LTE and the Evolution to 4G Wireless: Design and Measurement Challenges

Chapter 1

LTE Introduction

1.1 Introduction

The challenge for any book tackling a subject as broad and deep as a completely new cellular radio standard is one of focus. The process of just creating the Long Term Evolution (LTE) specifications alone has taken several years and involved tens of thousands of temporary documents, thousands of hours of meetings, and hundreds of engineers. The result is several thousand pages of specifications. Now the hard work is underway, turning those specifications into real products that deliver real services to real people willing to pay real money. A single book of this length must therefore choose its subject wisely if it is to do more than just scratch the surface of such a complex problem.

The focus that Agilent has chosen for this book is a practical one: to explain design and measurement tools and techniques that engineering teams can use to accelerate turning the LTE specifications into a working system. The first half of the book provides an overview of the specifications starting in Chapter 2 with RF aspects and moving through the physical layer and upper layer signaling to the System Architecture Evolution (SAE) in Chapter 5. Due to limited space, the material in Chapters 2 through 5 should be viewed as an introduction to the technology rather than a deep exposition. For many, this level of detail will be sufficient but anyone tasked with designing or testing parts of the system will always need to refer directly to the specifications. The emphasis in the opening chapters is often on visual rather than mathematical explanations of the concepts. The latter can always be found in the specifications and should be considered sufficient information to build the system. However, the former approach of providing an alternative, more accessible explanation is often helpful prior to gaining a more detailed understanding directly from the specifications.

Having set the context for LTE in the opening chapters, the bulk of the remainder of the book provides a more detailed study of the extensive range of design and measurement techniques and tools that are available to help bring LTE from theory to deployment.

1.2 LTE System Overview

Before describing the LTE system it is useful to explain some of the terminology surrounding LTE since the history and naming of the technology is not intuitive. Some guidance can be found in the Vocabulary of 3GPP Specifications 21.905 [1], although this document is not comprehensive. The term LTE is actually a project name of the Third Generation Partnership Project (3GPP). The goal of the project, which started in November 2004, was to determine the long-term evolution of 3GPP’s universal mobile telephone system (UMTS). UMTS was also a 3GPP project that studied several candidate technologies before choosing wideband code division multiple access (W-CDMA) for the radio access network (RAN). The terms UMTS and W-CDMA are now interchangeable, although that was not the case before the technology was selected.

In a similar way, the project name LTE is now inextricably linked with the underlying technology, which is described as an evolution of UMTS although LTE and UMTS actually have very little in common. The UMTS RAN has two major components: (1) the universal terrestrial radio access (UTRA), which is the air interface including the user equipment (UE) or mobile phone, and (2) the universal terrestrial radio access network (UTRAN), which includes the radio network controller (RNC) and the base station, which is also known as the node B (NB).

Because LTE is the evolution of UMTS, LTE’s equivalent components are thus named evolved UTRA (E-UTRA) and evolved UTRAN (E-UTRAN). These are the formal terms used to describe the RAN. The system, however, is more than just the RAN since there is also the parallel 3GPP project called System Architecture Evolution that is defining a new all internet protocol (IP) packet-only core network known as the evolved packet core (EPC). The combination of the EPC and the evolved RAN (E-UTRA plus E-UTRAN) is the evolved packet system (EPS). Depending on the context, any of the terms LTE, E-UTRA, E-UTRAN, SAE, EPC, and EPS may get used to describe some or all of the system. Although EPS is the only correct term for the overall system, the name of the system will often be written as LTE/SAE or even simply LTE, as in the title of this book.

Figure 1.2-1 shows a high level view of how the evolved RAN and EPC interact with legacy radio access technologies.

Figure 1.2-1. Logical high-level architecture for the evolved system

(from 23.882 [2] Figure 4.2-1)

The 3GPP drive to simplify the existing hybrid circuit-switched/packet-switched core network is behind the SAE project to define an all-IP core network. This new architecture is a flatter, packet-only core network that is an essential part of delivering the higher throughput, lower cost, and lower latency that is the goal of the LTE evolved RAN. The EPC is also designed to provide seamless interworking with existing 3GPP and non-3GPP radio access technologies. The overall requirements for the System Architecture Evolution are summarized in 22.278 [3]. A more detailed description of the EPC is given in Chapter 5.

1.3 The Evolution from UMTS to LTE

The LTE specifications are written by 3GPP, which is a partnership of standards development organizations (SDOs). The work of 3GPP is public and, as will be described in Section 1.6, it is possible to gain access to all meeting reports, working documents, and published specifications from the 3GPP website: www.3gpp.org. The organizational partners that make up 3GPP are the Japanese Association of Radio Industries and Businesses (ARIB), the USA Alliance for Telecommunications Industry Solutions (ATIS), the China Communications Standards Association (CCSA), the European Telecommunications Standards Institute (ETSI), the Korean Telecommunications Technology Association (TTA), and the Japanese Telecommunications Technology Committee (TTC).

Table 1.3-1 summarizes the evolution of the 3GPP UMTS specifications towards LTE. Each release of the 3GPP specifications represents a defined set of features. A summary of the contents of any release can be found at www.3gpp.org/releases.

Table 1.3-1. Evolution of the UMTS specifications

The date given for the functional freeze relates to the date when no further new items can be added to the release. After this point any further changes to the specifications are restricted to essential corrections. The commercial launch date of a release depends on the period of time following the functional freeze before the specifications are considered stable and then implemented into commercial systems. For the first release of UMTS the delay between functional freeze and commercial launch was several years, although the delay for subsequent releases was progressively shorter. The delay between functional freeze and the first commercial launch for LTE/SAE was remarkably short, being less than a year, although it was two years before significant numbers of networks started operation. This period included the time taken to develop and implement the conformance test cases, which required significant work that could not begin until the feature set of the release was frozen and UEs had been implemented.

