Francis J. Smith CTO Finesse Wireless Inc.

Transcription

1 Impact of the Interference from Intermodulation Products on the Load Factor and Capacity of Cellular CDMA2000 and WCDMA Systems & Mitigation with Interference Suppression White Paper Francis J. Smith CTO ABSTRACT Intermodulation products that fall inband of the receive signal of a CDMA cellular handset cause the handset to request more power from the base station to maintain link closure. This has the effect of raising the noise floor in the cell/sector and results in the single user consuming additional system resources with no increase in throughput. At some point, the percentage of the system capacity consumed by one user becomes unacceptable and the user s service is dropped. has developed patent pending technology that provides for up to of suppression of inband intermodulation products and thus recovers a significant portion of the lost system capacity and improves the sensitivity of the cell phone db for db down to the noise floor of the receiver. This paper describes the interference environment, the impact to system capacity and receiver sensitivity, and the mitigation of the impact of intermodulation products. The simulated results of the interference are provided herein. Voice services can be significantly degraded and still provide some utility. When data packets are lost, retransmission will consume even more of the system capacity. 1.0 Introduction: This paper presents an approach for the analysis of the impact of intermodulation products on the load factor and capacity of cellular CDMA systems and the mitigation of that impact. Capacity calculations for cellular CDMA Systems are a complex topic and not one that lends itself easily to closed form solutions. In most capacity models, the environment is assumed to consist of AWGN and cell-to-cell interference and statistical methods are often employed. This paper addresses the impact from intermodulation products and AWGN. Nominal values for cell-to-cell interference, orthogonality and activity factors are also included in the analysis. [Reference 1] Cellular mobile units (or handsets) must provide for the reception of an entire allocated telephony band because the handset does not have a priori knowledge of the channels to which it may be assigned. The allocated telephony bands are typically from 40 to 60 MHz. There will be multiple service providers, with transmitting base stations, within the allocated spectrum; and these signals can cause intermodulation products in the RF front end of the handset receiver. In the analysis presented herein, a single cell sector is analyzed in the presence of AWGN and intermodulation products, produced by other service provider base stations within the allocated telephony bands. This model incorporates average impact from other cells, but not the impact that the cell under analysis (own cell) will have on other cells as the own cell base station transmitter increases the transmission power to overcome the interference from intermodulation products (IMP) experienced in the handset. The impact to cell capacity from IMPs can be significant, but there will be additional capacity degradation in the adjacent cells as they compensate for the increased power transmission in the own cell. In this analysis, the impact on the load factor is computed for a single user, as a function of tones producing intermodulation products with power levels of the tones ranging from dbm to -21dBm. As will be shown, in CDMA systems with two way power control, when a single handset is subjected to high power intermodulation products (within the specifications of CDMA2000), the single handset can consume a very high percentage of the cell/sector available traffic capacity. This will have significant impact on the traffic capacity of the cell or the user will experience a dropped call. The analysis presented herein is a conservative assessment of the impact on the capacity because the cell to cell interference will create an additional impact to the capacity of the cluster of cells as one base station transmitter increases the transmitter power to overcome the interference seen by the impacted handset.

2 Practical CDMA systems typically operate with maximum load factors (LF) (described below) between.5 and.7. In this analysis, the load factor of a single user is computed and then used to determine the percentage of the cell traffic capacity that is used by the single user based on a maximum LF of.5 and.7. The load factor is the percentage of the theoretical maximum capacity of the cell that is being used. As an example, when a WCDMA service is a 64Kb/s service and the handset experiences high levels of IMP interference, the load factor of the cell/sector can be driven by only one user for one 64 kb/s service, thus reducing the cell/sector capacity to almost zero if the impacted user is not dropped. In most cases, the impacted user will be dropped. In the analysis presented herein, the impact to the cell/sector load factor is computed beyond where the user would most likely be dropped, but from the curves presented, that point can be selected. Five cases are presents herein, one for CDMA2000 voice, two for 3G WCDMA, one voice and one data. An additional two cases are considered when WCDMA is designed for IMPs from tones no higher than -49dBm wherein the design does not support an LNA bypass mode as in CDMA2000. There are an almost unlimited number of combinations of data and voice services that may exist simultaneously in 3G WCDMA. For simplification, one data and one voice case are presented (along with the No LNA Bypass Mode ). While 3G WCDMA is not required by standards specifications to handle two-tone tests as high as that of CDMA2000, WCDMA will, in reality, be subjected to the same level of interference in the field, but it will be more sensitive to intermodulation products due to the wider bandwidth of the signal. If WCDMA handsets are not designed to handle the greater interference, the impact on capacity is even greater. Within the model, many parameters can be varied such as the NF and IIP3 of the handsets. This model assumes one IMP inband of the SOI, but in reality, there can be multiple interfering IMPs. Note that the Finesse Wireless technology can remove multiple IMPs simultaneously. 2. The Intermodulation Interference Problem Intermodulation products (IMP) are produced in the nonlinear elements of the handset receiver and directly impact the sensitivity of the receiver. As shown in Figure 1 for the PCS band of North America, high power base stations of other service providers transmit high power signals in the same 60 MHz receive band, which the cell phone must accommodate at the receiver RF front end. These transmitters are nominally on the order of 45dBm, but can be as high as 50 to 60dBm. The location of the other service providers within the 60 MHz channel does not impact the magnitude of the 3 rd order IMP as long as the relative frequency spacing will result in an IMP inband of the signal of interest (SOI). As shown in figure 1, if X = Y, then the 3 rd order IMP will fall inband. The source signals for the 3 rd order IMPs can be close in to the SOI or anywhere in the allocated receive band. The specifications for CDMA2000 require a two tone test with the tones at -21dBm with a SOI sensitivity of -94dBm. When the tones are absent, the SOI sensitivity requirement is -119dBm. The CDMA system gives up 25dB in sensitivity when the high power tones (typically AMPS signals) are present. As will be shown below, the power control adjustment required by the base station to support the handset being jammed by the 3 rd order IMP will significantly impact the traffic capacity of the cell/sector. The source signals which generate the 3 rd order IMPs may not be AMPS signals, but can be wider band signals such WCDMA, GSM, EDGE and or TDMA IS-136 or similar telephony standards. If there are companion signals with the appropriate frequency spacing and power levels, the power in the inband IMP will be the same as for AMPS. The power in the IMP will be 2 of 22

