Accounting for Antenna and MIMO Channel Effects Using Agilent SystemVue

Transcription

1 Accounting for Antenna and MIMO Channel Effects Using Agilent SysteVue Application Note Introduction The Multiple-Input-Multiple-Output (MIMO) technique was introduced to address the growing deand for higher data-rate applications in counication standards like 3G, WLAN, WiMAX. In MIMO systes, both the radio channel propagation conditions and antenna characteristics influence the end user s quality-of-service (QoS). MIMO channel odels, with the ability to siulate the ultipath fading that incorporates antenna patterns, are therefore iportant. Agilent SysteVue offers a MIMO Channel Builder that provides just such a capability. This application note discusses two MIMO channel odels, naely geoetry-based and correlation-based odels, which were realized using Agilent s W1715 MIMO Channel Builder, to show how antenna effects and channel characteristics are coupled in the channel odels. It will also present soe MIMO Channel Builder applications as a eans of describing how to use the odels step-by-step and apply the in syste siulation and easureents. MIMO Channel Modeling Approaches Agilent SysteVue MIMO Channel Builder consists of a geoetry-based WINNER II MIMO channel odel that is standards copliant, and a correlation-based MIMO channel odel that is flexible enough for user custoization.

2 Approach 1 Geoetry-based Stochastic Channel Modeling The WINNER II channel odel is a geoetry-based stochastic odel. [1] It also called a double directional channel odel. The odel does not explicitly specify the locations of the scatters, but rather the directions of the rays, like the well-known Spatial Channel Model (SCM). [2] Both IMT-advances and consequently, LTE- Advanced, apply this kind of odel as the MIMO channel odel for perforance tests. Geoetry-based odeling of the radio channel enables separation of propagation paraeters and antennas. The tie variant ipulse response atrix of the UxS MIMO channel is given by H N ( t; τ ) = H n( t; τ ) n= 1 where t is tie, τ is delay, N is the nuber of paths, and n is path index. The ipulse response atrix is coposed of the antenna array response atrices F tx and F rx for the transitter (Tx) and the receiver (Rx) respectively. The channel fro the Tx antenna eleent s to the Rx eleent u, for cluster n, is expressed as H Where: Frx,u,V and Frx,u,H are the antenna eleent u field patterns for vertical and horizontal polarizations, respectively; α n,,vv and α n,,vh are the coplex gains of vertical-to-vertical and horizontal-to-vertical polarizations of ray n,, respectively; λ 0 is the wavelength of the carrier frequency; φn. u, s, n ( t; τ ) = M = 1 F F rx rx exp exp T, u, V ( ϕn, ) α n,, VV α n,, VH Ftx, s, V ( φn, ), u, H ( ϕ ) n, α n HV a,, n,, HH Ftx, s, H ( φn, ) 1 1 ( j2πλ0 ( ϕn, rrx, u )) exp( j2πλ0 ( φn, rtx, s ) ( j2πυ t) δ ( τ τ ) n, is the AoD unit vector; ϕ n. is the AoA unit vector; r tx, s and r rx, u are the location vectors of eleent s and u, respectively; and V n, is the Doppler frequency coponent of ray n,. WINNER II inherits the drop concept fro the SCM. During a drop, the channel undergoes fast fading according to the otion of the user equipent (UE). However, paraeters in the channel odel, such as nuber of paths, nuber of sub-paths, ean angular spread at the base station (BS) and obile station (MS), AoA/AoD distributions, delay spread, path loss, and shadowing are kept constant. Each drop reflects a short snapshot of the fading channel. Setting a longer drop tie can lead to an intrinsically sooth channel odel. Two consecutive drops are independent, which akes the channel odel discontinuous. The syste perforance is usually evaluated over ultiple drops. The siulation results for different drops ay be quite different. In order to get a stable link-level siulation result, it is better to average the siulated results over tens of drops, with each drop lasting tens of channel-coherent tie. This siulation condition can be guaranteed by setting the proper siulation tie and drop interval. WINNER II odels cover a wide scope of propagation scenarios and environents. They are antenna independent and well suited for evaluating transceiver techniques (e.g., adaptive radio links, equalization techniques, coding, and odulation). [1] Additionally, WINNER II odels provide both large-scale and sall-scale fading paraeters, which are useful for evaluating the perforance of ultiple cells and users. The odels also provide a useful tool for link budget calculation and network planning. n, 2

