[Tinkerphones] Status GTA04A5

joerg Reisenweber joerg at openmoko.org
Thu Mar 16 22:31:33 CET 2017

On Thu 16 March 2017 13:19:25 H. Nikolaus Schaller wrote:
> Hi Andreas,
> > Am 16.03.2017 um 12:20 schrieb Andreas Kemnade <andreas at kemnade.info>:
> > 
> > Hallo Nikolaus,
> > 
> > On Tue, 14 Mar 2017 21:24:18 +0100
> > 
> > H. Nikolaus Schaller <hns at goldelico.com> wrote:
> >> Hi all,
> >> you may be wondering how progress is.
> >> 
> >> First of all, I have currently 11 GTA04A5 working boards on my
> >> desk (one did work initially but failed since now) which boot
> >> fine.
> >> 
> >> It turned out that our hw-test script was not testing all
> >> components as good as possible, so I have worked on it to
> >> be much better. Now it also tests WLAN, Bluetooth, GPS
> >> and some other components. In total it was much more effort
> >> to put into kernel and hw-test to really be able to do
> >> such an automated test. Reason is that you also have to test
> >> the tester... And make the kernel and user-space work stable
> >> for almost all chips and interfaces.
> > 
> > What I am wondering about is the backup battery test. I do not think it
> > is a good idea to test only once. My old semi-broken gta04 gives flaky
> > values there. So sometimes the voltage might look ok, sometimes wrong.
> > So the test might miss problems.
> Yes. It is very difficult to write tests that have 100% coverage of
> all potential failures. Especially for rare failures.
> Most of the tests are a little primitive and hence incomplete. They
> mainly test if the PCB has been soldered properly, i.e. if peripheral
> components can be accessed and respond reasonable to all stimuli.
> Sometimes they don't even check this. For example the tvout test does
> not test hardware (unless a tv set is connected to the headset jack
> and somebody has a look at the result). So it is not even an automatic
> test.
> The backup test only checks if the supercap is being charged and has
> reached a reasonable voltage since power-on. If the supercap would be
> missing or had a bad solder joint it would no charge at all.
> It does not even attempt to test the capacity of the supercap and if
> and how long it maintains RTC during battery switching.

That's probably a problem here: even an unsoldered polyacene cap would leave 
the charger circuit with a - however minimal - capacity vs GND, so it would 
charge up very quickly and for sure reach nominal voltage. For a way more 
thorough test, the charger circuit in TWL4030 should get switched off prior to 
voltage test (if possible, I don't have the TRM at hand, with the register 
specs) or alternatively RTC get checked if it has a "power failed" flag set, or 
keeps time, after brief removal of main power supply


()  ascii ribbon campaign
against html e-mail - against proprietary attachments
http://www.gerstbach.at/2004/ascii/ (German)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 198 bytes
Desc: This is a digitally signed message part.
URL: <http://lists.goldelico.com/pipermail/community/attachments/20170316/e8c7859e/attachment.asc>

More information about the Community mailing list