3 2*S1(dBm) + S2(dBm)- 2*IIP3 where S1(dBm) is the power in the first source signal and S2(dBm) is the power in the second source signal and IIP3 is the system IIP3 of the handset (the cascaded IIP3 of all components). If the source signals produce a wideband IMP, then only that portion of the energy that falls inband of the signal of interest will interfere with the SOI. For this paper, the source signals will be assumed to be tones to simplify the concepts presented, but as can be shown, the inband energy is the same. The Finesse Technology works for any band width IMP, even if it is wider than the SOI. GSM and 3G WCDMA: The interference requirements for GSM and WCDMA are shown in figure 2. 3 Rd Order Intermodulation Product Interference Signal Generation Home Basestation 3 rd Order Mixing Produces Inband Interference Other Service Provider Basestation Other Service Provider Basestation G 1850 MHz Transmit Band 3 G WCDMA Signal of Interest PCS Band Band Class 1-94 dbm 1910 MHz CDMA IS-95, IS-98 CDMA X, 2X and 3X 1.25 MHz 2 x = 100KHz IF X = Y Intermod is inband Transmit Receive Guardband 20 MHz Signal of Interest 1930 MHz 10 MHz 900 KHz 800 KHz 1700 KHz 10 MHz Receive Band -21 dbm -21 dbm TDMA or AMPS -21 dbm -21 dbm TDMA or AMPS Figure 1. Cellular Telephony Interference Environment X 3 of 22 Y AMPS, TDMA and CDMA exist in Band Class 0 TDMA, GSM and CDMA exist in Band Class 1 GSM GPRS EDGE 1990 MHz Note: First jammer is 900KHz offset for Band Class o and offset 1270 for Band class 1 GSM GPRS EDGE 3 rd order IMP Interference can be Narrowband or Wideband WCDMA 3G is three times as vulnerable to multiple IMPs because it is 3 times the width of CDMA 2000 The two tone, 3 rd order IMP, tests for GSM and WCDMA only require the systems to be designed for tones up to -49dBm, and do not account for the very close in blocker at dbm. As can be seen in figure 2, there are specified blockers as high as -26dBm specified within the 60 MHz PCS receive band for North America. In reality, as shown in Figure 1, WCDMA will be exposed to the same high power interfering signals as specified for CDMA2000. The GSM and WCDMA specifications assume that the handset will be on the order of at least 250 meters from the offending base stations. As can clearly be demonstrated by driving down most freeways in North America, this is not always the case and multiple transmitters are often less than 20 to 30 meters off the roadway. There are often numerous service providers on the same tower and multiple towers close together. This is to be expected, because if the location is

