2-7 Calibration of SAR Probe

Transcription

1 Reserch nd Development of Clrton Technology -7 Clrton of SAR Proe Lr HAMADA nd Soch WATANABE The clrton of spefc sorpton rte proe s nmely clrton of the electrc feld mesurement sensor n the phntom lqud of hgh electrc loss. The stndrd electrc feld method tht uses the rectngulr wvegude flled wth phntom lqud s wdely used for the proe clrton. NICT s only nsttute to serve the clrton for the proe. In ths text, the clrton procedure of ths wvegude clrton technque s descred, long wth the detled procedure, the clrton result, nd the uncertnty evluton. Moreover, ths text ntroduces other clrton technques. 1 Introducton Spefc Asorpton Rte (SAR) s one ndex for ssessng electromgnetc feld exposure from devces nd equpment tht emt electromgnetc wves such s mole wreless devces nd se sttons. In Jpn nd overses, SAR lmtton vlues re estlshed sed on rdo-rdton protecton gudelnes [1] [6], nd re mnly ppled to wreless communcton devces used ner the humn ody, such s mole phones. Therefore, for wreless communcton devces such s mole phones, complnce evluton to rdo-rdton gudelnes re legslted under the stndrds nd regultons n Jpn nd overses [7] [10]. Mesurement equpment used n tests of wreless communcton devces must e clrted perodclly, sed on regultons. In tests to ssess complnce wth SAR lmt, SAR proe s used to mesure SAR, nd the SAR proe must e clrted. NICT s the only nsttuton performng SAR proe clrton servce n Jpn. Ths pper descres SAR proe clrton t NICT. Severl SAR proe clrton methods hve een stndrdzed n Jpn nd overses. Explned here re the prnples of three mjor methods, ncludng detled steps nd exmples for mesurement nd uncertnty evluton for the wvegude clrton method used n the clrton work of NICT. Defntons [7].1 Rdo-rdton protecton gudelne [1] Recommended gudelnes for sfe condtons under whch the exposure level of electromgnetc wves s consdered not to cuse possle undesrle ologcl effects on the humn ody (consderng the frequency rnge from 10 khz to 300 GHz).. Spefc Asorpton Rte (SAR) [1] SAR s electrc power sored y ologcl tssues (lossy delectrc). SAR n lossy med (such s phntom lqud) s relted to oth the electrc feld (E) nd the grdent of chnges over tme of temperture (dt/dt) n med. Therefore, the equton elow s provded sed on ths reltonshp. E dt SAR ck t0 (1) dt Here, the vrles re: σ Conductvty ρ Densty of the med C k Spefc het From Equton (1), n electrc feld n lossy med cn e mesured ndrectly y mesurng the temperture grdent n tht med. A temperture proe tht hs fst response tme (less thn 1 second) t hgh sptl resoluton ut does not dstur the electromgnetc feld (optcl fer proe or thermstor proe usng resstnce wre) cn e used..3 Locl exposure gudelne [1] The gudelne s used for electromgnetc felds tht hve prtl sorpton on prt of the humn ody, cused y electromgnetc wves emtted from, for exmple, wreless devces used very close to the humn ody. 135

2 Reserch nd Development of Clrton Technology.4 Locl SAR [1] SAR s provded s vlue per very smll volume. It dstrutes sptlly dependent on the locton n the otssues nd exposure condtons of the electromgnetc wves. For ths, SAR verged over the loclzed volume contnng certn mss of tssue such s 1 g or 10 g s clled locl SAR; the mxmum vlue s clled locl pek SAR. However, n ssessng complnce wth gudelnes to protect electromgnetc feld exposure n Jpn, the verge volume s defned s 10 g cue of tssue, nd ssessed s tme -verge durng 6 rtrry mnutes..5 Phntom [1] Ths s qus-humn model used to estmte SAR expermentlly. If the unform mterl s used cross the entre model, then t s clled unform phntom. If t models the electrcl chrcterstcs for ech correspondng tssue, then t s clled n nhomogeneous phntom. SAR mesurements use unform phntom comprsed of n outer shell (contner) to model the humn ody shpe, nd the lqud (phntom lqud) s flled nsde the shell. 1 Delectrc constnts of phntom lqud: Exmple of stndrd vlues Frequency Conductvty σ εr (MHz) (S/m) Numers n tlcs re otned y nterpolton etween frequenes Prese scnnng rm Proe Smll e-feld sensor Phntom shell DUT (Hndset etc.) FFg. Bse stton smultor Rdo nechoc chmer or shelded room Phntom lqud Holder 1 Dgrm of SAR mesurement equpment Therefore, SAR proe clrton lso requres determnton of the clrton fctor n phntom lqud..6 Phntom lqud [1] Electrcl chrcterstcs of phntom lqud flled n the phntom re n ccordnce wth stndrd vlues of frequenes of SAR mesurement methods such s n IEC [8], whch estlshes stndrd vlues t severl frequenes n 30 MHz 6 GHz. Vlues of frequenes not n the stndrd re gven y lner nterpolton. Tle 1 shows delectrc constnts t clrton frequenes of NICT..7 Boundry effect [11] Chnge of the senstvty of electrc proe cused y the exstng oundry etween dfferent delectrc medum. When, for exmple, the proe tp s ner the delectrc contner surfce, ths effect ppers s the proe senstvty chnge. 3 Structure nd clrton fctor of SAR mesurement proe [1][13] Frst, Fg. 1 shows n ctul SAR mesurement system for wreless devces, nd Fg. shows typcl proe structure. A SAR proe s n electrc feld mesurement sensor used n lossy phntom lqud. Therefore, the clrton fctor (or senstvty ) s the vlue tht reltes the ndent electrc feld nd output voltge of the SAR proe; ths s not dfferent from ordnry ntenn clrton n free spce. However, s the clrton fctor of the SAR proe s ffected y the delectrc constnt of surroundng med, the clrton fctor n the phntom lqud hs to e determned n SAR mesurements. The detls of the SAR mesurement methods re left for Chpter 3-3, nd plese refer to the chpter to see how the SAR proe s used n the mesurement. A typcl SAR proe structure hs sensors comprsed of Schottky dode nserted etween elements of three tny dpole ntenns ( = 0 to ), rrnged n trngulr structure for 3-xs orthogonl ntenn [11]. The output voltge V of the SAR proe n proporton to the squre of the electrc feld strength E n the poston of ech sensor due to the squred detecton chrcterstcs of the dodes. The electrc feld strength E n centers of the sensor s gven y the sum of the squres of the electrc feld strength tht these three sensors receve. Here, K s the clrton fctor of ech sensor. 136 Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)

