GTA02 GPS chip (ATR0635) VDD 1.8 power rail

[AlvieBoy]

CC: hardware @ openmoko.

Hi all,

According to GTA02V5 schematics, the GPS (ATR0635) VDD_1.8 power rail can be connected to PMU output or to the chip's 1.8V LDO, by choosing which resistors (0R)
to be connected during PCB assembly.

Right now (according to schematics) it is being powered by its internal LDO. We (the #gta02-core team) would like to remove the unnecessary pads from the layout
and schematics, but we have some questions:

* Is VDD_1.8 actually connected to the LDO ?
* Why was this chosen over PMU 1.8 power rail (IO_1V8) (or the other way around) ?
* Shall we keep it as is or switch to PMU ? (or the other way around)

Best regards,

Álvaro

_______________________________________________
hardware mailing list
[hidden email]
http://lists.openmoko.org/mailman/listinfo/hardware

[Werner Almesberger]

?lvaro Lopes wrote:
> * Why was this chosen over PMU 1.8 power rail (IO_1V8) (or the other
> way around) ?

I could see the following consequences of connecting GPS to IO_1V8:

- this may create an EMI path, so at least a bead should be there.
  (I.e., instead of R7601)

- IO_1V8 stays on in suspend while RF_3V (VDDIO, VCCx, VBP) can be
  cut. Not sure if this is an operating condition the ATR0635 would
  be happy in.

- IO_1V8 comes from a step-down converter, which is more efficient
  than an LDO. So if GPS draws a significant current on its 1.8 V
  rail, that may make a difference.

Ideally, someone with a bit of spare time on their hands would measure
GPS performance and power consumption on a GTA02, then rework it to
use IO_1V8 instead of the internal LDO, and then measure again, and
also check what happens if RF_3V is cut.

- Werner

_______________________________________________
hardware mailing list
[hidden email]
http://lists.openmoko.org/mailman/listinfo/hardware

[Joerg Reisenweber]

Am Mo  22. Juni 2009 schrieb Werner Almesberger:
> ?lvaro Lopes wrote:
> > * Why was this chosen over PMU 1.8 power rail (IO_1V8) (or the other
> > way around) ?
>
> I could see the following consequences of connecting GPS to IO_1V8:

> - IO_1V8 stays on in suspend while RF_3V (VDDIO, VCCx, VBP) can be
>   cut. Not sure if this is an operating condition the ATR0635 would
>   be happy in.

How are you supposed to switch off GPS then?
/j


_______________________________________________
hardware mailing list
[hidden email]
http://lists.openmoko.org/mailman/listinfo/hardware

[Werner Almesberger]

Joerg Reisenweber wrote:
> How are you supposed to switch off GPS then?

I would expect there to be some way to accomplish this over the
serial interface. But you're right, this would have to be
confirmed.

- Werner


_______________________________________________
hardware mailing list
[hidden email]
http://lists.openmoko.org/mailman/listinfo/hardware

[Rene Harder]

Werner Almesberger wrote:

> ?lvaro Lopes wrote:
>  
>> * Why was this chosen over PMU 1.8 power rail (IO_1V8) (or the other
>> way around) ?
>>    
>
> I could see the following consequences of connecting GPS to IO_1V8:
>
> - this may create an EMI path, so at least a bead should be there.
>   (I.e., instead of R7601)
>
> - IO_1V8 stays on in suspend while RF_3V (VDDIO, VCCx, VBP) can be
>   cut. Not sure if this is an operating condition the ATR0635 would
>   be happy in.
>  

I see a big problem with shutting down the RF_3V, we'll shutdown the
VBAT as well. That means no SRAM and backup RTC. Without backup RTC
we'll increase the start up time.

The good news is we do not need to shutdown RF_3V because if you bring
the GPS chip into standby mode, NSHDN connected to LDO_EN will go down
and will shutdown the internal 1.8V LDO as well.

One of the more important questions is, if you shutdown the chip by
software, does the core shutdown as well or is it still running if
VDD1.8V is active.
I'm not sure but i guess it will only go into sleep mode where it still
needs more power than in shutdown mode.


Rene

_______________________________________________
hardware mailing list
[hidden email]
http://lists.openmoko.org/mailman/listinfo/hardware

[Rask Ingemann Lambertsen-2]

On Mon, Jun 22, 2009 at 03:26:47PM +0200, Joerg Reisenweber wrote:

> Am Mo  22. Juni 2009 schrieb Werner Almesberger:
> > ?lvaro Lopes wrote:
> > > * Why was this chosen over PMU 1.8 power rail (IO_1V8) (or the other
> > > way around) ?
> >
> > I could see the following consequences of connecting GPS to IO_1V8:
>
> > - IO_1V8 stays on in suspend while RF_3V (VDDIO, VCCx, VBP) can be
> >   cut. Not sure if this is an operating condition the ATR0635 would
> >   be happy in.
>
> How are you supposed to switch off GPS then?

   Add a FET or another RT9702APB between IO_1V8 and GPS VDD18, gated by
RF_3V?

--
Rask Ingemann Lambertsen
Danish law requires addresses in e-mail to be logged and stored for a year

_______________________________________________
hardware mailing list
[hidden email]
http://lists.openmoko.org/mailman/listinfo/hardware

[Werner Almesberger]

Rask Ingemann Lambertsen wrote:
>    Add a FET or another RT9702APB between IO_1V8 and GPS VDD18, gated by
> RF_3V?

I still wonder if the GPS subsystem doesn't have a suitable
software-activated power-down mode already, i.e., the "Sleep State"
and the "Backup State". You wake up from them via a hardware signal
(EXTINT or reset, respectively), so it does seem rather low-level.

Adding switches to cut power is rather brute-force and it may have
unintended consequences, such as current creeping in over other
paths, and loss of "expensive" internal state.

- Werner

_______________________________________________
hardware mailing list
[hidden email]
http://lists.openmoko.org/mailman/listinfo/hardware