Dear Philip and Dear David,
Thank you for your inputs.
I have been testing the power consumption of CTP-LPL
by increasing the beaconing period of CTP. Following
were my results.
Current
^
| <-T1 -><T2>
| |-----| | | |-------|C1
| | | | | | |
| ----| |---|---|...N spikes ......|
|-----C2
--------------------------------------------> time
T1 = time period when current consumption is high due
to sending a packet
N = number of Spikes due to periodic waking up of
radio as directed by Local sleep period
T2 = time period between two spikes
Tp = time Period of spikes
C1 = Current when mote is in wake up mode
C2 = Current when mote is sleeping
I choose 500ms sleep period as it will reduce the
worst case wakeup time during sending a packet.
RxSleep = Localsleep = 500ms
Periodic packet sending period = 10 sec
CTP min beaconing period = 10 sec
Power consumption:
100ms <= T1<= 900ms
T2 ~= 490ms
N ~= 19
Tp ~= 10ms
C1 ~=29mA
C2 ~= 300uA
So average power consumtion ( worst case) ~=
((T1 + N*Tp)*C1) + (N*T2*C2)
----------------------------
T1 + N*(T2+Tp)
~= 3.6mA SO BIGGGGGG :-(
So it seems the radio duty cycling is working fine.
But the average current is very high.
Following are some factors making the average high:
1. The big duration T1 when mote consumes lot of
current is actually a deciding factor for the average
current consumption.
Is there a way to reduce this time period??
By decreasing the sleep period further can be a
approach, but it also have tradeoff.
What can be a good solution for this??
2. Sleep current is still very high. I tested with
TinyOS1.14 and I was able to get 50uA sleep
current.But using TinyOS2.0.2 I tested with Null
application and it came out to be 260uA.
What can be the reason behind this big current
consumption here??
Regards
Ravi
--- Philip Levis <[EMAIL PROTECTED]> wrote:
> On Aug 6, 2007, at 9:48 PM, Philip Levis wrote:
>
> >
> > defines the minimum beacon interval of the routing
> engine to be 1s.
> > If you increase it to, say, 30, then nodes will
> send one beacon
> > every 30s. They can send beacons more often than
> that in order to
> > repair from topology failures such as routing
> loops.
>
> Whoops, I was a bit imprecise. I mean "nodes wait at
> least 30s
> between sending beacons." If the network is stable,
> nodes send
> beacons every 1024s (the third parameter).
>
> Phil
>
____________________________________________________________________________________
Fussy? Opinionated? Impossible to please? Perfect. Join Yahoo!'s user panel
and lay it on us. http://surveylink.yahoo.com/gmrs/yahoo_panel_invite.asp?a=7
_______________________________________________
Tinyos-help mailing list
[email protected]
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
