Today T-Mobile officially announced the launch of its 4G LTE network. T-Mobile might be the last of the major carriers to launch a LTE network in the US, but this has allowed T-Mobile to implement some cutting edge LTE technology into its new network. T-Mobile is using two different equipment vendors, Ericsson and Nokia-Siemens, to power its LTE network. In the Ericsson markets, T-Mobile is deploying a brand new technology called active antennas. Active antennas are the evolution of cell site architecture, and offer the potential for substantial improvements in LTE performance and capacity.
T-Mobile's deploying the Ericsson AIR21 Active Antennas. This shows two building mounted sectors. Note the large depth of the antennas and how there are only two small cables to the antenna (one power and one fiber).
Traditional cell sites placed the base station radios at the bottom of the tower and used thick coax cables to transmit the signal to the antennas at the top. The issue with this solution is that long runs of coax cable cause attenuation. This means that a base station might output 20W of power, but by the time this signal reaches the antennas it is now only 15W. The same concept is true in the uplink direction when a mobile device is transmitting to the cell site. The uplink is usually the limiting factor in cellular communications as mobile devices can only transmit at a fraction of a watt compared to the multiple watts of a base station. One solution to improve the uplink performance is to mount an uplink amplifier at the top of the tower, and this is known as a tower mounted amplifier or TMA. Once the uplink signal is received by the antenna it is boosted by the TMA to help overcome the attenuation of traveling through the coax to the base station. TMA’s are widely used by AT&T and T-Mobile on their 3G UMTS networks.
A conventional cellular antenna design. Source: Commscope
The evolution of the TMA was the remote radio head or RRH. The RRH moved the entire transmit and receive radios and power amplifiers to the top of the tower. The benefit of moving to RRHs is that signal attenuation is greatly reduced, which increases both the downlink and uplink performance. Instead of thick coax cables, fiber and power cables are run up the tower taking up less space. RRHs still require a small coax jumper to connect to the antennas which adds a small amount of attenuation. RRHs are being used by most of the LTE industry, and they can be easily identified by big boxes near the antennas on top of towers.
A RRH cellular antenna design. Source: Commscope
The most recent technological advance, which T-Mobile is utilizing, is an integrated LTE radio inside an antenna. This technology is known as active antennas. The obvious benefit of this concept is completely removing coax cables from the equation, minimizing any signal attenuation. T-Mobile is using the Ericsson AIR antennas which claim to have a 1dB improvement in the uplink over a RRH solution. Another benefit of active antennas is the ability to better control the antenna’s beam pattern. This allows cellular operators to more accurately define coverage areas, which can improve performance especially at the cell edge. Additionally, in the future active antennas will allow for a concept known as beam forming. Beam forming “steers” the antennas beam via a concept known as phase shifting. Instead of the antenna’s beam providing coverage for an entire area, the beam is focused in the direction of each user utilizing the service at that instant. Beam forming has the ability to greatly improve wireless performance and capacity, and it is currently used in some WiFi access points. The issue with beam forming is that it requires a large number of radios and antennas to work to its full potential. With today technology this results in large antenna arrays which are expensive. While beamforming is still in its infancy, T-Mobile’s use of active antennas is paving the way on how cellular networks will be built in the future.
An Active Antenna cellular antenna design. Source: Commscope
A little over six months ago, Sprint-Nextel laid out its strategy for revamping its wireless network and called the plan “Network Vision.” If you have read any of my previous articles about Sprint, you would know that Sprint has not had any real network strategy since purchasing Nextel back in 2005. Today Sprint still has numerous sites where they have yet to combine their iDEN, CDMA/EVDO, and Clearwire’s WiMax network which has resulted in poor coverage and high maintenance and real estate costs. Well this is all about to change with Network Vision. After seven years without any true network plan, Sprint-Nextel has something that actually makes sense.
Sprint Network Vision Tower (Alcatel-Lucent Equipment)
Here is a brief overview of what “Network Vision” entails. The website, http://s4gru.com, has some excellent detailed information on what “Network Vision” really means from a technical perspective.
- Consolidate its cell sites, by removing sites that are not needed. Sprint currently has 68,000 sites and will reduce this by 44% to eventually remain with 38,000 sites.
* AT&T claims they have 55,000 cell sites so once Network Vision is completed its nationwide coverage will still lag behind that of AT&T.
- Shutting down iDEN and reusing the spectrum to support at least one 800MHz CDMA 1X Advanced carrier.
