Agilent ESG-D Series, Option UN5

Using the ESG-D RF SignalGenerator’s Multicarrier,Multichannel CDMA Personality for Component TestProduct Note

2

2334556678

10101010111212141515161617181818192122

IntroductionThe two baseband generators in the Agilent ESG-DHow to use this noteCDMA component test with multichannel CDMAGenerating multichannel CDMA waveformsPredefined channel configurationsCustom channel configurationsCode-domain power displaySignal generation detailsComplementary cumulative distribution functionTypical base-station and mobile component testsCode-domain powerWaveform quality (ρ)Spurious emissionsAdjacent channel power optimizationBaseband outputsMulticarrier CDMA Intermediate-frequency measurementsAdjusting CDMA parametersEvaluation of emerging standardsBasic CDMA parametersFIR filtersChip rateReconstruction filterWaveform lengthOversample ratioConstraintsSummaryRelated literature

IntroductionThe Agilent Technologies multichannel CDMA per-sonality, Option UN5, is ideally suited for adjacent-channel-power (ACP) and code-domain powermeasurements for CDMA base-station and mobilecomponent test for single and multiple carriers.This personality offers preconfigured settings forpilot, 9-channel, 32-channel, 64-channel, andreverse channel. For additional control of CDMAsignals, a channel editor provides the power todefine up to 256 Walsh-coded channels per carrierwith selectable channel type, Walsh code, power,PN offset, and data. Four IS-95 filters are providedas well as user-defined FIR filtering. To test CDMAvariants, the chip rate is selectable from 10 chips/second to 20 megachips/ second. All these capabili-ties are combined with superior ACP performanceto test high performance active components.

Option UN5 requires the dual arbitrary waveformgenerator, Option UND, to be installed.

Table of Contents

3

The two baseband generators in the Agilent ESG-DThe Agilent ESG-D family of RF signal generatorsoffers two optional baseband architectures withcomplementary features for generating complexdigitally modulated signals. These are Option UN8,real-time I/Q baseband generator and Option UND,internal dual arbitrary waveform generator. Theseoptions provide the flexibility to simulate existingcommunications standards, modify existing digitalprotocols, define and create digitally modulatedsignals, and intentionally impair these digital sig-nals. Both baseband generators are programmableso new digital modulation formats can be added bysimply loading new firmware into the Agilent ESG-D.

Option UN8 simulates a single communicationschannel with a variety of modulation types, FIR fil-ters, and symbol rates. Data sources are externalserial data, internally generated data, or down-loaded data patterns.

Option UND provides completely arbitrary I/Qwaveform generation capability, but does not pro-vide modulation of real-time data. Typical applica-tions with Option UND include:

• simulating digitally modulated signals with upto 20 MHz bandwidth

• generating two or more CW tones with just oneESG-D

• simulating multiple communications channels• simulating noise or other impairments.

How to use this noteThis product note discusses using Option UN5 forCDMA base station and mobile component test. Itbegins with the steps involved in generating a mul-tichannel CDMA signal, followed by a descriptionof the IS-95 physical layer and statistical nature ofthat signal. Next is a description of some commonCDMA component measurements for one or morecarriers and techniques for getting the most out ofthose measurements. Finally, this note concludeswith a discussion of the various parameters avail-able to the user who wants to test non-standard IS-95-like signals.

The minimum instrument configuration used inthe examples in this note is:

Agilent ESG-D signal generator, equipped withOption UND (internal dual arbitrary waveformgenerator), Option UN5 (multichannel CDMA personality), and Option H99–recommended forimproved adjacent channel power (ACP).

Serial Data

UN7 UND

Data In GPIB

Dual arbitrary

waveformgenerator

I/Qbasebandgenerator

I/Qmodulator

Datagenerator

Bit errorrate tester

RF Out

ESG-D family with Options

UN8

Figure 1. Complementary digital baseband generation in the ESG-D family

4

The examples in this product note use the follow-ing convention: softkeys (redefined by context) aredenoted by bold type. Hardkeys are denoted byunderlined type.

This product note assumes a basic knowledge ofdigital modulation concepts, CDMA, and instru-ment programming with Standard Commands forProgrammable Instruments (SCPI). It describeshow to set up multichannel CDMA signals, techni-cal details of signal generation, and advanced tech-niques for using the full potential of the ESG’s multicarrier, multichannel CDMA capability.

For additional information about these and othertopics, the following Agilent application and prod-uct notes are available through your local Agilentsales office or at www.agilent.com:

Creating Digitally Modulated Signals with theAgilent ESG-D and Option UND, Dual ArbitraryWaveform Generator, literature number 5966-4097E.

Customize Digital Modulation with the AgilentESG-D Series Real-time I/Q Baseband Generator,Option UN8, literature number 5966-4096E.

Digital Modulation in Communication Systems —An Introduction, literature number 5965-7160E.

Agilent application note 1307, Testing CDMA Base-station Amplifiers, literature number 5967-5486E.

