--320

.THP3 AC MEASUREMENT 51 MESURES EN CA

IN'J ERCOHPARI80lf 01' THERMAL CONVERTERSAT NIM, N :ST, PTB, SIRI, AND VSL FROM 1 TO 100 KHZ

o. Rebuldela-Hatl. Inst.

of stds. , Tech.Boulder, CO USA

D.-X. Huang.Shaoxing Xndust.

Res. Xnst.Shaoxing, Chilla

J .R. Kinard-Nat!. Inst.

of Stds. , TecJ..Gaithersburg, ~I USA

D. JanikPhys. Techn.

BUDd. Anst.Braunschweig, cermany

Abstract

Coaxial, thermal voltage convErters(TVC's)have been intercompared betweer. NIM, NIST,PTB, SIRI, and VSL in the frequenc:y range from1 to 100 MHz. This paper briefly describesthe methods and underlying princi .les on whichRF-dc difference determinations !re based ineach laboratory and gives the rEsults of theintercomparisons.

Introduction

Coaxial TVC's are commonly u,.ed for themeasurement of voltage in the fro,quency rangefrom 1 to 100 MHz. An intercomparison ofRF-dc differences in this fre)uency rangebetween the participating lab,.ratories isparticularly informative becaUSE the methodsand hardware differ significcntly. Theagreement between methods adds considerableadditional confidence to the process. In theseinter-comparisons, NIST-Gai thers burg was thepilot laboratory.

Methods of Individual LaboJatories

National Institute of Metroloqv(JIML. At NIM,TVC's have been constructed in ., special wayin order to achieve flat frequ. :ncy responsewithout component compensatiol. In thesespecial TVC's the thermoelement js placed nearthe input connector, to measu -e the inputcurrent approximately, and the r!nge resistoris connected to the shield(1] . The RF-dcdifferences of the TVC's have bEen determinedfrom 30 MHz up to 1 GHz by rEference to abolometer bridge with dual, thin-filmthermistors . The uncertain ~y of thisbolometer bridge at NIM is <!l\ "rom 10 MHz to3 GHz(2]. The special TVC's hive also beencompared to conventional TVC's at lowerfrequencies.

Hatl. Inst. of Stand. and rech. (HIST-Boulder) . Two techniques have been used toestablish the RF-dc difference of the NIST-Boulder working standard set 0" TVC' s. Thefirst of these involves the use of thebolometer bridge (3]. Although the bridge isno longer used, the RF-dc di fferences anduncertainties derived have beer. confirmed by

comparison with voltages a;signed fromimpedance measurements based on impedance

bridges and power measurement; based on amicrocalorimeter from 100 MHz to 1000 MHz.The second technique establishe~the frequencyresponse below 100 MHz experimentally andtheoretically as described in (4] and [5].These results derived at low-f~equency agreewith the characterization of a I-V TVC basedon the bolometer bridge from 1 10400 MHz. Onthe basis of the derivation o' voltage frompower and impedance measureDents and theagreement with theoretical pJ"edictions forTVC's directly, working standard TVC's fordetermination of RF-dc differe1ce from <1 to1000 MHz have been characteriz.d.

.Uectronfc and tlK. (ng. lab.. u.s. o.."t.. 0' CD!8... TrchnoLogyAOi8in.

.Currentty wo,.king for hUlll"ltine labs. Inc. end II . hlurch Auoci.te8t NIS'.. U.S.. Co rrwlent wor" not prot..Cled by I .$. copyright.

Z. ZhenHatl. Inst.

of MetrologyBeijing, China

J. de VreedeVan SwindenLaboratory

Delft, Netherlands

Nat!. Inst. of Stand. and Tech. INIST-Gaithersburq). Using low frequency standardsand nearly frequency-independent structures,the RF-dc differences of coaxial TVC's havebeen determined at 1 MHz (5]. To extend thefrequency range, TVC's having cylindrical,carbon film range resistors of 1 to 5 kn wereconstructed using designs and materials chosento permit detailed modeling and analysis. Forthese TVC's, the major contributions to RF-dcdifference which are range independent are thevoltage standing wave in the input connectorstructure and the current standing wave in thethermoelement section. Major range dependentcontributions are the transimpedance of therange resistor and skin effect of the overallstructure.

Phvsikalisch-Technisc~e BundesanstaltlPTBI.

At PTB transfer difference measurements in theRF range are made in two steps. Because thereis a separate set-up at PTB for measuring theac-dc transfer difference up to 1 MHz withvery small uncertainties, the RF-dc transferdifference is calculated from the RF--acdifference and the ac-dc difference measured

at 100 kHz. The RF voltage in the deviceunder test is compared to an ac voltage of 100kHz. For frequencies above 1 MHz, acommercial coaxial calorimeter of the dual-load type is used as the primary voltagestandard. With this standard containing twonearly identical resistance mounts the RFvoltage is compared with an equivalentadjustable DC voltage. The effectiveefficiency and the input conductance of thecalorimeter has been determined as a functionof frequency up to 1 GHz, thereby the RFvoltage can be derived from power andimpedance standards. The automated primaryvoltage standard for voltages between 0.2 Vand 1 V is described in (5].

For the RF-ac transfer difference calibrationof thermal converters in the range between 0.5V and 10 V and the frequency range up to 100MHz an automated measuring set-up is used (6].The set of working standards between 0.5 V and10 V is calibrated with a step-up procedurestarting with the 1 V converter which wascal ibrated referred to the primary vol tagestandard.

