Reading and understanding it and the expected path loss is one thing, and then real world experience is a whole another thing. You can then compound your confusion on expected performance by realizing that wide band spreading modes are yet another layer to the mix.
I built a simple analog 900 MHz repeater so that I could get a better idea of how signal propagation on the band. I was rather impressed. On the same token I was also impressed using the slower Aerocomm FHSS units on 900 MHz. But trying to draw comparisons between narrow to wide is a bad plan.
The XR-9 is a logical step for the average ham, with a modest tower. The versatility in data rates and ability to change the channel width from the default of 20 MHz spreading to a half-rate of 10 MHz and even a quarter rate of 5 MHz is ideal.
Overall in all the experimentation of wide band spread spectrum at 900 MHz and 2.4 GHz, from the Proxim Symphony, origional 900 MHz WaveLan, Ubiquiti Bullet to the XR-9, I feel the average ham will be mildly disappointed. (Also see David Rowe, VK5DGR's blog on Wifi is hard.)
All this stuff works great if you can get a central node/AP up high on a commercial tower. Not so great if you live in an established neighborhood with medium to mature size trees.
I have experienced the difference between 1 watt and adding a 10 watt BDA's on both bands. For the average ham, living in the city with a modest tower, this does improve things significantly, and provides something usable.
In light of the recent spread spectrum rule change, I would highly appreciate if TAPR would consider getting behind a BDA project.
I would appreciate if the ARRL would clarify how spread spectrum fits into the existing band plans on 902 and above.
To help things along here is a recent article I stumbled into detailing construction and theory on a switchless bidirectional amplifier for 2.4 GHz using circulators.
The theory can be applied to BDA development for less crowded bands like 3 GHz, 5 GHz or even 900 MHz.
Most commercial bidirectional amplifiers use radio frequency (RF) switches which switch between two signal paths. Their switching time can cause data loss at a high transfer rate. For instance, with a data rate of 54 Mbit/s the time duration for 1 bit is 17.66 ns. Switching times of commercial available bidirectional amplifiers can be as high as 2 ms, causing a loss of more than 100 bits. There, bidirectional amplifiers realized as monolithic microwave integrated circuits (MMIC) bias a transistor circuit in such a way that it amplifies either in the one or the other direction......
In the following, a switchless bidirectional amplifier is discussed. It is compared against a bidirectional amplifier concept using RF switches as commonly used in commercial wireless communication systems....
Back in June 2007, L-com acquired HyperLink Technologies of Boca Raton, Florida.
Below are some internal pictures of their 900 MHz 802.11 amplifier. Both their HA910I-APC and newer HA905I-DM are essentially the same.
The same variable attenuator input, M68772 RF power amplifier, and MD003 PIN diode switch. The newer model has much better heat dissipation, and runs at 12 volts in stead of 13.5 volts. So even the newer advertised 5 watt model is capable of more.
I don't like the idea of using Amplifiers on 2.4 GHz as the channels overlap and there is the potential to disrupt to many other users. However there is a cheap amplifier 5-8 watt for that band, the EDUP EP-AB003 2.4GHz Amplifier for about $50-60.