Dear all,

I have been working on a time sync protocol between the pc and the
motes. I think I have solved most of the problems on the java side and I
would like to improve things on the motes.

My test for the time-sync consists of running the sync protocol every
i.e. 60 seconds and sending messages to 2 motes every 2 seconds which
say: "at time X, change the DAC". I use a digital oscilloscope to
monitor the outputs and see how good the response is (difference between
both responses). As for now, I have noticed that for each resync the
delay in the changes of the DAC outputs remains constant, but there are
great differences between two resyncs: one might give a difference of
1ms and the next a difference of 20ms. I am trying to reduce this
variability as much as possible

As Janos suggested, I have looked for a way to do the timestamping with
the SFD interrupt. Looking through CC2420RadioM, I have noticed 
async event result_t SFD.captured(uint16_t time).

Looks like there is a timestamping going on already, and I have read
somewhere that it is recorded on rxbufptr->time, so I've been trying to
extract it on message reception (In ReceiveMsg.receive) like this:

event TOS_MsgPtr ReceiveMsg.receive( TOS_MsgPtr msg ) {
    struct TOS_Msg* r_time;
    r_time = (struct TOS_Msg *)msg->data;
    t2 = (r_time->time)/0.032;
}


But it doesn't seem to be working. I get three or four fixed values every time 
a synchronisation is run (0 and 10013) come to mind...

I have tried vu's timestamping as well but on a call to get timestamp t2 is 
always 
0. Does anyone know a simple way to get a timestamping with the SFD interrupt 
handler?
Or a suggestion as to to why a resynchronisation has such a big variability 
while the delays
are so constant (which would suggest that the problem is not on the timers, but 
on the send/receive 
interfaces) and how to improve it?

Any comments are appreciated.

Thank you,

- Jose.


_______________________________________________
Tinyos-help mailing list
Tinyos-help@Millennium.Berkeley.EDU
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to