[Aiko] Pebble thoughts

[Paul Szymkowiak]

Hi folks,

Looks like I'll miss CCHS tonight :(

I thought I'd share a few thoughts from my experience building and using the Pebble:

I assembled my prototype Pebble board on Saturday with hardware guidance from Luke and software assistance from Andy (many thanks).

Assembly

All up it took me ~6 hours to assemble, and about another hour for the software config (Loading code to the Arduino, Xbee config, etc). Mostly the process was uneventful, if you discount the fiddly work involved with the prototype board and its problems. The work specific to the prototype board accounted for about an hour or so of my effort (this shouldn't be relevant to LCA). There were a few assembly tasks that could have gone wrong or been problematic if Luke hadn't been "on tap" to consult with: some of those might be useful to document or have pre-built before hand to help reduce build time/ ensure success at LCA.

Temperature Sensor anomalous readings

The way I have the temperature sensor mounted currently, its always reading a few degrees hotter than actual.

I assume that's due to heat coming off the LCD or other Pebble components.

I'm thinking about remounting the temp sensor (and possibly the LDR) significantly away from the board - maybe a 3-pin header on the board and a 3-wire cable protected/ stiffened using heat shrink.

I also wondered whether in the future the Aiko Sensor code might incorporate sensor offsets/ calibration offsets as part of local node configuration info: that might work for different LDR stock, but I don't think it helps so much with the Temp Sensor issue.

Power consumption

I tried running the Pebble on a custom Li-battery pack I have (Li-ICR18650 rated @ 3.7v, 2200mAh).

It ran about 5.5 hours before the battery gave out. I was a little surprised it only ran for that long: it made me wonder about alternative pebble/ node designs that consume less power: e.g. smaller, two-line LCD; LCD with no powered backlight or a backlight that can be turned on only as needed.

My 5 cents ....

Paul

[Sam Sabey]

Hey Paul,

That's amazing you went for so long on that teeny battery pack!  My measurements have 0.4amp at 5v going through the WRT, which would mean your gear is probably running a little less than this.  

The majority of the electronz seem to me to get consumed in the CPU of the WRT - making it get quite warm.  I've  been stacking them one on top of another, and I've come to the conclusion they need moar ventilation.  I noticed Lukes the other day and the placement of his temperature sensor was in the prime location for the "chimney" and his temp was 8-9 degrees warmer than it probably should be.

Have you tried standing the WRT up, or even mounting it on a wall?  This may fix your heat measurement problems.

Sam.

--

You received this message because you are subscribed to the Google Groups "Aiko distributed platform" group.
To post to this group, send email to aiko-p...@googlegroups.com.
To unsubscribe from this group, send email to aiko-platfor...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/aiko-platform?hl=en.

[Paul Szymkowiak]

Hi Sam,

>That's amazing you went for so long on that teeny battery pack!

Well, it was the Aiko Pebble (not the Router - so no Router CPU, etc. :)

... and I don't have the relay's mounted yet

... which is why I was a little surprised it was so hungry ;)

guess it will get moar hungryz as it starts doing more (relay's, etc)

PS - I am keen to try the Router under battery too, so I'll probably follow your steps there.

Thanks & Best Regards,


Paul

2009/12/2 Sam Sabey <sam....@esskware.com.au>

[Sam Sabey]

icicic...

My energy meterz are pulling about 1.5amp hour per day from a 6v nominal metal hydride pack at the moment.  These are running through the Deuemelinove, complete with it's hungry FTDI chip etc.  My proz use a lot less, however Zigbee does pull a fair bit out of them.  I can run a 1mw zigbee with a pro with temperature and light on  2.4v pack 5amps capacity for around 2-3 days.

With deep sleep, this should get better.  I'm also planning on getting some cheapy ebay solar panelz (DYI) and installing them on my rooftop to start collecting real data from my solar panel system that will be installed on Friday.  :D

[Luke Weston]

Well, 5.5 hours from 2200 mAh is 400 mA, which doesn't sound too bad.

I wonder... if you were feeding 3.7V into the Arduino's voltage
regulator, what kind of output voltage are you getting, and since that
output voltage is less than 5 V, does that mean that the electronics
consumes more current than it does with a 5 V supply?

The LED backlight on the LCD is probably consuming 100 mA itself.

The whole thing isn't designed, overall, for extra-low power
consumption - that's a design requirement that your hardware has
really got to be specifically designed from the ground up to support.