3 -7 Clrton of SAR e Smll dpole sensor Dode Hghly resstve lne Element: Shorter thn 5 mm 54.74o -em I -em FFg. Structure of SAR proe E 0 E 0 V K The voltge U ctully detected n ech sensor s ffected y the non-lnerty of dode compresson DCP. Thus, s n exmple, correcton (clrton of DCP ) y Equton (3) s performed s per elow nd converted nto the voltge V ssumed n Equton (). U V U 1000 DCP Also, the clrton fctor K cn e expressed s the senstvty NF of ech sensor n free spce, multpled y the converson fctor whch s the rto of the senstvty n free spce to tht n phntom lqud. K () (3) NF Fctor (4) The converson fctor s ssumed to not depend on ech sensor. Thus, y tkng the detecton voltge U of the SAR proe t the clrton poston nd pplyng t nto Equton (5) elow, the Fctor cn e clculted. Fctor E 0 NF 1 U U 1000 (5) DCP If the SAR proe hs three sensors s shown n Fg., these steps cn e performed: frst the senstvty NF n free spce of the three sensors n rectngulr wvegude or t the TEM cell s otned, then the senstvty n phntom lqud s otned, nd then fnlly the clrton fctor K s determned. NICT provdes K s the clrton fctor. 4 Prnples of proe clrton Here, the prnples of typcl SAR proe clrton methods re explned. 4.1 Proe clrton system tht uses clrton wvegude [8] [10] In SAR proe clrton, the wvegude method s wdely used. A stndrd electrc feld s generted n rectngulr wvegude flled wth phntom lqud, the proe s nserted n tht, nd the electrc feld nd proe output voltge re compred. Fgure 3 shows n outlne of the clrton system. The lower prt of the wvegude s flled wth r, nd sc mode (TE 10 ) s propgted upwrds from the coxl wvegude converter of the lowest prt. There s delectrc sl etween the lqud nd propgtng prt of the wvegude. It s desrle tht the depth of the phntom lqud e t lest three tmes of the skn depth δ t the wvelength n the phntom lqud. The phntom lqud s hghly lossy med, nd therefore the delectrc constnts nd thckness re djusted so the delectrc sl ecomes qurter-wve plte n order to suppress reflecton etween the r nd the phntom lqud. The clrton wvegude currently used s mnufctured to hve 10 to 0 db or less reflecton t the nput port of the coxl wvegude converter, whle n stte flled wth phntom lqud. The delectrc sl cn hve wde vrety of structures nd mterls, however, NICT sclly uses sl mde of sngle lyer of polyetheretherketone (PEEK) resn to otn mpednce mtchng wth thckness of round 137

