D-STAR, Wifi, MESH networks

Griff Hamlin <griff@...>

Hi Fellow CMRA hams:

Michael Hauan AC0G recently wrote:
..So, we can combine D-STAR, MESH, and other networks in whatever fashion
makes sense and delivers the best service in emergencies.
... i'm picturing an in-town MESH connecting the D-STAR repeater with Joint
Communications, the CMRA club station, and one or more of the Columbia
hospitals. If we can get longer hauls to work, or establish a suitable
chain of MESH nodes, we can shoot for links to Hallsville and further...

To setup such a wireless network as Michael describes, we'll probably
need three different types of network "nodes" (digital repeaters):

1. Short range node:
Linksys WiFi HSMM-MESH node. Data rate: 1.5 to 11 megabits/second.
Michael has successfully installed HSMM-MESH firmware on his Linksys
WRT54G WiFI units. Using the built-in omni-directional antennas (about 5db
gain) I'd guess the outdoor range will be about 1 or 2 blocks.
Using a high-gain (24db) directional antenna mounted a on
a rooftop, I'd guess the range will be about 1 mile, depending upon
terrain, antenna height, etc. If range proves to be a problem, I have
three 1-watt bi-directional 2.4ghz amplifiers that could could be used
for testing to boost range.

Due to large coax cable loss at 2.4ghz, WiFi units should be mounted
outside, next to the antenna, unless a bi-directional amp is used.
RG-8 coax loses 12db/100'. Even LMR-400 coax loses 6db/100'. I have used
PVC weatherproof enclosures ($25 from Home Depot) in the past to mount
Linksys and other WiFi units outdoors on the antenna mast.

A mesh network of such nodes with omni-directional antennas could be quickly
constructed by simply placing a node every block or so. This could
cover a neighbourhood sized area. All nodes operate on the same RF channel.
The nodes will locate each other automatically, and data traffic will be
automatically routed among the nodes. Each node momentarily stores each
data packet and then re-transmits it (like a VHF packet radio network ).
This causes the data-rate to decrease in proportion to the number of
wireless mesh nodes the data passes through. But there should still be
plenty of data-rate. A node on the edge of this mesh could connect to a
near-by medium-range node (see below).

2. Medium range node:
High power (800mw) WiFi, such as Ubiquiti NanoStation2 or Bullet-2HP,
or EnGenius-2610, plus high-gain (24db) antennas mounted
as high as possible.
Data rate: 3 megabits/sec.
Maximum range is mostly determined by antenna heights, which must be high
enough to maintain line-of-sight over distances of several
miles. The "Fresnel zone" must maintain line-of-sight, which usually means
the antennas must be 10' to 20' above all obstructions in the path. My
range experiments over the years with various antennas, heights and
transmitter powers are:
1. Between 9db omni-directional antenna 50' above ground, and a 24db antenna
20' up, 100mw transmitter: 1.5mile range.
2. Between a 9db omni antenna 50' up, and a 5db omni-antenna just above
rooftop, 100mw transmitter: 0.3 mile range.
3. Between two 24db directional antennas (8 degree beamwidth), one 50'
up and the other 20' up, 400mw transmitter: 2.2 miles.
4. Between 24db antenna 85' up with 600mw transmitter and an unknown
antenna about 100' up, 5.1 miles.
Others have reported hill-top to hill-top distances of 30 miles and more.

3. Long range node: D-STAR.
2m or 70cm D-STAR digital voice + 1.2 kilobits/sec data.
1.2ghz D-STAR (expensive) data rate is 128 kilobits/sec.
I guess a 2m or 70cm D-STAR repeater range might be around 15 to 35 miles,
depending upon antenna heights, etc.