4 good for one provider, it is good for all. Real estate for placing cellular base stations is at a premium. If the free space loss calculations are run, the 250 meters explains the difference between the CDMA2000 two tone specification of -21dBm and that of GSM and WCDMA at - 49dBm. In reality, all signals will experience the same level of interference. While the impact of 3 rd order interfering signals on GSM is not explicitly covered in the scope of this paper, it is interesting to note if a GSM handset has a system IIP3 of +10dBm (probably optimistic) and the tones are at -21dBm, the inband 3 rd order IMP will be at -83dBm, or 20 + db above the GSM sensitivity requirement for GSM. At the specified level of -49dBm, the IMP generates no significant interference. Field measurements readily show blocking signals are common at to -35dBm. Signal levels between dbm and -21dBm are seen, but not as often. As the density of basestations increases, the elevation angle of the antennas is lowered and the potential interference becomes greater. Out-of-Band (i) (ii) Blocking Profile for North America PCS 1900 Inband (iii) Out-of-Band (iv) m m -12 dbm -26 dbm - 33 dbm - 43 dbm - 43 dbm - 33 dbm -26 dbm -12 dbm m m dBm MHz 1910 MHz fo -3.0 MHz fo -2.8 MHz fo -1.6 MHz fo -1.4 MHz fo khz f 0 fo khz fo +1.4 MHz fo +1.6 MHz fo +2.8 MHz fo MHz 2010 MHz 2070 MHz Out-of-Band (i) (ii) Blocking Profile for Europe DSC 1800 Inband (iii) Out-of-Band (iv) m... m -12 dbm 20 MHz -12 dbm -26 dbm -26 dbm 40 MHz - 33 dbm - 33 dbm - 43 dbm - 43 dbm dBm m m MHz 1910 MHz fo -3.0 MHz fo -2.8 MHz fo -1.6 MHz fo -1.4 MHz fo khz f 0 fo khz fo +1.4 MHz fo +1.6 MHz fo +2.8 MHz fo +3.0 MHz Figure 2 Blocking Requirements for GSM and 3G WCDMA 1920 MHz 1980 MHz 3.0 IMP Interference Mitigation Technology, System Simulations A MATLAB system simulation was built for a CDMA2000 receiver with performance specifications that just met the high power two tone test. Fig. 3 shows the SOI and the IMP generated by the system that meets the sensitivity requirement of -94dBm. Fig. 4 shows the improvement when the IMP is cancelled by the Finesse Technology. In the simulations, the SOI is shown as a tone so that the relative strength can be seen. In reality, it is below the noise floor and spread by 21 db. In the real world the source signals, that generate the IMPs, are not tones but are modulated signals. An additional simulation was run with simulated FM AMPS signal with a 30 khz 4 of 22

5 bandwidth. Fig.5 shows the SOI plus the IMP prior to IMP suppression. Fig. 5 shows the SOI after IMP suppression. As shown in Fig. 6, there is a small residual of the modulated IMP. This can be eliminated when the differential group delay is taken into account. Even at the present level, when the IMP is spread by the CDMA2000 spreading gain of 21dB, it will have no impact on the signal to interference ratio. The simulations are predicting 18 to of suppression for the modulated and tone IMPs respectively. As will be shown below, significant sensitivity and system capacity improvements can be achieved with IMP suppression in the handset or MS. Fig. 3: SOI Plus IMP from High Power Two Tone Test Fig. 4: SOI After IMP Suppression of IMP from High Power Two Tone Test Fig. 5: SOI Plus AMPS IMP from High Power Two Tone Signal Test Fig. 6: SOI After IMP Suppression of IMP from High Power Two Tone Signal Test 4.0 Pole Capacity and Load Factor [reference 1] The capacity of a CDMA sector cell is measured by a term called the loading factor. The loading factor is the percentage of the theoretical maximum traffic capacity or the pole capacity. Practical CDMA systems have a cell capacity between 5 and 7 of the pole capacity, so the typical 5 of 22

6 load factor is between.5 and.7. CDMA systems operate at a negative signal to noise ratio prior to the de-spreading coding gain. In a single sector/cell isolated case, the energy of each CDMA traffic signal adds to the noise floor and when the total traffic signal energy is equal to the noise floor, the load factor is.5 and the noise floor is raised by 3 db. This would be expected when the signal energy, which for CDMA signals looks like noise, is added to an equal thermal noise power. Any energy added, either in the form of additional CDMA traffic signals or other interfering signals, adds to the signal power and is effectively consuming cell capacity. When a given CDMA handset user experiences interference, the two way power control increases the power in the CDMA coded signal of that user. This increases the power in the transmitted spectrum and now the user in question looks like multiple users from a capacity point of view even though the data rate to that user is not increased. In multiple service systems like WDCMA, this user now looks like a user with a more resource intensive service that requires a higher Eb/No. The impact of the increased base station transmitter power is that more of the available cell capacity is used as the total signal power is increased without any additional traffic throughput. Any increase in the power spectrum of the downlink will consume system capacity. It is for this reason that CDMA systems are most effective when all transmitters transmit at the minimum power required to close the links. 5.0 Capacity Calculations for CDMA a Single Cell Sector The average load factor for the downlink in a CDMA system will be analyzed for a signal sector of a single cell. CDMA uses a 1:1 frequency reuse and when one sector transmitter raises the transmit power to support a handset which is experiencing interference, the effective loading of all adjacent cells is increased and thus the traffic capacity of the surrounding cells is also negatively impacted. In CDMA systems, the optimal traffic capacity is realized when the transmitter power is kept at a minimum. In this analysis, the average load factors for a single sector of a single cell are determined for varying levels of interference. The impact to the traffic carrying capacity for the service provider will be more severe than shown herein because there will be additional cell to cell interference that is not explicitly covered in this paper, however, nominal values for the impact on own cell are incorporated. The Finesse Wireless technology, when deployed in the impacted handset, will significantly mitigate this impact. For CDMA2000, an average cell capacity is around 13 to 16 users per sector with three sectors. This provides the cell with around 50 users depending on the environment with a sectorization gain of 2.6. As will be shown in the analysis that follows, if one user experiences high power IMPs and the base station increases power to support that user, a single user can consume 30 to 5 and up to 10 of the cell capacity (assuming the user is not dropped at some point). If multiple IMPs are present, or more users are impacted, the impact to capacity can be severe. Average Load Factor: (reference1 The average load factor can be computed by the equation below N Eq 1) LF avg = {AF j *(E b /N o )/(W/R j )}{(1-OF avg ) + CCI avg } j=1 Where: N = the number of users j per cell sector j = 1 to N 6 of 22

