APMP KEY COMPARISON APMP.EM.RF-K3.F. Bilateral Comparison of horn antenna gain

Transcription

1 APMP KEY COMPARISON APMP.EM.RF-K3.F Bilateral Comparison of horn antenna gain Final Report Jin-Seob Kang Korea Research Institute of Standards and Science (KRISS) 67 Gajeong-ro, Yuseong-gu, Daejeon , Korea July 015 APMP.EM.RF-K3.F (v4.3) 1/46

2 Bilateral comparison of horn antenna gain PARTICIPANTS Jin-Seob Kang/No-Weon Kang Korea Research Institute of Standards and Science (KRISS) 67 Gajeong-ro, Yuseong-gu, Daejeon Korea Masanobu Hirose/Koji Komiyama Metrology Institute of Japan (NMIJ) Tsukuba Central 3, Umezono 1-1-1, Tsukuba-shi, Ibaraki Japan ABSTRACT A key comparison of Ka-band (6.5 GHz to 40 GHz) antenna far-field gain, APMP.EM.RF-K3.F, has been carried out between KRISS (Korea Research Institute of Standards and Science) and NMIJ (National Metrology Institute of Japan) at 6.5 GHz, 33 GHz and 40 GHz. The bilateral comparison used the protocol based on that used in the key comparison CCEM.RF-K3.F to link this comparison to the CCEM.RF-K3.F. The KRISS participated in CCEM.RF-K3.F acted as the pilot and linking laboratory for the comparison. APMP.EM.RF-K3.F (v4.3) /46

3 CONTENTS 1. INTRODUCTION. TRAVELLING STANDARDS 3. PARAMETERS FOR MEASUREMENT 4. COMPARISON SCHEDULE 5. METHODS OF MEASUREMENT 6. MEASUREMENT RESULTS 7. APMP KC REFERENCE VALUES AND LINKING TO CCEM KC 8. CONCLUSIONS 9. REFERENCES 10. UNCERTAINTY BUDGETS 11. MEASUREMENT PROTOCOL APMP.EM.RF-K3.F (v4.3) 3/46

4 1. INTRODUCTION A key comparison of power gain for Ka-band horn antennas, CCEM.RF-K3.F (GT-RF 9-1), had been carried out from 1998 to 000 for evaluating consistency between participating national metrology institutes in gain measurements for horn antennas [1]. An APMP key comparison of Ka-band antenna gain, APMP.EM.RF-K3.F, has been carried out between KRISS (Korea Research Institute of Standards and Science) and NMIJ (National Metrology Institute of Japan). The purpose of the comparison is to determine the level of consistency of calibration results given by the two national metrology institutes and to link the antenna gain reference of the NMIJ to that of the KRISS participated in CCEM.RF-K3.F in the frame of the APMP-RMO key comparisons. The protocol is based on the key comparison CCEM.RF-K3.F and its pilot laboratory is the KRISS. Travelling standards are two commercially produced pyramidal horn antennas. The on-axis (boresight) far-field gain and the real and imaginary parts of the reflection coefficient of the antennas are determined at a number of prescribed frequencies, together with an appropriate uncertainty statement.. TRAVELLING STANDARDS Two comparison standards are used: 1) An MI technology model MI-1A-6, with nominal mid-band gain of 4.4 db. Marked INT and serial number 156 ) A Narda model V637, with nominal mid-band gain of 16.5 db. Marked INT and serial number 8709 The dimensions of the antennas are given below: APMP.EM.RF-K3.F (v4.3) 4/46

5 Table 1: Nominal Antenna Dimensions. Horn A [mm] B [mm] C [mm] MI-1A Narda V A C B Figure 1: Dimension of the travelling standards. The antennas are the same types of products as used in CCEM.RF-K3.F (GT-RF 9-1). * Note: An SA 1A-6 was used in CCEM.RF-K3.F. However, MI-1A-6 has the same dimensions as the SA 1A PARAMETERS FOR MEASUREMENT Based on CCEM.RF-K3.F, the following parameters are measured. 1) Gain For MI-1A-6 and Narda V637, the boresight far-field gain is measured at 6.5 GHz, 33.0 GHz, and 40.0 GHz. The boresight line is defined as the normal to the input flange of each antenna and it is not the normal to the aperture. APMP.EM.RF-K3.F (v4.3) 5/46