3 Approach 2 Correlation- based Channel Modeling Correlation-based odeling is another approach to spatial MIMO channel odeling. It is widely accepted for its elegant atheatical for and ease of odeling in WiMAX and WLAN. Unlike the WINNER odel, in a Correlation-based odel, the correlation atrices are explicit. Several coparisons have shown that correlation- and raybased odeling are equivalent in ters of channel coefficients and Block Error Rate (BLER). [3] For antennas with radiation patterns, the spatial correlation atrix can be expressed in [4] as ρ Tx, n = π π π π e j2π PAS Tx d n sin( θ ) λ ( θ ) G PAS Tx, Tx ( θ ) ( θ ) dθ G π π Tx, PAS ( θ ) G Tx Tx, n ( θ ) G ( θ ) dθ Tx, n ( θ ) dθ ρ Rx, n = π π π π e d n j2π sin( θ ) λ PAS Rx ( θ ) G PAS Rx, Rx ( θ ) ( θ ) dθ G π π Rx, PAS ( θ ) G Rx Rx, n ( θ ) G ( θ ) dθ Rx, n ( θ ) dθ Under the Kronecker assuption that the transitter and receiver side are uncorrelated, the spatial correlation atrix, Rs, can be calculated using the following equation R S 1 = tr{ R Rx R } Tx R In the standard, the joint correlation atrix of spatial and polarization can be expressed by [5] as R = R S R P where Rs is the spatial correlation atrix and Rp is the polarization correlation atrix. The channel coefficients can be generated using H s = RH u Rx where Hu is a spatially uncorrelated Rayleigh fading saple and teporal correlated by the Doppler spectru. For the generic antenna array, the Kronecker assuption ay not hold. Agilent Technologies has therefore developed a proprietary algorith that calculates the spatial-polarization and joint correlation properties as a eans of iproving the correlation odel based on the basic concept detailed above. Experiental validation has shown that this algorith is an excellent fit for realistic antennas. [6] This ethod is copatible with the well-known Kronecker odel, and escalates it to include the effect of different types of antenna eleents. It provides a flexible way of evaluating the perforance of antenna and channel effects. Moreover, since the correlation odel allows users to custoize all paraeters, including: Power Delay Profile (PDP), AoA and AoD, Angular Spread, different power angular spectrus and antenna patterns, it is a useful tool for users evaluating the syste perforance of their designs and antenna perforance under different channels environents. 3

4 Syste Siulation Using SysteVue Agilent SysteVue provides a convenient syste-level counications design platfor. To better understand how to set the channel odels in SysteVue, let s exaine an exaple that deonstrates swept throughput versus Signal-to-Noise Ratio (SNR) easureents for the LTE downlink in WINNNER II channel- and correlation-based channel environents. Exaple 1: WINNER II Channel Model Siulate the syste throughput using <SYSTEMVUE INSTALL DIR>\EXAMPLES\ MIMO CHANNEL MODEL\CHANNELTHROUGHPUT_LTE.WSV. Step 1 Set the Channel Siulation Environent Figure 1. Syste setting profile in WINNER Figure 2. Channel odel setting profile in WINNER First, set the syste paraeters, which are consistent with the syste siulation conditions (e.g., carrier frequency and sapling rate (Figure 1). Rando Seed is the seed of rando integer nuber generator. For each non-zero setting, the Channel Ipulse Response (CIR) can be repeated. Otherwise, the CIR is generated randoly for each siulation run. Output files can be applied for further analysis. Second, set the channel odel paraeters (Figure 2). In a WINNER II MIMOchannel odel, the channel scenario decides the path paraeters. Ten scenarios can be selected, fro the indoor to the outdoor environent. The WINNER II channel odel is fully consistent with the LTE specification. More coplex channel siulation environents can be set by choosing different options, such as UseFixedCdlPar, UseDualPolarise, UseIntraClusterDelays, UseManualPropCond, UseShadowModel, and UseLOS. A detailed explanation of the paraeters is found in reference [1]. Step 2 Antenna Setting Antenna patterns that include orthogonal polarizations (typically linear theta (θ) and phi (ф) polarizations) can be easured in an anechoic chaber. They can also be siulated using Agilent s 3D electroagnetic (EM)-solver EMPro software or soe other EM siulation tool. Of course, ideal antenna patterns like the oni-directional pattern, 3-sector pattern and 6-sector pattern are options as well. While Agilent s algorith supports 2D and 3D MIMO channel odels, the Channel Model Builder only considers 2D antennas. Approxiate EM wave propagation is regarded as in one panel, a practice which is widely accepted in channel siulation. 3D radiation patterns can be sliced into a 2D pattern according to the assigned paraeters TxPhiRotation and TxElevationAngle in Antenna Pattern profile. A detailed description regarding antenna file forat can be found in the odel help enu. 4

