As of 6-14-2005, Do not run Rockbox battery down!

[narf]

Let me start out by saying what a great project rockbox is! Even in the rough unfinshed stage it is such an improvement over the iRiver firmware! That said, here is my warning.

I wanted to find out just how effecient the current Rockbox firmware was on my 120, so I loaded up a playlist of about 800 mp3s at various bitrates from 128 to 320 CBR and VBR. I set the audio level to 60, turned on shuffle, reset the runtime counter, fully charged the battery and let it play away. The player went for about 12 hours and 35 minutes before it ...well it didn't exactly shut down cleanly. It looks like low power shutdown hasn't been added yet! The unit kept playing out of the ram buffer while the battery voltage dropped down to about 3.3 volts. Then the unit tried to spin up the drive to rebuffer. Not good. There wasn't enough battery capacity to run the drive, so it just got into this loop where it kept trying to spin up the drive, the display would dim, the drive would shut down, and then it would try again. I went running for my reset paperclip but even this didn't help! To get it out of this loop I had to put the unit back on the power supply and shut it down. It looks like all is OK, although I worry that I may have drained the battery past the safe limits. I don't know how much longer this loop would have kept up and if it would have eventually toasted the drive or battery.

[NPswimdude500]

that sounds....not fun. definitely won't be doing that for a while...

[dapureplaya_04]

no that's not cool. the closest i've ever got it down to was 20% battery life remaining.

[fred]

I drained my battery a week ago (I can't say which build I used - it was one of the first official builds with mp3-playback) and when RockBox recognised low battery, the icon began to blink and after five minutes or so it shut down without any problems.

[dapureplaya_04]

hmmm.......probably one of the latest builds then. But spoke to someone on IRC chat and they said the voltage shouldn't have gone down to 3.3V.

Also Narf and Fred, which bootloader are you guys using atm? V1 or V2?

[bagder]

Originally Posted by dapureplaya_04

someone on IRC chat and they said the voltage shouldn't have gone down to 3.3V.

No, that is not what preglow said. He said:

"i don't know if the hardware allows us to get to around 3 volts but if it does, we seriously need to cap it"

And before that, amiconn explained that:

"LiIon/ LiPoly batteries should never be drained below 3 volts per cell"

[qwert-e]

i agree that draining the battery is a bad idea at the moment. i accidentally did this during a file transfer and the player was stuck until i put it back on the charger. i've noticied that since then, the hard disk spins more loudly than it did before... maybe paranoia, but still,

DONT TRY DRAINING THE BATTERY!

[dapureplaya_04]

Originally Posted by bagder

No, that is not what preglow said. He said:

"i don't know if the hardware allows us to get to around 3 volts but if it does, we seriously need to cap it"

And before that, amiconn explained that:

"LiIon/ LiPoly batteries should never be drained below 3 volts per cell"

my bad i misinterpreted that.

[vinylivo]

hi folks, had similar experience!

actually it went down even further (well below 3V) and didn't shut down!

FYI I'm running CVS-050613 at the moment and have a 1600mAh battery from dapstore.com.
I just observed missing shutdown because the battery symbol was flashing and I switched over to the battery view (via Info/Debug/View Battery), it shows a graph of battery voltage over time. Because I've read on other places (good battery infos at http://www.batteryuniversity.com) it's not recommended to go below 3.0V, at 2.5V normally the PCM (li-polymer Protection Circuit Module) shuts off the battery from the rest of the system for safety reasons. Li-Polymer can explode if drained or charged too much!
After a deep discharge switchoff by the PCM you probably can't recharge the battery with the normal charger and need a special charger to get it work again. So take care!

With this knowledge in mind I switched the player off manually when the voltage reached 2.6V, actually it dropped quite fast after going below 3.3V. For the power supply chip (buck-bosst dc/dc converter, LTC3440) this isn't a problem, it works from 2.5 to 5.5V.