6 ) Reflection Coefficient The real and imaginary parts of the reflection coefficients of the antennas are measured at the same frequencies as the gain measurement. 4. COMPARISON SCHEDULE The travelling standards were circulated to the participating laboratories in the following order: KRISS (1, Korea) March/April 007 NMIJ (Japan) June/July 007 KRISS (, Korea) April /May METHODS OF MEASUREMENT 1) KRISS (1) & KRISS () The three antenna method based on the extrapolation technique [, 3] was used to measure the gains of the antennas. Measurements are performed on a planar near-field antenna measurement system, consisting of a planar near-field scanner, an extrapolation range, and a microwave subsystem, in the chamber of the dimensions 7 m (L) x 10 m (W) x 6 m (H) covered with 4 inch radiation absorbing material. The transmitter of the microwave subsystem, as shown in Figure, based on an external mixer configuration is modified to form a reflectometer for measuring the impedance of DUT (Device Under Test). The impedance of the receiver connected to a receiving antenna, L, is measured with the reflectometer calibrated using the extended cross-ratio reflection correction method [4] while that of the transmitter connected to a transmitting antenna, can be found by using the fact that the impedance seen looking from the transmitting G, APMP.EM.RF-K3.F (v4.3) 6/46

7 antenna toward the transmitter is the same as the source match, which is determined in oneport calibration procedure. The impedance of transmitting and receiving antennas, Tx and Rx, is measured with a network analyzer calibrated using the TRL method. Insertion loss and separation distance between the transmitting and receiving antennas are measured with a microwave receiver and a laser interferometer, respectively. An additional horn antenna (Narda model V637) is used in the three antenna technique. The initial and final separations between the transmitting antenna (Tx) and the receiving antenna (Rx) are, respectively, 1.0 cm and cm for Narda-Narda antenna pair, and 36.7 cm and cm for MI technology-narda antenna pair. Figure : Block diagram of the measurement set-up at KRISS. For the measurement set-up described above, the product of the gains of the antennas is given by APMP.EM.RF-K3.F (v4.3) 7/46

8 G G G 1 1 G G G r P P r P P R1 D R13 D r P P R3 D M M M (1) where r is the distance between the apertures of the transmitting and receiving antennas, P D is the received power when the transmitting and receiving waveguides are connected directly, P Rij is the received power when the two antennas, antenna i and antenna j, are connected to the transmitting and receiving waveguides, respectively, and M ij is the mismatch correction factor. This form of Equations (1) implicitly assumes that the antennas are polarization matched for their prescribed orientations. The gains of the AUTs can be obtained by G ( db) R 1 G ( db) R G ( db) R 3 M M M M1 M M ( db) R ( db) R ( db) R 3 M1 M M 13 M13 M M ( db) R ( db) R ( db) R 1 ( db) M ( db) M ( db) M 1 G 1 L (1 ) 1 ( db) ( db) ( db) 1 3 L 1 G (1 ) G 1 1 L (1 ) 1 1 G G G L L 1 3 G L L () where APMP.EM.RF-K3.F (v4.3) 8/46

9 R ij lim 4 r P P, r M i mismatch correction factor, 1 3 G L Rij D reflection coefficient of the AUT 1 (Scientific Atlanta 1A-6, marked INT ), reflection coefficient of the AUT (Narda V637, marked INT ), reflection coefficient of the AUT 3 (Narda V637, belongs to KRISS), reflection coefficient of the transmitter, reflection coefficient of the receiver. ) NMIJ The gain transfer method [5, 6] was used to measure the gains of the antennas. The chamber, whose dimensions are about 8 m long, 6 m wide, and 6 m high, is covered with various types of radiation absorbing materials (RAM) for millimeter waves. The RAM of the height of 13 cm was placed on the floor around the equipment and rails on the floor were covered by the RAM of the height of 1 cm. An antenna calibrated by National Physical Laboratory (NPL) of UK was used as the reference antenna in the transfer method. A vector network analyzer (VNA, Agilent E8364A) was used to measure the S parameters between one of the three antennas (two bilateral-comparison antennas and the reference antenna) as a transmitting antenna connected to the port 1 of the VNA and a receiving antenna (another model MI-1A-6) connected to the port of the VNA, which was used as the Rx antenna throughout the measurements. The VNA was calibrated by the TRL method. S 11 was used as the reflection coefficient for the three antennas. S 1 was used to obtain the gain of each antenna. The measurement set-up is shown in Fig. 3. The initial and final separations between two antennas were 1516 mm to 1538 mm (1.1 mm steps or 1 points) to reduce the multiple reflection effects in S 11 and S 1 by using a moving average method. APMP.EM.RF-K3.F (v4.3) 9/46

