Making VNA What you need measurements Also in this issue …nagui/Appnotes/Agilent/Network... · 2002. 11. 29. · 2 These test modules are supplied by Oleson Microwave Laboratories

HP 8510C

HP 8517B H06

HP 83650B

HP 83620A

Drawer

9

10

11

5

6

3

Also in this issue Test & Measurement Solutions Now available: Custom frequency converter test system Tips & Techniques Increasing throughput Tuning cavity resonator bandpass filters—faster Product News Now up to seven times faster! Calibrate electronically with HP’s new ECal modules New, lower pricing on HP 8510 instruments and systems

October 1998 Volume 9—Number 2

http://www.tmo.hp.com

Making VNA measurements above 110 GHz

ou can now makevector networkanalyzer measure-

ments in waveguide atfrequencies up to 220 GHzusing the HP8510C VNA.Previously, S-parametermeasurements above110 GHz required the userto develop the equipmentnecessary for multipli-cation to the desired

frequency, signal separation, andfinally downconversion to the IF required by the VNA. TheHP 85106D E12 system—in con-junction with components suppliedby two other vendors—greatlysimplifies this process. The systemas shown here can function in twomodes: one which uses the externalwaveguide test modules to makemillimeter wave measurements,and one in which the system functions as a standard 50 GHz coaxial system.

What you needFigure 1 shows the basic elementsof the system. (Not shown arecables and power supplies.) Allcomponents shown are suppliedby Hewlett-Packard as part of the HP 85106D E12 test system,except the two waveguide testmodules shown in gray on thesystem worksurface.

Continued on next page

Figure 1. The HP 85106D

E12 system with OML

test modules.

Y

Bud Noren, Custom System Specialist,HP’s Santa Rosa Systems Division

More details on page 12

Find us on the Web beginning Spring 1999!

Continued from previous page

2

These test modules are suppliedby Oleson Microwave Laboratories(OML) of Morgan Hill, CA. Currently,OML produces test modules inWR08 (90–140 GHz), and WR05(140–220 GHz). TRL calibrationdevices are also required. Thesemay be obtained in any of theabove bands from Aerowave Inc.,of Medford, Massachusetts. SeeProduct Note 8510-8A, Applying

the HP 8510 TRL Calibration for

Non-coaxial Measurements, pub-lication #5091-3645E, for detailson this calibration technique.

How does it work?Millimeter wave measurementsare accomplished by using theOML test modules in conjunctionwith the HP 85106D E12 microwavesubsystem. The interface to theHP 8510C is made through the HP 8517B H06 test set, shown inFigure 2. A standard HP 8517Btest set may be retrofitted withoption H06. When the HP 85106DE12 system is connected to a pair ofOML high-frequency test modules,the system is capable of function-ing in two modes: the HP 8517B(50 GHz coaxial) mode and theExternal Converter mode. The frequency coverage in the ExternalConverter mode will depend uponwhich OML test modules are con-nected. The mode of operation isselected by means of a modeswitch on the front panel of thetest set.

Typical system performance is shown in Table 1 for the HP 85106D E12 system when used with a pair of OML WR08(90–140 GHz) test modules.

Who to contactOleson Microwave

Laboratories

355 Woodview Dr. Suite 300Morgan Hill, CA 95037Tel: (408) 779-698Email: [email protected]: http://olesonmicrowave.com

Aerowave, Inc.

344 Salem St.Medford, MA 02155Tel: (781) 391-1555Fax: (781) 391-5338

Figure 2. The HP 8517B H06

Test Set serves as the interface

between the HP 8510C VNA

and the OML high-frequency

test modules.

Table 1. Typical

performance of the

HP 85106D E12

system with

WR08 modules.

Typical Performance with WR08 Modules

Port Dynamic Raw Raw Residual ResidualPower Range Directivity Match Directivity Match

–10 dBm 75 dB 35 dB 17 dB 50 dB 50 dB

”“ Today’s leading

technologies are moving towards higher

frequencies evenbeyond 110 GHz.


Test

& M

easu

rem

ent S

olut

ions


3

Now available: Custom frequency converter test system for the satellite and LMDS market

esigned specificallyfor the uniqueneeds of the

expanding satellite andLMDS market, the newHP Z2005A/H92 systemprovides superior mea-surement accuracy forfrequency-translationdevices such as mixersand frequency converters.The system is most valu-

able for designers, manufacturers,and system integrators of commu-nication systems (such as telemetryapplications, satellite and LMDS)in which high-accuracy groupdelay measurements are of great importance.

