Preliminary investigation in to the terrain profile using google earth, indicated that a direct link between the Brighton Aid station location and the Scotts Peak Aid station location would be difficult. It was determined that an intermediate node on Scotts peak was needed.
Brighton – based on results outlined in the July 2013 QST article “A Broadband Ham Network Crosses the Finish Line”, by Lynn Jelinski, AG4IU. It was decided that testing would start without the use of amplifiers,then if needed we would add amplification. The basic configuration is as shown (see 1 and 2):
A group of us wanted to do a pre-race mesh test to verify that what we had calculated would work to connect Scotts Peak Aid station with the Brighton Aid station. The members of the group were Brad Rupp (AC7BR), Steve Baxter (K7SRB), Joseph Gray (Ke7MQZ), Ed Sim (N7RTA) and Charles Gray (KE6QZU). We decided to do the testing on Labor Day, as we had that day off work. We had planned to start using just the routers without amplification and then add the amps if needed. We were planning on a mesh node at the Scotts Peak Aid station. This node would have a directional antenna pointing to Scotts Peak where we would have another mesh node that would connect Scotts Aid station to the Brighton Aid station. The Brighton Aid station node would have a 24 dbi parabolic antenna pointing to Scotts Peak. Ed set up the Brighton Aid station node in the parking lot near the location that would be used during the Wasatch 100 Ultra Marathon. This gave us a somewhat clear LOS (Line Of Sight) to Scotts Peak. Here is a picture of that node.
The rest of the group headed to Scotts Peak and Aid station to setup our nodes. Steve set up his node at the Scotts Peak Aid station location. His node used a 4″ 12 dbi cube antenna aimed at Scotts Peak. Brad Joseph and I headed to Scotts Peak to setup the relay mesh node. We found that on the peak it was very windy and looked for the best location to setup the mesh node. We found a location down from the peak where we could see both the Brighton Aid station and the Scotts Peak Aid station. We set up our node using a 4′ 12 dbi cube antenna and an 8 dbi omnidirectional antenna. Here is a picture of the Scott’s Peak node. I could not find a picture from the Labor Day test, but found a picture of the node during the event.
We of course used our HTs to communicate. Boy were we surprised with the results. We discovered that the Brighton node could see our node at Scott’s Peak just as good as the node at the Scotts Peak Aid station. Even with the cube pointed at Scotts Peak and not at Brighton. That meant that we did not need the relay on Scotts Peak. Ed then moved his node to where it would be used during the race. After the move the signal increased and we had an even better link. We decided to set up a TeamTalk server and try it out. I set a Rpi node with TeamTalk on the ground with a 5 dbi antenna. To our surprise, Brighton could also see the Rpi node. The signal was weak and only showed up as an IP address at Brighton, but we could now use the TeamTalk server to communicate over the mesh network. At this point we decided to take down the Scott’s Peak node. With this testing done we decided to go home. We had established a good usable mesh link between the Scott’s Peak Aid station and the Brighton Aid station that had a good LQ and signal strength without any amplifiers.
On Friday evening around 5:00 PM, September 6th we set up for the race. I had created a Rpi mesh-server that provided an Access Point for the Brighton node. This allowed us to connect to the Brighton node via WiFI from our computers. This mesh-server also ran a TeamTalk server and had many of the mesh and TeamTalk files. We also added an IP camera to the Brighton mesh node. We were all set at the Brighton Aid station. The Scotts Peak Aid station took some time to get setup. When they finally did get the mesh node up, they found that the PC they were using did not have a TeamTalk client installed. They connected to the mesh-server and down loaded the TeamTalk client to the PC. They then installed and configured TeamTalk and we were in business. From then on everything worked well…. worked well….. worked well. Yep we did have some problems with the Access Point on the mesh-server. Every now and then the WIFI would stop working and then would start back up at some later time. This was just the AP WIFI. The mesh-server did not die and we did not lose the mesh connection. It was just the PCs using the AP WIFI at Brighton were affected. The node at Scott’s Peak could use the IP camera connected to the Brighton mesh node. TeamTalk worked well also. We did a demo for the Wasatch 100 directors and they were very impressed.
Weeks later we concluded that the four PCs with VOIP and Video running at the same time caused the WIFI adapter on the Rpi over heated and would stop working until it cooled down and would start working again. This happened several times and we spent a lot of time trying to figure what was going on. We checked the mesh-server and the mesh connection and everything looked good.
One of the things that we learned from this experience was that we needed an automatic way to gather the LQ and the signal to noise ratio from the mesh nodes as we had forgotten again to record the LQ and signal to noise ratio during the race.
Here are more of the pictures from the Monday testing.