Chesapeake NetCraftsmen

It has occurred to me that the slides are a bit literal, and that might lead some folks astray. Let me explain the issue. If you take the linear dimensions and put AP's in a grid, you'll end up with a good many more than if you take the total square feet and divide by the square feet per AP. Why the discrepancy?

The answer is pretty simple. When you estimate the number of AP's by square feet, you are in effect working with non-overlapping AP circles. But deployments overlap! So if you budgeted one AP per 3000 sq ft, for linear placement you have to adjust the square feet per AP number (or the linear distance) upwards, i.e. each AP is going to cover something more like 5000 sq ft, i.e. the spacing is NOT going to be what I have in the linear distance table for a "pure" 3000 sq ft circle calculation. It looks like you need to bump the linear distance up by about 1.3, as another way of approaching it.

The key here is realizing you're best off working it exactly one way: pick linear dimensions (for no less than say my 5000 sq ft number of 57') and then lay out a grid and count. Or work budgetarily with X sq ft per AP, but then come up with linear dimensions that are appropriate when you take overlap into account.

If you're measuring in the field you might wonder where I got the 43' number from. That was from a Cisco VoWLAN doc where I apparently missed the qualifying portion a bit below there: they were turning the power down to 40 mW to match a 7921 G phone. So with lower power the 43' number makes more sense. (Other Cisco docs show 54 Mbps at 80' distances..., presumably at 100 mW power.)

The VoWLAN deployment number of 1 AP per 3000 sq ft or so works out to an AP spacing around 60', where the RRM power management will probably end up turning down the RF power a good bit, to avoid massive overlap. Due to wall materials and other factors, YMMV (Your Mileage May Vary).