7 AF j = Activity factor of user j at the physical layer; E b /N o = Signal energy per information bit Defined by the service W = CDMA chip rate R j = Information rate of user j OF j = orthogonality of channel of user j Nominally.58 for speech Nominally 1.0 for data CDMA2000 voice 9.6kb/s; 7dB WCDMA voice 12.2Kb/s; 7 db WCDMA data; 5dB and 1.5 db Dependent on class of service 3.84 Mcps for WCMA Mcps for CDMA 2000 Dependent on service Dependent on multipath 1 = full orthogonality 0 = no orthogonality OFavg = average orthogonality factor nominal =.5 CCI j = Cell to Cell interference CCI avg = average cell to cell interference Ratio of other cell power to own cell power received by user j Average ratio of other cell Power to own cell power Nominal =.65 The following nominal values were used for voice and data Variable Data Voice AF j E b /N o 5dB 7dB Block Error Rate BLER 1 1% Bit Rate of User R j 64kb/s 12.2 kb/s WCDMA 9.6 kb/s CDMA2000 Chip Rate WCDMA 3.84Mcps 3.84Mcps Chip Rate CDMA Mcps 1.23Mcps Noise Figure of Mobile 7dB 7dB System IIP3w/ LNA 0dBm 0dBm System IIP3 w/ LNA Bypass 8.5dBm 8.5 dbm Table 1: Nominal Values for Loading Analysis 7 of 22

8 The term Eb/No in equation 1 is the required Eb/No to meet the specified Block Error Rate (BLER) for the service of the signal of interest (SOI) or user j. When a user experiences interference, the effective Eb/No increases and the user now contributes more to the load factor of the cell and thus uses more of the system capacity. In the analysis that follows, the values in Table 1 are used and when interfering IMPs are present, the effective required Eb/No is determined and then the load factor contribution for the impacted user is computed. The effective required Eb/No for a single user service is computed by starting with the Eb/No required in a pristine environment, or the specified value for the required BLER. When the user is subjected to an interfering signal, the interfering signal adds to the noise floor of the system, which consists of the thermal noise floor plus the power from all user signals in the passband. (The user signals are all of the CDMA traffic channels plus the paging, pilot and control channels.) The sum of the thermal noise floor + the traffic channels + the IMP power will be called the system noise floor = SNF. For the target signal, or signal of interest, (SOI) the minimum receive power = SNF coding gain + required Eb/No. When interfering IMP signals are significantly below the SNF, the impact on the SOI is not significant. When the power of the IMP is well below the SNF, the sum of the two is approximately the power of the SNF, and the required Eb/No for the SOI is approximately the effective required Eb/No. However, when the power of the IMP approaches or exceeds the SNF, the effective required Eb/No for the SOI increases by the difference between the SNF and the SNF plus the IMP. This is required to deliver the required Eb/No for the SOI service to achieve the required BLER. When the effective required Eb/No is used in equation 1, the load factor contribution of a single user can be determined. When this number is added to the starting load factor, the cumulative load factor is determined. The typical load factors for functional cells range between.5 and.7. By computing the load factor contribution of one user experiencing jamming, we can compute the percentage of system capacity consumed by that one user. 6.0 CDMA System Architectures In the case of CDMA2000, the -21 dbm two tone test creates a very difficult set of performance requirements. The standards allow for the SOI signal sensitivity to degrade from -119dm (pristine environment) to -94dBm when the high power interfering blocking signals are present. CDMA2000 gives up 25dB in sensitivity. When the power in the interfering IMP exceeds the noise floor in the LNA, the typical solution is to bypass the LNA. In this situation the handset will have a higher noise figure in the bypass mode but also a superior IIP3. This however has the impact of raising the SNF and thus requires higher transmitter power from the base station and thus the system traffic capacity is impacted in the own cell as well as the adjacent cells. If the LNA is not bypassed, the IMPs get even larger and have a greater impact on the loading factor and the capacity of the cell. IIP3: The system IIP3 of a handset, as far as the two tone test is impacted, is the cascaded IIP3 of all the analog components in the receiver. For CDMA2000, the LNA requires a minimum IIP3 of 7.6dBm to meet the signal blocking tone (at dbm) desense test which results in the 3 rd order cross modulation products from the transmitter feed thru. The transmitter feed thru is filtered in the SAW filter following the LNA and is not impacted by the down stream IIP3 of the amplifiers, mixers, baseband filters and amplifiers. The system IIP3 as seen by the two tone test will probably be about lower than the LNA alone. The analysis herein assumes a maximum 8 of 22