* Deploying a 1X carrier in the 800MHz spectrum will greatly improve the voice performance along with coverage for Sprint, especially inside buildings.
Frequency plan for new 1X advanced carriers. Source: s4gru.com
- Deploying a LTE carrier in a 5x5MHz channel configuration in their 1900MHz (PCS) spectrum.
PCS Band Plan. Source: howardforums.com
Sprint is using three RRH per face (1 for 800MHz CDMA, 1 for PCS EVDO, 1 for PCS LTE)
At the end of the day Sprint will have a CDMA/EVDO/LTE network, just like Verizon Wireless. By consolidating its cell sites and turning off iDEN, Sprint will save a ton of money on operating expenses. It is interesting that Sprint is investing a lot of time and money upgrading CDMA/EVDO instead of just focusing on deploying LTE. Additionally, with MetroPCS, AT&T, and Verizon Wireless all committing to VOLTE it is interesting that Sprint is planning on deploying CDMA 1X Advanced for voice calls. Sprint must have believed that its CDMA/EVDO networks could be greatly improved with Network Vision and that both these technologies will be around for some time. Sprint has been successful at finding ways to monetize its old networks, such as offering Boost Mobile prepaid service over its iDEN network. As postpaid customers move to LTE, Sprint could offer competitively priced but slower data services overs its CDMA/EVDO network to maximize its investment.
The one element that was left out of Network Vision is Clearwire which Sprint owns 54% of the company. If Clearwire partnered with Sprint, like Lightsquared attempted before all their GPS interference issues, Clearwire’s network consolidation could save a great deal of money for the small carrier. Clearwire will be upgrading its network to LTE, but it will be based on TD-LTE technology instead of FDD-LTE that all the other US carriers are using. Clearwire’s 2.5GHz spectrum limits its usefulness to urban areas and the high cell density needed for good coverage makes network expansion expensive. Clearwire is hoping to sell extra LTE capacity to the major wireless carriers, but using a different LTE technology and a separate frequency band than everyone else will make this difficult. While Sprint’s issues with Clearwire remain, Network Vision is a huge step in the right direction for Sprint. One complete it will offer much greater voice coverage, improved EVDO performance, and most importantly bring Sprint into the LTE game.
A single dual band antenna supports all three technologies
Today AT&T launched its 4G LTE network in five cities. While this is a big step for AT&T, its main competitor, Verizon Wireless, already launched its 4G LTE network almost ten months ago. Verizon’s 4G LTE network now has service in 143 markets and covers over half the US population. Wireless carriers that utilize EVDO technology for 3G services such as Sprint-Nextel and Verizon Wireless have been the first to move to 4G technologies as EVDO offers slower theoretical speeds than the competing HSDPA+ technology. A single EVDO rev. A channel’s theoretical speed is 3.1Mbps in the downlink and 1.8Mbps in the uplink while a single HSDPA+ (not utilizing channel bonding or MIMO) channel can offer 21Mbps in the downlink and 5.8Mbps in the uplink. In real word applications the speed advantage for HSDPA+ is much less because HSDPA+ carries voice calls over the same channel which reduces the data speeds. Verizon Wireless and AT&T have always used different network technologies with Verizon choosing the CDMA (3GPP2) path while AT&T electing the GSM (3GPP) route. While both technologies offer their share of advantages and disadvantages, going forward both companies will use the same network technology – Long Term Evolution or LTE.
LTE is not a one size fit all technology, but instead a technology that allows for a variety of different configurations which greatly impact how it is deployed and its performance. Both AT&T and Verizon Wireless utilize frequency division duplex (FDD) mode which means that the upload and download channels are on two separate frequencies. LTE also offers the capability to use time division duplex (TDD) which allows for both the download and upload channels to use one frequency with the download and upload being allocated different time slots.
Another similarity between AT&T and Verizon Wireless is that they are both utilizing 2×2 MIMO antenna technology. While LTE supports MIMO is an extremely complicated topic, but basically it allows double the amount of data to be transferred in a single channel by utilizing two transmit and receive antennas instead of one.
LTE release 8 supports the options for one, two, or four antenna configurations where the highest performance is achieved utilizing a 4×4 MIMO solution. Almost all wireless carriers are choosing the 2×2 MIMO route as it offers the best performance/price ratio. To go with a 4×4 MIMO solution over that of 2×2 MIMO means that double the number of antennas and amplifiers are needed along with more powerful processors in mobile handsets and base stations to decode the additional data streams. Additionally according research by Ericsson Communication, MIMO only provides performance improvements when a receiver has a signal to noise ratio (SNR) of approximately 10dB or better. This means that MIMO is most beneficial when a user is close to the cell site, which for most cell sites is only a small percent of the users.