Using vector-error correction tomeasure these challenging devices,the custom HP system achieveshighly accurate and low-ripple mea-surements of magnitude and phase.Absolute group delay accuracy isspecified at an impressive ±150 ps,whereas alternative techniques mayhave typical accuracy of a few ns.

Advanced measurement technique This advanced absolute group delaymeasurement technique uses amixer calibration standard andvector error correction to achievehigh-precision absolute group delaymeasurements of frequency-translation devices. The advan-tages of this system include:

• Improved absolute and relativegroup delay measurement accu-racy of ±150 ps• Lower measurement ripples for highly accurate S-parametermeasurements• Vector-calibrated performance(by characterizing and removingtest system errors)

Test

& M

easu

rem

ent S

olut

ions

My Le Truong, Market DevelopmentEngineer, HP’s MicrowaveInstruments Division

Figure 1: Basic

System Configuration.

D


The basic Z2005A/H92 systemconsists of a network analyzer,customized hardware, computerwith customized software, powermeter, and synthesized source, asshown in Figure 1. The techniqueimplemented in this system con-sists of two steps. The first stepinvolves the creation of a mixercalibration standard using a vectornetwork analyzer (VNA). The mixercalibration standard, fully charac-terized by HP, is provided with thesystem. The second step involvesthe use of this mixer calibrationstandard, along with a standardcoaxial calibration kit, to calibratethe VNA-based test system. Oncethe significant error terms arecharacterized, a frequency trans-lation device can be connected (in place of the calibration mixer),and its vector-error correctedresponse is measured.



4

Figure 2. Measured Results.

The creation of a mixer calibrationstandard is achieved through anindirect measurement technique.There is no direct way to measurethe absolute group delay charac-teristics of a mixer. Instead, bymeasuring three similar mixers inpairs, the simple mathematicaltechnique for solving three equa-tions with three unknowns can beapplied. This approach has beenused by Hewlett-Packard for manyyears when characterizing the phasenoise of unknown oscillators. Thetechnique can be used to extractthe absolute response of eachmixer, thus providing the responseinformation needed (absolute groupdelay and conversion loss) whenthe mixer is used as a calibrationstandard. This mixer calibrationstandard is further characterizedby direct measurements of itsinput and output match.

The calibration mixer, along withadditional calibration standards, isused to correct for errors withinthe test system. A computer is usedto store data files (of the calibra-tion mixer and other components),extract error terms from the VNA,

modify the error terms, and importthe modified error terms back intothe VNA for error-term calculations.After calibration, absolute groupdelay and S-parameter measure-ments can be made on the fre-quency translation device undertest (DUT) with the added advan-tages of high accuracy, ease-of-use,and faster measurement throughput.Figure 2 compares the differencein measurement ripple betweenthis technique and a golden stan-dard substitution technique.1

Need more information?This measurement technology isdetailed in a new HP ApplicationNote, 1287-7, Improving Network

Analyzer Measurements of

Frequency-Translating Devices. Itis available on the World Wide Web(http://www.tmo.hp.com/tmo/

Notes/English/5966-3318E.html)

or by asking your HP sales office forHP literature number 5966-3318E.Your HP sales engineer can alsoprovide more information aboutthe HP Z2005A/H92 customizedsystem and should contact DennisPoulin at T-221-5053.

1. In the golden standard technique, you performa thorough response calibration with a referencemixer in place. Next, you enter the absolute groupdelay of the reference mixer via the electrical delayfeature of the VNA. Finally, you connect your DUTand measure its response relative to that of thereference mixer. In this technique, you must knowthe absolute group delay of the reference mixer.This may be provided by the manufacturer, but is not a typical specification. Depending on themeasurement technique used by the manufacturer,the uncertainty of their group delay measurementwill add to the uncertainty of your measurement.

Conversion Loss Absolute Group Delay

Test

& M

easu

rem

ent S

olut

ions

”

“ Absolute group delayaccuracy is specified atan impressive ±150 ps,

whereas alternativetechniques may have

typical accuracy of a few ns.

5

Your questionsanswered…

At my company we have setsome challenging goals for

reducing manufacturing costs byincreasing the throughput of ourproduction lines. How can I improvethroughput in test processes thatuse network analyzers?

For network analyzer applications, many people

focus on the sweep time of a network analyzer. It is important,however, to realize that sweep timeis only one part of the test process.And optimizing it may not be thebest way to improve throughput.