9 LNA IIP3 of +10dBm and a system IIP3 of m. In the LNA bypass mode, a system IIP3 of +8.5 dbm is assumed given that the bypass mode switch will have an IIP3 of around +15 to +20 dbm. For this analysis, these same numbers for the CDMA2000 two tone tests are used for WCDMA even though the standards do not call for the same high level two tone test. This is done because in reality, both systems will experience the same level of interference. 6.1 Case No. 1: CDMA2000 Voice Case (9.6kb/s) For the CDMA 2000 Voice Case, the parameter values used are those shown in Table 1. The system IIP3 is assumed to be m with the LNA having an IIP3 of +10dBm (possibly optimistic). The thermal noise floor is computed as -174dBm/Hz x 1.23 MHz = -174dBm/Hz + 61dB-Hz = -113dBm. With a NF of 7 db, the system noise floor is -106dBm. With a required Eb/No of 7 db and a coding gain of (1.23E6/9.6E3) 21dB, the SOI sensitivity = = - 1m with the LNA in and no interfering signals (standards specification is dbm). In the Table 2 below, the highlighted band in row 15 is the point at which the LNA bypass mode is used because the IMP energy now is high enough to impact the noise floor of the LNA and the handset receiver sensitivity will be better in the bypass mode. In the bypass mode, the noise figure is assumed to be +12dB and now the system noise floor without the interfering IMP is db = -101dBm. The SOI sensitivity = = - 115dBm without interfering IMPs. The system IIP3 in the bypass mode is assumed to be 8.5dBm. Table 2 gives the impact of the IMPs on the load factor for a signal or single? user. In column AJ, Table 2, the load factors for a single user that exceed.05 are highlighted in blue and those that exceed.1 are highlighted in red. These numbers indicate the percentage of the pole capacity of the system that will be consumed by one user with no other traffic on the channel. This is the percentage of the theoretical maximum capacity. In table 3, the percentage of the available system capacity consumed by one jammed user, with a system maximum load factor of.5 and.7 is presented. The cases wherein the single user consumes in excess of 5% of the total available system capacity are highlighted in blue. The cases where the signal user consumes in excess of 5 of the available system capacity are highlighted in red. As the percentage of the available system capacity consumed by one user approaches 10, this becomes a single user system. Figure 7 shows the plots of the percentage of the system capacity that one CDMA2000 voice user will consume as a function of the power in the signals (two tone) that generate intermodulation products and maximum load factors of.5 and.7. Load factors for most practical CDMA systems fall between.5 and.7. Figure 8 shows the percentage of system capacity consumed with a maximum load factor of.5 and interference suppression between 0 and. Figure 9 shows the percentage of system capacity consumed with a maximum load factor of.7 and interference suppression between 0 and. For Table 3 and figures 7 through 11, the maximum available system capacity is the maximum load factor times the pole capacity. 9 of 22

10 D G H I M N O T U V W X AB AD AH AJ LNA in System LNA By Pass Mode Optimal Selection w or w/o LNA PowerM Thermal MS Ioc Spreading Target Target Thermal Bypass Ioc Target Target Target Increase Tone N Noise IM3 Equals Gain Signal Signal Noise IM3 Equals Signal Signal with in Eb/No Effective Load Factor MS IMP Thermal db Based Based Bypass IMP Thermal Based Based IMP With req Eb/No per jammed dbm dbm dbm floor dbm on Ioc on IM3 dbm on Ioc on IM3 interfer interfer User Table 2: CDMA2000 Voice Case IMP Impact on Load Factor with LNA Bypass Table 3: CDMA2000 Voice Case: Percentage of Available System Capacity Consumed by a Single Jammed User with Maximum System Load Factors of.5 and.7 as a Function of the Power in the Source Tones that Generate the 3 rd Order IMP. 10 of 22

11 CDMA2000 Voice Single User Percentage of Available System Capacity Solid Line Maximum Load Factor =.5 Dotted Line Maximum Load Factor = Figure 7: CDMA 2000 Voice: Single User Consumption of Available System Capacity for Maximum Load Factors of.5 and.7 with No Interference Suppression CDMA2000 Voice: Percentage of Available System Capacity Consumed by one user with varing levles of IMP Suppression for a Maximum Load Factor of Figure 8: CDMA2000 Voice: Percentage of System Capacity Consumed with a Maximum Load Factor of.5 and Interference Suppression Between 0 and with LNA Bypass Capability CDMA2000 Voice: Percentage of Available System Capacity Consumed by one user with varing levles of IMP Suppression for a Maximum Load Factor of Figure 9: CDMA2000 Voice: Percentage of System Capacity Consumed with a Maximum Load Factor of.7 and Interference Suppression Between 0 and with LNA Bypass Capability 11 of 22