10 The measured S 1 is given by S G1 GC( ) 1 R 4R (3) where R is the distance between the apertures of the antenna 1 (Tx antenna) and the antenna (Rx antenna), i is the reflection coefficient of the antenna i (i=1,), G i is the farfield gain of the antenna i, and C(R) is a near-field correction factor. Following Chu et al. [7], the near-field correction factor is given by C( R) * * G G Re E H 1 n ds Re E H n ds 4R E H 1 1 H 1 E 1 1 n ds (4) where E i and H i are the electromagnetic field at the aperture of the antenna i when it is used as a transmitting antenna, and ( )* denotes the complex conjugate of ( ). C (R) was calculated numerically using the aperture field distributions used in Ref. [7]. The factor goes to 1 if R goes to infinity. Therefore the gain of the AUT by the transfer method can be obtained by R1 R S1 AR R1 S R CR GA GR (5) C A A 1R where C i (R) is the near-field correction factor and S i ( ) is the S 1 when the Tx antenna is i (i = R for the reference or i =A for the AUT). 1 R APMP.EM.RF-K3.F (v4.3) 10/46

11 Tx antenna Rx antenna Rail Vector Network Analyzer GPIB Computer GPIB Figure 3: Measurement set-up at NMIJ. Traceability The reflection coefficient and the gain were determined using the S parameters measured by the vector network analyzer, which was calibrated by the TRL method. The dimension of the quarter wavelength straight section was measured by equipment calibrated by NMIJ. The attenuation of the vector network analyzer is traceable to the attenuation standard of NMIJ through a stepped attenuator calibrated by NMIJ. The distances between antennas were measured by a laser interferometer calibrated by NMIJ. The gain of the reference antenna was calibrated by NPL. Attenuation Length NMIJ NPL S-parameters Distance Reference Antenna Gain Reflection Coefficient Gain Figure 4: Traceability chain at NMIJ. APMP.EM.RF-K3.F (v4.3) 11/46

12 Summary of Correction Applied Laboratory Near-zone Correction Mismatch Correction KRISS Yes Extrapolation method used Corrected for mismatch NMIJ Yes Theoretical calculation used (with moving average method) Corrected for mismatch 6. MEASUREMENT RESULTS The results presented from the participants are given in the tables below. The weighted mean is given in the tables for the gain and the unweighted mean is given for the real and imaginary parts of the reflection coefficient. Calculation of the mean values and the rationale for a weighted mean for the gain and an unweighted one for the reflection coefficient is described in the next section. Table : Gain results for MI horn 1A-6 s/n 156. Laboratory 6.5 GHz 33.0 GHz 40.0 GHz Value [db] u [db, k=1] Value [db] u [db, k=1] Value [db] u [db, k=1] KRISS (1) NMIJ KRISS () Weighted Mean Table 3: Gain results for Narda horn V637 s/n Laboratory 6.5 GHz 33.0 GHz 40.0 GHz Value [db] u [db, k=1] Value [db] u [db, k=1] Value [db] u [db, k=1] KRISS (1) NMIJ KRISS () Weighted Mean APMP.EM.RF-K3.F (v4.3) 1/46

13 Table 4: Real part of the reflection coefficient for MI horn 1A-6 s/n 156. Laboratory 6.5 GHz 33.0 GHz 40.0 GHz Value u (k=1) Value u (k=1) Value u (k=1) KRISS (1) NMIJ KRISS () Unweighted Mean Table 5: Imaginary part of the reflection coefficient MI horn 1A-6 s/n 156. Laboratory 6.5 GHz 33.0 GHz 40.0 GHz Value u (k=1) Value u (k=1) Value u (k=1) KRISS (1) NMIJ KRISS () Unweighted Mean Table 6: Real part of the reflection coefficient for Narda horn V637 s/n Laboratory 6.5 GHz 33.0 GHz 40.0 GHz Value u (k=1) Value u (k=1) Value u (k=1) KRISS (1) NMIJ KRISS () Unweighted Mean Table 7: Imaginary part of the reflection coefficient for Narda horn V637 s/n Laboratory 6.5 GHz 33.0 GHz 40.0 GHz Value u (k=1) Value u (k=1) Value u (k=1) KRISS (1) NMIJ KRISS () Unweighted Mean APMP.EM.RF-K3.F (v4.3) 13/46

14 Gain [db] Gain [db] MI Technology Horn Model MI-1A-6 s/n 156 at 6.5 GHz KRISS-1 NMIJ KRISS- Figure 5a: The gain of the MI Technology horn at 6.5 GHz. (The weighted mean is shown as a solid line. The uncertainty is for k=.) 5.00 MI Technology Horn Model MI-1A-6 s/n 156 at 33.0 GHz KRISS-1 NMIJ KRISS- Figure 5b: The gain of the MI Technology horn at 33.0 GHz. (The weighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 14/46