4 Reserch nd Development of Clrton Technology PC Power source Amplfer Attenutor Shelded room Postonng system R Power sensor Power meter Input port Coupler F Power sensor Power meter Scnnng rm DAU WG Proe electrc feld. As long s the reflecton of the lqud-flled prt due to the delectrc sl s suppressed suffently, round the center of the clrton wvegude flled wth phntom lqud, the n the perture surfce s lmost the sme s the TE 10 mode of the rectngulr wvegude of the lower prt. There s lso ttenuton due to the loss of the phntom lqud, therefore t hs n exponentl ttenuton grdent n the drecton of ts depth (z drecton n Fg. 3). The mgntude of the stndrd electrc feld to the depth drecton z long the center xs cn e otned from nput voltge P on y the followng equton. 4 Pn z Ez exp (6) σ FFg. Proe Phntom lqud Delectrc sl 3 SAR proe clrton system usng clrton wvegude, nd schemtc dgrm qurter wvelength. However, n order to ctully otn suffent mpednce mtchng, the effects of the phntom lqud delectrc constnt, must e lso consdered nd therefore prototypng nd numercl smulton nd so on must e performed to mnutely djust ts thckness. Consderng the envronment of how t s used, the sl must lso resst wter nd corroson, nd therefore oth sdes re coted wth polymde flm. As scrtches nd ules on the flm surfce cn cuse nccurcy of the clrton result. Consderng the cutoff frequency, lower frequenes of clrton need lrger wvegude sze. Thus, consderng hndlng of the mesurement equpment, ths method suts 700 MHz or hgher frequenes. In SAR clrton, the SAR proe mesures the electrc feld n phntom lqud, nd clrton s performed y correltng the output voltge of the proe wth stndrd Pn >3δ DAU:Dt Acquston Unt z x Here, s the wvegude long sde, s the wvegude short sde, δ s penetrton depth of the phntom lqud, σ s conductvty of the phntom lqud, nd z s dstnce from the surfce of the delectrc sl. As the physcl quntty SAR hs no prmry stndrd, trcelty to the ntonl stndrd s ensured y usng clrted power meter to the mesurement of nput power P on. Also, ner the center of the wvegude, electrc feld of TE 10 mode s domnnt; thus, electrc feld n the short sde (y) drecton s lmost constnt. Consequently, poston z cn e determned usng E (z) n Equton (6) nd y pproxmtng the n the long sde drecton (x) usng Equton (7) elow. π x E ( x) E ( ) cos (7) z 4. Procedure for proe clrton usng clrton wvegude The clrton procedure performed t NICT re shown elow. The dt ws otned usng commerllyvlle SAR mesurement system (DASY5 mnufctured y Schmdt & Prtner Engneerng AG). 1 Delectrc constnt mesurement of phntom lqud The delectrc constnts (permttvty nd conductvty) of the phntom lqud re mesured. If the mesured vlue devtes from the trget vlue y ±3 % or more, then the delectrc constnts of the phntom lqud must e djusted. Mesurement system preprton As shown n Fg. 3, the clrton wvegude s nstlled ner the scnnng root rm so the perture surfce s prllel to the floor. In ths cse, the 138 Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)

5 -7 Clrton of SAR e wve propgton drecton n the wvegude s the z drecton n Fg. 3. The clrton wvegude s flled wth phntom lqud. 3 S 11 Mesurement of clrton wvegude A network nlyzer s used to mesure S 11 t the nput port (see n Fg. 3) sde of the clrton wvegude. Ths S 11 vlue s used to determne the nput power of the clrton wvegude when the clrton fctor s clculted s descred n 6 nd 8. 4 Adjustment nd determnton of nput power Frst, connect power meter drectly to the wvegude-sde port of the drectonl coupler, nd djust the output of the sgnl genertor so the output power ecomes the requred level. Then, record the vlue (offset) of power meter t the port F (output of power n the forwrd drecton from the drectonl coupler), then remove the power meter tht s drectly connected to the wvegudeport, nd connect the wvegude drectly to the drectonl coupler. Durng clrton, the output power s fne-tuned so the red vlue of power meter F connected to the drectonl coupler s the sme s the offset vlue otned ove. Input power nto the clrton wvegude P on s 4.0 dbm (1 GHz or lower) or.0 dbm (1 GHz or hgher) ± 0.1 dbm. The S 11 of the wvegude s used to otn the nput power of the wvegude long wth the red vlue of power meter F. The red vlue of the power meter of the R port s not drectly used to control the nput power, rther t s used to check whether excess power s not reflectng durng clrton. 5 Instructon to the SAR mesurement system of the poston of the clrton wvegude Frst, the SAR mesurement system s lunched, the SAR proe s normlzed (vertcl djustment of the proe xs) y usng the lser dsplcement meter n the DASY system. Next, nstructons re gven to the SAR mesurement system usng the reference pont t the upper end of the clrton wvegude. 6 Preprton of clrton dt sheet For dervton of the clrton fctor K y curve-fttng e (descred n 8 elow) to the mesured vlues nd theoretcl vlues (Equton (6)), dedcted dt sheet s used for post-processng of dt. Frst, n preprton, the followng prmeters re flled n the clrton dt sheet. Delectrc constnts of the Phntom lqud: Use mesured vlue. S 11 of clrton wvegude: Use mesured vlue. rto for ech 3-xs sensor (free spce): Input 1 s ntl vlue. Boundry effect correcton prmeter: Use the vlue n the mnufcturer s proe clrton certfcte. It s lso possle not to pply oundry effect correcton, however, f correcton s not used, then mesured vlues my tend to e excessve ner the surfce [14]. Dode compresson prmeter (DCP) correcton fctor: Use the vlue n the clrton certfcte of the mnufcturer. The DCP correcton fctor does not depend on condtons of the surroundngs (r, phntom lqud); thus, clrton n free spce (TEM cell or wvegude) s possle. 7 Runnng the clrton jo The clrton jo sequence descred elow s run fve tmes, nd the mesurement dt (output voltge vlues for ech sensor) s exported to the dt sheet. () Proe s moved to the spefed poston (detled descrpton n 8) of the wvegude center. () The touch sensor connected to the proe s used to detect the delectrc sl surfce (lowest surfce of phntom lqud = z = 0 mm). Here, t must e consdered tht the ctul mesurement poston s offset n the z drecton from the proe tp of the sensor (1 mm on the EX3 DV4 proe). (c) The RF output s turned on. (d) Axl sotropy t z = 5 mm (mesure t ech 15 degree rotton ngle) s mesured. (e) Z-drecton (wvegude depth drecton) scn (ech 1 mm) of root rm s executed. Scn rnge of the z scn s chnged ccordng to the frequency (up to mxmum 80 mm). (f) After mesurements re completed, system nose wthout RF nput s mesured. 139