12 When one user requires more power from the base station to overpower inband IMPs, the load factor of the sector is increased and this increases the percentage of the system capacity consumed by the one service. The available system capacity will be 1- percentage of capacity consumed by the one user in question. Figures 10 and 11 show the available system capacity when one user experiences interference from IMPs generated by signals from to -21 dbm. The curves show the ability to recover system capacity as a function of the level in interference suppression from 0 to. % of System Capacity Available CDMA 2000 Voice Capacity with Varing Levels of IMP Suppression and Over a Range of Source Signal Power: Max LF = Figure 10: CDMA2000 Voice: Percentage of System Capacity Available for Two Tone Power from dbm to -21dBm for a Maximum Load Factor of.5 and Interference Suppression from 0 to m % of System Capacity Available CDMA 2000 Voice Capacity with Varing Levels of IMP Suppression and Over a Range of Source Signal Power: Max LF = Power in Two Tones in dbm -22 Figure 11: CDMA2000 Voice: Percentage of System Capacity Available for Two Tone Power from dbm to -21dBm for a Maximum Load Factor of.7 and Interference Suppression from 0 to m Each of the four following cases address WCDMA for Voice and Data Services with and without the LNA bypass capability required in CDMA2000 due to the high blocking tones of -21dBm. The results presented were developed as shown above for CDMA2000. For efficiency, the cases for a maximum load factor of.7 will be presented since this is the most optimistic. 6.2 Case No. 2: WCDMA Voice Case with LNA Bypass Capability (12.2kb/s) For the WDMA Voice Case, the parameter values used are those shown in Table 4. The thermal noise floor is computed as -174dBm/Hz x 3.84 MHz = -174dBm/Hz dB-Hz = -108dBm. With a NF of 7 db, the system noise floor is -101dBm. With a required Eb/No of 7 db and a coding gain of (3.84E6/12.2E3) 25dB, the SOI sensitivity = = -119dBm with the LNA in and no interfering signals. In the Table 4 below, the highlighted band in row 44 is the 12 of 22

13 point at which the LNA bypass mode is used because the IMP energy now is high enough to impact the noise floor of the LNA and the handset receiver sensitivity will be better in the bypass mode. In the bypass mode, the noise figure is assumed to be +12dB and now the system noise floor without the interfering IMP is db = -96dBm. The SOI sensitivity = = - 114dBm without interfering IMPs. In column AJ, Table 4, the load factors for a single user that exceed.05 are highlighted in blue and those that exceed.1 are highlighted in red. These numbers indicate the percentage of the pole capacity of the system that will be consumed by one user with no other traffic in the channel. This is the percentage of the theoretical maximum capacity. In Figure 12, the percentage of the available system capacity consumed by one jammed user, with a system maximum load factor or.5 and.7, is presented. As the percentage of the available system capacity consumed by one user approaches 10, this becomes a single user system at 12.2 kb/s. Figure 13 shows the percentage of the system capacity that is available as function of the interference levels and the interference suppression D G H I M N O T U V W X AB AD AH AJ LNA in System LNA By Pass Mode Optimal Selection w or w/o LNA PowerM Thermal MS Ioc Spreading Target Target Thermal Bypass Ioc Target Target Target Increase Tone N Noise IM3 Equals Gain Signal Signal Noise IM3 Equals Signal Signal with in Eb/No Effective Load Factor MS IMP Thermal db Based Based Bypass IMP Thermal Based Based IMP With req Eb/No per jammed dbm dbm dbm floor dbm on Ioc on IM3 dbm Noise on Ioc on IM3 interfer interfer User Table 4: WCDMA Voice Case IMP Impact on Load Factor WCDMA Voice Single User Percentage of Available System Capacity Solid Line Maximum Load Factor =.5 Dotted Line Maximum Load Factor = Figure 12: WCDMA Voice Case: Percentage of Available System Capacity Consumed by a Single Jammed User with Maximum System Load Factors of.5 and.7 as a Function of the Power in the Source Tones that Generate the 3 rd Order IMP. 13 of 22