15 Gain [db] Gain [db] MI Technology Horn Model MI-1A-6 s/n 156 at 40.0 GHz KRISS-1 NMIJ KRISS- Figure 5c: The gain of the MI Technology horn at 40.0 GHz. (The weighted mean is shown as a solid line. The uncertainty is for k=.) Narda Horn Model V637 s/n 8709 at 6.5 GHz KRISS-1 NMIJ KRISS- Figure 6a: The gain of the Narda horn at 6.5 GHz. (The weighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 15/46

16 Gain [db] Gain [db] Narda Horn Model V637 s/n 8709 at 33.0 GHz KRISS-1 NMIJ KRISS- Figure 6b: The gain of the Narda horn at 33.0 GHz. (The weighted mean is shown as a solid line. The uncertainty is for k=.) Narda Horn Model V637 s/n 8709 at 40.0 GHz KRISS-1 NMIJ KRISS- Figure 6c: The gain of the Narda horn at 40.0 GHz. (The weighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 16/46

17 Imaginary Part of Reflection Coefficient Real Part of Reflection Coefficient MI Technology Horn Model MI-1A-6 s/n 156 at 6.5 GHz KRISS-1 NMIJ KRISS- Figure 7a: The real part of the reflection coefficient of the MI Technology horn at 6.5 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) MI Technology Horn Model MI-1A-6 s/n 156 at 6.5 GHz KRISS-1 NMIJ KRISS- Figure 7b: The imaginary part of the reflection coefficient of the MI Technology horn at 6.5 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 17/46

18 Imaginary Part of Reflection Coefficient Real Part of Reflection Coefficient MI Technology Horn Model MI-1A-6 s/n 156 at 33.0 GHz KRISS-1 NMIJ KRISS- Figure 8a: The real part of the reflection coefficient of the MI Technology horn at 33.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) MI Technology Horn Model MI-1A-6 s/n 156 at 33.0 GHz KRISS-1 NMIJ KRISS- Figure 8b: The imaginary part of the reflection coefficient of the MI Technology horn at 33.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 18/46

19 Imaginary Part of Reflection Coefficient Real Part of Reflection Coefficient MI Technology Horn Model MI-1A-6 s/n 156 at 40.0 GHz KRISS-1 NMIJ KRISS- Figure 9a: The real part of the reflection coefficient of the MI Technology horn at 40.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) MI Technology Horn Model MI-1A-6 s/n 156 at 40.0 GHz KRISS-1 NMIJ KRISS- Figure 9b: The imaginary part of the reflection coefficient of the MI Technology horn at 40.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 19/46

20 Imaginary Part of Reflection Coefficient Real Part of Reflection Coefficient Narda Horn Model V637 s/n 8709 at 6.5 GHz KRISS-1 NMIJ KRISS- Figure 10a: The real part of the reflection coefficient of the Narda horn at 6.5 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) Narda Horn Model V637 s/n 8709 at 6.5 GHz KRISS-1 NMIJ KRISS- Figure 10b: The imaginary part of the reflection coefficient of the Narda horn at 6.5 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 0/46

21 Imaginary Part of Reflection Coefficient Real Part of Reflection Coefficient Narda Horn Model V637 s/n 8709 at 33.0 GHz KRISS-1 NMIJ KRISS- Figure 11a: The real part of the reflection coefficient of the Narda horn at 33.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) Narda Horn Model V637 s/n 8709 at 33.0 GHz KRISS-1 NMIJ KRISS- Figure 11b: The imaginary part of the reflection coefficient of the Narda horn at 33.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) 1/46

22 Imaginary Part of Reflection Coefficient Real Part of Reflection Coefficient Narda Horn Model V637 s/n 8709 at 44.0 GHz KRISS-1 NMIJ KRISS- Figure 1a: The real part of the reflection coefficient of the Narda horn at 40.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) Narda Horn Model V637 s/n 8709 at 40.0 GHz KRISS-1 NMIJ KRISS- Figure 1b: The imaginary part of the reflection coefficient of the Narda horn at 40.0 GHz. (The unweighted mean is shown as a solid line. The uncertainty is for k=.) APMP.EM.RF-K3.F (v4.3) /46