CDMA component test with multichannel CDMACDMA active components are tested best by actualmultichannel CDMA signals. The Agilent ESG-Dcan generate a variety of customizable signals thataccurately test active components. The nonlineari-ties of active components in CDMA base stationsand mobile stations make them susceptible toeffects such as gain compression, spectral regrowthand phase distortion. These effects degrade in-channel characteristics like code-domain powerand out-of-channel characteristics such as adja-cent-channel power ratio (ACPR). CDMA compo-nent test involves measuring in-channel and out-of-channel characteristics to determine whether anaccurate signal is transmitted while interferencewith other signals is minimized.

The phenomena that cause in-channel and out-of-channel distortion are highly dependent on boththe amplitude and phase of a transmitted signal.Measuring distortion requires a signal that closelyresembles a fully coded CDMA signal to character-ize CDMA components. The ESG-D with OptionUN5 generates a wide variety of statistically cor-rect multichannel and multicarrier waveforms forCDMA component test. For component test, the“statistical correctness” of the test signal is vital.This concept is explained more completely underthe section describing the complementary cumula-tive distribution function (CCDF) of a multichannelCDMA signal.

5

Generating multichannel CDMA waveformsThe Agilent ESG-D multichannel CDMA personalityuses the onboard digital signal processor (DSP) tocompute up to 256 CDMA channels on each of upto 12 carriers. A user can simulate the signals ofeverything from a single mobile station to one ormore fully loaded base stations. Methods to per-form these simulations include one-button configu-rations as well as custom setups, and the ability tomodify basic CDMA parameters to test emergingtechnologies, such as wireless local loop.

To simplify the task of configuring the highly flexi-ble CDMA signal parameters, the user interface issegregated into CDMA Select and CDMA Define.

The CDMA Select menu gives the user a “one but-ton” choice of default or user-defined channel set-ups. The CDMA Define menu allows tremendousflexibility in defining the CDMA signal parameters.For convenience, CDMA channel information ispreloaded into the CDMA channel editor for modi-fication.

Predefined channel configurationsTo simulate a CDMA forward link, use the soft keys to provide configurations for pilot, 9 channels,32 channels, and 64 channels. For a reverse link, a single offset quadrature phase shift keying(OQPSK) channel is generated. The CDMA Selectmenu is shown in Figure 2. The preconfiguredchannel setups are shown in Table 1.

Channel NumbersPilot Paging Traffic Sync FIR Filter

Pilot 0 - - - IS-95 filter modified for ACP with equalization9 Channel 0 1 8 to 13 32 IS-95 filter modified for ACP with equalization32 Channel 0 1 8 to 31, 32 IS-95 filter modified for ACP with equalization

33 to 3764 Channel 0 1 8 to 31, 32 IS-95 filter modified for ACP with equalization

33 to 63Reverse - - - - IS-95 filter modified for ACP

Figure 2. CDMA Select menu

Table 1. CDMA channel configuration defaults

6

Custom channel configurationsThe CDMA define menu gives the user maximumflexibility to modify the CDMA signal. The user caneither modify a preconfigured CDMA selection (the9-channel default, for example) or, starting from ablank screen, define a multichannel signal configu-ration with the multichannel CDMA table editorshown in Figure 3.

The variables for each channel that the user cancontrol are:

• Channel type: pilot, paging, traffic, synchron-ization

• Walsh code for each channel: 0 to 63

• Power for each channel: 0 to –40 dB, independ-ently

• PN offset: 0 to 511• Data feeding Walsh code: random data or binary

data 00000000 to 11111111

The ability to define a custom channel configura-tion is especially useful for tuning the complemen-tary cumulative distribution function (CCDF), ameasure of the stress on a component.

Code-domain power displayTo verify power levels or enabled Walsh codes,Option UN5 offers a code-domain display based onthe channel and power levels specified by the user,shown in Figure 4. To look at channel configura-tions at additional PN offsets, rotate the knob onthe signal generator.

Figure 3. CDMA channel-table editor preloaded with 64-channel configuration

Figure 4. Code-domain power display for a 9-channel configuration

7

Signal generation detailsFor multichannel CDMA waveforms to be playedback by the dual arbitrary waveform generator,they must be generated and then stored in RAM.This is done in the DSP and is controlled by theUN5 personality’s firmware algorithms.

The Agilent ESG-D with Option UN5 does not gen-erate fully coded multichannel CDMA signals.Instead, it generates partially coded signals thatare statistically equivalent to fully coded signals,with equivalent power spectral density, peak-to-average power statistics, and phase response.However, certain elements of real CDMA signalgeneration are not performed, such as long coding,64-ary modulation, interleaving, and convolutionalencoding.

This level of coding is sufficient to achieve theappropriate CCDF, as well as correlation of theWalsh-coded channels in a code-domain powermeasurement. This means this signal will stressamplifiers and other components exactly as a fullycoded multichannel signal. The stress applied to aCDMA component can be described by a signal’sCCDF.