5 In this step, nubers, positions and radiation patterns of Tx and Rx antennas should be assigned separately (Figure 3). When antenna patterns are loaded fro data files, it is assued that the calibration was copleted when easuring or siulating the antenna patterns. As a consequence, the positions of the antennas are unnecessary. Figure 3. Load antenna patterns in a WINNER channel odel Iporting 3D Radiation Pattern Data fro Agilent EMPro Two advantages of using Agilent s EMPro software to predict antenna pattern data are 1) earlier R&D access to physical perforance data, and 2) precise control. EMPro odels the antenna directionality using preliinary echanical CAD and industrial designs. It adds earlier physical degradations into the overall link-level syste perforance, well before any of the baseband signal processing, RF transceiver or echanical hardware designs have been integrated into a hardware prototype. This early validation allows diagnostics and soe syste-level perforance to be easured at the algorithic stage, where architectural changes and syste repartitioning are still easy to adjust. Using EMPro with the SysteVue MIMO-channel odels and LTE library akes it possible to achieve uch higher confidence when verifying full, working hardware prototypes on the test bench, or in Over-the-Air (OTA) real-tie testing. EMPro can also odel antennas in various positions and loadings, such as SAM reference head positions, or left versus right hand-held positions. Due to environental factors, as well as the differences between the physical antenna eleents theselves, it is recoended that each radiating eleent be exported individually. The Correlation and WINNER-II odels in SysteVue are able to accoodate up to 8 unique patterns, for either transit or receive arrays. To export 3D radiation data fro Agilent EMPro, go to its Results post-processing tab, choose the Gain fro one of the Far Zone Sensors, and right-click to export the data. Export the.uan ASCII type file for each MIMO radiating eleent (Figure 4). Then, fro either the Correlation or WINNER-II MIMO-channel odels in SysteVue, assign the pattern files to each NxN MIMO eleent (Figure 3). 5

6 Figure 4. Using Agilent EMPro to export MIMO antenna pattern files for each radiating eleent. These files add antenna directionality due to physical degradations and loading, and replace the ideal patterns built into the SysteVue MIMO Channel odels. Step 3 Syste Siulation The ability to evaluate the antennas influence in different MIMO-channel scenarios is a unique value provided by the SysteVue LTE siulation library plus MIMO channel odel. Now, SysteVue provides a flexible platfor that can be used throughout the coplete baseband design flow for baseband odeling, debugging and verification. The top-level syste platfor structure is shown in Figure 5. The ain odels in the block diagra include LTE MIMO signal source, channel odel and downlink baseband receiver. Designers can substitute the odels with their own to easily verify the algorith using SysteVue. 6

7 Figure 5. Design of LTE downlink coded 2 2 MIMO with throughput versus noise/fading 7

8 Correlation-based Channel Model Step 1 Set Channel Siulation Environent 1. Set the syste paraeters (Figure 6). 2. Set the channel odel paraeters (Figure 7). 3. Set the path paraeters (Figure 8). Because the Correlation-based channel odel is a user defined odel, the path profile (including path PDP, Doppler shape, type of PAS, angular inforation (e.g., angle of arrival, angle of departure and angular spread), and cross polarization ratio, should be set by the users requireents. Figure 6. Syste setting profile in correlation channel odel Figure 7. Channel odel setting profile in correlation channel odel Figure 8. Path profile in correlation-based odel 8

9 Step 2 Antenna Setting In the Correlation-based odel, there are two ways to get the correlation atrices: either assigned by the user or calculated by considering antenna effects. In order to enable the antenna pattern effects, the CorrSource should be set fro Antenna Pattern, otherwise, the correlation atrix ust assigned by the user path-by-path (Figures 9 and 10). Figure 9. Load antenna pattern in a correlation-based channel odel Figure 10. Correlation atrix setting 9

