should, right?

It should take a few days before any sort of irreversible damage starts, and weeks if not months before that damage reaches a hazardous point. Don't leave it plugged in for more than 48 hours and you'll be fine. If you unplug it after a day or two, let it run down about 5-10% to let it "cool" before plugging it in, you should be fine, more or less.