A13 WIFI discharging LiPo battery

[kizzoalfa]

Hi,

I have attached a LiPo (purchased from the site as well) battery to my A13 WIFI board. The battery is charging well in case I am running from external adapter. In case I left my board running from the battery it became empty in couple of Hours while in Sleep. The same happens in case I turn off the board: in the next morning the battery is empty (running Android + 7" TSLCD). As far as I understand the board need to be capable to stay in sleep for days from the LiPo battery.

Does any of you have any idea what is the reason of the very fast discharge? It is looks like the system is consuming too much current while in sleep also when turned off.


Thanks,

Zoltan

[olimex]

it's only 1400 mAh battery after all

[kizzoalfa]

it's only 1400 mAh battery after all

Hi,

Yes, you are right. But the consumption of an android based tablet / Phone in sleep is much less than the A13 board. If you take your phone how long can it run with screen / LCD off? If I choose shut down under Android than nothing should run on the board except some circuits responsible for button wake up.

Did you ever measured sleep current for the A13 board? Or the consumption from battery in case Android system goes into sleep? I still think this consumption in sleep / shut down mode is too high. Something is discharging.

An other topic, sorry if I should not include it here, but maybe related to power management as well: in case the board is coming back from sleep / reset the LCD has a big flash. I do not know if is power management or problem of the LCD handler. In my understanding LCD wake ups should be smooth instead of a big flash.

Thanks,

Zoltan

[murerten]

The cause of the flash on lcd screen is missing circuit on this board. If you look in allwinner tablet schemas on the internet you can see the difference. I wrote this in a previous subject.

[kizzoalfa]

Hi,

I see this is a HW error. Did you find any fix for it? Does it possible to wire it to any of the free I/O's and drive the pin to control LCD? As I checked the schematics the pin is floating, but named as PIN0 what I found as non existing pin on the A13 (may I missed something).

I tried to check the datasheet of the charger IC but the version I found not really in English.

I still do not understand how the board can discharge the 1400mA battery in hours while in sleep.

Does somebody from Olimex investigated the LCD backlight or the battery discharge issue in the past?

Are you aware about any suggestions they may communicated (I did not find anything on the forum).

Thanks,

Zoltan

[murerten]

There are 2 signals that control lcd power and backlight. You can see them in the fex file:

Code Select Expand

lcd_bl_en_used           = 1
lcd_bl_en                = port:power1<1><0><default><1>

lcd_power_used           = 1
lcd_power                = port:power0<1><0><default><1>

As you see they are controlled by power0 and power1. They mean axp gpio pins called PIN0 and PIN1. PIN0 is lcd power and it is connected on the lcd extension connector. But the PIN1 is not. You can find PIN1 on gpio extension connector.

There is a reference tablet design on olimex github:

https://github.com/OLIMEX/OLINUXINO/blob/master/HARDWARE/A13-PDFs/a13-sch.pdf

You can trace LCD_BL_EN signal on page 5.

[kizzoalfa]

Hi,

thank you very much for the explanation! In this case as I checked the GPIO-1 PIN1 should be than connected to the LCD board U1 pin 4 and the trace between U1 pin4 and pin 6 has to be cut.

Thanks,

Zoltan

[kizzoalfa]

Hi, I made the change and rewired the BL Ena feature. The LCD now works without flashing at startup!

[murerten]

Hi again,

I want to say something about discharging too. When the board is powered off but lipo is connected and 12V is not connected the lipo discharges in one day. I think it is because the lipo voltage is supplied to 7208 chip. 7208 chip's enable is 0 when powered off but the output is ~4.10V (The lipo voltage = the input voltage). I think the problem is here. Does anyone have something to say about this issue?

[nfespindolar]

Hi, I've got some problems with the battery. Once I plug a 3.7 V 1000mA LiPo (No matter whether booting from SD or NAND), I get the following messages, I can't type a single command. Is it a hardware problem? Do I need a specific driver? Thanks in advance.

Code Select Expand


[   42.320000] ---[ end trace bef2ebc82b873f32 ]---
<1>Fixing recursive fault but reboot is needed!

