Actually raise the antennas a bit more/reduce interference and we get perfect quality at 800-2000m. Many factors apart from the actual physical distance.

On the mailing list Rudolf Meijering counted 120 dwellings in a single 100m square area in a squatter camp. If the density is very high we will fill our mesh capacity quickly and need to deploy separate meshes on different frequencies. So we may find we need to pull our antennas off to reduce the range to 50m to avoid interference with adjacent mesh networks on the same channel

- David

In my neighbourhood, a suburban block is about 100M long. That makes 800M about 8 blocks distant, a sizeable amount.

At 800M, you say that the “quality was usable, but not perfect”. Let’s guestimate that at 400M between nodes, the call quality would be on the edge of “perfect”. That gives us about a maximum 4 block radius for “ideal” placement of MPs. Let’s reduce that to 2 blocks, in a rectilinear grid (so that the diagonal nodes are less than 400M apart), and you have a picture of (perhaps) cellphone-like coverage.

Now, /that/ image I can deal with. An MP on every other street corner, for minimum “proper” coverage? That’s a /great/ range.