X5L Battery Meter with Rockbox

[paeuk]

Hi,

Hope this is the right forum for posting. Having read about RockBox in the Personal Computer World Mag (UK) I loaded it up yesterday on my X5L. Was really impressed. Thanks for all the work.

I just have a question about the battery meter and battery life on the X5L. I searched the forums and read some threads about it but didn't answer my question.

I understand the X5L port is only 10% complete on the power management.

What I discovered on my X5L when loading Rockbox (daily build 11th Nov) was that the original firmware battery meter had been showing about 25% remaining but when I loaded Rockbox is was showing about 80% full which was a bit concerning. I had set the battery capacity to 2250mAh - although changing the battery capacity seemed to make no difference on what the meter showed.

Hence my questions

1. Is the incorrect battery meter reading due to the current 10% implementation (i.e. something I can look forward to improving as the % implementation goes up)
2. With some people saying that battery life on the X5L using Rockbox is significantly shorter than with the original firmware is there any way to know what the battery % is in reality? I want to avoid "missing a charge requirement" and ending up with a heavily discharged battery which I read about in another forum. Is this really only possible by keeping an eye on the battery voltage (the boot up screen) and re-charging when it gets to about 3.17v? (correct voltage?)
3. Is it expected that when the power management is 100% that the battery life on the X5L is likely to be as good as with the original firmware? - or is it difficult to say?

Thanks again. Apologies about the long post. I've temporarily removed Rockbox from my X5L at the moment because of the above questions but look forward to using it again.

(I'm not a software developer so can't help - sorry).

Paul

[LinusN]

The battery meter is incorrect because we haven't performed the needed measurements for the battery discharge curve.

The battery life with Rockbox on the X5L should be on par, if not better than the original firmware.

[paeuk]

Ok. Thanks LinusN

Paul

[CapnBry]

Actually after the 30 Oct 11:33   Mike Sevakis commit, my X5 battery meter reads 100% for a long time (from 4.19v all the way down to around 3.90v).  

My current Debug -> View IO Ports shows:
ADC_BATTERY = 29C
Batt: 3.91v 95%

EDIT:  Ah now I get it.  I thought you were saying "we" as in the rockbox software hadn't performed the needed discharge curve measurements, but you meant "we" as in the rockbox developers.  Code says:
    /* iAudio x5 series  - still experimenting with best curve */
// Lithium ion discharge curve
    { 355, 356, 357, 359, 362, 365, 369, 374, 380, 387, 395 }

Is there information we (users) can provide to help?

[ehud42]

When I upgraded to the 2006-11-12 from the 2006-11-05 daily build I noticed that the machine would give me a low battery warning when in my opinion I still had a couple of hours or more of play back left on it.  My unscientific method of noting the displayed voltage on the boot up screen seemed to be that 3.5 v or so was the cut off.  

One of the reasons that I switched to rockbox was to see if I could improve the battery life and I think it did help.  At least that daily build has degrade the battery performance for me.  

I will be updating to the latest build to see if that helps and if there is any data that I can provide I would be happy to.  I am using the iaudio x5.

Thanks for the great work on this firm ware.

[bundy]

Recently I performed two battery tests on my X5L with build from 31. Oct.

Both these test were performed on same set of files. The difference between those two was that there were some options enabled/disabled and remote plugged in/without remote. http://rasher.dk/rockbox/battery/

In both cases the battery lasted above 30 hours. With original firmware you _should_ be able to get up to 35 hours, but I really doubt anyone repeated this outside of Cowon's labs. Accordig to MP3.com, they got with OF only 27 hours:
http://www.mp3.com/features/stories/3646.html

I did not see any recent results with normal X5 - if the battery dircharge curve for the non L version is approximately alright, maybe there is need to have an individual curve for the X5L (i.e. depending on the battery capacity)

[the_house]

Is there information we (users) can provide to help?

So is there anything we can do to help?  What measurements do you need to take?

I just finished running the battery benchmark for my new 60GB X5, and it reported 100% battery for the entire 16.5 hours before it shut off (from 4220 mV all the way to 3490 mV).  So it seems like I will have essentially no information about the battery from the battery level indicator, which is somewhat annoying.  Digging through the menus everytime I want to know the battery status will be a pain.  Not that I'm complaining, though.  Keep up the good work.  Just let us (non developers) know if there's anything we can do.

On a related note, is there an option for displaying the actual voltage instead of the % remaining?

[LinusN]

The bug that caused the battery level to stay at 100% was fixed yesterday. The resulting file from the benchmark would be helpful.

[the_house]

The bug that caused the battery level to stay at 100% was fixed yesterday. The resulting file from the benchmark would be helpful.

Oh, thanks, I didn't know that.  Do you mean you want the benchmark file from while the bug was still there?  If so, here it is.

[attachment deleted by admin, too old]

[bernych]

Hi the_house,
What kind of file/bitrate you used for your 16.5h battery test?

I'm just curious to see what a new X5 can get.

Cheers

Bernych

[the_house]

It was done by repeating the album Lateralus by Tool, encoded at 128kbps mp3.  Dircache was enabled, which I think takes up some RAM, and so perhaps affects the battery life.  The volume was at -20dB, and Etymotic ER-6 headphones where attached.  I"m not sure what other settings matter.  

 I charged the battery overnight (the green light was on, indicating a full charge), hit play, started the battery benchmark plugin, checked to see that the file was created in the root directory,  unplugged the X5, and left it until it turned off by itself.   So the screen was off essentially the entire time, and therefore I doubt you'd ever get quite that much playing time in normal use.  Also, I have no idea if my experience is typical.  I'm certainly happy with it, though.

[CapnBry]

I did a battery test too on --alt-preset-std LAME ripped music (Jimmy Buffett's latest), on repeat shuffle.  12h 29m with my own custom WPS, -20dB volume, ReplayGain enabled, all bands on the EQ set non-zero, and using the headphones that came with the X5.

EDIT:
This makes the curve something like this:
  { 355, 367, 374, 377, 379, 382, 387, 394, 401, 408, 422 }
Which presents the obvious problem that there are values separated by only 2 or 3mV for 10%.

[attachment deleted by admin, too old]