T-Mobile and a number of other industry companies contend that T-Mobile will suffer interference to its UMTS/HSPA system being built out on the ASW-1 spectrum (2110-2155 MHz) if someone were to build out a nationwide system as proposed by the FCC for the AWS-3 spectrum (2155-2180 MHz). The reason for concern is that ASW-1 spectrum is configured for typical cellular operation with all of the cell site’s transmitters (high power with directional antennas) being in one portion of the band and the low-powered handsets that have to communicate with these cell sites in another portion of the band. The spectrum that is to be auctioned in the AWS-3 band will employ a different form of wireless where cell sites and handsets operate in the same frequency band using a time division technology.

Not to get too technical, but the cellular/PCS/AWS-1 and 700 MHz spectrum is designed for FDD or Frequency Division Duplex, which, as explained above, keeps all of the cell site transmitters in one portion of the band and the receivers at those sites in another portion of the band. Wireless phones operate or, depending on the technology, appear to operate in full duplex mode, which means both sides of the conversation can be heard by both ends of the call. There are a number of different ways to provide for this full duplex system, including hardware, antenna separation and other engineering tricks, but the best way is to make sure that the cell site receivers that are listening for a very weak signal are in a different portion of the band. The more separation between the band segments, the better the system will operate.

As you can see, there is a lot of frequency separation between the base or cell site transmitters and the mobile transmitters. It is this separation, and the fact that ALL AWS-1 systems have the same separation, that helps minimize interference to the cell site receivers and the wireless device receivers.

The AWS-3 spectrum that will probably be auctioned next year is located just above the AWS-1 F band, and it is a single swath of spectrum that requires the cell site and the mobile transmitters to operate on the same frequencies. But, you say, look how far away the AWS-3 spectrum is from the AWS-1 mobile device spectrum. That is true, however, the reason for concern is the same reason Nextel is spending several billion dollars to “reband” the 850 MHz Special Mobile Radio band. The high-powered transmitters in the AWS-3 band could interfere with the device receivers when they are close to AWS-3 cell sites. Is this a big deal? Yes it is. The FCC has a set of rules limiting the out-of-band emissions that can be generated by a transmitter, but due to the nature of the technology, all transmitters will transmit energy outside of their licensed frequency. There is another issue here as well. If I have an AWS-1 device on the T-Mobile system in the AWS-1 F band, for example, and I get near a cell site that is operating in the AWS-3 band using WiMAX, my device receiver could be swamped by the AWS-3 transmitter and I would not be able to use my phone.

Since both of these systems require a large number of cell sites, it is not the case of a T-Mobile customer being near an AWS-3 cell site once in a while, the interference issues could be widespread. This is one reason I was interested in reading the results of the FCC’s tests in Seattle that were conducted for all of three days, in a single location at Boeing’s test facility.

I think a question about interference that is this important deserves more than three days of testing in a single, controlled environment—there is much to be said for real-world tests. The test results were recapped in a Memorandum from Julius Knapp, Chief, Office of Engineering and Technology, and made public. There is no conclusion attached to the memo, only the details of the tests, who was present and the results. Already, we have seen both sides declaring victory and the folks at M2Z have decided that the tests were conclusive and that TDD and FDD technologies can co-exist.

There is nothing wrong with a TDD technology when it works in spectrum where it will not cause interference to FDD systems in adjacent spectrum. Someone at M2Z was quoted as saying that if T-Mobile would experience interference from the AWS-3 band, it would also experience interference from the 2.4 GHz Wi-Fi band and other portions of spectrum below the mobile spectrum. But the receivers in that part of the band located at cell sites have extra filters to block out or minimize interference. The receivers that are susceptible to interference are in handsets and wireless devices that are moving around, being carried by people who know nothing about “interference.” However, they do know when their calls are dropped or they cannot make an outgoing call or connect at a decent data rate.

Other experts in the wireless industry don’t agree with M2Z’s conclusions. M2Z is the company that hopes to build the AWS-3 network, and the way I read the test results, it is unclear to me whether they were really conclusive. Several of the tests were not even run due to time limitations. It seems to me they should have taken as much time as was necessary to complete all of the lab tests and then move to a real-world environment to run other tests. A ruling by the FCC based partly on these results could come back to haunt M2Z, just as the Nextel interference to public safety frequencies in the same band came back to haunt Nextel.