After Release 99, 3GPP stopped naming releases with the year and opted for a new scheme starting with Release 4. This choice was driven by the document version numbering scheme explained in Section 1.6. Release 4 introduced the 1.28 Mcps narrow band version of W-CDMA, also known as time division synchronous code division multiple access (TD-SCDMA). Following this was Release 5, in which high speed downlink packet access (HSDPA) introduced packet-based data services to UMTS in the same way that the general packet radio service (GPRS) did for GSM in Release 97 (1998). The completion of packet data for UMTS was achieved in Release 6 with the addition of high speed uplink packet access (HSUPA), although the official term for this technology is enhanced dedicated channel (E-DCH). HSDPA and HSUPA are now known collectively as high speed packet access (HSPA). Release 7 contained the first work on LTE/SAE with the completion of feasibility studies, and further improvements were made to HSPA such as downlink multiple input-multiple output (MIMO), 64QAM on the downlink, and 16QAM on the uplink. In Release 8, HSPA continued to evolve with the addition of numerous smaller features such as dual-carrier HSDPA and 64QAM with MIMO. Dual-carrier HSUPA was introduced in Release 9, four-carrier HSDPA in Release 10, and eight-carrier HSDPA in Release 11.

The main work in Release 8 was the specification of LTE and SAE, which is the main focus of this book. Work beyond Release 8 up to Release 12 is summarized in Chapter 8, although there are many references to features from these later releases throughout this second edition. Within 3GPP there are additional standardization activities not shown in Table 1.3-1 such as those for the GSM enhanced RAN (GERAN) and the IP multimedia subsystem (IMS).

1.4 LTE/SAE Requirements

The high level requirements for LTE/SAE include reduced cost per bit, better service provisioning, flexible use of new and existing frequency bands, simplified network architecture with open interfaces, and an allowance for reasonable power consumption by terminals. These are detailed in the LTE feasibility study 25.912 [4] and in the LTE requirements document 25.913 [5]. To meet the requirements for LTE outlined in 25.913 [5], LTE/SAE has been specified to achieve the following:

Increased downlink and uplink peak data rates, as shown in Table 1.4-1. Note that the downlink is specified for single input single output (SISO) and MIMO antenna configurations at a fixed 64QAM modulation depth, whereas the uplink is specified only for SISO but at different modulation depths. These figures represent the physical limitation of the FDD air interface in ideal radio conditions with allowance for signaling overheads. Lower peak rates are specified for specific UE categories, and performance requirements under non-ideal radio conditions have also been developed. Comparable figures exist in [4] for TDD operation.

Scalable channel bandwidths of 1.4 MHz, 3.0 MHz, 5 MHz, 10 MHz, 15 MHz, and 20 MHz in both the uplink and the downlink.

Spectral efficiency improvements over Release 6 HSPA of 3 to 4 times in the downlink and 2 to 3 times in the uplink.

Sub-5 ms latency for small IP packets.

Performance optimized for low mobile speeds from 0 to 15 km/h supported with high performance from 15 to 120 km/h; functional support from 120 to 350 km/h. Support for 350 to 500 km/h is under consideration.

Co-existence with legacy standards while evolving toward an all-IP network.

Table 1.4-1. LTE (FDD) downlink and uplink peak data rates (from 25.912 [4] Tables 13.1 & 13.1a)

The headline data rates in Table 1.4-1 represent the corner case of what can be achieved with the LTE RAN in perfect radio conditions; however, it is necessary for practical reasons to introduce lower levels of performance to enable a range of implementation choices for system deployment. This is achieved through the introduction of UE categories as specified in 36.306 [6] and shown in Table 1.4-2. These are similar in concept to the categories used to specify different levels of performance for HSPA.

Table 1.4-2. Peak data rates for UE categories (derived from 36.306 [6] Tables 4.1-1 and 4.1-2)

Categories 6, 7, and 8 were added in Release 10 for the support of LTE-Advanced (see Section 8.3). There are other attributes associated with UE categories, but the peak data rates, downlink antenna configuration, and uplink 64QAM support are the categories most commonly referenced.

The emphasis so far has been on the peak data rates but what really matters for the performance of a new system is the improvement that can be achieved in average and cell-edge data rates. The reference configuration against which LTE/ SAE performance targets have been set is defined in 25.913 [5] as being Release 6 UMTS. For the downlink the reference is HSDPA Type 1 (receive diversity but no equalizer or interference cancellation). For the uplink the reference configuration is single transmitter with diversity reception at the Node B. Table 1.4-3 shows the simulated downlink performance of UMTS versus the design targets for LTE. This is taken from the work of 3GPP during the LTE feasibility study [7]. Table 1.4-4 shows a similar set of results for the uplink taken from [8].

Table 1.4-3. Comparison of UMTS Release 6 and LTE downlink performance requirements

Table 1.4-4. Comparison of UMTS Release 6 and LTE uplink performance requirements

From these tables the LTE design targets of 2x to 4x improvement over UMTS Release 6 can be seen. Note, however, that UMTS did not stand still and there were Release 7 and Release 8 UMTS enhancements that significantly narrow the gap between UMTS and LTE. The evolution of UMTS continues through Release 12. Although the figures in Tables 1.4-3 and 1.4-4 are meaningful and user-centric, they were derived from system-level simulations and are not typical of the methods used to specify minimum performance. The simulations involved calculation of throughput by repeatedly dropping ten users randomly into the cell. From this data a distribution of performance was developed and the mean user throughput calculated. The cell-edge throughput was defined as the 5th percentile of the throughput cumulative distribution. For this reason the cell-edge figures are quoted per user assuming 10 users per cell, whereas the mean user throughput is independent of the number of users.

When it comes to defining minimum performance requirements for individual UE, the simulation methods used to derive the figures in Tables 1.4-3 and 1.4-4 cannot be used. Instead, the minimum requirements for UMTS and LTE involve spot measurement of throughput at specific high and low interference conditions, and for additional simplicity, this is done without the use of closed-loop adaptive modulation and coding. This approach to defining performance is pragmatic but it means that there is no direct correlation between the results from the conformance tests and the simulated system performance in Tables 1.4-3 and 1.4-4.

