One backhaul rb433rosv5.0 running on 5600Mhz centre frequency.
Test is done running in 5, 10 and 20Mhz channel bandwidth. nv2 protocol enabled
Directional mesh antenna 22dBi.
Call this unit router/antenna “A”
Other router on same mast, omni antenna 12dBi sits just above this directional. 30cm spacing?
Router is rb433rosv.5.0running
This router is used to run the scan to see what possible cross channel interference might be seen.
Call this unit router/antenna “B”
Scanning is done in “B” each time:
Scan in 20Mhz chnnl width, “A” running 20Mhz channel width, “B” ´sees´ “A” at 5600Mhz/-44dBm + 5620/-86dBm + 5580/-85dBm
Scan in 10Mhz chnnl width, “A” running 20Mhz channel width, “B” cannot ´see´"A"
Scan in 10Mhz chnnl width, “A” running 10Mhz channel width, “B” ´sees´"A" at 5600Mhz/-41dBm + 5640/-82dBm + 5560/-87dBm
Scan in 20 or 10Mhz chnnl width, “A” running 5Mhz channel width, “B” cannot ´see´"B"
Scan in 5Mhz chnnl width, “A” running 5Mhz channel width, “B” sees “A” at 5560/-85dBm + 5580/-59dBm 5600Mhz/-36dBm + 5620/-58dBm + 5640/-80dBm
One remark. This double, or quadruple listing in the scan is only seen when all data rates on “A” are allowed. (Default setting)
But the moment I set all data rates to one rate only the double or quadruple listings are gone…
Issue 1.
How is it possible that when the transmitting mask is narrowed, signal is spread at 20Mhz and 40Mhz distant up and down from the centre frequency? This would mean that choosing 10 or 5Ghz channel bandwidth to create more available channels or lesser interference to adjacent radios in fact we create a wider spread interference, special for units that are on the same tower?
Issue 2.
Although the link that is established by “A” to another distant router/radio “C” can achieve up to 48Mbps data rates (in 5Mhz), 36Mbps data rates (in 10Mhz) and 24Mbps (in 20Mhz), and the CCQ’s are each time in the near 100% regions, than when running a tcp bandwith test from “C” to another router “X” behind “A” the bandwidth in 20Mhz is some 18Mb (data rate 24Mbps), in 10Mhz some 11Mb (´real´ data rate is 18Mbps) while in 5Ghz (real data rate is 12Mbps) I only get 2,5Mbps over the link! I would have expected something between 6 and 8Mbps…)
Side remark on this little test I made:
a. I could note that nv2 in all bandwidth settings increased the throughput of the link slightly up to 5Mb in 20Mhz compared to normal 802.11 mode.
b. I also noted that indeed as some already stated, 802.11 has benefit by at least 3 data rate settings. (Highest reliable, next step lower for backup and lowest for basic rate coverage) and if rates are set to ´all´ (default) then link is not as stable as when set to limit maximum. In ´default´ the overall maximum throughput is a bit lower although some spikes were higher. CCQ was much lower because link was less stable, this resulted in a more unstable data stream.
In case of using nv2 best results were achieved when I set one data and similar basic rate only! This had to be tested by trying the highest data rate the link sustain and than fix it to that one. For peace of mind (and if you don’t need the capacity) I would suggest to go one step lower.
Test was done in the middle of the night, not to harm my clients…