23 7. APMP KC REFERENCE VALUES AND LINKING TO CCEM KC The mean values given in the tables and graphs above have been calculated according to CCEM.RF-K3.F [1] for calculating reference values of APMP.EM.RF-K3.F. The most contentious issue is whether to use a weighted mean or an unweighted one [1]. In CCEM.RF-K3.F, the weighted mean has been calculated for the gain values and the unweighted one for the real and imaginary parts of the reflection coefficients. The rationale for this choice is that the reported uncertainties for the gain values vary greatly between participants and also not all participants performed a near-zone correction, either by extrapolation or calculation. This means that some of the reported values have a significant systematic uncertainty component that would unduly influence the mean if account were not taken of the associated uncertainty. The reference values for an unweighted mean had also been calculated and was given in the tables below for completeness. For the real and imaginary parts of the reflection coefficient, the uncertainties quoted are all broadly similar, so an unweighted mean is more appropriate, though in fact it makes little difference whether the weighted or unweighted mean is taken. The prime objective of the comparison was to measure the gain of the standard antennas and therefore degrees of equivalence have only been calculated for the gain. In this report, the weighted and unweighted reference values, the degree of equivalence, and their uncertainties are calculated according to CCEM.RF-K3.F [1]. For all evaluations of the reference values, the two KRISS values were averaged and the result treated as a single measurement: KRISS APMP = (KRISS(1) + KRISS()) / (6) As stated in the measurement protocol based on that of CCEM.RF-K3.F [1], KRISS is acting as a linking laboratory that participated in CCEM.RF-K3.F. The two traveling antennas are the same or very similar in the two comparisons and the measurement APMP.EM.RF-K3.F (v4.3) 3/46

24 frequencies are the same. KRISS used almost the same measurement method at the two comparisons and the differences (KRISS CCEM KRISS APMP ) of the results from the two comparisons are within the measurement uncertainties (U(KRISS APMP )) of KRISS at APMP.EM.RF-K3.F, except for MI-1A-6 (SA-1A-6) at 40 GHz, as in Table 8. The differences between the reference values (KCRV CCEM and RV APMP(CCEM) ) of the two comparisons, where the reference value, RV APMP(CCEM), of APMP.EM.RF-K3.F is obtained using the results from KRISS and NMIJ calculated according to CCEM.RF-K3.F, is within U(KRISS APMP ) for both the unweighted mean and the weighted mean except for MI-1A- 6 (SA-1A-6) at 40 GHz, as in Tables 9 and 10. These can be supporting evidence. So the results of APMP.EM.RF-K3.F can be directly linked to those of CCEM.RF-K3.F with the measurement uncertainties reported by both laboratories even with the KCRV of CCEM.RF-K3.F. Under the assumption that the measurement methods of KRISS at the two comparisons are the same, one may link APMP.EM.RF-K3.F to CCEM.RF-K3.F as the following linking procedure (Tables 11 and 1): 1) Refer the following quantities from CCEM.RF-K3.F; - KCRV CCEM : KCRV of CCEM.RF-K3.F - KRISS CCEM : KRISS value of CCEM.RF-K3.F - d KRISS i,ccem (= KRISS CCEM - KCRV CCEM ): Degree of equivalence of KRISS at CCEM.RF-K3.F - U(d KRISS KRISS i,ccem ): Uncertainty of d i,ccem ) Refer the following quantities from APMP.EM.RF-K3.F; - KRISS APMP : KRISS value of APMP.EM.RF-K3.F - NMIJ: NMIJ value of APMP.EM.RF-K3.F - U(NMIJ): Uncertainty of NMIJ 3) Determine the reference value, RV APMP, for each antenna and at each frequency for APMP.EM.RF-K3.F by taking the KRISS value (KRISS APMP ), with a small offset (d KRISS i,ccem ) to account for the difference between KRISS and KCRV of CCEM.RF- K3.F. APMP.EM.RF-K3.F (v4.3) 4/46

25 KRISS - RV APMP = KRISS APMP - d i,ccem 4) Determine the degree of equivalence, d NMIJ i,ccem, of NMIJ at CCEM.RF-K3.F by subtracting the RV APMP at APMP.EM.RF-K3.F. - d NMIJ i,ccem = NMIJ - RV APMP 5) Determine the uncertainty of the degree of equivalence, U(d NMIJ i,ccem ), from the uncertainties of the degree of equivalence of KRISS at CCEM.RF-K3.F and those of the measured value of NMIJ and KRISS at APMP.EM.RF-K3.F in the sense of Root- Sum-Square. - U (d NMIJ i,ccem ) = U (d KRISS i,ccem ) + U (NMIJ) + U (KRISS APMP ) Finally the data, d NMIJ i,ccem and U(d NMIJ i,ccem ), in Tables 11 and 1 are used to complete the previously published matrix of equivalence and uncertainties for CCEM.RF-K3.F which of course will not be changed by APMP.EM.RF-K3.F results. Tables 13 to 17 show these full completed matrices for the six participating laboratories. The degrees of equivalence with respect to the KCRV from CCEM.RF-K3.F (d i and U(d i ) in Tables 13 to 14) are also shown under the form of graph (Figures 13 and 14). APMP.EM.RF-K3.F (v4.3) 5/46