The channel coding structures for both forwardand reverse channels are shown in Figure 5. Forthis discussion, we use the phrase “forward chan-nel” to include the forward traffic channels, pilotchannel, sync channel, and paging channels, all ofwhich are transmitted from the base station to themobile station. For forward channel generation,random data, or a defined sequence of repeatingdata, is supplied to a Walsh code modulator. Theoutput of the Walsh coder undergoes short-codescrambling (with the appropriate PN offset) forboth I and Q channels, and baseband filtering.

In the case of reverse-channel (mobile station tobase station) coding, data is supplied directly tothe I and Q short-code scramblers. After scram-bling, the Q channel is delayed by 1/2 chip to effectOQPSK modulation.

After individual channels are generated, they aresummed (forward link only) for a complete multi-channel CDMA waveform and scaled for optimaladjacent channel power, or spurious emissions.Finally, the I and Q waveform data is stored inRAM to be played back and modulated on an RFcarrier.

Walsh code

DataForward link

Reverselink

Q short code

I short code

QPSK

t/2 FIR

OQPSK(1/2 chip delay)

FIR I

Q

Figure 5. Physical layer implementation for forward and reverse CDMA channels

8

Complementary cumulative distribution functionThe instantaneous power level of a signal is animportant factor in the distortion it will induce ina nonlinear component like an amplifier. As theinstantaneous power level of a signal increases, itis more likely to exceed the 1 dB compressionpoint of an active device, causing an increase inintermodulation distortion (IMD) and degradingadjacent channel power (or spurious emissions)performance. With highly dynamic input signals,such as CDMA, active component distortion canoccur even if the average power level is well belowthe compression point of the component. Figure 6demonstrates this effect.

There are several measures of peak-to-averagepower that are common in the wireless industrytoday:

• Instantaneous peak-to-average power ratio• Crest factor

• Complementary cumulative distribution func-tion (CCDF), also known as probability distribu-tion function (PDF), or cumulative distribution function (CDF).

Peak-to-average power is a measure of the instan-taneous power relative to the time-averaged powerin a signal. Instantaneous peak-to-average powerchanges over time (depending on number of Walsh-coded channels, data on each channel, and relativepower levels in each channel), and is a poor indica-tor of a signal’s ability to push a component intocompression. The crest factor of a signal is a ratioof the peak power versus the average level oversome interval. Unfortunately, instantaneous peak-to-average power and crest factor do not completelydescribe the percentage of time the peak exceedsthe average power by a given amount.

Prms

Pout

Pin

XPpeak

Dynamic signal

1 dBcompression

point

Figure 6. Effect of high peak-to-average powerlevels on an active component

9

The CCDF gives a more complete statistical pictureof the peak-to-average power ratio for a given sig-nal. Expressed as a ratio with a given probability,the CCDF illustrates the chance of any instanta-neous power measurement being at a certain levelrelative to the average power of the signal. A sam-ple CCDF curve for a 9-channel CDMA signal gen-erated on the Agilent ESG-D with Option UN5 isshown in Figure 7.

The CCDF of a signal is an especially importantmetric for multichannel CDMA. Multiple forwardchannels are summed together before a basestationoutput amplifier, and can add constructively ordestructively. The peak power (completely con-structive combination) increases at a higher ratethan the average power as more channels areturned on (Figure 8). The average power is subjectto destructive and constructive combination. Thisresults in higher peak-to-average power ratios.

Increasing active component stress is visible in aCCDF curve as it moves outward to the right, cor-

responding to higher probabilities of higher peak-to-average power ratios.

IS-97 requires the use of nine forward channels for base-station test, but does not mention specificchannel setups or CCDF requirements. This leads toa potential discrepancy among CDMA base-stationcomponent tests.

Since independent pairs of Walsh codes are corre-lated in portions, certain channel setups, for thesame number of channels, result in higher CCDFsthan others. For example, the peak-to-averagepower that characterizes the largest 0.01 percent ofsamples in two different 9-channel configurationscan vary by more than 2 dB. To determine theappropriate level of stress to place on componentsunder test, the predicted traffic should be evaluated.Experimentation with multiple channel configura-tions is the best method of achieving one that opti-mally stresses the component under test or match-es system requirements. CCDFs for a pilot channeland three different 9-channel configurations areshown in Figure 9.

The next section describes some tests that a usermight perform on mobile- or base-station compo-nents.

1 channel 1 channel

+ =

2 channels

PeakAverage

+ =

Peak

Average

1 channel

Q

I

Q

I

Q

I

Q

I

Q

I

3 channels

Figure 7. A CCDF plot for the default 9-channel CDMAconfiguration in the Agilent ESG-D with Option UN5

Figure 8. Effect of multiple forward CDMA channels on peak-to-average power

10

Typical base-station and mobile component testsCommon component tests include adjacent channelpower ratio (ACPR) or spurious emissions, code-domain power, and waveform quality (ρ). Most ofthese tests involve standard setups of the AgilentESG signal generator. An explanation of each ofthese tests follows.

Code-domain powerCode-domain power is a base-station componentmeasurement of relative correlated power in eachof the Walsh code channels. This measurementassesses a component’s ability to process a multi-channel signal without distorting its code-domainresponse. For code-domain power and other in-channel measurements, use the standard IS-95 fil-ter for optimal modulation quality.

