On Feb 3, 2007, at 10:01 PM, David Moss wrote:
Interesting results. Hardware wise, there are a few possibilities
that I can think of:
* Antenna type - the telosb inverted-f antenna is a very good
antenna. The whip antenna attached to the mica's are much worse.
The impedance matching on older mica2's is very poor, which will
significantly decrease the range.
* Interference - there's plenty of interference in the 900 MHz
band, especially with FSK on those CC1000's. The 2.4 GHz CC2420
uses DSSS, which is less likely to be affected by interference.
This is not quite true; DSSS is more resistant to noise (that is,
AWGN), as its encoding can recover individual bit errors. However,
it's important to note that the 2.4 GHz band is shared by 802.11 and
Bluetooth, both of which interfere significantly. The real problem
with 802.11b is that because its packets are so much faster (and its
transceivers don't care about 15.4 much for clear channel assessment)
is that an 802.15.4 node can detect a clear channel, start
transmitting, have the middle of its packet overwhelmed by an 802.11
packet, and then complete transmission to what seems like a clear
channel.
The 900 MHz cell phone bands surround the 900 MHz band, and spurs
coming from the cell phone bands have been known to affect the link
performance. Also, any other 900 MHz device will negatively affect
performance. The solution is to add SAW filters to the antenna
network, but that's pretty much a no go with off-the-shelf hardware.
If you're consistently finding that the 2.4 GHz spectrum performs
much better in your environment, then go for it.
On one hand, 2.4 GHz has a lot of users; on the other, they tend to
be digital packet networks, rather than analog sources (one major
exception being microwave ovens). Therefore the noise/interference is
a lot easier to deal with than in the 900MHz band. I also think that
one of the big advantages of the 802.15.4 over prior data-link
technologies is simply that it's a packet-based radio: software
doesn't cause the radio to misalign or drop bytes, greatly reducing
the number of packet errors.
Phil
_______________________________________________
Tinyos-help mailing list
[email protected]
https://mail.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help
