I've Another Idea For The Spare Md Slot In C1k

[louigi600]

I've noticed that the nokia BL-5C Original battery is:
it's a Li-Ion just like the Z battery and has the same nominal tension,
it's nominally 850mA/h,
very light (will not impact much on overall weight),
just slightli thicker then a standard CF,
without the plastic enclosuer it's also about 3 mm longer then a CF.

So overall it should fit in the space left bay the missing MD and add an extra 50% battery life which is not bad.

Next time I open it up I'll check ... I also need to get another BL-5C battery to dismantle out of the plastic casing.

Now some more conciderations:
is it safe to have another Li-Ion in parallel with the standard Z battery ?
will the different capacities of the batteries cause any trouble ?
can the internal charge circuitary handle the 2 batteryes in parallel ?
how long will it take to charge the 2 batteries (currently my C1k charges overnight ... I fear that the night will not be enough for bothe) ?

Any conciderations are wellcome ;-)

[adf]

maybe you'd want a switch to handle which battery is online/chrging?
that'd be seriously  cool, though--and give the 1000 its own unique advantages.

[louigi600]

maybe you'd want a switch to handle which battery is online/chrging?
that'd be seriously  cool, though--and give the 1000 its own unique advantages.
[div align=\"right\"][a href=\"index.php?act=findpost&pid=165782\"][{POST_SNAPBACK}][/a][/div]

how about this:
with a low voltage drop diode the extra battery could allways deliver power,
use a spare GPIO to control whether the extra battery also gets charged (I don't want to drill holes in my C1k case and glue switches around the place).

This would allow you to:
change the standard battery without having to connect external power,
potentially decice what type of charge (if it's safe for the other conciderations I made)  
  fast (only the extra battery)
  normal (only the standard battery)
  slow (bothe batterys)

[adf]

Your solution sounds much more elegant.

[dougeeebear]

I think a separate battery applet that could software control the battery's on/off and charging/not charging status as well as show the current charge level would be ideal if it was possible.

If the battery always delivered power, you couldn't do a nand backup/restore or reset the z by taking out the original battery.

[InSearchOf]

If the battery always delivered power, you couldn't do a nand backup/restore or reset the z by taking out the original battery.
[div align=\"right\"][a href=\"index.php?act=findpost&pid=165802\"][{POST_SNAPBACK}][/a][/div]

VERY good point

Late

Ps. 2 posts shy of 1000

[louigi600]

Balls... I did not think about that....

But I've a solution:
deciding whether the standard battery gets charged is just a matter of putting it in the slot or not so I do not think it's necessary to have a second spare GPIO pin for choosing this,
drop the stuff about allways deliver and select when charge .... we could have connected or disconnected trough software (the applet is out of my programming skills ... if I do this it will be trough a command line {leave the applet to some devel guru}),
charge status of the single battery is difficult because the detent circuitary is only for one battery and not for two ... but if you disconnect the standard battery and only leave the extra one in you may get relatively accurate reading of the extra battery.

There is some side affects:
if you are running from the extra battery alone and you soft switch off the extra battery you will crash the system along with maybe other undesired effects,
if you forget to soft switch on the battery before replacing the standard battery (and external power is not connected) you would have a system crash anyway.

[craigtyson]

Thinking about batteries,  a clock backup battery would be good.  I never figured out why Sharp didnt keep this from the 5500. Having just killed my C750 and returned to my 5500 I was reminded how good the 5500 was when the battery ran out.  There is no need to reset the clock on bootup with a 5500.

[louigi600]

The 5500 is even worse in certain conditions:
it has an internal battery but when that runs out you loose everything and yo go back to the bare factory default (or whatever you flashed to the physical flash).
The mass storage thet the 5500 sees is a flash emulation in ram and when the battery is drained totally you loose everything that is not on the 16Mb physical flash.
The funny thing is that some modern PDA (like some Fujitsu-Siemens devices) still have the same problem. The 5500 has been around for some time and when it came out it might have been acceptable, but the new devices that suffer the same "loose all customization + all personal data"  problem on total power drain is pretty bad.

Anyway if it is possible to just power up the RTC leaving everything else not powered I'd immediately find space for a CR2016 or CR2032 (all Cxx00 models have space for that in the blask plastic separatot for the SD/MMC pcb. If anyone knows howto power the RTC alone I'm willing to give it a try.
Looking at the pxa270 tech doc the RTC is inside the PXA270 and is powered from an internally generated power domain ... so at first glance it's most likely not possible (without other concequences).

[Da_Blitz]

i was under the impresion that you needed to have a spare 32khz osscilator for teh rtc if you were running it fram a backup battery

[louigi600]

Hum ....  32kHz quarts can most likely be retreaved from an old mobo or an old mobile phone or any piece of scrap will do .... the problem would be placing the oscilator circuit near where it needs to be.

I stopped looking itnto the thing when I saw that the RTC is inside the PXA270 and that it's power domain is internally generated (VCC_RTC RTC and power manager supply Switched between VCC_OSC and VCC_CORE) looks like you need to power up core to get RTC powered ... so that would be kind of pointless (and you would also not be able to power down properly the pxa270) .... so unless I'm way wrong it's not acheivable this way.

Waht could be done is maybe find 2 spare gpio pins and have a separate home mede RTC that comunicates serial with the system. Tthis could be done with a PIC microcontroller for instance ... i've seen amongs the application notes something that would keep time more or less as accurately as a standard RTC just by using the internal XT oscillator  and an external xtal, te serial routines can be implemented by the pic itself. It could me bade to fit very confortably in the spare MD slot or even in the black mmc pcb separator ... but writing a kernel driver for it is a totally different story ;-) .... it takes much less effort to just set the clock manually when you need it.

[Da_Blitz]

why not just buy a i2c rtc chip and wire it trough some converters (or get a chip that does i2c at 3.3v as the pxa i2c line is 3.3v)

then just write a user space app to open the i2c device, grab the time and use ntp or somthing to input the time (i belive there is support for external time keeping devices)

[louigi600]

Yea ... that could cost less effort on the whole ... and if you cannot find a I2C RTC then I've seen I2C comunication implemented on pic application notes ... so I'ts possible anyway ;-)

Da_Blitz ... you're the hardware guru ... save me from having to write pic assembly stuff ...

[Da_Blitz]

when i dought refrence somone who has done it

i know for a fact that the nslu has an onboard rtc connected via i2c and there is a good chance its 3.3 v (nslus use an xscale processor) and its battery backed up

if you want the chip just tell me and i will send it to you. i use ntp for all my time stuff on my home network

links: http://www.nslu2-linux.org/wiki/Info/RTCChipInfo

[louigi600]

Intresting .... I'm going on vacation tomorrow .... but we can talk more about I2C rtc when I get back.

NTP yea I use it on all my pc that stay in a fixed place and on my laptot that frequently gets connected to somewhere that can get access to ntp service .... most of the time I use Z not connected to any network so NTP is not a solution on my Z :-(