Page 19 of 26 FirstFirst ... 91415161718192021222324 ... LastLast
Results 361 to 380 of 507
  1. #361  
    Quote Originally Posted by villiums View Post
    My dumpreg:
    00: 4f 27 00 1d 00 .....
    Pre switches off exactly at 3.4 Volts. Script alert and broadcast message at the same time..

    THANKS!
    I think somline got it

    He found out the register that tells the battery to go into learn mode.

    Quote Originally Posted by somline
    Will try to recalibrate tomorrow. The VAE is normally in the palm set to:
    hex 0xae * 19,5 = 3393mV
    This is what we know. Below 3,4 V

    I set it already to:
    hex 0xb0 * 19,5 = 3432 mV
    echo 0x66b0 > setreg

    May be this gives us the extra 10s to connect the charger.
    We'll see what happens ;-)

    This would be a solution. Doesn't seem to be dangerous.
    I've already tried this with somline's calibration script and it gave me about 15-20 secs to plug the charger.

    To return the register to its original state, after calibrating:

    echo 0x66ae > setreg

    Try it, you shouldn't have trouble.
    And report if it works for you (it did for me).
  2. #362  
    OK, with all these different hex values floating around, what do we do in what order?

    Play a game with mV to get extra plug in time:
    echo 0x66b0 > setreg

    Then reset the UVF:
    echo 0x01fb > setreg

    Then we tell the battery how much its supposed to hold:
    echo 0x1480 > setreg

    Then we sit and wait for the battery to almost die, wait for the Script (or Dr. Battery!) to tell us to plug it in.

    Then we don't touch it until morning.



    Is the above correct? I'm getting confused...


    M.
  3. #363  
    Argh! I set the UVF 0x66b0, got the extra time toplug in the charger. But then after charging for a minute or two the script complains that the charger was disconnected (not) and calibration terminated and ends. Tried it a couple of times with same result. Also, seems that the UVF gets reset after the script runs so you have to re-poke the value again to get the extra time.
  4. #364  
    Quote Originally Posted by villiums View Post
    Argh! I set the UVF 0x66b0, got the extra time toplug in the charger. But then after charging for a minute or two the script complains that the charger was disconnected (not) and calibration terminated and ends. Tried it a couple of times with same result. Also, seems that the UVF gets reset after the script runs so you have to re-poke the value again to get the extra time.
    mmm... weird, in fact I was about to test it, I'm at 0%, 3505mv, waiting for it to reach 3432.

    Will post what happens.

    Xanadu73: Those are basically the steps.
    I don't have the registers in hand so I can't confirm, but those are the steps

    *Edit* - Charging now, going to walk the dog
  5. #365  
    Quote Originally Posted by NuttyBunny View Post
    Xanadu73: Those are basically the steps.
    I don't have the registers in hand so I can't confirm, but those are the steps

    Fantastic! Thank you, man.


    M.
  6.    #366  
    Quote Originally Posted by NuttyBunny View Post
    mmm... weird, in fact I was about to test it, I'm at 0%, 3505mv, waiting for it to reach 3432.
    *Edit* - Charging now, going to walk the dog
    Sounds really good NuttyBunny ;-)

    Quote Originally Posted by villiums View Post
    Argh! I set the UVF 0x66b0, got the extra time toplug in the charger. But then after charging for a minute or two the script complains that the charger was disconnected (not) and calibration terminated and ends. Tried it a couple of times with same result. Also, seems that the UVF gets reset after the script runs so you have to re-poke the value again to get the extra time.
    I had the same in the early beginnings of my recalibrate tests.
    Here is my old post describing how complicated I did it the first times because the LEARNFlag always got cleared when the charger was connected.
    For some unknown reasons it was gone later. And all the others here didn't report such behavior.
    Try pulling out the battery for some seconds.


    Quote Originally Posted by NuttyBunny View Post
    Xanadu73: Those are basically the steps.
    I don't have the registers in hand so I can't confirm, but those are the steps
    Agree and there is no needed sequence.
    It's not clear what the clearing of the UVF will help. It seemed to help my battery. (And all these steps will be included in the app ;-)

    BTW: I just got home my app still running since some hours and suddenly the pre rebooted. I analyzed the logfile afraid finding my app did this.
    And then: It was the "RebootSchedulerApp" I have installed ;-)
    Last edited by somline; 08/27/2010 at 05:30 AM.
  7. #367  
    By the way, I didn't mess with UVF, but with VAE, which is the other value somline found:
    Code:
    Active Empty Voltage (VAE)—VAE stores the 
    voltage threshold used to detect the active empty point.
    The value is stored in 1-byte with units of 19.5mV 
    and can range from 0V to 4.978V. VAE is located in 
    the parameter EEPROM block.
    So, basically I did:
    Set the correct (advertised) capacity to the battery
    Disable overcharge check (powerd configuration file), then restart
    Set VAE to trigger Learning mode at 3432mv (a bit earlier than Palm's readings)
    Run somline's script
    - Discharge, and after 0% it took about 10 minutes to get to 3432mv
    - Plug the charger when it asked
    - Let it fully charge (2 or 3 hours until the script says so, meaning the batt is at full capacity).
    The script may or may not change something in your batt registers.

    After this you may set VAE to it's original value, and enable the overcharge check again.