[   42.320000] Fixing recursive fault but reboot is needed!
<1>Unable to handle kernel paging request at virtual address fffffffc
.
.
.
<0>Internal error: Oops: 17 [#10] PREEMPT

[   42.330000] Internal error: Oops: 17 [#10] PREEMPT
<d>Modules linked in:[   42.330000] Modules linked in: rtl8192cu rtl8192cu rtl8192c_common rtl8192c_common rtlwifi rtlwifi
.
.
.

<0>Internal error: Oops: 17 [#11] PREEMPT

[   42.350000] Internal error: Oops: 17 [#11] PREEMPT
<1>Unable to handle kernel NULL pointer dereference at virtual address 000001a0

[   42.350000] Unable to handle kernel NULL pointer dereference at virtual address 000001a0
<1>pgd = c0004000


And a lot of messages from Kernel of this kind:

Code Select Expand


[   42.330000]
[   42.330000] FP: 0xd98a660c:
660c [   42.330000] 660c  205b0003 205b0003 32342020 32342020 3032332e 3032332e 5d303030 5d303030 c0050020 c0050020 10624dd3 10624dd3 3b9aca00 3b9aca00 00000003 00000003

662c [   42.330000] 662c  d98a664e d98a664e 0000000f 0000000f 00000005 00000005 c07c352a c07c352a 00000009 00000009 00000001 00000001 00000113 00000113 0000000a 0000000a

664c [   42.330000] 664c  d9898e40 d9898e40 0000000b 0000000b d98a671a d98a671a c007821a c007821a 00000000 00000000 c007821c c007821c 00000001 00000001 c068e258 c068e258

666c [   42.330000] 666c  d9898e40 d9898e40 0000000b 0000000b d98a671a d98a671a c007821a c007821a 00000000 00000000 c007821c c007821c 00000001 00000001 c068e258 c068e258

668c [   42.330000] 668c  c0061a30 c0061a30 00000002 00000002 c076d4c8 c076d4c8 d98a6000 d98a6000 c005f81c c005f81c c007821c c007821c 00000001 00000001 c068e258 c068e258

66ac [   42.330000] 66ac  c076d4d0 c076d4d0 00000000 00000000 d98a6000 d98a6000 00000000 00000000 d98a671a d98a671a c007821a c007821a 00000000 00000000 c007821c c007821c

66cc [   42.330000] 66cc  00000001 00000001 c068e258 c068e258 c003a538 c003a538 d98a62e8 d98a62e8 0000000b 0000000b 00000008 00000008 bf000000 bf000000 63000004 63000004

66ec [   42.330000] 66ec  65363730 65363730 20386264 20386264 32353063 32353063 38353938 38353938 35306320 35306320 36343632 36343632 35652034 35652034 32333039 32333039


[murerten]

Axp chip handles the charging. So axp kernel driver should be compiled. Your problem may be bad power supply. Change it and see if there is a difference.

[granite_crusher]

hello,

I found the same behaviour of LiPo battery on A13 without WIFI and NAND. I found this post and wanted to check if it is true. I am using rev.E board. All devices was deatached (USB devices, LCD-TS). I left just board with fully charged LiPo (in XFCE4 it showed 4.3 Volts) over night with OS shutdown. In the morning I found LiPo battery empty (2.2Volts)

I want to say something about discharging too. When the board is powered off but lipo is connected and 12V is not connected the lipo discharges in one day. I think it is because the lipo voltage is supplied to 7208 chip. 7208 chip's enable is 0 when powered off but the output is ~4.10V (The lipo voltage = the input voltage). I think the problem is here. Does anyone have something to say about this issue?

I want to ask does this leak occurs if DC is conected?

[granite_crusher]

Hi,
Summary
I did small test this weekend, with boot and shutdown sequence and I think battery issue is connected with it. If battery is plugged in and OS gets shutdown, board doesn't. It stays on and consumes same amount of energy or even gets twice bigger consumption of power. Some combination of power or reset button holding or pressing brings board in off state and drops consumption of power ten times. In such state battery do not discharge steadfastly, but discharges "slowly". Preliminary tested after 3 hours (Yes I know, I am impatient) - battery voltage droped from 4.2 to 3.9 volts. Estimated total discharge of battery would happen in 28hours.