For example, consider a manualtuning process for a base station

duplexer filter. This mightinclude connecting theTx and Ant ports to theanalyzer, recalling aninstrument state and calibration, adjustingscrews to tune the filter,connecting the Ant andRx ports, recalling a newinstrument state and calibration, doing moreadjustments, checking the

Tx-Ant measurement and re-tuningif necessary, then verifying allmeasurements and disconnectingthe filter. A single iteration throughthis process would have most ofthe time taken up by adjustments,with measurement time beingabout one-quarter of the totalthroughput time.

The next item in Tips & Techniques

describes a new filter tuning methodthat could be used to reduce theadjustment time, which is themost time-consuming part of thistest process.

To reduce measurement time, youmight consider using swept-listmode. This feature (available inthe HP 8753E, 8719D, 8720D, and8722D) allows you to set up a cus-tomized sweep, and is especiallyhelpful for testing filters. You canmeasure many data points inregions of interest, and only mea-sure a few points or none at all infrequency spans where specifica-tions are not critical. You can alsoset the power level and IF band-width independently for each segment of the sweep. This allowsyou to optimize dynamic range formeasuring filter stop bands with-out slowing down the measure-ment in the pass band.

There are many other ways toimprove sweep speed besidesusing swept-list mode. ApplicationNote 1287-5 explains these methodsin more detail. The applicationnote also discusses the impact onthroughput from other parts of atypical test process—such asinstrument state recalls, use oflimit lines or markers, data trans-fers to an external controller, deviceconnection, and measurementaccuracy. It provides suggestionsfor optimizing test time and improv-ing throughput in many areas.

Q

A

Mirin Lew, Product Manager, HP ’s MicrowaveInstruments Division

Recall 2%Connection 8%

Adjustment 62%

Measurement time 28%

”

“ It is importantto realize

that optimizing sweep time may not be

the best way to improve throughput.

Increasing throughput

For the duplexer tuning

example, measurement time

is only about one-quarter of

the overall process time.



Tips

& T

echn

ique

s



I use network analyzers totune cavity resonator band-

pass filters, and it takes even askilled person quite a while totune each filter. Is there a way tospeed up this process?

HP has developed an innovativeresonator filter tuning tech-

nique that may significantly reduceyour tuning time and allow tuningto be done by operators who haveless training and experience. Thisnew technique uses the time domainresponse instead of the traditionalfrequency domain view of the filter’sreflection characteristics for tuning.

Some of the challenge in tuningresonator filters is in identifyingwhich resonator needs to be tuned,and determining how to tune it to obtain the desired response.The concept behind the new tech-nique is that the signal passingthrough a resonator filter experi-ences delay through each resonator.In time domain, this delay makesit possible to distinguish theeffects of tuning each individualresonator for a specific responsethat results in a properly tunedfrequency response.

Tips

& T

echn

ique

s

Q

A

6

Start by setting up the frequencydomain measurement. The centerfrequency of the frequency sweepmust be set to the desired centerfrequency of the bandpass filter.Choose the span to be three tofive times the bandwidth of yourfilter. A full two-port calibrationshould be performed to properlyset the reference planes for theS11 and S22 measurementsand correct for the usual systematic errors.

After calibration, you can useinterpolated error correction to look at measurements overmore narrow frequency ranges.Now, measure S11 and turn ontime domain (press [SYSTEM][TRANSFORM MENU] [BAND-PASS] [TRANSFORM ON] on an HP 8753 or 8720 family network analyzer).

Next, to determine where to setthe start and stop limits in timedomain, we need to consider thedelay of the filter. A good estimateof the one-way delay through thefilter can be obtained by measur-ing a properly tuned filter. Measuretransmission (S21) in the frequencydomain using the delay format,and look for the maximum delaywithin the pass band of the filter.A reflection signal (S11 or S22)

would experience approximatelytwo times that filter delay: oncegoing out and once returning tothe input port. For a filter with Nresonators, the delay through anyone resonator would be approxi-mately the one-way delay divided by N.

If a properly tuned filter is notavailable, the delay can be estimatedfrom the 3 dB bandwidth of thefilter (BW). The delay is approxi-mately 1/[(BW/2π) x N]. Since thisapproximation varies in accuracyfor different types of filters, addabout 20% to this value and usethat as the estimated delay.