10 Step 3 Syste Siulation LTE Syste Perforance Evaluation Both siulations use the sae LTE-FDD configuration with the sae propagation condition a 2 2 MIMO configuration for transit diversity with a targeted axiu throughput of Mbps. Table 1. LTE transit diversity syste configuration Paraeter Unit Value Channel bandwidth MHz 10 Allocated resource blocks 50 Modulation 16 QAM Target coding rate 1 / 2 The channel configuration paraeters are listed in Table 2. Table 2. Channel configuration NuberofTx=2 NuberofRx=2 DistanceofTx=2λ TxAntennaPatternType=OniDir Rx Antenna Files: RadiationPattern2D_Port1.txt; ectional RadiationPattern2D_Port2.txt Scenario=C1; NLOS; UseFixedCdlPar=1; UseDualPolarise=1; ThetaBs=0; ThetaMs=0; MsVelocity=60k/h; MsDirection=120; Drop Interval=1s; siulation tie=40s As the MIMO channels are correlated, the easured signal power at the receiver side ay be dependent on the correlation of the channels. This correlation dependency prevents a channel eulator fro accurately configuring the MIMO syste for a desired SNR, using power easureents at the receiver. To overcoe this difficulty, the syste siulation uses easureents of the signal power at the transitter to appropriately set the required SNR. [7] When the siulation tie is long enough, the siulation results are quite close for both channel odels (Figure 11). Figure 11. Siulated throughput fractions versus SNR 10

11 Exaple 2: Channel Analysis Using SysteVue Understanding MIMO channel characteristics is iportant in a MIMO counication syste. SysteVue MIMO channel Builder provides not only the channel odels used in the syste siulation, but also the CIR and correlation atrix, which custoers can post-process for further analysis and verification of the syste siulation results. To siulate the channel capacity using <SYSTEMVUE INSTALL DIR>\EXAMPLES\ MIMO CHANNEL MODEL\CHANNELCAPACITY_MEASUREMENTS.WSV, follow these steps: 1. Set a siple design with channel odel (Figure 12). 2. Enable Output Correlation Matrix in channel odels. 3. Set the channel capacity calculation design. Load the correlation atrix generated in Step 2. Figure 12. A siulation using the MIMO Channel odel can choose to save the correlation atrix for later use. This reduces siulation ties in subsequent siulations. 11

12 Figure 13 copares the results of the channel capacity of the WINNER and correlation-based channel odels under the conditions used in Exaple 1. The siulated results are quite close, which agrees with the conclusion that the two odels are equivalent in ters of their channel coefficients and BLER. [3] Figure 13. Channel capacity coparison of WINNER- and correlation-based odels Siilarly, channel coefficients can be saved in data files. The channel coefficient is saved with a 100 KHz saple rate, regardless of the syste s saple rate setting. This SysteVue feature is quite useful for users interested in the channel statistical characteristics. The channel inforation is used, for exaple, to investigate things like the Level-Cross-Rate (LCR), Average Duration of Fades (ADF), correlation factor, and capacity of the channel all of which can be extreely helpful in understanding channel properties and iproving algoriths. 12

13 MIMO Antenna OTA Perforance Evaluation Using SysteVue MIMO OTA testing is a new test ethod that ay allow engineers to test the true perforance of obile terinals. Different approaches to this task are being discussed in COST 2100, 3GPP RAN WG4 and the CTIA. Currently, several ajor test ethodologies exist to address MIMO OTA test. They include the: Reverberation chaber ethod A reverberation chaber generates echoes to iic ultipath propagation. [8] Eulator + anechoic chaber ethod Several wireless channel eulators in an anechoic chaber provide a realistic propagation environent. [9, 10] Two-stage ethod MIMO antenna radiation patterns are easured within the anechoic chaber and cobined with the easured antenna radiation pattern with the ultipath fading eulation. This data is then used to perfor the OTA perforance test. [6] SysteVue s MIMO Channel Builder siulates the ulti-channel fading, which incorporates realistic antenna patterns. As an alternative to the OTA test, it is based on a two-stage ethod that is both econoic and repeatable (Figure 14). The stages in this ethod include: Stage 1: Test ultiple antenna systes in a traditional anechoic chaber or siulate antenna radiation patterns using the Agilent EMPro software or soe other EM siulation tool. Stage 2: Load the antenna patterns into SysteVue MIMO channel odels and eulate the MIMO channel odel with incorporated patterns. Two different approaches exist in the SysteVue Channel Model Builder to add the antenna patterns to the MIMO channel odel. One ethod is to use the WINNER II channel odel. The other ethod is to use a correlation-based odel. This two-stage OTA ethod can be used to easure the following figures of erit in SysteVue: throughput, CQI, BLER, antenna correlation, and MIMO channel capacity. [11] This ethod also has the capability to easure TRP, TRS, antenna efficiency, and MEG. 13