14 WCDMA Voice Capacity with Varing Levels of IMP Suppression and Varing Range of Source Signal Power and 0 to of IMP Suppression with LNA Bypass: Max LF =.7 % of System Capacity Available Figure 13: WCDMA Voice: Percentage of System Capacity Available for Two Tone Power from dbm to -21dBm for a Maximum Load Factor of.7 and Interference Suppression from 0 to m with LNA Bypass Capability 6.3 Case No. 3: WCDMA Voice Case Without the LNA Bypass Mode (12.2kb/s) WCDMA and GSM are specified with the two tone test to be conducted with the source tones at - 49 dbm. This requirement does not require an LNA bypass mode. If WCDMA handsets are designed to only tolerate the -49dBm tones, then there will not be an LNA bypass mode and when the power in the IMP exceeds the noise floor of the LNA, the impact to the capacity of the system will be even greater as shown in Figures 12 and 13. The computational results for this case are shown in table 5 In column AJ, Table 5, the load factors for a single user that exceed.05 are highlighted in blue and those that exceed.1 are highlighted in red. These numbers indicate the percentage of the pole capacity of the system that will be consumed by one user with no other traffic in the channel. This is the percentage of the theoretical maximum capacity. In Figure 14, the percentage of the available system capacity consumed by one jammed user, with a system maximum load factor or.5 and.7, is presented. As the percentage of the available system capacity consumed by one user approaches 10, this becomes a single user system at 12.2 kb/s. Figure 15 shows the percentage of the system capacity that is available as function of the interference levels and the interference suppression. 14 of 22

15 D G H I M N O T U V W X AB AD AH AJ LNA in System LNA By Pass Mode Optimal Selection w or w/o LNA PowerM Thermal MS Ioc Spreading Target Target Target Increase Tone N Noise IM3 Equals Gain Signal Signal with in Eb/No Effective Load Factor MS IMP Thermal db Based Based IMP With req Eb/No per jammed dbm dbm dbm floor dbm on Ioc on IM3 interfer interfer User No By LNA By Pass Table 5: WCDMA Voice Case: IMP Impact on Load Factor without LNA Bypass Mode WCDMA Voice Single User Percentage of Available System Capacity Without LNA Bypass Solid Line Maximum Load Factor =.5 Dotted Line Maximum Load Factor = Figure 14: WCDMA Voice Case without LNA Bypass Mode: Percentage of Available System Capacity Consumed by a Single Jammed User with Maximum System Load Factors of.5 and.7 as a Function of the Power in the Source Tones that Generate the 3 rd Order IMP. WCDMA Voice Capacity with Varing Levels of IMP Suppression and Varing Range of Source Signal Power and 0 to of IMP Suppression without LNA Bypass: Max LF =.7 % of System Capacity Available Figure 15: WCDMA Voice: Percentage of System Capacity Available for Two Tone Power from dbm to -21dBm for a Maximum Load Factor of.7 and Interference Suppression from 0 to m without LNA Bypass Capability 15 of 22

16 6.4 Case no. 4: WCDMA Data Case (64 kb/s) For the WDMA Data Case, the parameter values used are those shown in Table 1. The thermal noise floor is computed as -174dBm/Hz x 3.84 MHz = -174dBm/Hz dB-Hz = -108dBm. With a NF of 7 db, the system noise floor is -101dBm. With a required Eb/No of 5 db and a coding gain of (3.84E6/64.E3) 18dB, the SOI sensitivity = = -114dBm with the LNA in and no interfering signals. In Table 6 below, the highlighted band in row 100 is the point at which the LNA bypass mode is used because the IMP energy now is high enough to impact the noise floor of the LNA and the handset receiver sensitivity will be better in the bypass mode. In bypass mode, the noise figure is assumed to be +12dB and now the system noise floor without the interfering IMP is db = -96dBm. The SOI sensitivity = = -109dBm without interfering IMPs. In column AJ Table 6, the load factors for a single user that exceed.05 are highlighted in blue and those that exceed.1 are highlighted in red. These numbers indicate the percentage of the pole capacity of the system that will be consumed by one user with no other traffic in the channel. This is the percentage of the theoretical maximum capacity. In Figure 16, the percentage of the available system capacity consumed by one jammed user, with a system maximum load factor or.5 and.7, is presented. As the percentage of the available system capacity consumed by one user approaches 10, this becomes a single user system at 64 kb/s. Figure 17 shows the percentage of the system capacity that is available as a function of the interference levels and the interference suppression. D G H I M N O T U V W X AB AD AH AJ 85 LNA in System LNA By Pass Mode Optimal Selection w or w/o LNA 86 PowerM Thermal MS Ioc Spreading Target Target Thermal Bypass Ioc Target Target Target Increase 87 Tone N Noise IM3 Equals Gain Signal Signal Noise IM3 Equals Signal Signal with in Eb/No Effective Load Factor 88 MS IMP Thermal db Based Based Bypass IMP Thermal Based Based IMP With req Eb/No per jammed 89 dbm dbm dbm floor dbm on Ioc on IM3 dbm noise on Ioc on IM3 interfer interfer User Table 6: WCDMA Data Case IMP Impact on Load Factor with LNA Bypass Capability 16 of 22