The recommended start limit isone resonator’s delay on the nega-tive side (time < 0). Rememberingthat an S11 measurement includestwo times the actual delay, thatmakes the start = –1 x (delay/N) x 2.For the stop limit, we need to makesure we include enough time tosee the signal reflected from allthe resonators, so we recommendadding one resonator’s delay tothe delay through the full filter:stop = 2 x [(delay) + (delay)/N].

Tuning cavity resonatorbandpass filters—faster

7

Figure 1 shows the time domainresponse of a properly tuned filterthat has five resonators, with thecorresponding S11 measurementin the frequency domain. Noticethe almost periodic dips in theresponse. Beginning with the firstdip near time=0, the next N dipsrepresent the response of res-onators 1 through N in the filter.You can see this in Figure 2, wherewe have deliberately mistuned thethird resonator. The third dip is nolonger at a minimum, but the dipsfrom resonators 1 and 2 are stillvisible. Note that in the frequencydomain response, it is much moredifficult to see which resonator isnot tuned. With some experimen-tation, you will find that tuningeach dip for its minimum valueresults in a properly tuned filter in the frequency domain.

Using this information, you cannow try adjusting a filter that hasnot been tuned. Start with theresonator closest to port 1, andtune it until you see the first dipshowing up near time=0. Changeyour adjustment slowly becausethe dip in the time domain is verysensitive to tuning, and it is easyto tune right past it. Next, tuneresonator 2 until its dip reaches aminimum. There will probably besome interaction between the resonators so that the dip fromresonator 1 is no longer at a mini-mum, so go back to resonator 1and tune it again. Go back to res-onator 2 and minimize that dip, thengo on to resonator 3, and so on.

Tips

& T

echn

ique

s

Figure 1. Response

of tuned filter.

Figure 2. Filter with

resonator 3 mistuned.




8

Be sure to go back and re-tunethe previous resonator each time.Note that since energy gets lost asit goes through the filter, it maybecome too difficult to properlytune the later resonators. It is agood idea to tune half of the res-onators looking at S11, and thenswitch to measuring S22 to tunethe remaining resonators. Whenyou change to S22, you need tostart with the resonator closest to port 2 (which will now be thefirst resonator in the time domainresponse), and repeat the tuningprocess, working backward throughthe resonators from port 2 towardport 1. When all the resonator dipsin time domain have been mini-mized, turn time domain off toreturn to the frequency domainmeasurement. The filter shouldnow be properly tuned.

One difficulty with this techniqueis knowing where the dips for eachresonator should appear in thetime domain response, especiallywhen working with an unfamiliarfilter. This problem can easily besolved by starting with a properlytuned filter and storing its timedomain response as a memorytrace to create a template. Thememory trace can be displayedalong with the data trace whiletuning, so the operator can seewhere the dips should appear.

This tuning method works wellwith all-pole filters with simplecoupling, for instance, where everyresonator is coupled to the adjacentresonators. There are other filterswith more complex characteristics,such as cross-coupled filters, andfilters where the coupling betweenresonators can be tuned as well asthe resonators themselves. Forthese filters, simply minimizingresonator dips in the time domainresponse may not be sufficient.

More investigation is being doneto refine this technique for thesetypes of filters. You may want totry creating a template to definethe optimal response for operatorsto use, as described in the previous paragraph.

Since this technique is fairly simple, only moderate training isneeded for operators. The abilityto identify and optimally tuneeach resonator allows tuning ofcomplex filters to be done morequickly and accurately.

Note that although this techniquecan be used on any network ana-lyzer with bandpass time domaintransform, it is not practical unlessthe analyzer is fast enough to allowclose to real-time tuning in timedomain. The HP 8753E, 8719D,8720D, and 8722D with firmwarerevisions higher than 7.0 have anew faster CPU board that can be used for this type of tuning. HP 8753D analyzers and older 8720family analyzers can be upgradedby ordering Option 000 with theappropriate upgrade kit of HP 8753DU, 8719DU, 8720DU, or 8722DU.

Tips

& T

echn

ique

s

”

“ The ability to identifyand optimally tune

each resonator allows tuning of

complex filters to bedone more quickly

and accurately.


9


Now up to seven times faster!

ood news. TheHP 8719D, 8720Dand 8722D VNAs

now include: a new 68040CPU, two wider IF filterbandwidths (3.7 kHz and6 kHz), new faster DSPhardware, and codeenhancements whichprovide exceptionalsweep speed, error cor-rection, register recall

and data-transfer rates. Other pro-ductivity enhancements include:

Four-parameter displayMeasure and display all four S-parameters simultaneously(available October 1998). This feature can significantly reducemeasurement throughput time for tuning/testing duplexers, isolatorsand other devices. Display reflec-tion and transmission parameterswith magnitude, phase, group delay,Smith chart, polar, SWR, or time-domain formats. View results inoverlay or split-screen format on anLCD color display with one, two,or four graticules. Quickly recordor print all four S-parameters fordata archiving.