14 Figure 14. Antenna patterns for each MIMO radiating eleent can be siulated in 3DEM software, or easured directly, then added to the propagation and fading odels in the MIMO Channel odels. The overall result is siilar to live hardware over-the-air (OTA) testing, but predictive and uch earlier in the overall design process. 14

15 Exaple 3: MIMO Perforance With and Without SAM Head Evaluate the MIMO syste perforance, with and without the Specific Anthropoorphic Mannequin (SAM) head influence, using <SYSTEMVUE INSTALL DIR>\EXAMPLES\MIMO CHANNEL MODEL\CHANNELTHROUGHTPUT_ LTE.WSV. In real-world situations, the effects of the huan body can be coplicated. Never the less, the head is a critical aspect of successful ipleentation. Here, we try to evaluate the ipacts caused by the interaction of the head and the real antenna. Two antenna gain patterns released with SysteVue are shown in Figure 15. One of the antenna gain patterns is without the influence of the head. The other is an antenna gain pattern with SAM head, which iics the real scenario. As is evident, the SAM head will significantly change the antenna pattern. Figure 15. The farfield directionality of each individual MIMO antenna eleent changes with proxiity to body parts and other environental loadings, thus degrading the throughput perforance of the MIMO array. Shown above is the effect of a SAM head. 15

16 Figure 16 is the siulated throughput fraction using the different antenna patterns shown in Figure 15 under the 2 2 MIMO LTE-FDD spatial ultiplexing configuration with a targeted axiu throughput of Mbps. The LTE syste configurations are the sae as that in Table 1. The channel configuration is shown in Table 3. The results deonstrate that for the designed antenna array, regardless of whether or not there is head influence, the rotation of the antenna will cause throughput fluctuation. When there is head influence, however, the fluctuation range will be uch greater copared to the case with no head influence. In this way, the influence of the huan body can be conveniently evaluated and the antenna perforance easily investigated. The user siply changes the orientation or ultipath scenarios to observe the antenna s robustness under different cases. Table 3. Channel configuration PathIndex PathPowerindB PathDelay PathAoA PathAoD ASofAoA ASofAoD E E E E E NuberofTx=2; NuberofRx=2; DistanceofTx=4λ; TxAntennaPatternType=OniDirect ional; XPR=-9; NLOS; ThetaBs=0; ThetaMs=0; MsVelocity=10k/h; MsDirection=30; SNR=23dB In the sae siulation conditions, channel capacity is siulated in different rotation angles. The change trend of channel capacity is consistent with that of the throughput seen fro the results (Figure 17). Figure 16. Throughput coparisons of antenna patterns with/without SAM head influence Figure 17. Channel capacity with/without SAM head influence at reference SNR=23 db 16

17 We also find that the channel capacity at 60 degree is close to that at 60 degrees for the antenna with SAM head influence, but the throughputs at 60 and 60 degrees have a ore than 20% total throughput difference. Coparing the channel statistical properties fro the output files, it is apparent that while there is severe power ibalance at 60 degrees, there is very little power ibalance at -60 degrees. The channel capacity for the two channels are the sae, but since the siulation does not adjust the transitter power according to the channel, the channel capacity at 60 degrees is hard to achieve and results in a big throughput drop. The output of the channel statistical properties can be very useful tools in locating perforance issues and helping to explain throughput siulation results. The benefits of using the SysteVue LTE library and MIMO channel odel to evaluate the MIMO antenna perforance are as follows: evaluates MIMO antenna perforance under the custoized channel; evaluates MIMO antenna perforance under different MIMO working odes and can evaluate both open-loop and closed-loop perforance; and helps antenna designers optiize the antenna design even before the antenna is anufactured. 17