    We do hope this work for everyone. I've already checked out somline's Dr. Battery and let me tell you, it's coming around really nice!

    I just can't wait so everyone can fix their battery consumption, with this and the voltage tweaking settings on the testing kernels we will be able to get the most of our batts
  8. #368  
    Quote Originally Posted by NuttyBunny View Post
    After this you may set VAE to it's original value, and enable the overcharge check again.

    What's the original value?


    M.
  9. #369  
    Quote Originally Posted by Xanadu73 View Post
    What's the original value?


    M.
    0xae

    Quote Originally Posted by somline
    The VAE is normally in the palm set to:
    hex 0xae * 19,5 = 3393mV
    This is what we know. Below 3,4 V
    Restore it with

    echo 0x66ae > setreg
  10.    #370  
    Quote Originally Posted by NuttyBunny View Post
    The script may or may not change something in your batt registers.

    After this you may set VAE to it's original value, and enable the overcharge check again.

    We do hope this work for everyone. I've already checked out somline's Dr. Battery and let me tell you, it's coming around really nice!

    I just can't wait so everyone can fix their battery consumption, with this and the voltage tweaking settings on the testing kernels we will be able to get the most of our batts
    As I always said: The script is not writing somewhere. If the value gets back its something from the battery or more likly the pre.

    Thanks for the plaudit I really hope we can get a stable recalibrate mechanism.
    I may need some help writing the helppages for the app but still working on it.
    The app itself is almost ready but the new findings have to be included and tested.
  11. #371  
    Quote Originally Posted by NuttyBunny View Post
    0xae



    Restore it with

    echo 0x66ae > setreg

    Thanks, man. I *thought* I saw it posted here, but I guess my eyes are too tired to see it... Geesh, it's only 10:11pm here...

    Thank you!


    M.
  12.    #372  
    Quote Originally Posted by NuttyBunny View Post
    0xae

    Restore it with

    echo 0x66ae > setreg
    Short explanation about the registers. You can find the addresses in the manual. In this case 66 one byte.
    To find the current value of this register just cat dumpreg:
    Code:
    00: 0f 01 00 43 00 4b 0b 0c f8 ac 1c a0 5d 20 f1 ca
    10: 01 95 f0 80 6f 6e 7f d8 02 08 00 ac ff ff ff 00
    20: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    60: 62 ff 0b 80 d5 16 b0 05 0a 32 0e 60 01 05 0b 23
    70: 07 0f 22 66 07 08 08 08 04 00 00 00 1e 14 0a 1a
    b0: 04 18
    You see on the left the counting. So in row "60:" you find all register between 0x60 and 0x6f.

    In this case you see I changed it already to b0.
  13. #373  
    Quote Originally Posted by Xanadu73 View Post
    Thanks, man. I *thought* I saw it posted here, but I guess my eyes are too tired to see it... Geesh, it's only 10:11pm here...

    Thank you!


    M.


    • 0x66b0
    • 0x01fb
    • 0x1480
    • After told to, charge until full
    • 0x66ae


    Yes?

    (I'm sorry folks, I'm usually much more "with it" than this. I've just had a fairly rough week and a half or so at work and am mentally drained...)

    Thank you for your patience.



    M.
    Last edited by Xanadu73; 08/26/2010 at 09:30 PM.
  14. #374  
    Quote Originally Posted by somline View Post
    Short explanation about the registers. You can find the addresses in the manual. In this case 66 one byte.
    To find the current value of this register just cat dumpreg:

    You see on the left the counting. So in row "60:" you find all register between 0x60 and 0x6f.

    In this case you see I changed it already to b0.

    I want to be smart again. How do I be smart again?!? IT has made me dumb. Well, the Windows part anyway.

    Windows made me dumb.

    M.
  15.    #375  
    Quote Originally Posted by Xanadu73 View Post
    • 0x66b0
    • 0x1fb
    • 0x1480
    • After told to, charge until full
    • 0x66ae


    Yes?
    Almost.
    It's 0x01fb

    It's 4:30 in the morning here ;-)Can't sleep.
  16. #376  
    Quote Originally Posted by somline View Post
    Almost.
    It's 0x01fb

    It's 4:30 in the morning here ;-)Can't sleep.

    I have it typed in the sticky note correct, I just didn't type it here correct. I'll edit my post for future generations.

    Thank you!


    M.
  17. #377  
    Quote Originally Posted by somline View Post
    It's 4:30 in the morning here ;-)Can't sleep.
    That's the reason I took over the thread, I thought you were asleep
  18.    #378  
    Since so many new user seem to hack the battery now i want to repeat my warning here.
    (Xanadu73 is knowingly a nethead here and was already helping a'lot on the finings)

    If you mistype the two numbers after "0x" you are writing something into the Battery without knowing which parameter you change.
    This can (I think) result in a desaster and may even turns your palms battery into a bomb ;-)
  19.    #379  
    Quote Originally Posted by NuttyBunny View Post
    That's the reason I took over the thread, I thought you were asleep
    Waiting for the learnflag ;-)
  20. #380  
    Quote Originally Posted by NuttyBunny View Post
    That's the reason I took over the thread, I thought you were asleep

    I keep a couple extra clocks on my (KDE!) desktop. One for Berlin (somline) and one for South East Australia (Rod) so I know when they are asleep.


    M.

Tags for this Thread

Posting Permissions