Swept-list modeSpeed up your testing by measuringat only selected frequencies. Youcan specify up to 30 arbitrary CWfrequencies or frequency sweepsegments at which to test yourdevice. You may independently settest-port power levels, number ofdata points and IF bandwidth foreach segment. Reduce test time andincrease measurement throughputby optimizing each segment toyour specific test requirements.

Fast data transfer to your CAE programThe Touchstone©-compatible(S2P) format provides fast S-parameters data transfers for newdesign and simulation applications.

Faster performance for existing features• Automate measurements quicklyand easily with keystroke recording.• Recall measurement states withthe push of a button.• Save time and avoid recalibrationwhen changing frequencies byusing the interpolative error-correction mode.• Locate and resolve mismatchesin your test device, fixture orcable using time-domain analysis.

New features improve your measurements• Save more complex calibrationdata and instrument-states withincreased nonvolatile memory. • Use the built-in disk drive toquickly and easily upgrade theflash memory-based firmware.• Accurately measure non-insertable devices with adapter-removal calibration.

To upgrade your existing HP 8720D family analyzer...Option 000 for the HP 8719DU,8720DU, and 8722DU analyzersprovides all the new performanceand features for the HP 8720Dfamily of vector network analyzers.For more information, ask yourHP sales engineer or visit us on the World Wide Web(http://www.hp.com/go/8720).

Chad Gillease, Product Manager, HP’s Microwave Instruments Division

G

Prod

uct N

ews

Calibrate electronically with HP’s new ECal modules

10

ant to speed up your calibrations? Try HP’s new Electronic Calibration

(ECal) modules for your HP 8510,8720 and 8753 series vector net-work analyzers. ECal is ideal forthe manufacturing environment.A full two-port calibration can beaccomplished with a single con-nection to the ECal module, usingminimal operator intervention.

New microwave and RF ECal modules

The new HP 85090 seriesRF ECal modules providecalibration for your HP 8753C/D/E networkanalyzers. Calibrationperformance is specifiedfrom 300 KHz to 6 GHzand typical from 30 KHzto 300 KHz.

Operating on the 1 GHzto 18 GHz or 26.5 GHz

frequency range, the improved HP 85060 series microwave ECalmodules provide calibration forthe HP 8510B (firmware revision 6),8510C, 8719C/D, 8720C/D and8722C/D. Order Option 001 to addan RF module (30 kHz to 6 GHz)for the lower frequency range ofyour network analyzer.

Both the HP 85060 and 85090 ECalmodules are available in 7 mm,Type-N, and 3.5 mm. StandardType-N and 3.5 mm ECal modulesare equipped with male and femaleconnectors and options 00M and00F are available with male-to-maleor female-to-female connectors.

Also new from Hewlett-Packard,the HP 85097A PC interface—with control software—allowsyou to control the ECal modulewith a PC during the calibrationprocess.

Availability and pricingThe HP 85090/85060 series ECal modules and HP 85097A PC interface are available now.Prices begin at $3,000 (U.S.) forthe HP 85090 series of RF mod-ules, and $6,430 for the HP 85060series of microwave modules. The PC interface (HP 85097A) ispriced at $1,400. Delivery is esti-mated at four weeks for the HP 85090 series, and ten weeksfor the 85060 series.

Sal Caruso, Product Manager, HP’s Santa Rosa Systems Division

“

W

With ECal, a full two-port calibration

can be accomplishedwith a single

connection to the ECal module, using minimal operator intervention. ”


Prod

uct N

ews

11

his is your opportunity to save 10 to 25%! HP hasreduced the prices on more

than 20 products in the HP 8510vector network-analyzer family.

Modular solutionsFor complete measurement flexibility, our high-performanceHP 8510 family is modular. A typical system is assembled with a receiver, a test set that coversyour measurement frequency, and amicrowave source. Measurementscan be performed over a widerange of microwave and millimeterfrequencies. Accommodate yourchanging measurement needs bychanging or upgrading compo-nents of the system.