Introduction and methods
My board is without WIFI and NAND, ver.E.
It means Wifi have nothing to do with rapid battery discharge.
Android also have nothing to do with it, because I tested it with ubuntu linux (R18).
Disconnected near fully charged battery keeps charged weeks (checked) - it is not problem with battery.
I didn't attached LCD, no network dongle. Time to time I had usb keyboard plugged in to type in shutdown command. Board was attached through VGA sub to external monitor, but when I was testing battery everything was disconnected.
In previous post I mentioned that battery discharged (actually over-discharged) from 4.3 volts to 2.2 volts during night (6 hours). The fact that it got to 2.2 volts points to some hardware problem. In documentation there is written that AXP209 controls all things related with power (on /off sequences also), chip have to cut off battery if voltage drops bellow 3 volts. Well, previous test (2.2 volts) proves it doesn't work as supposed. This is very severe problem: discharging LiPo bellow 3 volts damages it, and reduces capacity!

Firstly I tried to insert amp meter between liPo and board. However it was unsuccessful, because board didn't wanted to start up with amp meter between. Then I thought maybe leak is general and can be measured from DC power.... well, before this test boot sequence looked weird for me (with or without battery different behaviour of power and reset buttons), but after this test shutting off looks even much more weird.
While I have just one multimeter I also checked voltage between contacts (gnd <-V-> P1(1.5) to P12(+5V_OTG_PWR)) from bottom side of board, but I couldn't be sure about board power states, because I had no amp meter between DC and board at that moments.

Results

• DC power, no LiPo
  
  
  
  • system on (~150 up to 200 mA)
  • system off with "# shutdown -P now" or through xfce GUI (current drops down to 15mA or 40mA (every second time))
    conclusions from above: logically 15mA_DC*12V/3.7V=~48mA_bat (or if 40mA_DC*12V/3.7V=~130mA_bat) is amount of current required from LiPo battery when board is off. Estimating 1400mAh should be sufficient for 28 (or just 10+) hours in off mode.
    Boot up and shutdown, power buttons works as expected. Reset, power pins in GPIO1 works right.
• DC power + fully or 99% charged LiPo
  
  
  
  • system on (150 to 220 mA depending from charging LiPo)
  • system off with:
    
    
    
    • "# shutdown -P now" or through xfce GUI (drops down to 150mA then after sec raises above 200mA and reboots)
          "# shutdown -P now" 2nd time (OS gets off, VGA signal off, looks it should be off, (for convenience  even de attaching keyboard and VGA)  - IT ISN'T: amp meter shows 300mA or above! (no surprise battery gets squeezed out so fast))
      Now I am not still sure about combinations sometimes it works, sometimes it not:
      
      
      
      • hold power button for few sec. Results
        
        
        
        • current drops to 15mA! This is important state. Unplugging DC and leaving board just on battery do not result in empty battery in few hours! I checked in 3 hours and battery still was near full (3.9V)
        • if it is another try after next or nth reboot it doesn't work (minimizes current to 200, or stays 300mA), until reattaching battery and DC and trying from beginning
      • hold power or reset button 2nd time for few sec, and board reboots
    • "# shutdown -H now" (no change in current: no lowering down, but also no rising up to 300mA; no following reboot. It is enought to shutdown once.).
      Methods as above sometimes works sometimes doesn't
      
  
  
• just on the battery (I couldn't measure current):
  
  
  
  • Voltage of 3.3V rail (for example UART) after shutdown keeps 3.3 volts, but after short reset or power button press (sometimes works, sometimes don't) drops down to ~1 volt (so if I will have some UEXT devices does it mean they will draw power out or will it cut off? (to low voltage))

I hope somebody find this useful. Is there anybody who don't have issue with this card + LiPo?

[jess]

... In this case as I checked the GPIO-1 PIN1 should be than connected to the LCD board U1 pin 4 and the trace between U1 pin4 and pin 6 has to be cut.

Question to Olimex - Is this the official fix confirmed by Olimex ?

[JohnS]

Because the PWR button is connected to the AXP209 and reduces the current, to similarly reduce the current when shutting an OS down almost certainly can be done if the shutdown sequence suitably commands the AXP209.  Possibly newer AXP209 driver code does it or else someone may wish to write it.

It doesn't look to be a hardware problem but rather that the software hasn't been written or isn't being used.

John