1.5 LTE/SAE Timeline

The timeline of LTE/SAE development is shown in Figure 1.5-1. This includes the work of 3GPP in drafting the specifications as well as the conformance test activities of the Global Certification Forum (GCF) and the trials carried out by the LTE/SAE Trial Initiative (LSTI). The work of GCF towards the certification of UE against the 3GPP conformance specifications is covered in some detail in Section 7.4. The LSTI, whose work was completed in 2011, was an industry forum and complimentary group that worked in parallel with 3GPP and GCF with the intent of accelerating the acceptance and deployment of LTE/SAE as the logical choice of the industry for next-generation networks. The work of LSTI was split into four phases. The first phase was proof of concept of the basic principles of LTE/SAE, using early prototypes not necessarily compliant with the specifications. The second phase was interoperability development testing (IODT), which was a more detailed phase of testing using standards-compliant equipment but not necessarily commercial platforms. The third stage was interoperability testing (IOT), similar in scope to IODT but using platforms intended for commercial deployment. The final phase was Friendly Customer Trials, which ran through 2010. GCF certified the first UE against the 3GPP conformance tests in April 2011. By November 2012 there were 102 FDD and 11 TDD commercial networks launched in 51 countries according to the Global Suppliers Association.

Figure 1.5-1. Projected LTE/SAE timeline

1.6 Introduction to the 3GPP LTE/SAE Specification Documents

The final section in this introductory chapter provides a summary of the LTE/SAE specification documents and where to find them.

1.6.1 Finding 3GPP Documents

A good place to start looking for documents is www.3gpp.org/specifications. From there it is possible to access the specification documents in a number of different ways, including by release number, publication date, or specification number. A comprehensive list of all 3GPP specifications giving the latest versions for all releases can be found at www.3gpp.org/ftp/Specs/html-info/SpecReleaseMatrix.htm. Each document has a version number from which the status of the document can be determined. For instance with 36.101 Vx.y.z, × represents the stability of the document, y the major update, and z an editorial update. If × is 1, then the document is an early draft for information only. If × is 2, then the document has been presented for approval. If × is greater than 2, then the document has been approved and is under change control. Once under change control, the value of × also indicates the release. Therefore a 3 is Release 1999, a 4 is Release 4, a 5 is Release 5, and so on. Most documents in an active release will get updated quarterly, which is indicated by an increment of the y digit. The document will also contain the date when it was approved by the technical specification group (TSG) responsible for drafting it. This date is often one month earlier than the official quarterly publication date.

To avoid confusion, individual documents should be referenced only by the version number. Groups of documents can be usefully referenced by the publication date—e.g., 2008–12—but note that the version numbers of the latest documents for that date will vary depending on how frequently each document has been updated. For example, at 2008–12, most of the physical layer specifications were at version 8.5.0 but most of the radio specifications were at version 8.4.0. It is therefore meaningless to refer to version 8.x.y of the specifications unless only one particular document is being referenced.

The set of specifications valid on any publication date will contain the latest version of every document regardless of whether the document was actually updated since the previous publication date. To access the specifications by publication date, go to ftp://ftp.3gpp.org/specs/. Within each date there will be a list of all the Releases and from there each series of specifications can be accessed. If only the latest documents for a Release are required, go to ftp://ftp.3gpp.org/specs/latest/. Newer, less stable, unpublished documents can often be found at ftp://ftp.3gpp.org/specs/Latest-drafts/, although care must be taken when making use of this type of information.

All versions of the releases of any particular document number can be accessed from ftp://ftp.3gpp.org/specs/archive/. This information can also be obtained from ftp://ftp.3gpp.org/Specs/html-info/, which provides the most comprehensive information. From this link the easiest way to proceed is to select a series of documents; e.g., ftp://ftp.3gpp.org/Specs/html-info/36-series.htm. This location will list all 36-series documents with the document numbers and titles. Selecting a document number will access a page with the full history of the document for all releases, including a named rapporteur and the working group (WG) responsible for drafting the document. At the bottom of the page will be a link to the change request (CR) history, which brings up yet another page listing all the changes made to the document and linked back to the TSG that approved the changes.

By tracing back through the CR history for a document it is possible to access the minutes and temporary documents of the TSG in which the change was finally approved. For instance, tracing back through a CR to 36.101 V8.5.0 (2008–12) would lead to a temporary document of the TSG RAN meeting that approved it stored under ftp://ftp.3gpp.org/tsg_ran/TSG_RAN/TSGR_42/. The change history of a document can also be found in the final annex of the document, but linking to the CR documents themselves has to be done via the website. The lowest level of detail is found by accessing the WG documents from a specific TSG meeting. An example for TSG RAN WG4, who develop the LTE 36.100-series radio specifications, can be found at ftp://ftp.3gpp.org/tsg_ran/WG4_Radio/TSGR4_50/. The link to this WG from the document can also be made from the html-info link given above.

The final way to gain insight into the work of the standards development process is to read the email exploders of the various committees. This capability is hosted by ETSI at http://list.etsi.org/.

1.6.2 LTE/SAE Document Structure

The feasibility study for LTE/SAE took place in Release 7, resulting in several Technical Reports of which [1] and [2] are the most significant.

The LTE RAN specifications are contained in the 36-series of Release 8 and are divided into the following categories:

36.100 series, covering radio specifications and eNB conformance testing

36.200 series, covering layer 1 (physical layer) specifications

36.300 series, covering layer 2 and 3 (air interface signalling) specifications

36.400 series, covering network signaling specifications

36.500 series, covering user equipment conformance testing

36.800 and 36.900 series, which are technical reports containing background information.

The latest versions of the 36 series documents can be found at www.3gpp.org/ftp/Specs/latest/Rel-11/36_series/.

The SAE specifications for the EPC are more scattered than those for the RAN and are found in the 22-series, 23-series, 24-series, 29-series, and 33-series of Release 8, with work happening in parallel in Release 9. A more comprehensive list of relevant EPC documents can be found in Chapter 5.

1.7 References

[1] 3GPP TR 21.905 V11.2.0 (2012-09) Vocabulary for 3GPP Specifications

[2] 3GPP TR 23.882 V8.0.0 (2008-09) 3GPP System Architecture Evolution: Report on Technical Options and Conclusions

[3] 3GPP TS 22.278 V12.1.0 (2012-06) Service requirements for the Evolved Packet System (EPS)

[4] 3GPP TR 25.912 V11.0.0 (2012-09) Feasibility study for evolved Universal Terrestrial Radio Access (UTRA) and evolved Universal Terrestrial Radio Access Network (UTRAN)

[5] 3GPP TR 25.913 V9.0.0 (2009-12) Requirements for Evolved Universal Terrestrial Radio Access (E-UTRA) and Evolved Universal Terrestrial Radio Access Network (E-UTRAN)

[6] 3GPP TR 36.306 V11.1.0 (2012-09) Evolved Universal terrestrial Radio Access Network (E-UTRA); User Equipment (UE) radio access capabilities

[7] 3GPP TSG RAN WG1 Tdoc R1-072578 Summary of downlink performance evaluation, Ericsson, May 2007

[8] 3GPP TSG RAN WG1 Tdoc R1-072261 LTE performance evaluation — uplink summary, Nokia, May 2007

Links to all reference documents can be found at www.agilent.com/find/ltebook.