[Sam Sabey]

Interestingly I;'ve observed that when supplying a duemelinionove (or however it's spelt...) with 5V, the regulator drops this down a working voltage of 3.3v.  I noticed when the batteries go flat, the battery voltage goes high before fail - I suspect this is related to the default AREF dropping from 5 to 3.3 volts.

Sam.

Well, 5.5 hours from 2200 mAh is 400 mA, which doesn't sound too bad.

I wonder... if you were feeding 3.7V into the Arduino's voltage
regulator, what kind of output voltage are you getting, and since that
output voltage is less than 5 V, does that mean that the electronics
consumes more current than it does with a 5 V supply?

The LED backlight on the LCD is probably consuming 100 mA itself.

The whole thing isn't designed, overall, for extra-low power
consumption - that's a design requirement that your hardware has
really got to be specifically designed from the ground up to support.

--

You received this message because you are subscribed to the Google Groups "Aiko distributed platform" group.
To post to this group, send email to aiko-p...@googlegroups.com.
To unsubscribe from this group, send email to aiko-platfor...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/aiko-platform?hl=en.

[Jonathan Oxer]

On Wed, 2009-12-02 at 13:13 +1100, Sam Sabey wrote:
> Interestingly I;'ve observed that when supplying a duemelinionove (or
> however it's spelt...) with 5V, the regulator drops this down a
> working voltage of 3.3v.

Hmmm, I suspect you're pushing the limits of the CPU there. The 7805
voltage reg on a Duemilanove is not a "low dropout" part, so it always
loses around 1.6V from the input voltage. If you're feeding 5V into the
barrel connector (as opposed to the USB connector) of a Duemilanove then
you're running the reg at below its rated minimum input of 6.6V. Luckily
the ATMega328P can handle running down around 3.3V (can't remember the
exact spec) so it's surviving on a starvation ration of power.

Also, remember that the 3.3V rail in the Duemilanove doesn't come from
the primary regulator but comes from the internal regulator of the FTDI
chip used for the USB interface. Are you talking about the VCC rail
(which should be 5V if there is sufficient input voltage) or the 3.3V
rail?

Cheers :-)

Jonathan Oxer
--
Practical Arduino: www.practicalarduino.com
Geek My Ride: www.geekmyride.org
SuperHouse Automation: www.superhouse.tv

[Sam Sabey]

Hey John,

This is more or less what Andy told me when we discovered this "feature".  Since then I've only been running pros on a 5v regulated supply.
Now I suppose what's interesting is I wonder if running a nominal 6v battery pack (5 metal hydrides) would drop the voltage to 3.3?  I should test this out.

In other news on AREF mods, I've been experimenting with EXTERNAL aref, bringing this down to 0.7v, it seems any lower than this I find the digital signal  clips at 1023.  Is there some lower limit for the EXTERNAL aref?

Sam.

-----Original Message-----
From: Jonathan Oxer <j...@oxer.com.au>
Reply-to: aiko-p...@googlegroups.com
To: aiko-p...@googlegroups.com

[Luke Weston]

Hmmm, I suspect you're pushing the limits of the CPU there. The 7805
voltage reg on a Duemilanove is not a "low dropout" part, so it always
loses around 1.6V from the input voltage. If you're feeding 5V into the
barrel connector (as opposed to the USB connector) of a Duemilanove then
you're running the reg at below its rated minimum input of 6.6V. Luckily
the ATMega328P can handle running down around 3.3V (can't remember the
exact spec) so it's surviving on a starvation ration of power.

It's not just a 7805, it is a low-dropout regulator.

But certainly it should be at least a little bit above 5V on the unregulated input.
Specifically, it's - let me check the schematic - an MC33269-5.0. It's 1.0 V dropout for that device - so 6V is a safe minimum for reliable operation.

[Luke Weston]

 Is there some lower limit for the EXTERNAL aref?

The ATmega238/168 series datasheet (the epic 500 page datasheet) gives it as 1.0V (page 323).

[Sam Sabey]

That may 'expliain' the peculiar behavior observed when pushing the
lower limits of arduino...

Interestingly, the honeywell hall effect sensors have a much wider
range of response compared with the Jaycar inductive test clamps.
Perhaps if we run the hall effect output through a voltage divider,
set aref as INTERNAL we could get enough resoulution with no amplifier?

...or use a greater bit ADC?

...from @samotage on the go.

On 02/12/2009, at 5:56 PM, Luke Weston <reindeer...@gmail.com>
wrote:

>
> Is there some lower limit for the EXTERNAL aref?
>
> The ATmega238/168 series datasheet (the epic 500 page datasheet)
> gives it as 1.0V (page 323).
>