While planning for the AREDN Mesh Network to support the Baker to Vegas Relay Race, I installed iPerf on some nodes and took them out to the desert for a test. (If you are not familiar with the AREDN Mesh Network for the Baker to Vegas Relay Race, here's a link: https://www.aredn.org/content/baker-vegas-2018-broadband-mesh-support)
First, I set up a test AREDN mesh network at home:
- I installed iPerf on a Rocket M2 Omni, a Nanostation M2 and a Nanostation Loco M2. I also borrowed a couple of NanoBeam M5 from Don Hill, KE6BXT, but since they weren’t mine, I didn’t install iPerf on them. I put the NSM2 and a NBM5 (without the dish, just the transceiver) in the living room, the Rocket M2 with an omni in the dining room, and the NSLM2 and the other NBM5 (also without the dish) in the kitchen. I dtdlink’d the two nodes in the living room and dtdlink’d the two nodes in the kitchen. Later, I could put the 2 GHz nodes on different channels so that iPerf would route the packets across the 5 GHz link
- I ran multiple signal to noise tests on 2GHz channels -2, -1, +1 and +6 at 5, 10 and 20 MHz bandwidths. The test results in my home “lab” were what I had expected: The signal to noise ratios were best on channel -2 and -1. The signal to noise ratio on channels +1, 6 were higher; on each channel, the signal to noise ratio was best at 5 MHz bandwidth, and worse at 10 and 20 MHz (but the difference between 10 and 20 MHz was barely distinguishable). I didn’t test the signal to noise on the 5 GHz nodes.
- The iPerf throughput on each channel was highest when the bandwidth was set to 20 MHz and lowest when the bandwidth was 5 MHz.
- The best overall iPerf speed was on Channel -1 when the bandwidth is 20 MHz (Ch -2 at 20 MHz is not usable because it falls outside the ham band).
- I did not have time to test all of combinations of channels and settings for the 5 GHz nodes; I only had time to make sure that I could dtdlink them to 2 GHz nodes that had iPerf and force the packets across the 5 GHz iink.
- The test results in my home “lab” were exactly what I had expected.
(continued)
I headed to the desert on Monday, October 23. Because the trip was on short notice I went by myself, but John Emig, W6CYS agreed to keep track of me and if I didn’t return on schedule, to search for me. My first stop was at Stage 2, where I set up a mast with a Nanostation M2 and a Nanobeam M5.
Then I proceeded to Ibex Pass, where I set up a mast, a Bullet M2 with an omni antenna and Nanobeam M5. The distance between the sites is 11.1 miles.
I set up the site at Stage 2 in less than an hour, but I took over two hours to get the Ibex site running because the site is about 200 yards from the parking area but more importantly, one of the nodes would not boot up; the status lights flickered and I did not see a steady red link light. The problem appeared to be due to the voltage drop between the battery and the node. I made a trip back to my Jeep where the node started OK using the Jeep’s battery, so I wired in a voltage booster to increase the voltage to 20 volts and returned to the site. By the time I got the network up, it was late in the afternoon so I only had a limited amount of time to test the performance of the link between Ibex Pass and Stage 2.
(continued)
Test #1. 2 GHz nodes on Channel -2 at 10 MHz bandwidth (2 GHz Bullet with 13db omni antenna to 2 GHz Nanostation) The signal to noise ratio was -75 db and the first four runs of the iPerf test showed an increase from 1.44 Mbps to 5.62 Mbps. I ran it two more times, and showed results of 2.08 and 4.55 Mbps. (note the time on my iPhone screen shot and my remaining iPhone battery life).
(continued)
(continued)
(continued)
Test #4. 2GHz on Channel -1 at 10 MHz bandwidth. The signal to noise ration on Channel -1 at 10 MHz bandwidth was roughly the same as at 20 MHz bandwidth, but the iPerf was much better. There were two test results which made no sense (see screen capture). At this time, the sun was about to set, and my iPhone battery was nearly exhausted, so I was unable to perform more tests.
Conclusion. I consider the test of connectivity and throughput between Stage 2 and Ibex Pass to be a success and the AREDN mesh network should be able to support the B2V race requirements. More testing is indicated to determine the optimum channel and bandwidth to use.