Marco Langerwisch wrote: > Hi Brano, > > theres still something I don't understand: > > > it has access to the SFD_tx timestamp; instead of transmitting T1^ and > > T_SFD_tx separately, it transmits their offset T_offset=T1^-T_SFD_tx > > (saving 4 bytes). > As I can see, T-SFD_tx is in LocalTime and happens after T1^, hence > T_offset would be negative, but is unsigned. it works correctly with unsigned ints (when you add it to T1 at the receiver)
> You might wonder why I come up with these questions. I am porting FTSP > to the cc1000 Stack. The motes synchronize to a certain global time, but > I don't know yet if they are synchronizing correctly... for cc1000, you may want to look at http://tinyos.cvs.sourceforge.net/tinyos/tinyos-1.x/contrib/vu/tos/platform/mica2/ ClockTimeStamping* and SysTimeStamping* implement what you need. it's a little hard to understand, as any performance-tweaked code written by miklos, but it basically timstamps 6 different bytes of the transmitted message and takes minimum of the first three, minimum of the last three and reports back the average of the two minima. this helped to minimize both interrupt related (always positive) as well as chip (gaussian) timestamping errors... brano _______________________________________________ Tinyos-help mailing list Tinyos-help@millennium.berkeley.edu https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help