We hired some help to come in and build us a complete system.  The whole
thing took about a month to put together, scripts and all, and there may be
easier ways to go.

 

The system is composed of a pixie box (computer) hosting dual 24-bit 50kSps
DAQ's which record voltage measurements to a hard drive.  The two voltage
probes from the DAQ's are placed across a 1-ohm precision resistor, which is
in series with the battery and the hardware being tested.

 

Custom Matlab scripts are then used to process the raw data, converting the
voltage samples into units of mA and plotting it. From the plots, we can
select an area to integrate under the curve and find out information such as
average mA's consumed, mAh/day (as if the highlighted section were to run
all day long), microjoules, minimum mA's consumed, maximum mA's, etc.

 

We'd like to open it up to the public (Matlab code at least), but due to the
custom nature of the system it is somewhat difficult to do so.

 

-David

 

 

 

 

  _____  

From: zhiyong yang [mailto:[EMAIL PROTECTED] 
Sent: Friday, December 21, 2007 2:49 AM
To: David Moss
Subject: Re: [Tinyos-help] tinyos 2.x radio

 

Hi: 

  I have a question. How did you measure  the energy  consumed by a node(or
its microcontroller or radio)? What's the tool(apparatus or instrument) you
used?
Thank you very much!
 

2007/12/19, David Moss <[EMAIL PROTECTED]>: 

The answer mostly depends on how noisy your network is.  

 

A CC2420 radio that is always left on (1.x) will consume ~462 mAh/day.  To
put this amount of power in perspective, your node will last about 3.6 days
on a 1700 mAh (cell phone) battery. 

 

Obviously if you have setup a network where all the CC2420 radios transmit
all the time, the radios are never going to get to sleep.  But by allowing
your nodes to duty cycle and quieting down that network, you can get your
power consumption down as low as 3-6 mAh/day pretty easily, somewhere around
1-5 second receive check intervals. At those limits, the microcontroller
becomes the largest consumer of energy on the node and not the radio.  In
one of my deployments, I managed to get the CC2420 radios to ~4 mAh/day -
that's about 425 days of lifetime for a 1700 mAh cell phone battery (not
taking into account internal losses of the battery, etc.). 

 

Take a look at the BMAC and XMAC papers for more information on the
trade-offs of power.

 

http://www.polastre.com/papers/sensys04-bmac.pdf

 

http://www.cs.colorado.edu/department/publications/reports/docs/CU-CS-1008-0
6.pdf

 

The CC2420 low power listening implementation is a combination of the BMAC
and XMAC strategies.

 

-David

 

 

 

 

 

 

 

 

  _____  

From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]
<mailto:[EMAIL PROTECTED]> ] On Behalf Of roberto
pagliari
Sent: Tuesday, December 18, 2007 11:10 AM
To: tinyos-help@millennium.berkeley.edu
Subject: [Tinyos-help] tinyos 2.x radio

 

Hi All,
I have a question concerning the radio stack of cc2420 in tinyos 2.x.
A low power listening is used, which means that the radio is turned off and
on periodically in order to detect a preamble sequence. Hence, the
transmitter broadcasts many times the packet to ensure reception to its
neighbors. The radio stack of tinyos 1.x, at least the version I've been
using, keeps the radio on all the time, and download the message from the
cc2420 as soon as a preamble is detected.

Is it possible to know, in terms of energy consumption and packet losses,
which is the best strategy? 

thanks

Roberto


_______________________________________________
Tinyos-help mailing list
Tinyos-help@millennium.berkeley.edu
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help 

 

_______________________________________________
Tinyos-help mailing list
Tinyos-help@millennium.berkeley.edu
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to