26 Table 8: Gain results of KRISS at both comparisons (APMP.EM.RF-K3.F and CCEM.RF-K3.F, k=). Standard MI-1A-6 Narda V637 Freq [GHz] CCEM.RF-K3.F APMP.EM.RF-K3.F KRISS CCEM U (KRISS CCEM ) KRISS APMP U (KRISS APMP ) KRISS CCEM KRISS APMP Table 9: Gain results of Reference Values at both comparisons (unweighted, k=). Standard Freq [GHz] KCRV CCEM RV APMP(CCEM) KCRV CCEM - RV APMP(CCEM) U (KRISS APMP ) MI-1A Narda V Table 10: Gain results of Reference Values at both comparisons (weighted, k=). Standard Freq [GHz] KCRV CCEM RV APMP(CCEM) KCRV CCEM - RV APMP(CCEM) U (KRISS APMP ) MI-1A Narda V APMP.EM.RF-K3.F (v4.3) 6/46

27 Table 11: Linking procedure (unweighted, k=). Standard MI-1A-6 Narda V637 Freq [GHz] CCEM.RF-K3.F APMP.EM.RF-K3.F Linking KRISS KCRV CCEM KRISS CCEM d i,ccem U(d KRISS NMIJ i,ccem ) KRISS APMP U(KRISS APMP) NMIJ U(NMIJ) RV APMP d i,ccem U(d NMIJ i,ccem ) Table 1: Linking procedure (weighted, k=). Standard MI-1A-6 Narda V637 Freq [GHz] CCEM.RF-K3.F APMP.EM.RF-K3.F Linking KRISS KCRV CCEM KRISS CCEM d i,ccem U(d KRISS NMIJ i,ccem ) KRISS APMP U(KRISS APMP) NMIJ U(NMIJ) RV APMP d i,ccem U(d NMIJ i,ccem ) APMP.EM.RF-K3.F (v4.3) 7/46

28 KCRV and Degree of Equivalence for Gain Table 13: KCRV (unweighted mean) and degree of equivalence for gain (k=) Standard Freq [GHz] x u U(x u ) NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ d i U(d i ) d i U(d i ) d i U(d i ) d i U(d i ) d i U(d i ) d i U(d i ) MI-1A NardaV Standard MI-1A-6 NardaV637 Table 14: KCRV (weighted mean) and degree of equivalence for gain (k=) Freq NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ x [GHz] w U(x w ) d i U(d i ) d i U(d i ) d i U(d i ) d i U(d i ) d i U(d i ) d i U(d i ) APMP.EM.RF-K3.F (v4.3) 8/46

29 Inter Laboratory Degree of Equivalence for Gain Standard MI-1A-6 NardaV637 Table 15: Degree of equivalence at 6.5 GHz for gain (k=1) LABj NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ LABi d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ Standard MI-1A-6 NardaV637 Table16: Degree of equivalence at 33 GHz for gain (k=1) LABj NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ LABi d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ APMP.EM.RF-K3.F (v4.3) 9/46

30 Standard MI-1A-6 NardaV637 Table17: Degree of equivalence at 40 GHz for gain (k=1) LABj NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ LABi d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) d ij U(d ij ) NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ APMP.EM.RF-K3.F (v4.3) 30/46

31 Degree of Equivalence, d i [db] Degree of Equivalence, d i [db] MI Technology Horn Model MI-1A-6 s/n 156 at 6.5 GHz NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ Figure 13a: Degrees of equivalence for gain for the MI technology horn at 6.5 GHz. The degree of equivalence with respect to the weighted mean is shown including the expanded uncertainty for k=. The uncertainty in the weighted mean, u(xw), is shown by dashed lines MI Technology Horn Model MI-1A-6 s/n 156 at 33.0 GHz NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ Figure 13b: Degrees of equivalence for gain for the MI technology horn at 33.0 GHz. The degree of equivalence with respect to the weighted mean is shown including the expanded uncertainty for k=. The uncertainty in the weighted mean, u(xw), is shown by dashed lines. APMP.EM.RF-K3.F (v4.3) 31/46

