Measuring Coverage Quality for Femtocell and Macrocell Broadband Data Services Jay A. Weitzen, and Theodore Grosch Airvana, Chelmsford, Ma, 01824, USA and University of Massachusetts Lowell, Lowell Ma 01854 ABSTRACT —This paper presents the results of field trials conducted to measure and compare indoor airlink data rates from femtocells to indoor coverage from macrocells and concludes that operators can expect to see significant (3-5x) improvement in average airlink data throughput relative to the existing macrocell network. Tests were conducted in a number of single family and multi-unit residences. The paper concentrates on UMTS/HSDPA measurements and presents the methodology used in tests and offers summary results. It shows that the case for using femtocells for broadband data services is stronger than for voice based services. Index Terms — Femtocells, Radio Propagation I. INTRODUCTION Much of the early attention on femtocells has focused on how they can be used to improve voice services, especially in coverage limited locations [3]; however broadband data services are increasingly becoming a significant source and percentage of the mobile operator’s business. While a single voice call can be supported using only about 12 kbps of airlink capacity, broadband data services require much higher throughput levels to deliver acceptable performance for users. A strong case can be made that the expected improvement in indoor user experience due to deployment of femtocells will be significantly greater for broadband data services than for voice services. A recent study [1] found that 36 percent of voice calling and 45 percent of mobile data usage happens within the homes of consumers. The argument that femtocells provide greatly improved user experience for broadband data indoors is based on two key factors when femtocells are used for broadband data: 1) Broadband data rates require high signal levels and signal to noise ratios 2) Broadband wireless data systems such as 1xEV- DO and HSDPA are shared media in which the total airlink bandwidth is shared among all the users who are in each sector The quality of coverage, that is the received signal level seen by a user at a given location within a building, is a function of four primary propagation mechanisms: large-scale attenuation, outdoor shadowing, indoor shadowing, and micro- spatial fading. A number of propagation models such as industry standards Okamura-Hata [4], and Cost-231 [6] predict the mean attenuation as a function of parameters such as receive and transmit antenna heights, frequency, general land use, etc. Typical cell radii in suburban areas range from about 1 to 5 or more kilometers, depending on propagation conditions, traffic density, and network deployment costs and strategies. At any given range within the cell radius, the actual signal level or pathloss will vary about the average, usually according to a log normal distribution. Additional margin must be built into to the link budget calculation to make sure that not only does the average user get the desired pathloss or signal, but a high percentage of the users receive the required signal threshold. After inserting shadowing margin into the link budget calculation, the next step is to insert building penetration loss. While the exact penetration is a function of the construction materials, etc, and varies from building to building, also log normally distributed, most standard models assume an average penetration loss of approximately 10 dB for suburban dwellings, and around 15 to 20 dB for more urbanized areas. Finally within the building, due to walls, obstructions, etc, on the average the signal strength will vary log normally around the mean value. Again, high reliability of the coverage requires that extra dB of link margin be built into the link budget. Requiring people to do their data downloads next to a window is not acceptable. The bottom line is that when all these losses are taken into consideration, the cell radius required to support high data rates indoors, with high reliability becomes economically impossible, except in very densely populated urban areas. The following link budget example illustrates the point: Table 1: Link Budget Calculations for Macrocell Network a) Thermal Noise Level -174 dBm/Hz b) Target Data Rate: 2 Mbps c) Receiver Noise Figure: 10 dB d) Required Signal To Noise Ratio: 10 dB e) Target RSSI=a+10Log10(b)+c+d -91 dBm f) Exterior Shadowing Standard Deviation 10 dB g) Suburban Building penetration loss 10 dB h) Margin @ 75% cell edge (0.675 σ) 7 dB i) Indoor Shadowing Standard Deviation 10 dB j) Indoor Margin (90% reliability,1.3σ) 13 dB k) Required RSSI= e + g + (h 2 +j 2 ) 0.5 -66 dBm l) Effective Radiated Power 56 dBm m) Maximum Path loss (k - l) 122 dB