Chapter 2

Air Interface Concepts

This chapter covers the radio aspects of LTE, starting in Section 2.1 with an overview of the radio frequency (RF) specifications. Sections 2.2 and 2.3 describe the downlink and uplink modulation schemes in some detail and, finally, Section 2.4 examines the way in which LTE uses multi-antenna methods to improve performance.

2.1 Radio Frequency Aspects

The RF specifications for LTE are covered in two 3GPP technical specification documents: 36.101 [1] for the user equipment (UE) and 36.104 [2] for the base station (BS), which is known in LTE as the evolved node B (eNB) although the more generic term BS is more commonly used. One of the first things to note about LTE is the integration between the frequency division duplex (FDD) and time division duplex (TDD) radio access modes. In the previous Universal Mobile Telephone System (UMTS) specifications, which also supported FDD and TDD, the RF specifications for the UE FDD, UE TDD, base station FDD, and base station TDD modes were covered in separate documents. However, the early decision by 3GPP to fully integrate FDD and TDD modes for LTE has resulted in only one RF specification document each for the UE and the BS. With the higher level of integration between the two modes, the effort required to support them should be less than it was in the past.

The structure of 36.101 [1] for the UE follows the UMTS pattern of minimum requirements for the transmitter and receiver followed by performance requirements for the receiver under fading channel conditions. The final section covers the performance of the channel quality feedback mechanisms. The structure of 36.104 [2] for the BS follows the same pattern as UMTS with transmitter, receiver, and performance requirements.

The purpose of this section is to highlight those aspects of the LTE RF requirements that will be new compared to UMTS. These include issues relating to LTE’s support of multiple bands and channel bandwidths as well as those RF specifications peculiar to the use of orthogonal frequency division multiple access (OFDMA) modulation on the downlink and single-carrier frequency division multiple access (SC-FDMA) on the uplink. The RF performance aspects will be covered in Sections 6.5, 6.7, and 7.2.

The first edition of this book was based on the December 2008 Release 8 specifications. Since then there have been substantial additions. In particular, the September 2012 version of 36.101 [1] for Release 11 has more than tripled in length from the Release 8 version in December 2008 as more than 450 change requests have been approved since then. Much of this additional content is related to the addition of 15 new frequency bands, which brings the total to 40 (28 for FDD and 12 for TDD).

The other major contributor to the increase in length is the introduction in Release 10 of carrier aggregation (CA), uplink multiple input multiple output (UL-MIMO), and enhanced downlink MIMO (eDL-MIMO). The background to these developments is described in Technical Report 36.807 [3], which like other 800 series reports is not published. However, it does contain a wealth of technical background information used to develop the specifications.

The incorporation of CA, UL-MIMO, and eDL-MIMO into 36.101 [1] is quite daunting and so a notation system has been derived whereby the new subclauses specific to the new features are named as follows:

Suffix A, additional requirements need to support CA

Suffix B, additional requirements need to support UL-MIMO

Suffix D, additional requirements need to support eDL-MIMO.

The C suffix is reserved for future use. This naming convention makes it easier to determine which subclauses in Sections 5, 6, and 7 of 36.101 [1] apply to UE supporting the new features.

2.1.1 Frequency Bands

Table 2.1-1 shows the IMT-2000 (3G) frequency bands defined by the European Telecommunications Standards Institute (ETSI) and 3GPP. Most of the frequency bands are defined in 36.101 [1] Table 5.5-1, meaning they are recognized by all three International Telecommunications Union (ITU) regions, although it should be noted that definition of a band does not imply its availability for deployment. The exceptions in Table 2.1-1 that are not defined in 36.101 [1] are Bands 15 and 16. These have been defined by ETSI in TS 102 735 [4] for ITU Region 1 (Europe, Middle East, and Africa) only. These bands have not been adopted at this time by ITU Region 2 (Americas) or Region 3 (Asia), which is why they do not appear in 36.101 [1]. Figure 2.1-1 shows how the terms bandwidth, duplex spacing, and gap are used for FDD in Table 2.1-1. The concepts of gap and duplex spacing don’t exist for TDD.

Figure 2.1-1. Explanation of frequency band terms

Table 2.1-1. Defined frequency bands for IMT-2000 (MHz)

Table 2.1-1 shows the large number of options that exist for IMT technologies, which now include LTE. When UMTS was first specified in 1999, only one frequency band was defined. This band became a point around which the industry could focus its efforts in developing specifications and products. In the years since then, bands have been gradually added, and when LTE was specified in 2008, it inherited all the existing UMTS bands plus some new ones added in Release 8. Moreover, with the integration of TDD into the LTE specifications, another eight bands were added to the list.

It is clear from Table 2.1-1 that many of the bands are overlapping or subsets of one another, so the actual RF coverage may not seem to present a problem for power amplifiers and receivers. Where the difficulty lies, however, is in handling the many combinations of filtering that are required to implement the different bands. The bandwidth, duplex spacing, and gap are not constant, which adds to the challenge of designing the specific band filters required for each implemented band. There are also issues in designing efficient antennas to cover the wide range of possible supported bands.

The possibility of variable duplex spacing is not precluded but as of Release 11 has not been developed. For a specified FDD band, variable duplex spacing would mean that the currently fixed relationship between the uplink and downlink channels could become variable. This would increase deployment flexibility but also increase the complexity of the specifications, the equipment design, and network operation.