32 Degree of Equivalence, d i [db] Degree of Equivalence, d i [db] MI Technology Horn Model MI-1A-6 s/n 156 at 40.0 GHz NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ Figure 13c: Degrees of equivalence for gain for the MI technology horn at 40.0 GHz. The degree of equivalence with respect to the weighted mean is shown including the expanded uncertainty for k=. The uncertainty in the weighted mean, u(xw), is shown by dashed lines Narda Horn Model V637 s/n 8709 at 6.5 GHz NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ Figure 14a: Degrees of equivalence for gain for the Narda horn at 6.5 GHz. The degree of equivalence with respect to the weighted mean is shown including the expanded uncertainty for k=. The uncertainty in the weighted mean, u(xw), is shown by dashed lines. APMP.EM.RF-K3.F (v4.3) 3/46

33 Degree of Equivalence, d i [db] Degree of Equivalence, d i [db] Narda Horn Model V637 s/n 8709 at 33.0 GHz NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ Figure 14b: Degrees of equivalence for gain for the Narda horn at 33.0 GHz. The degree of equivalence with respect to the weighted mean is shown including the expanded uncertainty for k=. The uncertainty in the weighted mean, u(xw), is shown by dashed lines Narda Horn Model V637 s/n 8709 at 40.0 GHz NPL(mean) NMi-VSL NIST BNM-LCIE KRISS NMIJ Figure 14c: Degrees of equivalence for gain for the Narda horn at 40.0 GHz. The degree of equivalence with respect to the weighted mean is shown including the expanded uncertainty for k=. The uncertainty in the weighted mean, u(xw), is shown by dashed lines. APMP.EM.RF-K3.F (v4.3) 33/46

34 8. CONCLUSIONS An APMP key comparison (APMP.EM.RF-K3.) of Ka-band antenna gain between KRISS and NMIJ has been carried out to determine the level of consistency of calibration results given by the two national metrology institutes and to link the antenna gain reference of the NMIJ to that of the KRISS participated in CCEM.RF-K3.F in the frame of the APMP-RMO key comparisons. The measurement comparison has been successfully completed with a range of measurement approaches having been used by the participants. The comparison has demonstrated very good agreements, in the most parts, between the participants. The agreements between the participants for the reflection coefficient measurements are fairly good except for the results at 40 GHz. In terms of the magnitude of the reflection coefficient, the agreement was very good at other frequencies, so the major cause of the difference might be from phase measurements. The two sets of results measured by the pilot laboratory show that the characteristics of the two travelling standards were stable throughout the comparison. 9. REFERENCES [1] Draft B Report of CCEM Key Comparison CCEM.RF-K3.F (GT-RF 9-1). [] A.G. Repjar, A.C. Newell, and D. T. Tamura, Extrapolation Range Measurements for Determining Antenna Gain and Polarization, NBS Technical Note 1311, August [3] IEEE Standard , IEEE Standard Test Procedures for Antennas, pp [4] U. Stumper, Extended cross-ratio reflection correction at microwave frequencies using waveguide air-lines, IEEE Trans. Microwave Theory Tech., Vol. MTT-50, No. 4, pp , Apr [5] G. E. Evans, Antenna Measurement Techniques, Norwood, MA, Artech House, [6] IEEE Standard , IEEE Standard Test Procedures for Antennas, pp [7] T. S. Chu and R. A. Semplak, Gain of Electromagnetic Horns, The Bell System Technical Journal, pp , March APMP.EM.RF-K3.F (v4.3) 34/46

35 10. UNCERTAINTY BUDGETS 1) KRISS Table 11: KRISS Gain uncertainty budget for MI-1A-6 with serial number 156 at 6.5 GHz. Table 1: KRISS Gain uncertainty budget for MI-1A-6 with serial number 156 at 33 GHz. APMP.EM.RF-K3.F (v4.3) 35/46

36 Table 13: KRISS Gain uncertainty budget for MI-1A-6 with serial number 156 at 40 GHz. Table 14: KRISS Gain uncertainty budget for Narda V637 with serial number 8709 at 6.5 GHz. APMP.EM.RF-K3.F (v4.3) 36/46

37 Table 15: KRISS Gain uncertainty budget for Narda V637 with serial number 8709 at 33 GHz. Table 16: KRISS Gain uncertainty budget for Narda V637 with serial number 8709 at 40 GHz. APMP.EM.RF-K3.F (v4.3) 37/46