Setup:• Preset, Mode, Arbitrary Waveform Generator,

CDMA, CDMA Select, 9 Ch Fwd (or 32 or 64),CDMA Define, Filter, Select, IS-95, IS-95 w/EQ,Return, Return, CDMA On

Waveform quality (ρ)Waveform quality is the measured cross correlationof the actual RF waveform to an ideal waveform.The standard IS-95 filter for optimal EVM (withequalization for the base-station test) is recom-mended for this in-channel test. Only the pilotchannel is required for base-station test, and thereverse channel is required for mobile stations.

Setup:• Base-station components: Preset, Mode,

Arbitrary Waveform Generator, CDMA, CDMASelect, Pilot, CDMA Define, Filter, Select, IS-95, IS-95 w/EQ, Return, Return, CDMA On

• Mobile components: Preset, Mode, ArbitraryWaveform Generator, CDMA, CDMA Select,Reverse, CDMA Define, Filter, Select, IS-95, IS-95, Return, Return, CDMA On

Spurious emissionsThe superior ACP performance of the Agilent ESG-D with Option UN5 (typically –75 dBc for mul-tichannel signals at 900 MHz) makes it ideal forspurious emissions measurements. With the con-venience of a preset 9-channel configuration thatmeets IS-97 requirements, these measurements are also easy.

Figure 9. CCDF curves for pilot channel and three separate 9-channel configurations

11

Setup:• Base-station components: Preset, Mode,

Arbitrary Waveform Generator, CDMA, CDMASelect, 9 Ch Fwd (or 32 or 64), CDMA On

• Mobile components: Preset, Mode, ArbitraryWaveform Generator, CDMA, CDMA Select,Reverse, CDMA On

Base-station component manufacturers with special requirements can modify channel config-urations to achieve a specialized CCDF, or use the multicarrier capability to test multicarriercomponents.

Adjacent channel power optimizationFor out-of-channel tests such as spurious emis-sions, the adjacent channel power (ACP) perform-ance of the signal source needs to be sufficientlybelow the ACP specification of the device undertest. The typical performance of a standard AgilentESG-D for multichannel CDMA on a 900-MHz carri-er is –73 dBc. However, performance as good as–79 dBc can be achieved with ACP-enhancing hard-ware options and adjustments.

The Agilent ESG-D offers three RF output sectionsthat provide different ACP performance:

• Standard electronic attenuator (same perform-ance with Option UNA, electronic attenuator for alternate timeslot power)

• Option UNB, mechanical attenuator• Option H99, output board modification and

mechanical attenuator for improved ACP.

The mechanical attenuator, Option UNB, offersimproved ACP over the standard electronic attenu-ator, as well as increased output power. OptionH99 offers the best ACP, but reduces the maximum

power level from that of the standard electronicattenuator. For the best possible ACP, the suggestedconfiguration for CDMA component test includesthe ESG-D with Options UND, UN5, and H99.

For any hardware configuration there is variabilityin the Agilent ESG ACP performance at differentpower levels and settings. The complex interactionof noise levels, amplifier limitations, and attenua-tor steps at varying power settings leads to vari-ability in ACP performance. Optimal settings forACP performance can be quickly determined experi-mentally. The settings that can affect ACP aredescribed below.

For high power applications above 0 dBm, the ESGACP will degrade significantly. For optimal ACPwith signal power above 0 dBm, the user shouldapply a higher power amplifier at the ESG RF out-put and use an RF power setting below 0 dBm onthe ESG.

Attenuator hold causes the step attenuator tomaintain the current attenuator level while powerlevels are adjusted. With this technique, the userhas nominally 10 dB control over the operatingpoint of the output amplifier, which may allow forimproved ACP performance at a given power level.

In order to optimize ACP performance, the multi-channel CDMA personality automatically activateshigh-crest mode. This lowers the operating level ofthe output amplifier by setting the attenuator acti-vation point to a lower power level. In this modeall output power levels are achieved at a lowerpoint in the amplifier’s operating range, furtherfrom the compression point.

12

The user can experiment with these settings inaddition to power level and frequency in order toget the best possible ACP for her or his application.

Baseband outputsThe Agilent ESG-D with the multichannel CDMApersonality automatically modifies the internallygenerated baseband CDMA signal for optimal ACP.One modification is a reduction in drive level forthe baseband signal. This reduces the contributionto distortion in the signal generator’s I/Q modula-tor and output amplifiers from the baseband sig-nal, improving performance. However, for userswho use the baseband outputs (“I Out” and “QOut”), this also reduces their drive level from nom-inal >1Vp-p. Users who need higher voltage levelsfrom the baseband signals will need to provideexternal amplification.

Multicarrier CDMA (available with firmware version B.02.20 andgreater)Many users are interested in multicarrier CDMAtesting to evaluate broadband power amplifiersthat transmit multiple RF carriers within a band tomeasure spectral regrowth between the carriers.The multicarrier CDMA capability in the AgilentESG-D allows customers who manufacture compo-nents for multicarrier base stations to fully testthem with just one ESG-D. It allows a user to placeup to 12 carriers in a 15 MHz bandwidth, eachwith a custom configuration of channels. A multi-carrier table setup with its corresponding signal isshown in Figure 10a.