6 Reserch nd Development of Clrton Technology 8 Dervton of clrton fctor On the clrton dt sheet, the optmzton functon such s n Excel Solver s used, nd the mesured dt s numerclly curve-ftted to mtch the theoretcl formul for electrc feld. For exmple, n the solver functon n Excel, the Generlzed Reduced Grdent method (GRG) s used s n optmzton lgorthm. The curvefttng steps re descred elow. () Output dt of three sensors s mported (z drecton nd xl sotropy dt). () Correcton of dode chrcterstcs (DCP). (c) Dt rnge used s set to ft curve (z drecton). Ner the delectrc sl, chnge of senstvty s s not neglgle due to the oundry effect cused y electrcl couplng etween the proe tp nd delectrc sl [11]. On the other hnd, t postons fr from the delectrc sl, receved feld strength s declned due to loss of phntom lqud. Therefore, t s mportnt to choose fttng rnge where these effects re smll. NICT currently uses the curve fttng dt rnges whch determned emprclly n Tle, nd SN rtos re evluted t postons n the mddle pont of these curve fttng rnges. The settng of these curve fttng rnges s dscontnuous etween ech frequency nd, however, there re plns to optmze t n the future. The stndrd szes of the wvegudes re descred n the Tle. (d) For the syntheszed electrc feld of 3-xs sensor output otned from Equton (), the clrton fctor K s optmzed nd the curve s numerclly ftted so t fts towrd the z drecton ttenuton curve of the theoretcl electrc feld n Equton (6), nd the clrton fctor K s determned so the curve fttng reltve errors re less thn (e) The clrton fctor K from the verge of fve mesurements s determned. 4.3 Clrton system usng temperture rse [8] [9][15] Temperture rse, electrc feld strength nd SAR hve proportonl reltonshp, s shown n Equton (1), f therml dffuson due to therml conducton nd therml emsson nd so forth s neglgly smll. Tht s, f the spefc het nd nput power of the phntom lqud re known, clrton of n SAR proe cn e performed usng temperture rse. A coxl type clrton system s n exmple of clrton system usng ths prnple [8][15]. Ths devce ws developed y the Ntonl Physcl Lortory (NPL) med t pplcton to proe clrton round 150 MHz or less. Fgure 4 shows dgrm of the clrton devce. Structure of ths devce s coxl wvegude of 50 Ω mtchng, nd ts end s shorted y metl plte nd prt of the center conductor s replced y lqud contner. In the trget frequency nd, phntom lqud ehves lke conductor, nd therefore reltvely unform electrc feld cn e otned ner the center of the contner flled wth phntom lqud. Actully, t s fesle to Proe nserton hole Short plte Phntom lqud contner Inner conductor Outer conductor Input N-type connector FFg. 4 Coxl clrton system [8][15][16] Frequency (MHz) Wvegude stndrd(eiaj) Curve fttng rnge z (mm) Rnges of curve fttng dt for wvegude clrton WRI 9 WRI 14 WRI WRI 40 WRI Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)

7 -7 Clrton of SAR e pply frequenes from severl MHz to up to out 450 MHz. 4.4 Clrton system n phntom lqud usng stndrd ntenn [13] A clrton method tht does not use clrton wvegude hve een developed the stndrd electrc feld method usng stndrd ntenn wth known structure nd rdton chrcterstcs ws used wth contner flled wth phntom lqud. Fgures 5 nd 6 re dgrms of clrton systems. Frst, gn of the stndrd ntenn s otned s follows. The steps re, frst, n dentcl pr of ntenns re nstlled n contner flled wth phntom lqud, the dstnce etween the ntenns s chnged, then S 1 s otned t ech dstnce, nd gn of the stndrd ntenn nstlled on the tnk ottom surfce s otned (Fg. 5). Next, the proe to e clrted s ttched to n opposng root rm, ts output s mesured, nd ntenn clrton s performed (Fg. 6). When determnng the gn of stndrd ntenn, use the reflecton loss nd chrcterstcs of dstnce ttenuton etween ntenns, nd use the two-ntenn method wth the Frs trnsmsson formul extended to lossy med. In phntom lqud, ts shrp ttenuton dffers from n free spce; thus, t s dffcult to perform clrton n the fr feld regon lke n usul ntenn clrtons. Therefore, consderng ttenuton n phntom lqud nd effects n very ner feld, the Frs trnsmsson formul ws extended untl the Fresnel domn. As the stndrd ntenn n phntom lqud, dpole ntenn tht hs smple structure wth smple theoretcl nlyss nd detched wvegude ntenn were proposed. Mesurement System Mesurement System Hed tssue equvlent lqud Proe Network Anlyzer r Antenn Hed tssue equvlent lqud Antenn SG + Amp Antenn Coupler Power Meter FFg. 5 Asolute gn clrton of stndrd ntenn n phntom lqud FFg. 6 Proe clrton usng stndrd ntenn n phntom lqud Source of uncertnty Repetlty of mesurements Devton from stndrd vlue of delectrc constnt (ε r or σ) Uncertnty of network nlyzer nd other Comned expnded uncertnty 3 Uncertnty of lqud delectrc constnt mesurement C u = (/) (c) Uncertnty vlue Stndrd uncertnty Norml 1 1 N-1 Unform 3 1 N-1 Unform