18 Real-World Verification Using SysteVue SysteVue provides strong instruent connectivity support. With the SysteVue LTE siulation library and the MIMO channel odel, designers can evaluate the MIMO antennas perforance using siulation, as well as a real-world easureent ethod. The SysteVue LTE siulation library generates and downloads the signal into the instruent and then sends it out with Agilent signal generators. The RF analyzer captures the received signal and sends it back to SysteVue for analysis. With this approach, the antenna influence through siulation can be further copared with the real-world easureent results. Figure 18 shows an exaple of the connections between SysteVue and various instruents. Figure 18. A 2x2 MIMO LTE downlink signal is coded in SysteVue, then faded using the MIMO Channel siulation odels, and downloaded to two RF signal generators for hardware testing. If the decoded data bits are the recovered and copared to the original bit sequence, a fully coded BER/FER test can be ade with only partially-ipleented algoriths. Figure 19 depicts an integrated test syste that is used to conduct the throughput test in MIMO OTA. The advantages of syste-level verification are: autoates tests using integrated hardware and software; generates specific wavefors that are not available using regular hardware instruents; and allows the designer to use a software receiver before a hardware receiver prototype has been built, especially during the developing phase of a new product. Figure 19. SysteVue siulation odels can be used to fill gaps in either a link-level PHY, or in the MIMO test equipent, helping to facilitate inexpensive, ulti-channel functional verification in early stages of R&D. 18

19 Suary The W1715 MIMO channel builder in Agilent SysteVue provides a flexible platfor for wireless MIMO systes design. Soe of its key design and test benefits are: cobines both antenna and channel effects in siulation; provides standard-copliant channel odels and universal channel odels to satisfy different application requireents; and provides R&D and anufacturing engineers with an inexpensive alternative to OTA evaluation. Reference [1] IST-WINNER II Deliverable v.1.2, WINNER II Channel Models, IST- WINNER2, Tech. Report, [2] 3GPP TR , Spatial channel odel for MIMO siulations. [3] VM Kolonen, et al, Coparison of Correlation-based and Ray-based Radio MIMO Channel Models, IEEE International Syposiu on Personal, Indoor and Mobile Radio counications (PIMRC 06). [4] Laurent Schuacher, et al, Closed-For Expressions for the Correlation Coefficient of Directive Antennas Ipinged by a Multiodal Truncated Laplacian PAS, IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, P , Vol. 4, No. 4, [5] IEEE C /062, Draft channel odel for advanced air interface, [6] 3GPP R , Perforance Analysis of Two-Stage MIMO OTA Method Versus SCM Approxiation Method, Agilent Technologies, [7] Agilent Application Note, EN, Agilent MIMO Channel Modeling and Eulation Test Challenges. [8] Yoshiki Okano, et al, Evaluation of OTA Perforance for Mobile Terinal Antennas Reflecting Practical Usage and Iproveent of Measureent Efficiency, P.17-25, NTT DOCOMO Technical Journal Vol. 11, No.2, [9] 3GPP R , Evaluating Channel Models for MIMO OTA, Spirent Counications, Noveber, [10] 3GPP R , Verification of the anechoic chaber and fading eulator based MIMO OTA ethod, Elektrobit, [11] 3GPP TR V , Measureent of radiated perforance for MIMO and ulti-antenna reception for HSPA and LTE terinals, WiMAX, Fixed WiMAX, Mobile WiMAX, WiMAX Foru, the WiMAX Foru logo, WiMAX Foru Certified, and the WiMAX Foru Certified logo are tradearks of the WiMAX Foru. All other tradearks are the properties of their respective owners. 19

20 For ore inforation about SysteVue, please visit us on the web: Product inforation Product Configurations Request a 30-day Evaluation Downloads Helpful Videos Technical Support Foru For ore inforation on Agilent Technologies products, applications or services, please contact your local Agilent office. The coplete list is available at: Aericas Canada (877) Brazil (11) Latin Aerica Mexico United States (800) Asia Pacific Australia China Hong Kong India Japan 0120 (421) 345 Korea Malaysia Singapore Taiwan Thailand Europe & Middle East Austria 43 (0) Belgiu 32 (0) Denark Finland 358 (0) France * *0.125 /inute Gerany 49 (0) Ireland Israel /544 Italy Netherlands 31 (0) Spain 34 (91) Sweden Switzerland United Kingdo 44 (0) Other European Countries: Revised: July 8, 2010 Product specifications and descriptions in this docuent subject to change without notice. Agilent Technologies, Inc Printed in USA, Septeber 24, EN