Hi all, If one reads through the actual FCC report findings, you'll see that unlicensed devices in the TV bands are not just workable, but that the data the FCC itself collected supports the view that they're a viable technology. Keep in mind that these devices are 1st generation prototypes. Meanwhile, here's some food for thought:
First, let me just say that the Microsoft prototype sucked -- there's no other way around that one -- it doesn't perform even close to spec and fails miserably on just about every single benchmark that was tested. I find it hard to believe that the device was performing as engineered, but then again, Microsoft has been known to lay an egg from time to time. That said, as it turns out, the Phillips prototype performed exceptionally well. One _major_ measurement decision directly affecting the results of the study needs to get drawn out -- there is no standard for the sensing threshold for these devices. The White Spaces Coalition created a standard of -114 dBm as the necessary level of measure. However, the 802.22 committee has also been working on a standard (which is not yet set) of -116 dBm. The prototypes were ngineered to the -114 dBm standard; however, the FCC researchers tested them based on the IEEE proposal of -116 dBm. In essence, they were tested outside of spec to begin with (you can read the FCC's statement to the effect in the second paragraph of section 3.1 of the report). Several figures included in the report include the gradients from -119 dBm to -113 dBm -- so one can find out how the Phillips prototype did at the -114 dBm that it was manufactured to -- and the results are _stunning_: Figure 3-4. Baseline Detection Threshold Results for Prototype B (page 14 of the report) -- measures how well the prototype detects a DTV signal on the same channel. At -114 dBm this prototype detected a signal 100% of the time (not a lot, majority, most, or almost all, but 100%). At -115 dBm is also detects TV signal 100% of the time (i.e., the prototype performs even better than it was manufactured to do). It's only at the -116 dBm mark (which is out of spec) that it only detects things about 97% of the time on two of the channels and less than 40% of the time on a third channel. Of course, only the out-of-spec -116 dBm results were widely disseminating while the in-spec -114 dBm home run wasn't mentioned at all. Figure 3-9. Two-Channel Detection Threshold Test Results for WSD Prototype B (page 18 of the report) -- measures how well the prototype detects DTV signal on adjacent channels. Once again, at -114 dBm the Phillips prototype detects adjacent channel signals 100% of the time. It does poorly at the -116 dBm mark, but it was never manufactured to measure at that sensitivity. The Phillips prototype was never used in field tests (at the request of the manufacturer). But taken together, it appears that unlicensed devices can work _extremely well_ within the TV-Band at the level of sensitivity they are manufactured to see. The problem is that they were tested completely out-of-spec (I suspect that the manufacturers did not know that the FCC would use a different testing metric than they were using). When tested in-spec, the Phillips Prototype scored a whopping 100% on both bench tests -- which is both remarkable and quite promising for the technology. It's rare that I find something so completely different from what has been reported. But in this case, the "news" has all been that the idea is a complete failure; yet it appears that the Phillips Prototype has demonstrated quite conclusively that unlicensed devices in unused bands are quite possible. --Sascha Meinrath Research Director Wireless Futures Program New America Foundation -------------------------------------------------------------------------------- WISPA Wants You! Join today! http://signup.wispa.org/ -------------------------------------------------------------------------------- -- WISPA Wireless List: [email protected] Subscribe/Unsubscribe: http://lists.wispa.org/mailman/listinfo/wireless Archives: http://lists.wispa.org/pipermail/wireless/