8 Reserch nd Development of Clrton Technology 5 Exmple of SAR proe clrton uncertnty evluton y clrton wvegude method Descred here s n exmple of n ctul SAR proe clrton result nd evluton of uncertnty of method usng clrton wvegude. The delectrc constnts of phntom lqud nd severl other tems re frequency dependent; thus, evluton of ech frequency s requred. Ths evluton method of delectrc constnt comples wth IEC Uncertnty sources nd method to evlute ech tem (1) Input Power The vlue of uncertnty n the clrton certfcte of the termnl power mesurement power sensor ws used. A norml prolty ws ssumed. () Mslgnment of Clrton Wvegude The tolernce vlue for mslgnment of the clrton wvegude nput end cn e derved from Equton (8) elow. Therefore, n ths study, mesurements were mde for the output termnl reflecton of the termnl type power sensor, devce nput end nd drectonl coupler, nd evlutons were mde on mslgnment. A U-shped prolty ws ssumed. (1 p ) (1 s c ) M 1 (8) (1 ) s Here, M s the mxmum msmtch tolernce vlue, Γ S s the reflecton t the output end of drectonl coupler, nd Γ P s the reflecton of the termnl power sensor. (3) Dt Acquston Unt (DAU) Here, the vlue of uncertnty n the clrton certfcte of the mnufcturer ws used. A norml prolty ws ssumed. (4) Permttvty nd Conductvty of Phntom Lqud For uncertnty evluton of delectrc mesurements of the phntom lqud, possle wy s to use the typcl ccurcy provded y the mnufcturer of the permttvty mesurement proe [17] [0]. Ths pper s sed on the evluton method of p IEC609 1 [8]. 1 Repetlty of Mesurements At ech frequency, permttvty nd conductvty of the phntom lqud s mesured 10 tmes nd the stndrd devton otned s dvded y the verge vlue mesured to otn the tolernce vlue. A norml prolty s used. Devton from Stndrd Vlue of Delectrc Constnts The devton of the verge mesured vlues form the trget vlues s consdered s the tolernce. A rectngulr prolty s used. 3 Uncertnty of Network Antenns nd Other Ths tme, to derve uncertnty of the network nlyzer, uncertnty evluton softwre ws employed tht uses the Monte Crlo method nd ws mnufctured y NPL n the UK. A rectngulr prolty s used. Tle 3 shows the uncertnty udget of IEC609 1 [8]. (5) Devton from Trget Vlues of Permttvty nd Conductvty of Phntom Lqud durng Clrton At ech frequency, the devton of the delectrc constnt of the phntom lqud used from the trget vlue s evluted. An djustment s mde so the delectrc constnt of the phntom lqud s wthn ± 3% of the trget vlue t ll frequenes n NICT, nd therefore 3 % s used here. A unform prolty ws ssumed. (6) Unformty of Electrc Feld Dstruton For the electrc feld, electrc feld ner the center of the clrton wvegude s mesured, nd evluton s conducted on the mxmum devton from the pproxmton formul of TE 10 mode, y squred nd cosgn squre functons of the electrc feld strength mesured. Here, for electrc feld, the pproxmton Equton (7) shown erler ws used. For mesurement of the electrc feld, evluton ws conducted t mesurement poston (z drecton) wth dstnce etween the proe tp nd delectrc sl t 10 mm (835 to,450 MHz) or 5 mm (5 GHz nd), nd mesurement rnge of ±0 mm from the center n the wvegude length drecton. A unform prolty ws ssumed. 14 Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)

