[Tinyos-help] Again!!!! Low-power CC2420 multihop in TinyOS2.x question????

[Philip Levis]

On Aug 5, 2007, at 7:13 PM, Daniel Widyanto wrote:

>
> Dear Ravi,
>
> I've also done some test for CTP currrent consumption using micaz.
> I have the same result as you are. It took around 30 min for a node
> to join a network, and the minimum current consumption is around
> 5-6mA (not shown in my graph).
>
> My test result (along with the firmware that I used) is in :
> http://gamedevel.wordpress.com/2007/08/06/micaz-platform-current- 
> consumption-with-tinyos-2x-multiple-hop-radio-and-power-management/
>
> A bit misuse of my personal blog. But I want to know whether others
> has the same experience with T2.0.1
>
> I think the 5-6mA represents IDLE sleeping mode, which is caused
> by the Timer1. The Timer1 is used to cycle the radio. Thus, according
> to getPowerState() in McuSleepC.nc in tos/chips/atm128, if Timer1
> overflow interrupt is activated the sleep mode should be IDLE.
>
> So, if we move the power cycle from Timer1 to Timer0 ASYNC,
> I think we can use POWER_SAVE idle mode (around 8uA according
> to the datasheet).
>
> Am I correct ?

I'll be 100% honest here; the group of us working on CTP have not yet  
starting working on its interactions with LPL in order to try to get  
it to be ultra low-power. Rather, we've been focusing on issues such  
as topology formation and link estimation. Two weeks ago we had a big  
breakthrough on this front, which means I think we can finally put  
all those issues to rest and start working on LPL. We wanted to  
settle on what we think the best packet formats and semantics are  
before optimizing its scheduling. And who knows -- low power  
operation might call for additional control bits.

Phil