Whether selecting predefined multicarrier CDMAconfigurations or explicitly defining the character-istics of each channel on a carrier, the user candetermine the level of customization required foran application. This allows the user to tailor a testto specific requirements like CCDF, and use carrierplacements that reflect the frequency allocation forspecific systems.

Figure 10a. Multicarrier CDMA custom configuration

Figure 10b. Spectral plot of multicarrier CDMA signal

13

Note: Those who want to perform multicarrierCDMA component tests can access updatedfirmware at www.agilent.com/find/esg/ or contact alocal Agilent sales representative. To verify thefirmware version in an existing ESG, press Utility,Instrument Info, Diagnostic Info and read the line enti-tled Firmware Revision.

Two effects may be evident when a multicarriersignal is generated, depending on carrier place-ment: LO feedthrough and carrier images. Botheffects can distort in- and out-of-channel measure-ments. LO feedthrough and single sideband mixingimages are results of an I/Q imbalance in the base-band generator and are evident in any multicarriersignal generated at baseband.

Figure 11 depicts a multicarrier CDMA signal withbaseband images. Multicarrier signals are generatedat baseband by creating individual multichannelcarriers and applying a complex sinusoid toachieve single sideband mixing. I/Q imbalancesadd an undesired mixing component as a mirror

image of the desired signal around the rf carriersetting (dc at baseband). The intended carriers arecentered at the frequency setting (1 GHz) as wellas –2.5 MHz and +1.25 MHz relative to the frequencysetting. The resulting images are centered at +2.5MHz and –1.25 MHz. These images can affect out-of-channel measurements such as spurious emis-sions. For these measurements, it is recommendedthat carriers be arranged such that these imagesdo not appear where the measurement is to bemade.

Image carriers can also affect in-channel measure-ments. Figure 12 depicts a two-carrier signal withsymmetrical carriers. A code-domain power meas-urement on one of the nine-channel carriers showsan increased noise level as a result of the image co-located with the carrier. For in-channel measure-ments such as code-domain power and waveformquality, best results are achieved when carriers arenot co-located with baseband images (i.e., asym-metrical waveform placement).

Figure 11. Multicarrier CDMA signal with baseband images

14

LO feedthrough is a result of dc offset in a wave-form or the I/Q path in the ESG. Figure 13 depictsa multicarrier signal with LO feedthrough. LOfeedthrough that is co-located with a carrier orpresent at the location of an out-of-channel meas-urement will have effects similar to those experi-enced with baseband carrier images, describedabove. LO feedthrough can be minimized by adjust-ing I and Q offset, and I/Q quadrature skew underthe I/Q adjustments menu.

Intermediate-frequency measurementsSome users may wish to use the Agilent ESG-Dwith Option UN5, multichannel CDMA personalityto simulate intermediate frequency (IF) portions oftheir system. For these users, it is important to beaware of a phenomenon that occurs in I/Q modula-tion for carriers below 250 MHz.

Figure 12. Two-carrier CDMA signal with code domain noise from baseband image

Figure 13. LO feedthrough in multicarrier CDMA

15

Carriers between 250 kHz and 250 MHz are in theESG’s “heterodyne band.” Carriers in this bandare created by mixing down from higher synthe-sized frequencies (Figure 14). In the case of theESG, this is done after any I/Q modulation isapplied to the synthesized carrier. To maximizespectral purity in the heterodyne band, the ESGuses synthesized carriers between 750 and 999.750MHz with a 1 GHz LO. Since the synthesized carri-er is at a lower frequency than the LO, the final RFcarrier output from the mixer is the “mirror image”of the original spectrum. In terms of negative fre-quency, this is the image of the original modulatedcarrier that existed between –999.750 and –750.0MHz. The final effect is a modulated carrier that isreversed in the frequency domain.

For spectrally symmetrical signals like IS-95CDMA, this effect is not normally noticeable forspectrum measurements. However, it is significant,and will cause failures in code-domain powermeasurements. In a time-domain sense, the effectof reversal in the frequency domain of a complexsignal is the same as reversing the I and Q chan-nels. To compensate for this effect, the user canreverse the I and Q channels. The steps for doingthis follow:

1. Connect the I and Q out connectors on the rearpanel of the Agilent ESG to the Q and I in con-nectors on the front panel, respectively. Note:The I and Q channels have been intentionallyswitched.

2. Select external I and Q inputs as the modulationsource for the ESG: I/Q, I/Q Source, Ext I/Q, I/Q ON.

Alternatively, the user can connect the I and Q out-puts to the I and Q inputs without switching themand use I/Q, More (1 of 2), Ext I/Q Phase Polarity Invertto achieve the same effect.

If you want to adjust the basic parameters of themultichannel CDMA signal at any frequency, thefollowing section explains some ESG capabilitiesand technical details that help use the full poten-tial of Option UN5.