9 -7 Clrton of SAR e (7) Proe Poston Proe poston determnton uncertnty cused y nccurte poston detecton n the z drecton of the root ws evluted y usng dstnce error (Δz) durng surfce detecton, nd penetrton depth δ, n Equton (9). z SAR (1 e ) (9) Δz: Dstnce etween the proe tp nd delectrc sl surfce, durng surfce detecton (z = 0) Here, Δz s the z = 0 poston detecton error of the root n phntom lqud, tht s, equvlent to the detecton error of the delectrc sl surfce poston. Here, the SAR mesurement system used t NICT employs sprng mechnsm n the proe connector of the output dt cquston unt (DAU), menng ths lso ncludes error cused y ply n the proe mounted to the DAU. For the evluton method, flt plte ws hung vertclly ove the phntom lqud, surfce detecton ws performed on tht lqud, nd feeler guge ws used to mesure the dstnce etween the surfce nd proe tp. A unform prolty ws ssumed. (8) Lnerty of SAR Proe The 0.6 % lnerty uncertnty n the clrton certfcte of the mnufcturer ws used. A norml prolty ws ssumed. (9) Dt S/N Rto The output voltge of the DAU wth/wthout RF sgnl nput ws used to otn the S/N rto. A norml prolty ws ssumed. (10) Vrlty of Clrton Fctor Clrton ws conducted fve tmes, nd reltve stndrd devton of these dt ws used. A norml prolty ws ssumed. Frst, Tles 4 to 15 elow show exmples for reltve uncertnty evluton. For uncertnty vlues ted from the clrton certfcte, for exmple, norml ws ssumed nd ws ssumed s the dvsor. Source of tolernce 4 Exmple of evlutng wvegude clrton uncertnty (733 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch.53 U DAU uncertnty(k=) 1.50 Norml constnt (conductvty) 1.10 Norml constnt (permttvty) 1.03 Norml lqud (conductvty) lqud (permttvty) Electrc feld unformty 0.13 Unform Proe postonng 0.30 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.63 Norml Mesurement dt vrlty (mesure fve 0.3 Norml N(N=5) Comned stndrd uncertnty 3.91 Coverge fctor k (95% confdence level) Expnded uncertnty

10 Reserch nd Development of Clrton Technology Source of tolernce 5 Exmple of wvegude clrton uncertnty (835 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 3.10 U 1.19 DAU uncertnty(k=) 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 0.86 Norml Norml Unform Electrc feld unformty 0.40 Unform Proe postonng 0.31 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.63 Norml Mesurement dt vrlty (mesure fve 0.41 Norml N(N=5) Comned stndrd uncertnty 4.13 Coverge fctor k (95% confdence level) Expnded uncertnty 8.11 Source of tolernce 6 Exmple of wvegude clrton uncertnty (900 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power 3.30 Norml Wvegude msmtch 1.4 U DAU uncertnty 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 1.3 Norml Norml Unform Electrc feld unformty 0.4 Unform Proe postonng 0.33 Unform Proe lnerty 0.60 Norml Output voltge S/N rto 0.70 Norml Mesurement dt vrlty (mesure fve 0.57 Norml N(N=5) Comned stndrd uncertnty 3.77 Coverge fctor k (95% confdence level) Expnded uncertnty Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)

11 -7 Clrton of SAR e Source of tolernce 7 Exmple of wvegude clrton uncertnty (1,450 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 1.03 U DAU uncertnty(k=) 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 0.8 Norml Norml Unform Electrc feld unformty 0.11 Unform Proe postonng 0.4 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.76 Norml Mesurement dt vrlty (mesure fve 0.30 Norml N(N=5) Comned stndrd uncertnty 3.69 Coverge fctor k (95% confdence level) Expnded uncertnty 7.6 Source of tolernce 8 Exmple of wvegude clrton uncertnty (1,64 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 3.94 U 1.79 DAU uncertnty(k=) 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 0.77 Norml Norml Unform Electrc feld unformty 0.34 Unform Proe postonng 0.46 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.95 Norml Mesurement dt vrlty (mesure fve 0.77 Norml N(N=5) Comned stndrd uncertnty 4.64 Coverge fctor k (95% confdence level) Expnded uncertnty

12 Reserch nd Development of Clrton Technology Source of tolernce 9 Exmple of wvegude clrton uncertnty (1,767.5 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch.77 U DAU uncertnty(k=) 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 0.7 Norml Norml Unform Electrc feld unformty 0.5 Unform Proe postonng 0.48 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.47 Norml Mesurement dt vrlty (mesure fve 0.33 Norml N(N=5) Comned stndrd uncertnty 4.1 Coverge fctor k (95% confdence level) Expnded uncertnty 8.11 Source of tolernce 10 Exmple of wvegude clrton uncertnty (1,950 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 0.48 U DAU uncertnty(k=) 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 0.95 Norml Norml Unform Electrc feld unformty 0.99 Unform Proe postonng 0.49 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.5 Norml Mesurement dt vrlty (mesure fve 0.8 Norml N(N=5) Comned stndrd uncertnty 3.88 Coverge fctor k (95% confdence level) Expnded uncertnty Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)

13 -7 Clrton of SAR e Source of tolernce 11 Exmple of wvegude clrton uncertnty (,018 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 0.50 U DAU uncertnty(k=) 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 1.4 Norml Norml Unform Electrc feld unformty 0.31 Unform Proe postonng 0.50 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.5 Norml Mesurement dt vrlty (mesure fve 0.55 Norml N(N=5) Comned stndrd uncertnty 4.06 Coverge fctor k (95% confdence level) Expnded uncertnty 8.05 Source of tolernce 1 Exmple of wvegude clrton uncertnty (,450 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 0.84 U DAU uncertnty(k=) 1.50 Norml constnt (conductvty) constnt (permttvty) lqud (conductvty) lqud (permttvty) 1.10 Norml Norml Unform Electrc feld unformty 0.80 Unform Proe postonng 0.64 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.93 Norml Mesurement dt vrlty (mesure fve 0.57 Norml N(N=5) Comned stndrd uncertnty 4.10 Coverge fctor k (95% confdence level) Expnded uncertnty