2.1.2 Channel Bandwidths

A trend in recent years has been for radio systems to be ported to new frequency bands, although typically these systems support only one channel bandwidth. The first release of UMTS, which supported both FDD and TDD modes, used a common CDMA chip rate of 3.84 Mcps and a channel spacing of 5 MHz. Release 4 of UMTS introduced a low chip rate (LCR) TDD option (also known as TD-SCDMA) that used the lower 1.28 Mcps with a correspondingly narrower channel spacing of 1.6 MHz. This was followed in Release 7 by the 7.68 Mcps option with its 10 MHz channel spacing. In Release 8 a dual-carrier version of HSDPA was introduced; however, the wider bandwidth comes from two separate 5 MHz channels. Four-carrier HSDPA was introduced in Release 10 and eight-carrier in Release 11.

The situation for LTE is very different. The OFDMA modulation scheme upon which UMTS is based has as one of its properties the ability to scale its channel bandwidth linearly without changing the underlying properties of the physical layer—these being the subcarrier spacing and the symbol length. The details of the LTE modulation schemes are discussed fully in Sections 2.2 and 2.3. It is sufficient to say at this point that LTE was designed from the start to support six different channel bandwidths. These are 1.4 MHz, 3 MHz, 5 MHz, 10 MHz, 15 MHz, and 20 MHz. Earlier versions of the specifications also supported 1.6 MHz and 3.2 MHz for interworking with LCR TDD, but these were removed when the LTE TDD frame structure was aligned with the FDD frame structure rather than the TD-SCDMA frame structure from UMTS.

The choice of many channel bandwidths means that LTE has more deployment flexibility than previous systems. The wide channel bandwidths of 10, 15, and 20 MHz are intended for new spectrum, with the 2.6 GHz and 3.5 GHz bands in mind. These wider channels offer more efficient scheduling opportunities, which can increase overall system performance. With the potential of having a much wider channel available, individual users might perceive that they have a high bandwidth connection when in fact they are sharing the bandwidth with many other users. An individual user’s perception of owning the channel comes from the fact that the demand typically is variable and what matters is the peak rate available at the time of the demand. This perception is known as the trunking effect, wherein the wider the channel, the greater the gains. Narrowband systems such as GSM with a channel bandwidth of only 200 kHz are not in a position to instantaneously offer more capacity, even if other users are not making full use of their channel.

The other benefit of a wider channel is the possibility of scheduling users as a function of the channel conditions specific to them. This topic is discussed in more detail in Section 3.4, but the essence is that OFDMA has the ability to schedule traffic over a subset of the channel and thus, with appropriate feedback of the instantaneous channel conditions, can target transmissions at frequencies exhibiting the best propagation conditions and lowest interference.

Table 2.1-2. Combinations of channel bandwidth and frequency band for which RF requirements are defined (36.101 [1] Table 5.6.1-1)

The 5 MHz channel bandwidth option for LTE is an obvious choice for re-farming of existing UMTS spectrum. This re-farming will not benefit from trunking gains over UMTS but still has the possibility of gains through frequency-selective scheduling. The 1.4 MHz and 3 MHz options are targeted at re-farming of narrowband systems such as GSM and cdma2000®. Even the 1.4 MHz option will have significant trunking gains over 200 kHz GSM as well as the ability to do some frequency-selective scheduling. The consequence of a system that has so much flexibility in terms of frequency bands and channel bandwidths is the complexity that is created. Several of the LTE RF requirements described in this section reflect this growth in complexity: requirements that in UMTS were expressed as single-valued figures are now represented by multi-dimensional tables.

Although the LTE system could be operated in any of the defined bands at any channel bandwidth, certain combinations are not expected in real deployment, and for such cases no RF performance requirements are defined. Table 2.1-2 shows the combination of channel bandwidths for which performance requirements exist (or do not exist) for the different frequency bands. Table 2.1-2 shows, for example, that no requirements are defined for the 1.4 MHz and 3 MHz bandwidths for several bands including E-UTRA Band 1 (the primary UMTS operating band at 2.1 GHz) as these deployment combinations are not likely. The table also shows that for some combinations there are relaxations in the requirements. For example, there are several bands for which the receiver sensitivity requirements are relaxed when the system operates at 15 MHz and 20 MHz channel bandwidths. At the time of this writing these relaxations are limited to reference sensitivity although the list of affected requirements may grow over time.

2.1.3 Reference Measurement Channels

Before describing the UE and BS RF requirements it is useful to introduce the concept of reference measurement channels (RMCs). These exist for both the downlink and uplink and are used throughout the RF specifications to precisely describe the configuration of signals used to test the UE and BS transmitters and also their receivers.

2.1.3.1 Uplink Reference Measurement Channels

The flexible nature of the uplink transmissions makes it important that the signal definition be explicit when performance targets are specified. As a result, many of the UE transmitter requirements in 36.101 [1] Subclause 6 and some of the receiver requirements in Subclause 7 are defined relative to specific uplink configurations. These are known as uplink RMCs. A similar principle was used in UMTS and the main difference for LTE is the use of SC-FDMA rather than W-CDMA for the air interface.

Since the uplink RMCs are primarily used for testing UE transmitter performance, many of the variables that will be used in real operation are disabled. These include no incremental redundancy (1 HARQ transmission), normal cyclic prefix only, no physical uplink shared channel (PUSCH) hopping, no link adaptation, and for partial allocation the resource blocks (RBs) are contiguous starting at the channel edge. Table 2.1-3 shows an example quadrature phase shift keying (QPSK) uplink reference measurement channel (RMC) for various allocations of up to 75 RBs (75%).

Table 2.1-3. Reference channels for 20 MHz QPSK with partial RB allocation (36.101 [1] Table A.2.2.2.1-6b)

For the purposes of testing the BS receiver, further uplink RMCs are defined in 36.104 [2] Annex A. These are referred to as fixed reference channels (FRCs).

2.1.3.2 Downlink Reference Measurement Channels

An example of a single antenna downlink RMC for use with 64QAM PUSCH and common (cell-specific rather than UE-specific) demodulation reference symbols (DMRS) is given in Table 2.1-4. This RMC will be used for performance testing under faded channel conditions.

Table 2.1-4. Fixed reference channel 64QAM R = 3/4 (36.101 [1] Table A.3.3.1-3)

It can be seen from Table 2.1-4 that these RMCs are for a fully allocated downlink, and the maximum throughput represented reaches a maximum of 61.7 Mbps for the 20 MHz channel bandwidth case. Note that this peak figure represents the maximum transmitted data rate and is in no way intended to indicate the performance of the downlink in real radio conditions. This peak figure is the reference used for specifying the expected performance, which will be specified relative to the maximum figures.