Adjusting CDMA parametersThe Agilent ESG-D with the multichannel CDMApersonality is easy to configure for most compo-nent test. However, some advanced features allowthe user to maximize testing capabilities andimprove performance, while expanding potentialapplications areas.

Evaluation of emerging standardsSome wireless-local-loop (WLL) standards use IS-95 coding at higher chip rates. The ESG-D withmultichannel CDMA personality lets a user testmobile-station and base-station components forWLL systems by simply adjusting the chip rate ofthe CDMA signal.

5 dB stepattenuator

RF Out

ALC reference

<250 MHz

>250 MHz

ALCmodulator

I

Q

1 GHz

Figure 14. I/Q modulation and frequency synthesis in the ESG “heterodyne band”

16

For example, to test a WLL based on IS-95 codingwith a chip rate of 5 MHz perform the followingsteps:

1. Mode, Arb Waveform Generator, CDMA, CDMA Select,9 Ch Fwd, CDMA Define, Chip Rate, 5, MHz to set thechip rate to 5 MHz.

2. Reconstruction Filter, 8.000 MHz to select the widerreconstruction filter needed to transmit thewider spectrum.

3. Return, CDMA On to generate the waveform andactivate it.

The ESG-D now will be generating a multichannelIS-95-based CDMA signal with a chip rate of 5MHz. Please consult the “Constraints” section inthis product note for more information on the nec-essary adjustments of the reconstruction filter oroversampling ratio to achieve usable signals.

The above example described changing the samplerate of a CDMA signal to emulate a WLL systembased on IS-95 coding. Adjustments to the basicparameters of a CDMA signal require some calcula-tions. These are outlined below.

Basic CDMA parametersThe multichannel CDMA personality offers thecapability to modify FIR filters (choose predefinedfilters or enter coefficients for custom filters), chiprate, reconstruction filter, waveform length, andoversample ratio.

Fir filtersThe multichannel CDMA personality provides fourIS-95 based filters, and the ability to define FIR fil-ter coefficients.

• IS-95 w/Eq (with equalization) is based on thestandard IS-95 filter for base-station componenttest. This filter will provide zero PN timing off-set (typically) as analyzed on an Agilent 8935Abase station test set or Agilent 8921A call sitetest set.

• IS-95 Mod w/Eq (modified with equalization)improves the ACP for base-station componenttest. Because this filter modifies group delay in-channel to improve ACP out-of-channel, a PNoffset of typically 2 microseconds will appear onan 8935 or 8921A.

• IS-95 is a version of the standard filter from theIS-95A specification.

• IS-95 Mod (modified) improves the IS-95 filter forbetter ACP. This is normally selected for reverse-link component test such as mobile poweramplifiers.

FIR filters with up to 1024 FIR coefficients with16-bit resolution and up to 32-times oversampleratio can be saved in the instrument. FIR filtercoefficients can be modified from existing filters,computed externally and downloaded into theESG-D, or entered on the front panel.

From the front panel, press Mode, Arb WaveformGenerator, CDMA, CDMA Define, Filter, Define User FIR.

The FIR table editor will be displayed (Figure 15).Enter coefficients for the first half of the filter. For symmetrical coefficients, press the Mirror Tablesoftkey to generate the second half of the coeffi-cients in reverse order. Otherwise, enter theremaining coefficients manually. Set the OversampleRatio (number of filter taps per symbol).

17

To verify the data, use the Display Impulse Responseand Display FFT soft keys, and compare it againstthe expected outcome (Figure 16).

Chip rateChip rate is adjustable from 10 Hz to 20 MHz,depending on the setting for oversample ratio. Tosimulate wider bandwidth systems such as WLLbased on IS-95 coding structures, increase the chiprate (for example, to 5 MHz).

You may need to adjust the reconstruction filter or oversample ratio to accommodate various chiprates. This is discussed below under “Constraints.”

Figure 15. FIR table editor

Figure 16a. Screen plot of impulse response of FIR filter data

Figure 16b. Screen plot of frequency response (FFT) of FIR filter data

Reconstruction filterOption UND offers three reconstruction filters: 250kHz, 2.5 MHz and 8 MHz. IS-95 defaults to the 2.5MHz reconstruction filter. Other reconstruction fil-ters can be selected or externally applied.

The reconstruction filter should have a bandwidthgreater than the baseband bandwidth of the signal.For CDMA, the baseband bandwidth equals one-half the chip rate. If possible, allow extra guard-band between the filter cut off and the signal’sbaseband bandwidth to prevent in-channel roll off.

Waveform lengthWaveform length can be set from 1 to 16 shortcodes. This lengthens the amount of random datathat feeds the Walsh code, thus lengthening thetime before the data repeats.

Lengthening the waveform will also increase theamount of memory storage required, limiting boththe maximum number of waveforms and the over-sample ratio for each waveform.

Question: What is meant by waveform length (inshort codes)?