14 Reserch nd Development of Clrton Technology Source of tolernce 13 Exmple of wvegude clrton uncertnty (,585 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 5.34 U DAU uncertnty(k=) 1.50 Norml Mesurement of lqud delectrc constnt (conductvty) Mesurement of lqud delectrc constnt (permttvty) Devton from stndrd vlue of lqud (conductvty) Devton from stndrd vlue of lqud (permttvty) 1.11 Norml Norml Unform Electrc feld unformty 0.94 Unform Proe postonng 0.69 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.65 Norml Mesurement dt vrlty (mesure fve 0.34 Norml N(N=5) Comned stndrd uncertnty 6.1 Coverge fctor k (95% confdence level) Expnded uncertnty 1. Source of tolernce 14 Exmple of wvegude clrton uncertnty (3,500 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 3.17 U 1.4 DAU uncertnty(k=) 1.50 Norml Mesurement of lqud delectrc constnt (conductvty) Mesurement of lqud delectrc constnt (permttvty) Devton from stndrd vlue of lqud (conductvty) Devton from stndrd vlue of lqud (permttvty) 1. Norml Norml Unform Electrc feld unformty 3.87 Unform Proe postonng 0.87 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.08 Norml Mesurement dt vrlty (mesure fve 1.7 Norml N(N=5) Comned stndrd uncertnty 5.17 Coverge fctor k (95% confdence level) Expnded uncertnty Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)

15 -7 Clrton of SAR e Source of tolernce 15 Exmple of wvegude clrton uncertnty (5,00 MHz, EX3DV4 Proe) Tolernce u=( / ) Input power(k=) 3.30 Norml Wvegude msmtch 1.57 U DAU uncertnty(k=) 1.50 Norml Mesurement of lqud delectrc constnt (conductvty) 1.09 Norml Mesurement of lqud delectrc constnt (permttvty).60 Norml Devton from stndrd vlue of lqud (conductvty) 1.94 Unform Devton from stndrd vlue of lqud (permttvty) Electrc feld unformty 1.94 Unform Proe postonng 1.43 Unform Proe lnerty (k=) 0.60 Norml Output voltge S/N rto 0.37 Norml Mesurement dt vrlty (mesure fve 0.5 Norml N(N=5) Comned stndrd uncertnty 4.55 Coverge fctor k (95% confdence level) Expnded uncertnty Uncertnty sources nd stndrd uncertnty n wvegude clrton Frequency (MHz) Wvegude msmtch(%) Delectrc constnt mesurement (conductvty) (%) Delectrc constnt mesurement (permttvty) (%) Expnded uncertnty(%) Exmple of clrton result (EX3DV4) Frequency (MHz) Clrton fctor K [μv/(v/m) ] Prevous clrton result Devton (%) Rto of devton nd expnded uncertnty Tle 16 shows lst of typcl uncertnty sources, uncertntes, nd expnded uncertntes, from the results descred ove. In ths evluton, expnded uncertntes overll tended to ncrese long wth hgher frequenes. Lookng t ndvdul sources, wvegude msmtch s not frequency dependent ecuse of djustments to mtch ech frequency, however, there s need to consder t t ech frequency. On the other hnd, uncertnty of permttvty nd conductvty ncreses long wth hgher frequenes; thus, there s need for very ccurte mesurements t hgh frequenes. As n exmple of clrton results, Tle 17 shows how results of SAR proe EX3DV4 clrton performed t NICT dffer from the results of lst yer. 149