2.1.4 Transmit Power

This section covers UE and BS transmit power requirements.

2.1.4.1 UE Transmit Power

Over time the transmit power requirements for the UE have become more complex. In earlier standards such as GSM and UMTS Release 99, the transmit power specification was a simple one-to-one relationship between the power class of the UE and the maximum output power. This relationship has become more complicated over time with relaxations that take into account the crest factor of higher-order modulation formats. The trend to specify power back-off started in Release 5 for UMTS with the introduction of a fixed back-off for 16QAM. In Release 6 the fixed back-off was superseded by a more advanced cubic metric that related the allowed back-off to a formula that included the cube of the voltage waveform relative to a standard QPSK waveform.

There are four power classes defined for the LTE UE. At the time of this writing, a maximum power requirement is defined only for Class 3 and is specified as 23 dBm ±2 dB for almost all bands. However, the flexibility of the LTE air interface requires consideration of additional dimensions including the channel bandwidth and the size of the power allocation within that bandwidth. The introduction of carrier aggregation and uplink MIMO in Release 10 further complicates the specifications, which now require 15 pages to define all the exceptions to the default UE maximum output power of 23 dBm.

Table 2.1-5 shows the maximum power reduction (MPR) that applies for Power Class 3 depending on the modulation being used and the number of resource blocks transmitted in each channel bandwidth. An RB is the minimum unit of transmission and is 180 kHz wide and 0.5 ms in duration.

Table 2.1-5. Maximum power reduction for power class 3 (36.101 [1] Table 6.2.3-1)

The trend shown in Table 2.1-5 is that with increasing modulation depth (meaning higher signal peaks) and increasing transmitted bandwidth, the maximum power is reduced.

The introduction of carrier aggregation (see Section 2.1.11) has resulted in further MPR allowances as shown in Table 2.1-6.

Table 2.1-6. Maximum power reduction for Power Class 3 (36.101 [1] Table 6.2.3A-1)

In addition to the maximum power reductions, which are specified for all operating conditions, there is another class of dynamic MPR known as additional MPR (A-MPR), which comes into play when the network signals the UE. Fourteen different network signaling values have been defined, as shown in Table 2.1-7. The behavior of the UE depends on which band it is using, which channel bandwidth, the number of resource blocks allocated, the modulation depth, the allowed A-MPR, and the specific spurious emission requirements that have to be met under these conditions

Table 2.1-7. A-MPR/spectrum emission requirements (36.101 [1] Table 6.2.4-1)

For example, a UE receiving NS_03 from the network when operating in bands 2, 4, 10, 23, 25, 35, or 36 with a 15 MHz channel bandwidth and > 8 RBs allocated is allowed to reduce its maximum power by up to an additional 1 dB above the normally allowed MPR in order to meet the additional spurious emission requirements identified in 36.101 [1] Subclause 6.6.2.2.1. Table 2.1-8 identifies the additional spurious emissions and spectrum emission mask (SEM) requirements for which the network might signal the UE.

Table 2.1-8. Additional spectrum emission requirements for NS_03 (36.101 [1] Table 6.6.2.2.1-1)

Table 2.1-8 shows another consequence of LTE’s channel bandwidth flexibility: the additional SEM requirements, which are a function not just of the frequency offset as in UMTS but also of the channel bandwidth. Table 2.1-9 is an example of one of the most complex A-MPR definitions, in this case for NS_12. The allowed power reduction is a function of channel bandwidth and the position and size of the RB allocation within the channel bandwidth.

Table 2.1-9. A-MPR for NS_12 (36.101 [1] Table 6.2.4-6)

It should be evident at this point how complex the rules are which govern the maximum power that the UE can use under different conditions and the requirements, both static and dynamic, that have to be met. Checking for correct behavior under all possible conditions will be a substantial verification exercise.

Accuracy for Maximum Output Power

Given the complexity of the maximum power specifications, several new terms have been defined. The number of terms and their definitions have evolved from Release 8 to their current form in Release 11.

PCMAX is defined in 36.101 [1] as the configured maximum UE output power. This is the nominal power the UE chooses to set as its maximum power based on all the requirements and applicable relaxations. The UE is allowed to set PCMAX between a lower limit PCMAX_L and an upper limit PCMAX_H such that

The definitions of PCMAX_L and PCMAX_H are affected by a number of parameters:

where

and

where ΔTC = 1.5 dB when an allowance applies for transmission bandwidths within 4 MHz of the band edge (see Note 2 in 36.101 [1] Table 6.2.2-1) and ΔTC = 0 dB when the band-edge allowance is not applied.

P-MPR is an additional allowance that can be applied to meet applicable electromagnetic energy absorption requirements and to address unwanted emissions and self-desense requirements in case of simultaneous transmissions on multiple radio access technologies (RATs) for scenarios not within scope of 3GPP RAN specifications. P-MPR may also be used in conjunction with proximity detection to meet electromagnetic compatibility (EMC) requirements. For cable-conducted testing P-MPR is set to 0 dB and is not considered further here. P-MPR was introduced in the PCMAX equation so that the UE can report to the BS the available maximum output transmit power, which may be less than otherwise expected due to EMC reasons. Lower available power might impact uplink performance and needs to be considered by the BS for scheduling decisions.

Having defined the range PCMAX_L to PCMAX_H within which the UE must set PCMAX, the specifications now establish the requirements for how accurate the actual maximum output power needs to be. When the UE is configured for its chosen PCMAX, which is a nominal or target power, the power that is actually transmitted is defined in 36.101 [1] as PUMAX. This is the measured configured maximum UE output power—that is to say, the power that is actually transmitted at the antenna connector (see Section 6.4.3.1) as would be measured assuming no measurement uncertainty.

The limits on PUMAX are defined by extending the range PCMAX_L to PCMAX_H by a tolerance which varies as a function of PCMAX. The tolerance is denoted as T(PCMAX) and is given in Table 2.1-10.

