Rockbox on sansa consumes more power

[911boy]

I installed the latest build on my sansa e250.
seems the rockbox consumes more power than the original firmware.
Is it true? Has anyone else this same problem?

Do we have plan to improve it?

Thanks.

[Llorean]

Yes, it's true. This happens on all PortalPlayer based devices. There's no "plan" to improve it because, although we'd really like to see it improved, it's not something one can develop a plan to tackle at this time. Optimizations will help, but until someone discovers what is causing the drain, it can't be fixed. A current suspicion is hardware that isn't properly being turned off when not in use.

[X86]

I that the problem is that the processor is running all the time,so it needs something to send it to temporary sleep,I think that,and I don't want anyone to get angry.

[LinusN]

I that the problem is that the processor is running all the time,so it needs something to send it to temporary sleep,I think that,and I don't want anyone to get angry.

We have several problems on the PortalPlayer platforms. One of them is that the CPU isn't very fast, so we need to optimize the code more. Then there would be time for the CPU to go to sleep.

Also, there is another problem. Even when the CPU is completely idle, the player draws far too much power. We still don't know what causes this.

[X86]

I think is the backlit or the lcd display,the lcd display doesn't turn off even if you turn off backlit,it stays on all the time and it appears black,but if you see the lcd display with a little of light you might see that the display is working

[Llorean]

Please, random guesses aren't going to solve this. As I said, it affects ALL PortalPlayer devices, and not other Rockbox devices, which strongly suggests it's something on the PortalPlayer chip, and not how Rockbox handles other things.

As well, the original Sansa Firmware does the same thing with the LCD.

[X86]

Llorean all the portalplayer chips have the same trouble,ohh... god,ok then is a common problem,ok and I'm going to analyze and think what's making that

[cvitullo]

by the way, i think that when the original firmware turns off the screen it pauses everything but the audio, the play time is always behind when you press a button and the screen turns back on. as far as i know, rockbox doesn't do this. forgive me if i'm wrong, i'm just a noob.  

[Toni]

I did some current measurements a few days ago with the Sansa:

Deactivating:
 - the lcd-device (when backlight off)
 - the ata-controller (if no read/write in progress)
Setting/resetting the correspondent bits in the DEV_EN register did the job.
In addition the usb device can be switched off (and on if needed).
I could reduce the current from 55mA to 41mA during mp3 playback.
The cpu frequency affects by ~1mA/4MHz.
Sleeping both cop and cpu has no effect. So a better solution might be to reduce the cpu/cop frequency if both cpu/cop are idling.

As a first step, frequency scaling should be enabled for the sansa.

[JdGordon]

what current is the OF pulling during mp3 playback?

[linuxstb]

Toni,

Did you measure the effects of ATA and LCD disabling individually?  I'm curious to know their relative contributions.

[Bagder]

In addition the usb device can be switched off (and on if needed).

You know how? Done any measurements on the impact of disabling it?