Answer: In generating data to build the multiplechannels simulated in Option UN5, random bitstreams are built and multiplied by the standardIS-95-defined 32,768 chip short code. The default isto use enough data to exhaust one short code(32,768 chips/64 chips per bit = 512 bits). Thus, thesame short code repeats, with the same randomdata, once every 26.667 ms for a 1.2288 MHz chiprate (32,768 chips/1.2288 Mchips per second). Ifwe use six short codes, then six 512-bit randomsequences are generated. This way, the entire base-band signal repeats once every 6 x 26.667 = 160 ms.

Oversample ratioOversample ratio is adjustable from 2 to 8, depend-ing on the settings for chip rate and waveformlength.

Question: How does the “oversample ratio” work forUN5?

Answer: There are two “oversample ratio” soft keysin the multichannel CDMA personality. They aredistinguished by their positions in the softkeymenu tree.

One type of oversample ratio is used for filter defi-nition, and exists under the Define User FIR softkey.For instance, if users choose to enter the IS-95 fil-ter, they would enter 48 filter-tap values. Thesewould correspond to the 4X-oversampled filterwith 12 symbols. The user would then use Mode,CDMA, CDMA Define, Filter, Define User FIR, OversampleRatio to set the ratio to 4. This informs the instru-ment that the filter-tap values correspond to a 4xoversampled filter. It is very important to enter theoversampling rate correctly to assure that furthercalculations for the filter will be consistent withthe filter design.

When the CDMA signal is generated, the user canchoose an oversampling ratio for playback, whichcan be found under the CDMA Define softkey. Thiswill adjust the smoothness and memory usage ofthe CDMA signal loaded in the dual arbitrarywaveform generator. If, as in the above example,users enter an IS-95 filter, they can go to the Mode,CDMA, CDMA Define, More (1 of 2), Oversample Ratiofunction to select, say, a 5X oversampling rate forplayback. In this case, the instrument would usethe definition’s sampling rate (4 taps/symbol)entered above to correctly re-sample the filter to a5X oversampling rate, which results in a similar fil-ter consisting of 12 symbols, but 60 taps. Thisincreases the number of data points, consuming25% more arbitrary waveform generator memory(5/4 = 1.25).

Oversampling rates can also be adjusted for thepredefined filters included in the multichannelCDMA personality. In this case, the instrumentknows the original sampling rate, and uses that inits calculations for re-sampling, if necessary.

18

19

ConstraintsCertain aspects of the dual arbitrary waveformgenerator, the multichannel CDMA personality, anddigital-to-analog conversion limit the application ofthe above parameters. The constraints imposed bythe hardware are:

• Memory• Maximum sample rate• Reconstruction filters.

Fortunately, for standard CDMA applications,many of the parameters are optimized by theAgilent ESG signal generator. However, when theuser wishes to change the chip rate, these con-straints must be taken into account. This discus-sion traces the decisions a user might make whensetting up a nonstandard CDMA signal such asWLL.

1. Set a chip rate. The basic design parameter thatdrives this analysis is chip rate. Most other par-ameters are set to optimize the output based onthe chip rate. In this example, select a 5 MHzchip rate for WLL.

2. Select a reconstruction filter. The reconstruc-tion filter should have a cut off higher than the

baseband bandwidth of the signal. In this case,the 2.5 MHz filter has a 3 dB roll off at the edgeof the baseband spectrum (5 MHz/2 = 2.5 MHz),so use the 8-MHz filter (Figure 17). Selecting theminimum filter cut off that still sufficientlytransmits the desired bandwidth allows morefreedom in later stages.

3. Select an oversample ratio (OSR). The oversam-ple ratio, in conjunction with the symbol rate,determines the sample rate (sample rate = OSRx symbol rate). In digital-to-analog conversion,sampling images are created at offsets equal tothe sample rate. These images need to be sufficiently attenuated by the reconstruction fil-ter to avoid distortion. With an oversample ratioof two, the sampling image will be centered at 10 MHz (5 MHz symbol rate x 2), resulting in asignificant interfering signal at less than 7.5MHz (Figure 18). This is below the reconstruc-tion filter cut off. An OSR of four, however,moves the sampling image to 20 MHz (Figure19). This is sufficient to reject the samplingimage with the reconstruction filter. For mostapplications, a minimum OSR of four is recom-mended. However, select the lowest oversampleratio possible to conserve memory.

Reconstruction filter

2.5MHz

8MHz

2.5MHz

8MHz

10MHz

Figure 17. Selecting a reconstruction filter for CDMA with a 5 MHz chip rate

Figure 18. WLL with a 5 MHZ chip rate, 8 MHzreconstruction filter, and OSR of 2

20

4. Select a waveform length. An increased wave-form length decreases the repetition rate of therandom data encoded on the Walsh channels ofthe CDMA signal. However, a longer waveformlength also results in more memory usage. Withan OSR of four, the maximum waveform length is:

There are some other constraints to keep in mind:

• The maximum sample rate can limit OSR or chiprate. For instance, with an OSR of four, the max-imum chip rate is 40 MHz/4 = 10 MHz.

• The presence of other waveforms in RAM willlimit the available space to store and play backCDMA waveforms.