Table 2.1-10. PCMAX tolerance (36.101 [1] Table 6.2.5-1)

The tolerance is evaluated for PCMAX_L and PCMAX_H independently. From these tolerances, the limits on PUMAX are defined as the following:

The considerable complexity in defining PUMAX does not include the additional complexity that will be applied when the test system uncertainties are taken into account.

2.1.4.2 Base Station Transmit Power

There are several parameters used to describe the BS output power:

Pout—the mean power of one carrier

Pmax—the maximum total output power for the sum of all carriers

Pmax,c—the maximum output power per carrier

Rated total output power—the manufacturer-declared available output power for the sum of all carriers

PRAT—the manufacturer-declared available output power per carrier.

Unlike the UE case, there are no requirements for BS maximum output power, either for all carriers or per carrier. The only requirements relate to the accuracy with which the manufacturer declares PRAT. Different PRAT limits can be applied to different BS configurations. Three different BS classes are defined based on their PRAT as given in Table 2.1-11. (Note that the term home BS in this case is equivalent to home eNB or HeNB.)

Table 2.1-11. Base station rated output power (based on 36.104 [2] Table 6.2-1)

Regional requirements can sometimes override the 3GPP specifications; for instance, Band 34 in Japan is limited to 60 W for the 20 MHz channel bandwidth, whereas no upper limit is defined for other regions.

There are restrictions on the home BS Pout for certain deployment scenarios. For protection of an adjacent UTRA (UMTS) deployment, Pout is limited to between 8 dBm and 20 dBm as a function of the received level of the adjacent UTRA common pilot indicator channel (CPICH) and Ioh, which is defined as the total power present at the home BS antenna connector on the home BS downlink operating channel excluding the power generated by the home BS on that channel. For protection of an adjacent LTE deployment there are similar Pout restrictions, but these are based on Ioh and the received level of the cell reference signals (CRS) from the adjacent LTE channel. In order to meet the requirements on Pout the home BS ideally needs to have the ability to measure the adjacent channel signal levels although the requirement to control Pout does not mandate how the downlink power measurement is achieved. Adjacent channel measurement is not a usual capability of a BS since frequency planning is used to control adjacent channel interference.

Restrictions on the home BS apply when the adjacent channel that needs to be protected belongs to a different operator. If the adjacent channel belongs to the same operator, then the issue of interference mitigation is left to that single operator to work out.

The most stringent interference requirement for the home BS applies in the co-channel case in which an operator chooses to deploy the home BS on the same channel as the macro network using a closed subscriber group (CSG). In a CSG, macro users are not allowed to use the home BS. The Pout restrictions for this scenario are a complex function of CRS, Ioh, and Iob. This last parameter is the uplink received interference power, including thermal noise, present at the home BS antenna connector on the operating channel.

2.1.5 Output Power Dynamics

This section covers UE and BS output power dynamics.

2.1.5.1 UE Output Power Dynamics and Power Control

The UE output power dynamics cover the following areas:

Minimum output power

Off power

Power time mask

Output power control (accuracy).

2.1.5.1.1 Minimum Output Power

The UE minimum output power is defined as −40 dBm for all channel bandwidths. This is the lowest power at which the UE is required to control the power level and meet all the transmit signal quality requirements. In UMTS, the transmit signal quality requirements apply only from maximum power down to −20 dBm for QPSK and −30 dBm for 16QAM, which is well above the −50 dBm UMTS minimum power requirement. The LTE requirement that signal quality not be degraded over the full operating range puts more demands on the fidelity of the digital-to-analog convertors of the transmitter than was the case with UMTS.

2.1.5.1.2 Off Power

When commanded to switch off its transmitter, the UE output power must be less than −50 dBm. This applies for all channel bandwidths.

2.1.5.1.3 Power Time Mask

The on/off requirements for slot-based transmissions are similar to UMTS. Figure 2.1-2 shows the profile for the general on/off time mask.

Figure 2.1-2. General ON/OFF Time Mask (36.101 [1] Figure 6.3.4.1-1)

The general requirement is used any time the signal turns on or off. The measurement duration is at least one subframe excluding any transient period. Note that the transient period is not symmetrical with the subframe boundary; the on power ramp transient period starts after the subframe boundary but the off power ramp starts after the end of the subframe. There is no ideal position for the transient period in an FDD system in which no gaps are defined, and the solution specified is a compromise. The choice made for FDD is to minimize any interference to adjacent subframes during the ramp up but allow the ramp down to be delayed until after the end of the subframe.

There is a similar mask for the physical random access channel (PRACH) shown in Figure 2.1-3, but the on period is one PRACH symbol and the on ramp is shifted to before the symbol starts, making it symmetrical with the off ramp.

Figure 2.1-3. PRACH ON/OFF time mask (36.101 [1] Figure 6.3.4.2-1)

The time mask for the sounding reference signal (SRS) is similar although a special case for TDD is shown in Figure 2.1-4 for the dual SRS transmission in the uplink pilot timeslot (UpPTS).

Figure 2.1-4. Dual SRS time mask for the case of UpPTS transmissions (36.101 [1] Figure 6.3.4.2.2-2)

The shift in the start of the on power requirement from the general requirement is particularly important for the SRS since the SRS symbol can be transmitted in isolation for approximately 70 μs and is used by the BS to estimate the uplink channel conditions. If this symbol were not stable for its nominal duration, the BS might get an incorrect estimate of the uplink channel. A symmetrical mask was not chosen for the general on/off case since there are times when it is important to protect the symbol just prior to a power change.

Another example of SRS protection is shown in Figure 2.1-5. This shows the time mask for FDD SRS blanking, which occurs when the UE is required to blank its output during the SRS symbol period.

Figure 2.1-5. SRS time mask when there is FDD SRS blanking (36.101 Figure 6.3.4.4-4)

Apart from the traditional on and off transitions described thus far, other transitions from one power state to another are sometimes necessary. These include changes to the transmit power and transitions into and out of the SRS. In addition, a change to the allocated frequency can trigger a power transient due to baseband compensation for known unflatness in the transmission path.