The main difference between AT&T’s and Verizon Wireless’ 4G LTE network is the bandwidth that each channel uses, and this is based on the spectrum allocation that both companies own. AT&T is using a mixture of 10MHz and 5MHz channels while Verizon Wireless is solely using 10MHz channels. In areas where AT&T uses 5MHz channels, Verizon Wireless’ network will theoretically offer double the performance of that of AT&T’s. Theoretically a 10MHz channel utilizing 2×2 MIMO supports peak downlink data rates of 73Mbps while a 5MHz channel will only support 37Mbps. As with any wireless technology reaching anywhere near these theoretically numbers is extremely unlikely. AT&T knows that its 5MHz channels will put it at a large capacity and speed disadvantage compared to Verizon Wireless, so in markets where it has both 700MHz and AWS spectrum it will try to utilize two 5MHz channels instead of just one. The two channels become beneficial when a large number of users are on the network and the load is distributed across two 5MHz channels instead of crowding everyone into one 5MHz channel. Currently, these two channels can’t be bonded for higher throughput, but this technology will come available in LTE Advanced and is known as carrier aggregation.
The final difference between the two 4G LTE networks is the base station radios used. A growing trend in the wireless industry is to mount the base station’s radio and amplifier at the top of the tower. This is known as remote radio heads (RRH) and this technology minimizes the cable attenuation experienced by an antenna system. In traditional base station deployments, the radios and amplifiers are mounted on the ground, where they can be easily upgraded and repaired, and thick coax runs up the tower to the antennas. The issue with this is that long runs of coax cable experience attenuation. According to this spec sheet, 100FT of 1 ¼ coax cable has a loss of 1.6dB or roughly 31% less power at the top of the tower compared to when the signal left the amplifier. Clearly reducing the antenna’s output power by 31% not only reduces coverage and degrades downlink throughput, but it also affects the uplink. The antenna at the top of the power sees 31% more power from the mobile handset that what actually makes it to the base station. This result in decreases coverage, reduced uplink throughput, and diminished battery life for handsets. By mounting the radio and amplifier at the top of the top the 1.6dB cable loss is practically eliminated, greatly improving performance over traditional base station deployments. Instead of running thick coax up the tower, a much thinner combined fiber optic cable and power cable are run to each remote radio head.
AT&T 700MHz LTE RRH
Clearwire’s 4G WiMAX network was the first wireless operator to solely use RRH and they can be easily spotted by the large boxes connecting to the antenna. AT&T is following in Clearwire’s footsteps by primarily using RRH for its LTE deployment. Currently RRH can only support one technology and frequency band, so with AT&T dual frequency (700MHz and AWS) LTE deployment this means two RRH are needed for each face of the tower (most towers have three faces). Most of the Verizon Wireless tower I have observed do not have any RRH mounted, so it is probable they are going with the conventional base station deployment model with the radios and amplifiers mounted at the bottom of the tower. The benefit of this solution is that equipment can be quickly and easily repaired and upgraded while staying protected from the outdoor elements. When a RRH goes bad or needs to be upgraded it requires someone to climb the tower which is time consuming and can become very costly. Given that RRH technology is still very new, it will take time to see whether the performance gains of RRH make up for the limitations in repair and upgradability.
Overall, the technology is similar for both Verizon Wireless’ and AT&T’s LTE networks. While the technology might be similar, Verizon Wireless is the clear leader already having half the US population covered compared to just five cities for ATt&T. In the end regardless of whether one chooses Verizon Wireless or AT&T, US consumers are the true winners having access to multiple advanced 4G LTE networks.
Last week Sprint-Nextel finally announced its LTE plans which involved teaming up with Lightsquared to build a 4G LTE network. This announcement was expected for some time, the only question that remains is what Sprint is going to do with Clearwire. Clearwire’s stock price dropped 20% on the news that it was not part of Sprint’s LTE plan. Clearwire’s market cap is now a little over $2 billion, which begs the question, who is going to buy Clearwire? One issue is that Sprint-Nextel owns a majority stake in Clearwire, which greatly complicates a takeover. Even so, Sprint-Nextel needs cash and would rather upgrade its own network than fund Clearwire to migrate from WiMax to LTE.