17 WCDMA Data Single User Percentage of Available System Capacity Solid Line Maximum Load Factor =.5 Dotted Line Maximum Load Factor = Figure 16: WCDMA Data Case: Percentage of Available System Capacity Consumed by a Single Jammed User with Maximum System Load Factors of.5 and.7 as a Function of the Power in the Source Tones that Generate the 3 rd Order IMP with LNA Bypass Capability. % of System Capacity Available WCDMA Data Capacity with Varing Levels of IMP Suppression and Varing Range of Source Signal Power and 0 to of IMP Suppression with LNA Bypass: Max LF = Power in Two Tone in dbm Figure 17: WCDMA Data: Percentage of System Capacity Available for Two Tone Power from dbm to -21dBm for a Maximum Load Factor of.7 and Interference Suppression from 0 to m with LNA Bypass Capability 6.5 Case No. 5: WCDMA Data Case Without the LNA Bypass Mode (64kb/s) WCDMA and GSM are specified with the two tone test to be conducted with the source tones at - 49 dbm. This requirement does not require an LNA bypass mode. If WCDMA handsets are designed to only tolerate the -49dBm tones, then there will not be an LNA bypass mode and when the power in the IMP exceeds the noise floor of the LNA, the impact to the capacity of the system will be even greater as shown in Figures 16 and 17. In column AJ table 7, the load factors for a single user that exceed.05 are highlighted in blue and those that exceed.1 are highlighted in red. These numbers indicate the percentage of the pole capacity of the system that will be consumed by one user with no other traffic in the channel. This is the percentage of the theoretical maximum capacity. In Figure 18, the percentage of the available system capacity consumed by one jammed user, with a system maximum load factor or.5 and.7, is presented. As the percentage of the available system capacity consumed by one user approaches 10, this becomes a single user system at 64 kb/s. Figure 19 shows the percentage of the system capacity that is available as a function of the interference levels and the interference suppression. 17 of 22

18 D G H I M N O T U V W X AB AD AH AJ 113 LNA in System LNA By Pass Mode Optimal Selection w or w/o LNA 114 PowerM Thermal MS Ioc Spreading Target Target Target Increase 115 Tone N Noise IM3 Equals Gain Signal Signal with in Eb/No Effective Load Factor 116 MS IMP Thermal db Based Based IMP With req Eb/No per jammed 117 dbm dbm dbm floor dbm on Ioc on IM3 interfer interfer User No By LNA By Pass Table 7: WCDMA Data Case: IMP Impact on Load Factor without LNA Bypass Mode WCDMA Data Single User Percentage of Available System Capacity Without LNA Bypass Solid Line Maximum Load Factor =.5 Dotted Line Maximum Load Factor = Figure 18: WCDMA Data Case without LNA Bypass Mode: Percentage of Available System Capacity Consumed by a Single Jammed User with Maximum System Load Factors of.5 and.7 as a Function of the Power in the Source Tones that Generate the 3 rd Order IMP. % of System Capacity Available WCDMA Data Capacity with Varing Levels of IMP Suppression and Varing Range of Source Signal Power and 0 to of IMP Suppression without LNA Bypass: Max LF = Figure 19: WCDMA Data: Percentage of System Capacity Available for Two Tone Power from dbm to -21dBm for a Maximum Load Factor of.7 and Interference Suppression from 0 to m without LNA Bypass Capability 18 of 22

19 7.0 Receiver Sensitivity CDMA2000 and WCDMA 7.1. CDMA2000 Sensitivity with and without IMP Suppression The sensitivity of the CDMA2000 receiver in the presence of high power IMPs generated by two tones ranging from dbm to -21 dbm with IMP suppression of 0, 10, 15 and is shown in Fig.20.The voice service for CDMA2000 is 9.6kb/s CDMA 2000 Sensitivity versus Two Tone Power Level Fig. 20: CDMA2000 Sensitivity as a Function of Two Tone Power Level and IMP Suppression 0 to Sensitivity in dbm 7.2. WCDMA Sensitivity with and without IMP Suppression WCDMA supports both data and voice services and a mixture thereof. For this analysis, an all voice case and an all data case were conducted. WCDMA is required by standards to operate with the two tone test at -49 dbm as apposed to the -21dBm for CDMA2000. If WCDMA receivers are designed to only handle the -49dBm tones, then an LNA bypass mode is not required as it is for CDMA2000. If the WCDMA receiver is subjected to signals as high as those specified for CDMA2000, the impact to receiver sensitivity and system traffic carrying capacity can be severe. In the analysis that follows, WCDMA is analyzed for a receiver designed to meet CDMA2000 two tone test requirements with the LNA bypass and for a receiver without the LNA bypass. This is done for both the voice and data cases WCDMA Voice with LNA Bypass Fig. 21 gives the WCDMA-voice sensitivity for two tone values from dbm to -21 dbm for IMP suppression values of 0, 10, 15, and. WCDMA voice services are at 12.2kb/s. Sensitivity in dbm Two tone Power in dbm WCDMA Voice Sensitivity versus Two Tone Power with LNA bypass Fig. 21: WCDMA Voice with LNA Bypass: Sensitivity as a Function of Two Tone Power Level and IMP Suppression 19 of 22