From: Matthias Klumpp <Matthias.Klumpp_at_gmx.de>
Date: Thu, 3 Feb 2005 18:28:30 +0100

Hi Linus,

first of all a short update of the latest test results:

Test of your SW during the evening/night starting at 19:00.
At 00:xx it got stuck again.
It seems pretty reproducable to me, that it hangs after more or less 5h (was
the same with last test I did).
I disconnected/reconnected the charger (with JB off/on) and I was running
the test during the night again (starting at 00:30 more or less).
I found the JB stuck this morning (of course not sure ;-)) when exactly it
hang up).
This was with your SW flashed.

During the day I tried with Archos SW flashed and your SW loaded from HDD.
The JB did not hang up, but after 4h 12min the "Battery Charging" Pop-Up
came.

I disconnected the charger and switched the JB on w/o charger
-> JB boots Archos FW, tries to spin up HDD (short "click" sound), then JB
shuts off
-> Batteries empty :-((

4h12m is (more or less) what I observed previously with V2.3 and it is also
more or less the time the JB stuck with flashed RB.
I guess that "stuck" problem with flashed Rockbox could maybe also be caused
due to low power.
The (flashed) FW was then possibly stopping and returned to the browse
screen.
That way it was continuing chargeing, so I found it in this state (I was
never seeing exactly the time it went to the browser).

What do you think ?
Please have also a look at my proposal more at the end of this mail !

I'll charge now again with Archos FW and I'll repeat that test to verify the
4h once again.

Pointing out the test conditions, so I don't have to comment that in every
mail from now on.
All tests are done with these conditions (always the same):
- Fully charging batteries with Archos FW flashed until it shows "Charging
completed".
- Using test SW
- 9V Car charger connected to power supply (input voltage 14V as in the car)
- Connecting charger
- Select always same MP3 directory and play it endless
- Stopping time until it looses juice (or get's stucked ;-?)

Below some more comments according to your last mail.

THX

Matthias

# -----Ursprungliche Nachricht-----
# Von: rockbox-bounces_at_cool.haxx.se
# [mailto:rockbox-bounces_at_cool.haxx.se]Im Auftrag von Linus Nielsen
# Feltzing
# Gesendet: Mittwoch, 2. Februar 2005 20:46
# An: Rockbox development
# Betreff: Re: AW: AW: Bug: Battery Charging with low voltage
#
# > In both cases I get a output voltage (regulated by a LM317)
# > of exactly 8.79V.
#
# That's pretty low, if you ask me.

Yeah, you're absolutely right.

My 1st mail was pretty long, so you maybe skipped a little detail - I'll
explain again:
I know, that it would be better to have 11V.
I adjusted the R's of the LM317 to get 11V and Rockbox charges w/o problems.
The disadvantage is, that I get audio distortions in the car as soon as it
starts charging (battery symbol starts running).
I tried to find out what is the source of that distortions, but until now
with bad luck (please let us NOT discuss that here - audio distortions in
the car are a different topic ;-)).

I then found out, that 9V and Archos FW works 100%.
So I thought it can't be a "principle" problem, but it must be a SW problem.
If Archos is able to do so, Rockbox also could be able (ok - charging
philosophie might be different).

With starting this topic, I wanted to find out the reasons why Rockbox
behaves different in that (sepcial) condition and (if possible) find a way
to fix it.

# 9V means a charging current of 140 mA, so it will take up to
# 10 hours to charge the regular 1500mAh batteries.

Just to point that out once again:
My intention is NOT to get batteries FULLY charged in the car.
I only would like to have at least a condition, that keeps my JB alive for
long time while being connected to the charger.

# A regulated supply should have at least 11V if you want the optimal
# charging current of 350mA. This is even more important when you have
# high capacity batteries (>1500mAh) with the Rockbox charging algorithm.

I'm using 2000mAh batteries.
Do you think it's worth to repeat the test of your SW with some lower
capacity batteries ?

# A too low charging current makes the charging take too long and the
# algorithm will time out, and the batteries will never become full.

Getting back to my proposal in one of my previous mails:
Let's assume your standby/sleep mod will not fix the problem.
Wouldn't it be a possible way (or better workaround) to implement something
like a "aggressive charging mode", user configurable in the settings:
- start charging immediately after the charger is detected
- we already have the bat. capacity in the settings
- there could be a 2nd parameter where the user specifies the charging
voltage (don't know if RB can get it from ADC)
- same way you calculated, the FW could estimate the current and that way
the total charging time (aproximatly)
- you then set the timeout (when charging stops) to that value

For sure this is not the best way to charge batteries and probably it is
able to kill the batteries, therefore it's called "aggressive" ;-)))

Any comments about that ?

# The Rockbox algorithm is designed for the original charger, or a
# regulated 11-12V charger.

I know.
As m.a. the algo works fine with 11V - no doubt about that ;-))

# It could still be a grounding problem, but let's put that
# aside for now.

Jepp ;-))

# I think the charging current problem is more important.

As increasing the charger voltage will not solve my problem (as m.a.), any
ideas what could be done ?


_______________________________________________
http://cool.haxx.se/mailman/listinfo/rockbox
Received on 2005-02-03