Clearwire has a massive amount of spectrum in the 2.5GHz band which has been estimated to be worth between $5-10 billion. The 2.5GHz band does not propagate nearly as well as the 700MHz bands which Verizon Wireless paid $4.7 billion for just a 22MHz nationwide license. This means to deploy a network at 2.5GHz with similar coverage to that of one deployed at 700MHz would require between 3x-4x the numbers of cell sites. While more cells are required to achieve similar coverage, the capacity of a 2.5 GHz network is far greater given the huge spectrum availability compared to what is available with existing 700MHz, 850MHz, PCS, and AWS bands. This makes Clearwire’s spectrum extremely valuable in dense urban environments such as New York City or sport stadiums where deploying a large amount of spectrum on the 2.5GHz band would give a wireless carrier a massive speed and capacity advantage over the competition.
In addition to Clearwire’s spectrum, its assets include a WiMax network and one of the largest microwave backhaul networks in the US. Sadly the WiMax network has not lived up to its expectations as a next generation network and performance wise it delivers similar results to that of HDSPA+ used by T-Mobile and At&t. There are few reasons to keep the WiMax network operational as it isn’t going to generate substantial revenues now that rivals are deploying LTE which has far better performance. Some of the newer the sites have equipment that can be easily upgraded to LTE, but this is still going to require a serious investment. Today Clearwire announced that it would move to LTE Advanced, but this would require an additional $600 million and something they didn’t mention is that technology wouldn’t even support LTE Advance for at least another year.
Clearwire’s microwave backhaul network paints a better picture. Clearwire has an advanced microwave network which allows it to offer a high amount of bandwidth at a fraction of the cost of traditional TDM services and in many cases fiber based solutions. This microwave network could be used to resell spare bandwidth to other carriers, as a majority of cells still use T1’s as fiber is not available at many cell locations. Overall, Clearwire’s microwave network could be quickly converted to lease spare bandwidth to other carriers which could bring in a fair amount of additional revenue. Time will tell whether Clearwire survives, but it is a potentially great buy for a wireless carrier or cable provider wanting to gobble up a huge amount of spectrum.
It has been about a year since I wrote my research paper on femtocells for my master’s degree. At the time when I wrote the paper, industry analysts were still optimistic in femtocells technology and believed the industry was on the verge of massive growth. Additionally, data traffic usage was about to explode driven by wide smartphone adoption and faster 4G networks being deployed worldwide. One year later, the analysts were correct about the massive data growth, but the popularity of femtocells has yet to materialize. This isn’t to say that wireless providers have lost faith in femtocells; in fact the opposite is true. 2010 saw the release of 3G capable femtocells (the earlier ones were only 2G) from At&t, Sprint-Nextel, and Verizon Wireless. At CTIA 2011 Sprint-Nextel was proud to announce that it had over 250,000 femtocell users which represents a majority of the total femtocell users in the US.
While Sprint-Nextel might be proud of this accomplishment, a quarter million femtocell users has done little to improve capacity and many Sprint-Nextel customers still live with poor coverage. As I wrote in my paper, femtocells need to be deployed on a large scale basis to improve capacity and coverage. Sprint-Nextel’s industry leading penetration is only .5%, where penetration needs to be minimally 10-20% to start seeing gains in capacity.
Capacity is crucial as wireless carriers are investing increasing capital to keep up with growing data demand. The solution to this is to invest in more advanced wireless technology which lowers the cost to deliver data services to customers. While this has worked successfully in the past, today’s explosive growth in data services mean that even 4G networks will require substantial investments to keep up with capacity. This means that the ROI on network infrastructure will diminish, unless a creative solution can be found.
Femtocells like technology can be this solution, but as I wrote in my research paper traditional femtocells will never be the answer. Traditional femtocells, small boxes similar to Wi-Fi routers are complex, costly to manufacture, and are only compatible with a single wireless carrier. Femtocell technology needs to utilize Wi-Fi already present at most homes and businesses. As 4G networks start to implement VOIP based calling such as VOLTE, Wi-Fi offload will be much easier to implement compared to today’s circuit switched based calls. If the 3GPP and IEEE worked together, seamless and universal LTE to Wi-Fi offload could become a reality. Wireless carriers need to realize that Wi-Fi is an extremely important technology at their disposal, and a crucial aspect in retaining profitability. Overall, today’s femtocells will never gain traction in the marketplace, and instead the future is intelligent Wi-Fi offload solutions.