16 Reserch nd Development of Clrton Technology The rtos of dfferences from the clrton results dt of lst tme nd expnded uncertnty were lso presented for reference. 6 Concluson Ths pper descred the clrton method, prnples, clrton results nd n uncertnty evluton exmple for the SAR proe. A prmry stndrd hs not een estlshed for SAR proe clrton yet, nd therefore t s desrle to e le to pply clrton tht uses multple prnples. In ddton, n order to ensure rellty, t s lso mportnt to hve mutul comprsons etween multple clrton nsttutons. It ws confrmed tht clrton results of NICT re wthn the rnge of clrton uncertnty (pproxmtely 10 %) of the NPL of the UK nd the proe mnufcturer (ISO1705 certfed clrton nsttuton) [1]. In the future more detled evluton of uncertnty s necessry, ncludng n optmzton of the curve-fttng rnge n clrton nd other sources such s oundry effect. There re lso plns to extend the clrton servce frequenes. Furthermore, uncertnty evluton of the methods tht use devces other thn clrton wvegude wll e provded. Acknowledgments Prt of ths reserch ws mplemented under Mnstry of Internl Affrs nd Communctons contrcted reserch (Electromgnetc wves sfety evluton technology). mentton, nd procedures - Prt 1: Procedure to determne the spefc sorpton rte (SAR) for hnd-held devces used n close proxmty to the er (frequency rnge of 300 MHz to 3 GHz), (016). 9 IEC 609-:010 : Humn exposure to rdo frequency felds from hnd-held nd ody-mounted wreless communcton devces - Humn models, nstrumentton, nd procedures - Prt : Procedure to determne the spefc sorpton rte (SAR) for wreless communcton devces used n close proxmty to the humn ody (frequency rnge of 30 MHz to 6 GHz), IEEE Stndrd IEEE Recommended Prctce for Determnng the Pek Sptl-Averge Spefc Asorpton Rte (SAR) n the Humn Hed from Wreless Communctons Devces: Mesurement Technques, Ktj Pokovc, Advnced Electromgnetc Proes for Ner-Feld Evlutons,, Swss Federl Insttute of Technology, Zurch Doctorl Dssertton ETH No , Ktj Pokovc, Thoms Schmd, nd Nels Kuster, Mllmeter-Resoluton E-Feld Proe for Isotropc Mesurement n Lossy Med Between 100 MHz nd 0 GHz, IEEE TRANSACTIONS ON INSTRUMENTATION AND MEASUREMENTS, vol.49, no.4, AUGUST Hmd et.l., Expnson of the SAR-proe clrton system usng the reference dpole ntenn n hed-smultng lqud, Tech Rep. IEICE, EMCJ 109(370), , 010 ( Jn. 010).(In Jpnese) 14 Iwtn et.l: A Study on Influence of Boundry Effect n Spefc Asorpton Rte Mesurement of Rdo Communcton Equpment, 015 Annul Conference of IEICE, B-4-4, 015 (In Jpnese) 15 Low-frequency ( MHz) Coxl SAR Clrton System co.uk/nstruments/products/rf-mcrowve/low-frequency-coxl-sr-clrton-system/ 16 Loder, B.G, Gregory A.P. Bownds, D., Coxl rtefct stndrd for spefc sorpton rte 100 khz to 400 MHz, Progress n Electromgnetc Reserch Symposum (PIERS), 3-7 Mrch, 009, Bejng, Chn 17 Akyo et.l, Proe clrton system for SAR mesurement nd evluton of clrton uncertnty MHz/1950MHz system nd depth effect of tssue equvlent lqud on clrton, Tech Rep. IEICE, EMCJ005-11, pp.33-38, Nov. 005.(In Jpnese) 18 Akyo et.l Proe clrton system for SAR mesurement nd evluton of clrton uncertnty : 900MHz nd system n complnce wth ARIB stndrd, Tech Rep. IEICE, EMCJ004-16, pp.55-59, Jn. 005.(In Jpnese) 19 Ichno et.l., Uncertnty of SAR mesurement of cellulr phones, Tech Rep. IEICE, EMCJ004-15, pp.49-54, Jn. 005.(In Jpnese) 0 Hmd et.l., Uncertnty of Spefc Asorpton Rte Proe Clrton Usng Wvegude Systems, Tech Rep. IEICE, EMCJ 111(49), 19-3, 01 (In Jpnese) 1 Loder et.l., SAR proe clrton: the results of n ntercomprson study, EMC Kyoto 009, pp , 009 RReference 1 Mnstry of Posts nd Telecommunctons, Telecommunctons Technology Counl, Report of Inqury No.38(1990)(In Jpnese) Mnstry of Posts nd Telecommunctons, Telecommunctons Technology Counl, Report of Inqury, No.89(1997)(In Jpnese) 3 ICNIRP Gudelnes for lmtng exposure to tme-vryng electrc mgnetc, nd electromgnetc felds (up to 300 GHz), Helth Physcs, 77(4), pp.494-5(1998) 4 Evlutng Complnce wth FCC Gudelnes for Humn Exposure to Rdofrequency Electromgnetc Felds, FCC OET Bulletn 65 OET Supplement C, (1997) 5 IEEE C IEEE Stndrd for Sfety Levels wth Respect to Humn Exposure to Rdo Frequency Electromgnetc Felds, 3 khz to 300 GHz, (005) 6 ICNIRP Gudelnes for lmtng exposure for lmtng exposure to tme vryng electrc nd mgnetc felds (1Hz 100 khz), Helth Physcs, 99(6), pp (010) 7 Mnstry of Internl Affrs nd Communctons, Informton nd Communctons Counl, Prtl Report of Inqury No 118(011)(In Jpnese). 8 IEC 609-1:016, Humn exposure to rdo frequency felds from hnd-held nd ody-mounted wreless communcton devces - Humn models, nstru- Lr HAMADA, Dr. Eng. Senor Resercher, Electromgnetc Comptlty Lortory, Appled Electromgnetc Reserch Insttute Boelectromgnetcs Soch WATANABE, Dr. Eng. Reserch mnger, Electromgnetc Comptlty Lortory, Appled Electromgnetc Reserch Insttute Boelectromgnetcs 150 Journl of the Ntonl Insttute of Informton nd Communctons Technology Vol. 63 No. 1 (016)