Hello,

Let me start out by saying that I am very new to the Broadband Hamnet solution and mesh networking in general.  My experience to date consists of converting 4 Linksys routers (I had to bring one back from a brick) and running them throughout the house.  I have been able to successfully connect an internet gateway and have setup a v5 router (running stock Linksys firmware) as a wireless access point to one of the nodes.  I plan on experimenting with IRC chat, integrating with the home LAN, adding a file server and other solutions that others have previously setup. Unfortunately there are no other active nodes in Minneapolis so experimenting with range and antennas is on the back burner until I get more familiar with the firmware aspects of the technology.

That said, in the summer I volunteer for a Search and Rescue organization in North Idaho that provides communication support for local multisport events (triathalon, marathon, mountain bike races).  The organization has only used VHF radios and these can be unreliable (especially the 5W HT's) in the remote areas of the events.  My thought is that a mesh hinternet would be just the thing to improve voice communication and allow for adding video communication.  It's a pretty ambitious idea.  

What I'm hoping for from the group is an assessment of and help with building a mesh plan.  In the attached Google Earth .kmz file I have mapped out the locations that I want to connect.  In the file the push-pins labeled AS #x are aid stations on the course.  The main location is the Command Post (CP) located at the finish line by a local resort.  The goal is to tie these 11 locations together.  

To facilitate the linking I have picked 5 additional locations to put nodes.  Three nodes are on high points that surround the event course.  Two nodes are on high points that look down on the CP and which are visible to the 3 perimeter nodes. I have plotted paths and checked if they have line of sight and all look pretty good (a couple are iffy and are labeled with ??).

Since I have no field experience I'd like feedback if people think this is even feasible.  Some things I am worried about

• Is 16 nodes too many?
• The perimeter nodes need to hit other nodes in a field of vision that is 60, 90 and, in one case, nearly 180 degrees - is this even feasible?  
• How do you keep from having a choke point? I'm worried about the CP node.  I figured if it was visible from only one perimeter node then that path has to carry all inbound traffic from all the course nodes.  For this reason I thought adding a couple of nodes with view to the CP and the other perimeter nodes would give alternate routes and help relieve congestion.  Is this a good idea or does it cause more problems that it would solve?

I will not be able to experiment on-site until next June.  The first event is the end of August so there is some time to try out antennas and locations.  The winter will be spent acquiring more nodes and assembling weatherproof, battery powered, portable mesh node/WAPs so there is plenty of work to do.

Thanks for any and all replies.  

Bill

AC0LV

Try Radio Mobile to do your link and coverage assessment or to plan your deployment.

Radio Mobile: http://www.cplus.org/rmw/english1.html

You will have to either pay the fee to use 2412 MHz or be satisfied with 2305 MHz substitution, but it gives credible results and truly excellent depictions of terrain and Fresnel Zone profiles.

I use 4.74 uVolts for receiver sensitivity which corresponds to the threshold of most consumer WIFI APs running at about 1 Mbps. You should get a link margin at the end of an assessment report, and you want that link margin to be 20 or 30 dB for good broadband performance.

73;

Bob KV4PC

Bill:

Im a big fan of Radio Mobile. It is awesome that such a good tool is essentially free to use.

Years ago I used OpenWRT/OLSR radios as an emulator/surrogate for the Harris HNR radio in a series of field demonstrations for a Defense contract, and I had a proprietary tool that used similar estimating algorithms (Longley-Rice at base, Fresnel Zone model, Foliage model, Fading model). It gave extremely good estimates of link behavior for "clear sunny day" conditions. We saw link margins that were probably within 3 or 4 dB of predicted in nearly every case. That is close enough to know when a link is going to work and when it is not.

The tool I had did only two more things than Radio Mobile - it had a precipitation model which was nice. But the biggie was that is it would generate kml files for coverage viewsheds that geo-registered semi-transparent JPEG images. You could then load the kml file into Google Earth and see the coverage viewsheds overlaid on the geometry, and look at them from "bird's eye view" on top of exaggerated 3D geography. Useful to really understand your coverage area.

The view sheds could be generated in different levels of Longley-Rice "reliability" or with precip so you could get a stack of coverages for different conditions. Or lay out repeater linkages like for trunk radio. Very valuable for Meshes.

73;

Bob KV4PC