For the developers I would recommend to switch off the unit at about 3.2V, or even better to have an emergency discharge option for going down to 2.9 or 3.0V for longer playing time. This is especially useful if you're not near a mains plug and you just need that extra power for some reason.

I love your firmware, this really rocks!

Bless 2 da Rockbox developers!

jvo

[narf]

Originally Posted by dapureplaya_04

hmmm.......probably one of the latest builds then. But spoke to someone on IRC chat and they said the voltage shouldn't have gone down to 3.3V.

Also Narf and Fred, which bootloader are you guys using atm? V1 or V2?

I have been updating the builds at least one a day, anytime I see anything interesting on the update list. I think I have bootloader v2. (It has the USB mode on it) I can't read the bootloader on startup, it goes by too fast.

Edit: Looks like it was bootloader V1. I've just upgraded to V2, but I'm not planning on any battery tests for a while.

[TCK]

I had the same experience about a week ago, when playing one of the games on the firmware. Pluging it in solved the problem, but it did provide me with a bit of a scare.

Out of interest, how are you getting these voltage totals?

[Febs]

Originally Posted by TCK

Out of interest, how are you getting these voltage totals?

Info --> Debug --> View Battery

[TCK]

Thanks, I'll be sure to check that out.

[coreman]

The built-in battery protection circuit cuts off the battery around 2.6V...it can't discharge deep enough to damage the cells unless the protection circuit is faulty. That is the reason there is both under voltage and over voltage protection...to stop damaging discharges and damaging overcharges (FYI, "Full" Cell voltage is 4.2V).

From what I can tell...a deep discharge to 2.5v won't damage the battery and prevent it from charging.

[narf]

After thinking about what happened, my best guess is that the problem lies in the shutdown procedure. When Rockbox shuts down, it appears to write the current playback position to disk before turning off. If you are at marginal battery voltage and it tries to shut down, the battery voltage takes a big drop as the disk spins up. The spinup fails, the drive shuts down, and the system recovers. The it all starts over again. Perhaps a special low voltage shutdown procedure needs to be written that skips all disk activity and just shuts down?

[coreman]

Originally Posted by narf

After thinking about what happened, my best guess is that the problem lies in the shutdown procedure. When Rockbox shuts down, it appears to write the current playback position to disk before turning off. If you are at marginal battery voltage and it tries to shut down, the battery voltage takes a big drop as the disk spins up. The spinup fails, the drive shuts down, and the system recovers. The it all starts over again. Perhaps a special low voltage shutdown procedure needs to be written that skips all disk activity and just shuts down?

Makes sense to me....

[gorcorps]

Originally Posted by narf

After thinking about what happened, my best guess is that the problem lies in the shutdown procedure. When Rockbox shuts down, it appears to write the current playback position to disk before turning off. If you are at marginal battery voltage and it tries to shut down, the battery voltage takes a big drop as the disk spins up. The spinup fails, the drive shuts down, and the system recovers. The it all starts over again. Perhaps a special low voltage shutdown procedure needs to be written that skips all disk activity and just shuts down?

That's a great idea! That's probably exactly what's happening too after thinking about it.

[fred]

I'm using bootloader v2 and accidently today I drained battery to zero with yesterdays build. Again, no porblem occured. Strange thing...

fred

[animatorgeek]

So would running down the battery be an alternative to hitting the reset button? I ask because currently my Rockbox isn't responding to anything, not even reset. I'm at a loss for how to fix the problem. My only thought is to run down the battery until the protection circuits kick in and cut off power. It probably wouldn't take too long, either, since I'm pretty sure the disk is running as well

So... would I damage my battery and/or make it unchargeable if just let it run down? For that matter, would I damage the disk?

[linus]

The reset button is a hardware mechanism, and it is not a matter of "responding to the reset". The firmware can't stop the reset button from working. Either your button is broken, or you aren't pushing it correctly.

[bugmenot]

Id say chances are you arent pushing the reset button correctly, mine took me a few times the first time to find just the right spot. Its a big hole for a small sweet spot. heh.

Keep trying.