A common example of a frequency-induced power transient occurs when the UE is transmitting the physical uplink control channel (PUCCH). The PUCCH generally transmits one timeslot at the lower end of the channel followed by another timeslot at the upper end. See the measurement example in Section 6.4.6.7. The PUCCH frequency hopping is shown graphically in Figure 3.2-13. The requirements for spectrum flatness in 36.101 [1] Subclause 6.5.2.4 specify that at the band edge (and for wide channels in narrow bands most of the channels are at the band edge), the UE is allowed to have a variation in power across the channel of +3 to −5 dB. This variation could be a slope of some 8 dB. In extreme conditions the allowance rises to 12 dB. When the UE transmits the PUCCH or narrow allocations of the PUSCH, it may be necessary to compensate for known flatness issues. This then creates the possibility of a power transient at baseband and RF even though the nominal power remains constant.

The requirements for maintaining PUSCH/PUCCH power accuracy apply to the second and subsequent subframes after the start of a contiguous block of subframes. This requirement also applies to non-contiguous transmissions provided the gap is less than or equal to 20 ms (two frames). There are also requirements for the PRACH that apply to the second and subsequent PRACH preambles.

2.1.5.1.4 Output Power Control

The requirements on UE output power accuracy are not particularly onerous and are shown in Table 2.1-12

Table 2.1-12. Relative power tolerance for transmission (normal conditions) (36.101 [1] Table 6.3.5.2.1-1)

From Table 2.1-12 it can be seen that even for no change to the nominal power, the relative power can vary by up to ±2.5 dB. This makes allowance for the case in which transmissions are contiguous in time but not in frequency. For changes to the configured power the allowance increases up to a maximum of ±6 dB for steps between 15 dB and 20 dB.

There is a growing trend within UE design to reduce cost through the use of multi-stage power amplifiers. These reduce the dynamic range that has to be covered in one section but also introduce the possibility of power and phase transients at the power level where the switching between gain stages takes place. This is a known issue being studied in the specifications, particularly for UL-MIMO, and it is likely that there will be requirements developed to allow for gain-stage switching with appropriate limits on hysteresis to avoid unnecessary transients.

2.1.5.2 Base Station Output Power Dynamics

The BS output power dynamics cover the following areas:

Resource element power control dynamic range

Total power dynamic range

Transmitter off power and transient period.

2.1.5.2.1 Resource Element Power Control Dynamic Range

This requirement, which is shown in Table 2.1-13, specifies the range over which the BS is required to control the output power of a single resource element (RE) relative to the average RE power.

Table 2.1-13. E-UTRA BS RE power control dynamic range (36.104 [2] Table 6.3.1.1-1)

Note 1: The output power per carrier shall always be less or equal to the maximum output power of the base station.

This is a necessary capability required to implement power boosting or de-boosting of particular parts of the downlink signal.

2.1.5.2.2 Total Power Dynamic Range

This requirement, shown in Table 2.1-14, is the minimum required total power dynamic range between a fully allocated signal at maximum power and a signal with only one RB allocated. The required dynamic range is 10 log (NDLRB). See Table 3.2-7 for the number of RBs per channel.

Table 2.1-14. E-UTRA BS total power dynamic range (36.104 [2] Table 6.3.2.1-1)

2.1.5.2.3 Transmitter Off Power and Transient Period

The off power is defined as a maximum power spectral density measured over a period of 70 μs in a square filter equal to the transmission bandwidth configuration. The off power is required to be less than −85 dBm/MHz.

There are no equivalents to the UE power time mask; however, for TDD operation the concept of transmitter transient period is defined. The transient period occurs twice during a TDD frame, first at the off-to-on transition from the uplink subframe to the downlink subframe and second at the on-to-off transition from the downlink subframe to the guard period or uplink pilot timeslot. In both cases the transient period is defined to be 17 μs.

2.1.6 Transmit Signal Quality

The transmit signal quality requirements specify the in-channel characteristics of the wanted signal. These are distinct from the out-of-channel requirements, which specify limits on unwanted emissions outside the wanted channel. This section covers transmit signal quality for the UE and BS.

2.1.6.1 UE Transmit Signal Quality

Subclause 6.5 of 36.101 [1] defines the in-channel signal quality requirements. These are split into five categories: frequency error, error vector magnitude (EVM), carrier leakage (IQ component), in-channel emissions, and EVM equalizer spectrum flatness. These measurements are fully defined in 36.101 [1] Annex F. Figure 2.1-6 shows the block diagram of the measurement points.

Figure 2.1-6. EVM measurement points (36.101 [1] Figure F.1-1)

The frequency error component is a residual result from measuring EVM and identical in concept to UMTS, so will not be discussed further here.

2.1.6.1.1 Error Vector Magnitude Definition

The EVM definition for LTE is similar in concept to UMTS but there are two new elements specific to SC-FDMA that need to be explained. The first relates to the presence of the TX-RX chain equalizer block from Figure 2.1-2 and the second relates to the time window over which EVM is defined.

The EVM requirements are shown in Table 2.1-15.

Table 2.1-15. Minimum requirements for error vector magnitude (36.101 [1] Table 6.5.2.1.1-1)

The requirements apply for UE output power ≥ −40 dBm. The measurement period is one timeslot for PUSCH and PUCCH and one preamble sequence for the PRACH. If an SRS symbol is inserted into the frame, the measurement interval is shortened accordingly. The measurement period is also shortened when the mean power between slots, the modulation depth, or the allocation between slots is expected to change. For the PUSCH, the reduced measurement period is the on period defined for the power time mask less a further 5 μs. The exclusion period is applied after the inverse discrete Fourier transform (IDFT) shown in Figure 2.1-6. For the PUCCH, the measurement interval is reduced by one symbol adjacent to the boundary where the power change is expected to occur.

The stated intention in early Release 8 to provide requirements for 64QAM has been dropped and no uplink 64QAM requirements have been developed.

EVM Equalizer Definition

In UMTS, the EVM measurement was defined through a root raised cosine (RRC) filter matched to the filter defined for UE transmissions. This same filter was assumed in the UMTS BS and was required in order to optimize the received signal quality. In LTE no such transmit filter is defined, which opens up a significant new challenge in determining how to specify transmitter performance. In real-life operation the BS will attempt to determine the amplitude and phase characteristics of the transmitted signal as seen through the imperfect radio channel. It is essential for accurate demodulation that this equalization process take place, but the LTE specifications for the BS do not define the method or a reference receiver. This has partly to do with the complexity of the problem, which is a function of noise and dynamic channel conditions. As a result the equalization process is considered proprietary to