Potential battery life improvement?


Binky

Death's Steed
Staff member
Joined
May 28, 2003
Messages
6,997
Location
16A (TO)
If someone who really understood the Pandora's hardware and low level programming got their teeth into it, could the Pandora’s standby/light usage battery life be improved much?


n900_power_management.png


Are we waiting for dsp support?


Do we have the necessary hardware?


Does anyone know?


Have I missed a obvious thread about this (I don't think so)
 
As I understand efficient standby mode is possible with current hardware but will require massive hackery into the kernel and most of the drivers.
 
I heard that there was a some money for paid developers around somewhere...


How long would it take to wake up? (from low power mode)
 
Last edited by a moderator:
I believe we need the new Linux kernel to get the OMAP power saving features working, right now the chip pulls almost full power when in a standby state, the only power saving is from turning the backlight off and WIFI off.


The new kernel supports this as far as I know, maybe not all the features but some of them. I think this might be in order for release 2 of the Pandora OS.


Wakeup time should be the same.
 
Last edited by a moderator:
It would be nice if low power mode saved more power. I like to charge my Pandora over night and usually wake up to find that it discharged at least 10% after it finished charging.

Ahh, that is due to a different phenomenon. I assume you are using HF6 alpha or beta? That is due to a new charging routine that protects the battery. There is no way currently to stop charging the battery and switch to run off only AC, so the battery gets trickle charged which is fine for NiCd and NiMh but HORRIBLE for LiPo (Like degradation, ballooning, fire and explosions bad... ), so instead when in standby mode, the battery will charge t0 100%, then drain to 90% - 80% then charge to 100%.


This has nothing to do with Low Power Mode, other then charging in Low Power Mode, it is a hardware limitation of the power control stuff.
 
I wonder if it wouldn't be feasible to make a timed charging program. Where you could input say 8 hours and it would maximize the charge cycles so that it would be near or at 100% when that time expired. I only thought of this due to most people saying they put it on charge at bedtime. Since different people sleep different lengths, it's hard to know where the charge cycle will be when you get up. Just a thought. Probably not too workable.
 
I assume you are using HF6 alpha or beta?

Nope, just HF5.
It was normal behavior for a lot of Pandoras to shut off at 100% and then restart at 80%, but it was really close to the trigger threshold: due to minor fluctuations in the resistance of the board, not every Pandora would shut off automatically. My first fix increased the sensitivity of the charge circuit, and Notaz took my numbers and put them to good use.
 
W.r.t. the discharge to 80% before re-starting charging, this will work much nicer when (hopefully) improved power saving features arrive. In practice it'll be charging to 100%, then stop. And then it would discharge very slowly, it would in practice still be at 100% in the morning if the power consumption of the N900 (with its much smaller battery) is anything to go by.
 
It is still a very bad idea to trickle charge a LiPo battery, though 30-80 mA is not bad, I still consider it a safety hazzard. If not then it still decreases the usable life of the battery, there is a reason they don't charge Li-Ion's in electric cars past 80% or below 20%, because doing so is damaging to the battery and the batteries in those cars cost a lot.
 
It is still a very bad idea to trickle charge a LiPo battery, though 30-80 mA is not bad,
Hotfix 6 doesn't trickle charge at all. Once the charge rate drops below 80ma (ie, the battery is no longer accepting as great a charge because it is full) it stops charging altogether, hence the drop to 90%
 
It is still a very bad idea to trickle charge a LiPo battery, though 30-80 mA is not bad,
Hotfix 6 doesn't trickle charge at all. Once the charge rate drops below 80ma (ie, the battery is no longer accepting as great a charge because it is full) it stops charging altogether, hence the drop to 90%

It seems to me when the unit is running, then it will continue to trickle charge, I am on HF6 A4


For instance, rightn ow my battery is at 100%, voltage is 4.201 volts, and the charge rate is at 56 mA. It is a differerent story when on standby though, the red charge light does go out.
 
Last edited by a moderator:
Back
Top