[Gta04-owner] Staus update GTA04A5

NeilBrown neilb at suse.de
Wed May 13 01:45:51 CEST 2015

On Tue, 12 May 2015 11:25:27 +0200 "Dr. H. Nikolaus Schaller"
<hns at goldelico.com> wrote:

> Hi,
> Am 12.05.2015 um 11:18 schrieb NeilBrown <neilb at suse.de>:
> > On Mon, 11 May 2015 09:37:07 +0200 Dr. H. Nikolaus Schaller
> > <hns at goldelico.com> wrote:
> > 
> >> Hi all,
> >> you may wonder what is going on with this project. Did it turn into a black hole? Or
> >> did someone do a “rm -rf /“ :)
> >> 
> >> No, nothing of that. Therefore, I think we owe you some short status update on this project.
> >> 
> >> There hasn’t been much to report in the last months, but now we have several news
> >> within some days.
> >> 
> >> 1. I have checked again and we can still get the GTM601W, but the distributor says
> >>    the risk is high that OPTION declares EOL any time and in that case we would have
> >>    to make an order of at least 3000 units so that they restart production just for us.
> >> 
> >>   This means we must secure these modules for us now, since there is no replacement
> >>   that easily fits into the space constraints of the GTA04 (resp. GTA01/02 board).
> >> 
> >>   Cinterion modules would be nice since they are planned for the Pyra and the Neo900,
> >>   but they are approx. 10% too big.
> >> 
> >> 2. we see some difficulties in getting the W2SG0084 GPS module, but that is some
> >>    paperwork the distributor has to do with Wi2Wi.
> >> 
> >>   Nevertheless, this is also a risk candidate for EOL because only one distributor shows
> >>   stock.
> >> 
> >> 3. There are also good news:
> >>    we participate in the risk buy of Samsung 1GB RAM+512MB NAND chips (as
> >>    planned for the Neo900). So the GAT04A5 gets twice as much RAM :)
> >> 
> >>    This chip has already been tested by reworking a BeagleBoard XM and modifying
> >>    the boot loader (kernel didn’t find the NAND yet, but U-Boot did. This is a problem
> >>    that will jointly be fixed with the Neo900 software team).
> >> 
> >> 4. Production: we still are a little away from the quorum of 100 preorders (at the
> >>    moment of this writing we have estimated approx. half = 45 units). So it is still
> >>    not possible to make final decisions about production dates and final pricing.
> >> 
> >>    So we need to get a little more support for our project. If you have ideas, please
> >>    let us know or start activities. Many small activities are as good as a big one!
> >>    And we know that this is a niche product that requires an existing GTA01/02
> >>    device. So it is not possible to “attract the masses”. We must find and convince
> >>    the previous GTA01/02 owners whose devices are collecting dust. There had been
> >>    produced ca. 15k such devices so that we just need another 1/3% or in other
> >>    words only 1 out of 300 GTA02 owners needs to decide for a GTA04A5.
> >> 
> >> 5. Regarding the risk buy of the GTM601W, RAM chips and W2SG0084, I plan to
> >>    take the budget (~5100 EUR) that we already have collected.
> >> 
> >>    This will transform the vouchers into ownership of components. So it has the
> >>    consequence that we can’t refund in money (cash, bank transfer) any more, but
> >>    only in unused hardware components (i.e. 100 EUR ~ 1*GTM601+1*RAM+1*GPS)…
> >> 
> >>    For future preorder vouchers you would also buy such a component set (which we
> >>    would keep in safe warehouse of course) until we can really build the whole devices.
> >> 
> >> 
> >> So quite some news and I would be happy if we can fill up the missing preorders [1]
> >> sooner than later so that we can start production. Please think about securing your
> >> set of components for the last production batch of the most open smartphone platform
> >> that we already have.
> >> 
> > 
> > I have been thinking about ordering a GTA04A5, but there is one barrier.
> > I need to be confident that it can draw less than 10mA in standby.
> > Currently the best we have seen with GTA04A3,4 is about 20mA (and 40mA on
> > Linux 4.0, but that must be a software issue).
> > 
> > This is probably a software issue, but until it's actually been demonstrated
> > that the hardware can sit in standby using under 10mA, I cannot be certain.
> There will be >= 2 changes that affect the standby current:
> a) the 26 MHz oscillator is not always on any more
> b) IrDA receiver and RS232 level shifters can be turned on/off separately (on A3/A4 either one is always on)
> c) the HMC5883l and ITG3200 will likely be replaced by newer Bosch sensors

I don't think the sensors use enough to be significant, particular once the
ITG has been placed in low-power mode.

I've played a bit with the IrDA and RS232, seeing what the power usage is
with one of the other, and with nothing plugged into the serial port I can
detect no difference whichever is switched on.  This leads me to suspect that
neither use very much when idle.
When something is plugged in, I assume the charge-pump has to keep running to
supply +/- 5V and this wastes power.  So having separate switches is useful
but I don't think it will make the difference.

The oscillator might be relevant.  I don't think that it by itself uses very
much current, but if the clock signal is being propagated very far through
other circuitry, that could certain draw more power.
I don't think the 26MHz clock should get beyond the twl4030 when the CPU is
suspended, but I'm not really sure.  I'd be great if there was some way to
test this...

I've been trying to think what components are left that I haven't examined as
closely as I can and one that has occurred to me recently is the memory.
I presume this needs to be put in a low power mode and I assumed that it was.
Maybe it is ... I should try to find the code that does that.

> If you know something else which could affect standby current, please let me know.
> Ah, a new unknown is the new SAMSUNG RAM/NAND chip. There we have no data sheet
> but know that it is used in the N9.
> > It’s all very frustrating...
> At least for me it is fun and a playground to learn :)

Yes, it is fun.  But when one fails to make progress for such an extended
period, the  fun seems to diminish.


> BR,
> Nikolaus
> _______________________________________________
> Openmoko community mailing list
> community at lists.openmoko.org
> http://lists.openmoko.org/mailman/listinfo/community

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 811 bytes
Desc: OpenPGP digital signature
URL: <http://lists.goldelico.com/pipermail/gta04-owner/attachments/20150513/1a885b53/attachment.asc>

More information about the Gta04-owner mailing list