A13 tablet with large LiPo battery - how to?

[jess]

The small size 3.7V/1400mAh battery is a bit of mystery to me. From what I read it just provides enough juice to support short time "backup" functionality when the main DC via 2.5mm power connector is lost. That is without full support of LCD screen and I guess it's due to the A13 on-board power management and DC/DC circuitry limitations.

My question is - will it help to replace that 1400mAh battery and connect say 3000mAh 3V7 LiPo battery to power up the A13 WiFi + LCD7 and use it as a battery operated tablet? What changes need to be made to achieve it? Did anybody try to build a battery operated tablet system and can share the insights?

Thanks,
John

[olimex]

LiPo1400mA battery was stocked at Olimex long time before A13-OLinuXino was designed
you can connect any LiPo battery and no need to do any changes, AXP209 will take care to charge it properly.

[jess]

I understand that to charge LiPo is the one thing that AXP209 does well - btw. what is the charging current set by AXP209 and can we change/adjust it?

But how about using the high capacity LiPO battery as the main one?
I need a stand alone A13 tablet without using DC jack power supply other than for re-charging my 3000mAh or higher LiPo?

Kind of contradictory as User Manual page 27 recommends otherwise (unless I miss something  )

" When using the battery connector keep in mind that it is an energy solution that wouldn't be able to
power the board and all the peripherals. The voltage of a 3.7V LIPO battery would be enough to
power the processor and the memory but won't be enough to power external touchscreen LCD."

[jess]

User Manual page 27
" When using the battery connector keep in mind that it is an energy solution that wouldn't be able to
power the board and all the peripherals. The voltage of a 3.7V LIPO battery would be enough to
power the processor and the memory but won't be enough to power external touchscreen LCD."

Bump - I'd like to hear an official comment from Olimex on this subject. Is your manual:
A13-OLinuXino and A13-OLinuXino-WIFI
Open-source single-board
Android 4.0 mini-computer
USER'S MANUAL
Revision F, March 2013

statement on page 27 - is it correct or NOT?

That text is either incorrect, misleading or both as it implies that I CANNOT use the VBAT line as a Lipo battery power supply for my A13 Olinuxino-WiFi? Or perhaps that "VBAT" is to be used exclusively for RTC at 30mA via LDO1...

Olimex - am I correct?

[granite_crusher]

At this moment of implementation AXP209 is the good just and only just at charging battery. AXP209 works partly autonomous and for some functionality software is required. I can confirm from my practise that LiPo is enought to power board with LCD, tested it.
Now if You would mind to go through forum and check out all posts about lipo, power, etc... you would see there is something wrong with circuit, or AXP, or there is no software drivers.
In schematics it looks bat is connected only to AXP, at revision E it discharges battery very fast. Even more it over-discharges it (I got it discharged to 2.2V) what is not healthy to battery and can result in damaged capacity of battery. (I tested it without LCD and any peripherals, it is far from functioning correctly)

However there is hope: I saw in linux-sunxi site, in latest linux-sunxi kernel 3.4 there is pmu drivers which should take care of AXP209.

[JohnS]

Why do you think it is contradictory?  It says it can't power ALL but can power SOME.

John

[jess]

My boards are Rev.C and I experience the same and that makes the 2 of us (i.e. me and granite_crusher) to confirm that the whole business around AXP209 is unfinished and not safe to use. The bigger battery won't fix it (contrary to May 29th comment from olimex).
@JohnS - not sure what they call SOME but I've seen plenty of A13 tablets using the same AXP209 arrangement yet none of that nonsense that SOME of their circuitry work on battery

[JohnS]

Those tablets will have working software in Android that it seems no-one has put in Linux for these boards or at least no-one seems to have chased the config.

At least one person reported board fully off after pressing a particular button and you'll find that button is connected to AXP209...

John

[jess]

At this moment of implementation AXP209 is the good just and only just at charging battery.

I doubt even that .... as when I try to charge my 3000mAh or 8000mAh batteries the olinuxino android gauge goes to say 89% (i.e within few percents of being fully charged). However, after running the A13 android board from the same "charged" battery it takes just minutes before the same battery goes back to were it was before being charged. And yes, the battery's good and tested ok on my other tablets. Go figure

[JohnS]

Is the software reading the level correctly?

John

[jess]

Oh yeah, android software shows quite rapid drop in battery health within minutes to something around 65%, then (sometime) signals (or not) low battery while on other occasions (read - most of the time) simply dies without any warning - while (again sometime) corrupting its own boot sector

All that misbehaviour is supposed to be taken care of inside AXP209, not?

[granite_crusher]

I did some test long ago (probably month ago) and think that software is "stupid", probably in both android and debian. Most easy to orientate if battery is fully charged is voltage of the battery, if it shows 4.2Volts it is fully charged if >3.7 somewhere between 95% -70%, if less... well, somewhere 70-0%. Mine reached 4.2 (AXP209 charged it and when switched off charging) so AXP is able to fully charge, poor Debian xfce gui showed 0%, probably android would show something else. I unplugged DC from board and let it work for 20 min with LCD plugged in, no problem, checked the voltage (was around 3.7).

The biggest problem of axp is inability to turn off battery then voltage drops bellow 3.2 3.3 volts. it is proclaimed in manual it should, but it doesn't. At first I thought maybe there is some circuit which drains the power. Even if there would be leakage in some SY(... whatever chips) the axp is supposed to cut off bat when bat voltage drops bellow 3.2, it doesn't. ... probably because it needs to be shutdown properly:
in my tests with full battery and ampermeter between DC line I saw that board in some cases after shutdown would double current consumption.
other test:
minimal software implementation with some u-boot, kernel in SD card, or no implimentation (No SD card) - different voltage on 3.3V rail after board shutdowned/turned off(with power button) - with SD/software ~1.2V, without ~2.8V.

this will sound a bit paranoid: Is it possible that AXP reads u-boot from SD card even if board is off?