Using a network of short and medium range nodes, one can attach a laptop
computer to any node via ethernet cable, and use your favorite
voice-over-IP (VOIP) software with your computer's sound card to talk using
digital voice. Two laptops on a WiFi network can talk to
each other using VOIP, with no external internet connection required.
At WiFi data rates, digital video is possible, as well as file-transfer,
email, web-browsing, etc., as long as at least one node in the wireless
network is connected to the Internet.
I often use the open-source EKIGA VOIP software which is peer-to-peer
(no central server computer required). It can do digital voice and
video and runs on all operating systems. It can also talk to
Microsoft's NetMeeting software.


The Ubiquiti outdoor WiFi units, and maybe others, can operate at 2397mhz
with signal bandwidth reduced from the standard 20mhz to 5mhz. 2397mhz is
in the 13cm ham band but NOT in the 802.11 US unlicensed (part 15) band. This
should reduce interference and quadruple the power spectral density, so the
range should be longer, at the expense of data-rate. I need someone to help
me test range under these conditions.

High-power bi-directional 2.4ghz amplifiers that incorporate a transmitter
power amp and receiver pre-amp are available to licensed hams, but
they're fairly expensive for an individual ($800 for 10 watt amp,
$2500 for 25 watt amp).

I have successfully created WiFi "digital repeaters" by connecting
two WiFi units together with a short run of ethernet cable. Each WiFi unit
has it's own antenna and uses a different RF channel. The SMESH network
(see http://www.smesh.org) uses a multi-radio repeater like this at each
node. For creating a long point-to-point link, a repeater with 2 high-gain
directional antennas pointing in opposite directions can be setup near
the center of the long point-to-point link. For point-to-multipoint,
the way hams typically use a VHF/UHF repeater, one antenna is
omni-directional. I've been able to connect together WiFi units from different
manufacturers, DLink, Ubiquity and EnGenius to make repeaters. Using multiple
radios at each "repeater" node theoretically allows one to maintain nearly the full

I need someone to help me test the range of such WiFi repeaters.
I've only done "proof-of-concept" by setting up such a repeater in my
yard. We may find the range is reduced because one transmitter
desensitizes the other receiver, even though the repeater doesn't
necessarily need to transmit and receive simultaneously.
Ideally we'd like to put the "repeater" on a tower or tall building.

Such a multi-radio WiFi repeater is inexpensive. It can be built for
$130, not counting the antenna mast, as follows:
Two NanoStation Loco (100mw) Wifi units @$45: $90
Two hi-gain yagi antennass @ $15: $30
Misc cat-5 network cable & connectors: $10
A high power (800mw) version can be built for $90 more.
I haven't yet received the 17 element yagi antenna I ordered for $15.
I can't believe the gain will be 25dbi as advertised, but I'll find out
soon. Normally, a 24db parabolic reflector antenna costs $40 to $60.


As Michael pointed out, we need to be able to make D-STAR and WiFi nodes
talk to each other. We could connect a computer to both a D-STAR radio and
a WiFi radio. It should be possible to obtain the 1.2 kb/sec
D-STAR data from D-STAR computer software, and then re-send it to the WiFi
Repeating D-STAR digital voice is harder because it is encoded using
proprietary AMBE vocoder. This means we won't be able to decode the D-STAR
digital voice signal back into ordinary digital voice data without a
decoder such as dv-dongle (www.dvdongle.com) that has paid the AMBE
license fee. To get around this we could use one of two approaches:
1. Let the D-STAR radio decode the digital voice signal and then
re-digitize the voice audio signal using the computer's sound card. The
computer could then re-send the unencoded voice data over a WiFi network
or over the internet.
2. Don't decode the D-STAR digital voice until it reaches it's final
destination. It should be possible to obtain the encoded D-STAR voice
signal (3.6kbits/second) from D-STAR software, and then encapsulate it
in IP packets and re-send it using VOIP software over a WiFI network
or the internet. At the final destination, a dv-dongle or D-STAR
radio or other AMBE decoder would be needed to decode the D-STAR
digital audio.

Griff W5VWP