Shaoxinq Industrv Research InstitutelSlRll-At SIRl, the RF-dc differences of TVC's havebeen determined by experimental analysis andtheoretical calculations (8]. TVC's rangingfrom 0.5 V to 5 V have been measured bythin-film bolometers (3] from 30 MHz up t~300 MHz. An empirical equation, d=do+AJf+B!where do is the ac-dc difference at audIofrequency and A and B are constants to ~edetermined, was used to make a non-linear fItemploying the least squares method. Thus, theRF-dc differences from 100 kHz to 100 MHZ for

individual TVCs were determined. D~rec~comparisons between adjacent ranges conflr~ethis modelling (9]. Additional confirmationof these primary TVC's was obtained .bYcomparison with UHF, thin-film multijunctlOnthermal converters having RF-dc differences

<0.5\ from 10 Hz to 1 GHz, and by comparison

with TVCs having multifi: ar ground-leadcompensation and also having F F-dc differences<0.5\ from 10 Hz to 300 KHz.

van ewin~en LAborAtorv(VSL). In VSL the RF-dc difference standards above 1 MHz have beenestabl ished in terms of a s,:t of calculablestandards and can be used up to 30 KHz (10).Recently a new set of 4-V standards, withsmaller but not necessarily ~egligible RF-dcdifferences, was constructec for use up to100 MHz (11). All standards; .re provided withtype-N male connectors and tte symmetry planeof a type-N tee is taken .tS the plane ofreference. In the automited measurementsystem, each thermoelement ou:put is amplifiedand then the difference betwe~n the outputs isamplified. The RF-dc dift erences of eachconverter have been . :alculated andintercomparisons between ~'C's yield goodagreement within the combine.. uncertainties.

Transfer Standards anS Results

This intercomparison was pezformed using twotypes of TVC's. The first type wascommercially manufactured wlth 1-, 5-, and10-V ranges containin) UHF-patternthermoelements, internal shield structures,and conventional series resi~tors. The secondtype was a VSL-constructed, coaxial, 5-V TVCcontaining a thin, straight lire for a seriesresistor.

The results reported by eacl, laboratory wereaveraged to produce a represEntative value for

that laboratory at each frequency. Therepresentative values, or de/iations, for allthe laboratories at each frequency, were

averaged and the data are pl(tted (see Fig. 1)relative to that overall overage which isindicated as 0\ RF-dc difference.

R~terences

[1) Z. Zhen, L.-J. Wang, J.-L. Yan, and S.-L.weng, "Precise RF Voltage Standard Using aCoaxial Thermal Voltage :onverter up to1000 KHz," IMEKO XI, Metrol >9Y Vol., p. 519,Oct. 1988.

(2) L.-J. Wang, Z. Zhen, C.' L. Guo, and J .-L.Yan, "A precision voltage ~ tandard at RF andMicrowave Frequencies," 'CTA Metroloaica

100MHz

~, vol. 1, no. 1, 1981.

(3) M.C. selby and L.F. Behrent, "A Bolometer

Bridge for Standardizing Radio-FrequencyVoltmeters," Natl. Bur. Stand.(U.S.\ J. Res.,

RP2055, vol. 44, Jan 1950.

(4) F.L. Hermach and E.S. williams,Voltage Converters to AccurateMeasurements to 30 Hegacycles perAlEE Trans (Comm. Electron.\, vol.200-206, July 1960.

"Thermal

VoltageSecond,"79, pp.

[5J J.R. Kinard and T.-X. Cai, "Determinationof AC-DC Difference in the 0.1-100 KHz

Frequency Range," IEEE Trans. Instrum. Meas.,vol. 38, pp. 367, April 1989.

(6) H. Gierke, L. Grno, D. Janik, and K.Mnter, "Automatic RF Voltage Calibration witha Primary Voltage Standard up to 1 GHZ," to bepublished in the CPEH '92 Diaest.

(7) D. Janik, K. Munter, and F. H. Lou,"Automatisierung des Hochfrequenz-Spannungsmepplatzes der PTB, in Fortschrittein der Hochfrequenz-meptechnik," PTB BerichtE-38, pp. 3-20, 1990.

(8) D.-X. Huang, M.-L. Chen and S.-Z. He, "RF-dc Differences of Coaxial Thermal Standards,"

IEEE Trans. Instrum. Heas., vol. IM-39, pp.313-317, April 1990.

(9) D.-X. Huang and S.-Z. He, "ExperimentalStudies of the RF-DC Differences of VoltageStandards," IEEE Trans. Instrum. Meas., vol.IM-39, pp. 10-14, Feb. 1990.

(10) M. Nomair and K.J.P.M. Harmans, " HighAccuracy Calculable AC-DC Transfer Standardsfor the LF-30 MHz Frequency Range," ~Trans. Instrum. Heas., vol. 38, pp. 342-345,April 1989.

(11) H. Nomair and K.J .P.M. Harmans, "A HighAccuracy Calculable AC-DC Transfer Standard,"VSL Report ES-88-04 (implementationin annex),1989.

50 MHz0.62

0.31

-0.31

-0.62NIM NIST NIST PTB SIRI VSl

Sou Gai

10 MHz0.16

0.12

0.08

tg 0.04

; 0:c0_0.04u~ .0.08u.cr

.0.12

.0.16NIM NlST NIST PTB SIRI VSl

Bcu Goi

,, ]0 ..: '- ..."" "

1.2

l .. l. 0.'t

uj 0 i

;eis 0u -0.4 u..,

;r,u.cr .0.8 cr

-1.2NIM NIST NIST PTB SIRI VSl

Bcu Gai

30 MHz0.35

_ 0.25I I! 015 I I I l.

j 0.05LL L I I I ;d -0.05uJ. -0.15cr

.0.25

.0.35NIM IST NIST PTB SIRI VSl

Bcu Gai