Price Reductions (in U.S. dollars)

Product Current U.S. Price New U.S. Price % Reduction

HP 8510C vector network analyzer $ 41,515. $ 31,800. 23%HP 8517B 50 GHz S-parameter test set $ 47,530. $ 35,035. 25%HP 83651B 50 GHz source $ 47,075. $ 37,000. 21%HP 85107B 50 GHz system $150,500. $120,000. 20%

T

New, lower pricing on HP 8510instruments and systems


Prod

uct N

ews

Millimeter Upgrade Promotion

If you currently own an HP 85106C/D or HP 85109C and want toachieve single sweep, on-wafer measurements up to 110 GHz, youcan trade in your used V/W8510As and HP 85105As for credittowards the purchase of an HP 8510XF or HP 8510XF upgrade. This promotion expires March 31, 1999.

HP 8510A or 8510B analyzer, youmay now upgrade to the higherperformance HP 8510C for just$16,250 or $12,463, respectively.You’ll save 23%.

Look on the Web for more product savingsA complete list of productsincluded in this price reduction isfound on the World Wide Web at: http://www.hp.com/go/8510. Or as always, you may call yourlocal HP sales office.

Complete rack systemThe HP 85107B, on the other hand,is a complete factory-integratedtest system mounted in a rack. Itincludes a calibration kit and testport cables, on-site installation,one day of engineering consulting,and one full year of on-site ser-vice. Frequency range is 45 MHzto 50 GHz.

Upgrades reduced, tooIf you have an existing HP 8510system, now is the time to upgradeat reduced prices. For instance,U.S. customers with 20 GHz or26.5 GHz systems may now upgradeto 50 GHz systems for $72,035 (a23% reduction). Or if you have an

Find us on the Web,beginning Spring 1999!

The HP 8510/8720 News is published regularly by the Santa Rosa Systems Division of Hewlett-Packard. Please send queries, submissions and comments to:

HP 8510/8720 News, MS 3US-M1400 Fountaingrove ParkwaySanta Rosa, CA 95403-1799, U.S.A. FAX (707) 577-2108

Please note Hewlett-Packard reserves the right to use or edit submissions.

For more information

about Hewlett-Packard test and

measurement products, applications

or services, and for a current sales

office listing, visit our Web site,

http://www.hp.com/go/tmdir. You

can also contact one of the following

centers and ask for a test and

measurement sales representative.

United States:

Hewlett-Packard CompanyTest and Measurement Call CenterP.O. Box 4026Englewood, CO 80155-40261 800 452 4844

Canada:

Hewlett-Packard Canada Ltd.5150 Spectrum WayMississauga, Ontario L4W 5G1(905) 206 4725

Europe:

Hewlett-PackardEuropean Marketing CentreP.O. Box 9991180 AZ AmstelveenThe Netherlands(31 20) 547 9900

Japan:

Hewlett-Packard Japan Ltd.Measurement Assistance Center9-1, Takakura-Cho, Hachioji-Shi,Tokyo 192, JapanTel: (81-426) 56-7832Fax: (81-426) 56-7840

Latin America:

Hewlett-PackardLatin American Region Headquarters5200 Blue Lagoon Drive, 9th FloorMiami, Florida 33126, U.S.A.(305) 267 4245/4220

Australia/New Zealand:

Hewlett-Packard Australia Ltd.31-41 Joseph StreetBlackburn, Victoria 3130, Australia1 800 629 485

Asia Pacific:

Hewlett-Packard Asia Pacific Ltd.17-21/F Shell Tower, Times Square,1 Matheson Street, Causeway Bay,Hong KongTel: (852) 2599 7777Fax: (852) 2506 9285

Copyright © 1998

Hewlett-Packard Company

Printed in U.S.A. 10/98

5968-1911E

Dear reader:Good news! Beginning with our March 1999 issue of HP 8510/8720

News, you’ll be able to find us on the World Wide Web. This meanseasy access for you, and quicker delivery of those important productarticles. You may even notice our new look— a new design, and lots of color!

When you’re at the newsletter web site, you’ll also be able to “link” to:• Previous issues of the 8510/8720 newsletter• Free firmware upgrades• Training news…and much, much more. Come visit us on the “WEB” and you’ll see!

We’ll send you an e-mail twice a year, letting you know the new issueis available. You will find the newsletter at:

http://www.hp.com/go/8510-8720news

Making VNA What you need measurements Also in this issue …nagui/Appnotes/Agilent/Network... · 2002. 11. 29. · 2 These test modules are supplied by Oleson Microwave Laboratories

Documents