[MVMA] The North and the South

[Maurice Riggins]

Part 1 - The North
 
After recovering the DARA omni from a failed upgrade, I noted it had eight consistent neighbors (sporadically more)--including practically all the high profile omnis. What they had in common was DARA saw them all with much lower LQ than their LQ of DARA. Most notable was the nearby W8GUC which received DARA at 100% while DARA's LQ of it was typically in the fifties. AE8I-1 was in a similar situation.
 
Only two of the eight neighbors linked well enough to transfer data, including with the primary OLSR second hop neighbor, WA8APB-Services. Those two were W8GUC and AE8I-1. I tried at least dozen times to upgrade the Helke Rd omni from W8GUC through DARA with a failed transfer every time. Sometimes it would get almost all the way done, but never completed. 
 
The northern nodes are a microcosm of the entire Channel -2 omni to omni mesh, where links suffer low LQs, often more so in one direction than the other. So just to see, and acknowledging their node receivers would still be hammered by the high-gain omnis from other sites, I moved DARA and W8GUC to a different SSID, AREDN-N (N for North). BAM! 100%/100% LQs at both sites. 
 
So I contacted Ken, AE8I, who moved AE8I-1 to AREDN-N. BAM! Instant improvement! DARA's LQ of it wasn't quite what AE8I-1's was of DARA, but it was 90-100%. However, DARA's LQ of W8GUC was down from 100% to just a bit above AE81-1. Ken moved his node in and out of AREDN-N and the results were consistent. DARA's LQ of two neighbors wasn't perfect, but was still pretty good. There's more going on now at AE8I that deserves separate discussion.
 
To keep a long story short, I added and removed Helke Rd and Vandalia to and from AREDN-N. Consistently, the third neighbor caused DARA's LQ of all neighbors to drop to a less than happy level, one acceptable until a non-2.4 link can reduce the number of neighbors in one SSID. The fourth neighbor dropped DARA's LQs down by a greater amount; however, I was able to upgrade Helke Rd from W8GUC on the first try. There may be something still wrong with Vandalia (other than needing filtering) that impacts DARA's LQ of it. There seems to be a larger drop in DARA's LQ of W8GUC when Vandalia is added back in.
 
DARA now has half the number of neighbors as before and its LQ of all is better than before. The unusable neighbors it lost also have a bit better LQ of their remaining neighbors. Half DARA's "-N" neighbors are what I'd consider impractical for data transfer. I believe all nodes still suffer from their RF "hearing" non-neighbored, other-SSID nodes on the same channel, but there's not a lot we can afford to do about that. It is clear DARA is still seeing too many neighbors for desirable LQs, so splitting off Helke and Vandalia into a -NW (NorthWest) SSID, linked to DARA with non-2.4 directionals is in order.
 
 
Part 2 - The South
 
Some of our omnis had Wilmington nodes as unusable neighbors as did their omnis of ours. Our experience up north indicated both ours and Wilmington's nodes suffered reduced LQs of all their received nodes as a result. When I shared our findings at DARA, Mark Atwell, KD8DGH, graciously moved all the Wilmington nodes to AREDN-S. Thank you, Mark! I didn't examine their nodes for improvement, but I expect they saw at least some. 
 
Our omnis that were uselessly neighbored with them saw some improvement, but as all our omnis, they still have a lot of useless neighboring--something typical of our omni-to-omni network, and the result of that speaks for itself.
 
TO BE CONTINUED