Below are the critical equations needed to calcu-late viable combinations of chip rate, OSR, recon-struction filter, and waveform length:

sample rate = OSR x chip rate ≤ 40 MHz

memory (samples) = waveform length x OSR x 32,768 ≤ 1,048,576

reconstruction filter cutoff < sample rate - chip rate - guard1

2

reconstruction filter cutoff > chip rate2

1. The “guard” term allows for filter rolloff. See Figure 20.

1,048,576 samples32,768 chips/short code x 4 samples/chip

= 8 short codes

2.5MHz

8MHz

20MHz

guardband

Figure 19. WLL with a 5 MHz chip rate, 8 MHz recon-struction filter, and OSR of 4

Figure 20. Guardband to allow for filter rolloff in recon-struction filter selection

21

SummaryThe Agilent ESG-D with Option UN5, multichannelCDMA personality, offers a flexible high-performancesignal for CDMA base-station and mobile compo-nent test. Ready-made channel and carrier configu-rations allow the user to quickly generate a testsignal, while the channel and carrier editors pro-vide the capability to tailor signals to very specificneeds. The realistic CCDF of these signals ensuresan accurate stimulation for tests like code-domainpower, waveform quality, and spurious emissions.By using the performance-enhancing capabilities of the Agilent ESG signal generator the user canoptimize critical specifications like ACP to testeven the most advanced active components.

The ability to adjust chip rates and FIR filtersallows the user to simulate new IS-95-based stan-dards in addition to IS-95. Knowledge of the inter-action of parameters like reconstruction filters andchip rates enables the user to effectively generatethese signals.

With the wide range of capability illustrated above,and the path to expand and enhance these capabil-ities as new standards evolve, the Agilent ESG-D isan ideal solution for component test both todayand in the future.

22

For further information

Related literature:Agilent ESG Family of RF Signal Generators, Data Sheet, literature number 5965-3096E

IntuiLink Software, Data Sheet,literature number 5980-3115EN

Agilent ESG Family of RF Signal Generators,Configuration Guide, literature number 5965-4973E

Generating and Downloading data to the ESG-D RFSignal Generator for Digital Modulation, Product Note,literature number 5966-1010E

Customize Digital Modulation with ESG-D Series Real-Time IQ Baseband Generator, Option UND,Product Note, literature number 5966-4096E

Multi-channel CDMA Personality for Component Test,Option UN5, Product Note, literature number 5968-2981E

Generating Digital Modulation with the ESG-D SeriesDual Arbitrary Waveform Generator, Option UND,Product Note, literature number 5966-4097E

Using the ESG-D Series of RF Signal Generators and the8922 GSM Test Set for GSM Applications, Product Note,literature number 5965-7158E

Agilent Technologies’ Test and Measurement Support, Services, and AssistanceAgilent Technologies aims to maximize the value you receive,while minimizing your risk and problems. We strive to ensurethat you get the test and measurement capabilities you paid for and obtain the support you need. Our extensive supportresources and services can help you choose the right Agilentproducts for your applications and apply them successfully.Every instrument and system we sell has a global warranty.Support is available for at least five years beyond the produc-tion life of the product. Two concepts underlie Agilent’s overall support policy: “Our Promise” and “Your Advantage.”

Our Promise“Our Promise” means your Agilent test and measurement equip-ment will meet its advertised performance and functionality.When you are choosing new equipment, we will help you withproduct information, including realistic performance specifica-tions and practical recommendations from experienced testengineers. When you use Agilent equipment, we can verify thatit works properly, help with product operation, and providebasic measurement assistance for the use of specified capabili-ties, at no extra cost upon request. Many self-help tools areavailable.

Your Advantage“Your Advantage” means that Agilent offers a wide range of additional expert test and measurement services, which you can purchase according to your unique technical and businessneeds. Solve problems efficiently and gain a competitive edge by contracting with us for calibration, extra-cost upgrades, out-of-warranty repairs, and on-site education and training, as well as design, system integration, project management, and otherprofessional services. Experienced Agilent engineers and tech-nicians worldwide can help you maximize your productivity,optimize the return on investment of your Agilent instrumentsand systems, and obtain dependable measurement accuracy for the life of those products.

By internet, phone, or fax, get assistance with all your test and measurement needs.

Online Assistancewww.agilent.com/find/assistPhone or FaxUnited States:(tel) 1 800 452 4844

Canada:(tel) 1 877 894 4414(fax) (905) 206 4120

Europe:(tel) (31 20) 547 2323(fax) (31 20) 547 2390

Japan:(tel) (81) 426 56 7832(fax) (81) 426 56 7840

Latin America:(tel) (305) 269 7500(fax) (305) 269 7599

Australia:(tel) 1 800 629 485 (fax) (61 3) 9272 0749

New Zealand:(tel) 0 800 738 378 (fax) (64 4) 495 8950

Asia Pacific:(tel) (852) 3197 7777(fax) (852) 2506 9284

Product specifications and descriptions in this document subject to change without notice.

Copyright © 1999, 2000, 2001Agilent TechnologiesPrinted in U.S.A. May 29, 20015968-2981E