The bottom line is that this is too important an issue to simply spend a few days testing in a lab. If the FCC is determined to proceed with the AWS-3 spectrum auction, there should be some wording included in the bid documents clearly stating the winning bidder’s obligations if it turns out the lab tests were inaccurate and future tests on the commercial systems prove there is, in fact, interference. In the “old days” of radio it was much simpler because radio channels were small slivers of spectrum 30, 25 or 15 KHz wide and most systems (with the exception of GSM) used only a few channels. It was easy to build a receiver front-end that was very sensitive on the intended channel(s) but filtered to reject signals on either side.

In today’s world of broadband radios (wireless), channel widths are measured in multiple megahertz, with CDMA using 1.25 MHz, UMTS 5 MHz and WiMAX at least 8 MHz. When LTE comes to market, it will operate in any amount of spectrum from 1.25 MHz up to 20 MHz. Building receivers to receive signals over such a broad portion of spectrum means they will probably be less able to reject interference from adjacent channels, and they certainly will “see” interference in their receive spectrum as noise, which tends have a negative impact on a receiver. Even though CDMA (UMTS is also based on CDMA) is good at working with signals “down in the noise” as it is called, there are still issues when noise is introduced from other sources.

The bottom line for me here is that the FCC continues to open up more spectrum for auction to raise money for the federal government without consideration of the consequences that may befall the winners.

Extending the Tests to 700 MHz

This of course brings up the issue of the public/private partnership D Block of 700 MHz spectrum. As I noted in my blog last week, I think it is quite possible for WiMAX to become the standard for this network, which would be a real disservice to the first responder community for reasons I have already laid out and because the issue of mixing TDD and FDD spectrum may have to be faced in this band as well.

Currently, WiMAX is a TDD-only technology, although I found this posting on the WiMAX Forum website to be interesting:

“When will the WiMAX Forum begin work certifying FDD mobile WiMAX products?”

“The IEEE 802.16 standard includes both FDD and TDD. Mobile WiMAX TDD is currently included within the IMT-2000 family of standards, and Mobile WiMAX FDD is to be proposed for inclusion later this year as enhancement of the existing air interface. With an increased interest in the 700 MHz band and more service providers around the world choosing WiMAX as their next generation mobile broadband technology, the WiMAX Forum has added this frequency band to our technology roadmap. Current market demand indicates that the priorities for additional TDD and FDD profiles are 700 MHz and FDD profiles for the US “AWS” band. The WiMAX Forum is currently working on FDD certification profiles for this spectrum. Work on the technical specifications for 700 MHz WiMAX Forum certification is already underway in the association's working groups. The specifications will be published as they are completed and they will support both TDD and FDD certification profiles.”

It appears as though the FCC is willing to bet WiMAX will get its FDD act together in time to start building out the public safety network. That is a bet I am not sure I would want to take, and if the WiMAX community thinks it can use TDD in the 700 MHz D Block, as well as the public safety 5X5 block, it is not facing reality. The impact of mixing TDD with FDD in the 700 MHz spectrum would be not only to create interference for the C and perhaps A and B Blocks, but also for the 14 MHz of spectrum the public safety community will be using for standard narrowband operations employing APCO 25 and perhaps even analog FM.

I believe mixing TDD and FDD technologies on adjacent bands is asking for trouble. If we have not learned from the Nextel “rebanding” issues, what will it take for us to learn? Interference is a way of life. It is what destroyed EarthLink’s muni-Wi-Fi systems in Anaheim, New Orleans and Philadelphia and turned muni-Wi-Fi into a nightmare for those that deployed it. It has cost Nextel and then Sprint a lot of money in rebanding efforts, and I cannot believe three days in the Boeing lab—a controlled environment—proves or disproves T-Mobile’s contention that the AWS-3 spectrum will adversely affect its network.

If we have learned anything the past thirty years, it should be that interference, even out-of-band interference, can degrade service. Wireless companies have paid the federal government a lot of money for spectrum that is useable, not for spectrum that is sometimes usable.

Andrew M. Seybold