I am just wondering why it takes so long to program a mica2 (or mica2dot) over the air, via a TOSBase node? I am just programming a single mote, so there is not multihop stuff getting in the way. My program image size is 58KB, and it originally took about 25 minutes to program using mica2dots. That's about 300 bits/s, which seems a bit small compared to the 12Kbits/s that they claim the dot's radio can handle. Using a mica2, it was worse. I added some debug statements into the Deluge java tools and noticed that it was dropping a lot of packets, resulting in 5-6 retries for each page (15-19pkts / 48 pkts). I added a 80ms delay between packet transmission and that reduced the number of dropped packets to 1-4 per page, and I can now program it in
7.5 minutes (1030 bits/sec). I can't seem to do any better than that, though it seems that I should be able to, with a 12kbps radio.
I also noticed that in between each transmit cycle, there is a 1-3 second delay, which is accounting for
3.6 minutes of the total programming time. Is there any way to reduce this delay? Any ideas would be great. Deluge is a great idea and I would love to use it more, but 7 minutes is a long time to wait for programming a mote.
Thanks,
Jacob Sorber
Computer Science Dept,
University of Massachusetts, Amherst
_______________________________________________
Tinyos-help mailing list
[email protected]
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