38 ) NMIJ Table 17: NMIJ Gain uncertainty budget for MI-1A-6 with serial number 156 at 6.5 GHz. Source of Uncertainty Value of Standard Uncertainty ui ci ci*ui Degrees of Freedom System drift Waveguide connection repeatability Near-zone correction effects Inf. Mismatch Antenna alignment Inf. Residuals of multiple reflection between antennas Inf. Reflections from RAM Linearity of S TRL Calibration Gain Uncertainty of the reference antenna Total (6.5 GHz) Table 18: NMIJ Gain uncertainty budget for MI-1A-6 with serial number 156 at 33 GHz. Source of Uncertainty Value of Standard Uncertainty ui ci ci*ui Degrees of Freedom System drift Waveguide connection repeatability Near-zone correction effects Inf. Mismatch Antenna alignment Inf. Residuals of multiple reflection between antennas Inf. Reflections from RAM Linearity of S TRL Calibration Gain Uncertainty of the reference antenna Total (33 GHz) APMP.EM.RF-K3.F (v4.3) 38/46

39 Table 19: NMIJ Gain uncertainty budget for MI-1A-6 with serial number 156 at 40 GHz. Source of Uncertainty Value of Standard Uncertainty ui ci ci*ui Degrees of Freedom System drift Waveguide connection repeatability Near-zone correction effects Inf. Mismatch Antenna alignment Inf. Residuals of multiple reflection between antennas Inf. Reflections from RAM Linearity of S TRL Calibration Gain Uncertainty of the reference antenna Total (40 GHz) Table 0: NMIJ Gain uncertainty budget for Narda V637 with serial number 8709 at 6.5 GHz. Source of Uncertainty ci ci*ui Degrees of Freedom System drift Waveguide connection repeatability Near-zone correction effects Inf. Mismatch Antenna alignment Inf. Residuals of multiple reflection between antennas Inf. Reflections from RAM Linearity of S TRL Calibration Gain Uncertainty of the reference antenna Total (6.5 GHz) APMP.EM.RF-K3.F (v4.3) 39/46

40 Table 1: NMIJ Gain uncertainty budget for Narda V637 with serial number 8709 at 33 GHz. Source of Uncertainty ci ci*ui Degrees of Freedom System drift Waveguide connection repeatability Near-zone correction effects Inf. Mismatch Antenna alignment Inf. Residuals of multiple reflection between antennas Inf. Reflections from RAM Linearity of S TRL Calibration Gain Uncertainty of the reference antenna Total (33 GHz) Table : NMIJ Gain uncertainty budget for Narda V637 with serial number 8709 at 40 GHz. Source of Uncertainty ci ci*ui Degrees of Freedom System drift Waveguide connection repeatability Near-zone correction effects Inf. Mismatch Antenna alignment Inf. Residuals of multiple reflection between antennas Inf. Reflections from RAM Linearity of S TRL Calibration Gain Uncertainty of the reference antenna Total (40 GHz) APMP.EM.RF-K3.F (v4.3) 40/46

41 11. MEASUREMENT PROTOCOL This section contains the Measurement Protocol that was used in the comparison. 1) Introduction This protocol is based on that used in the key comparison CCEM.RF-K3.F. One of the participants, the Korea Research Institute of Standards and Science, took part in CCEM.RF-K3.F. The purpose of the comparison is to determine the level of consistency of calibration results given by two national standards laboratories and to link the antenna gain reference of the NMIJ (National Metrology Institute of Japan) to that of the KRISS in the frame of the APMP- RMO key comparisons. Travelling standards are two commercially produced pyramidal horn antennas. The on-axis (boresight) far-field gain and the real and imaginary parts of the reflection coefficient of the antennas are determined at a number of prescribed frequencies, together with an appropriate uncertainty statement. ) Travelling standards Two comparison standards will be used: a) An MI technology model MI-1A-6, with nominal mid-band gain of 4.4 db. Marked INT and serial number 156 b) A Narda model V637, with nominal mid-band gain of 16.5 db. Marked INT and serial number 8709 The dimensions of the antennas are given below: Nominal Antenna Dimensions Horn A [mm] B [mm] C [mm] MI-1A Narda V APMP.EM.RF-K3.F (v4.3) 41/46

42 A C B Figure 13: Dimension of the travelling standards. The antennas are the same types of products as used in CCEM.RF-K3.F (GT-RF 9-1). * Note: an SA 1A-6 was used in CCEM.RF-K3.F. However, MI-1A-6 has the same dimensions as the SA 1A-6. 3) Participants The contact person at the coordinating (pilot) laboratory for this comparison is: Jin Seob Kang Korea Research Institute of Standard and Science (KRISS) Division of Physical Metrology P.O. Box 10, Yuseong, Taejon Republic of Korea Tel: Fax: jinskang@kriss.re.kr APMP.EM.RF-K3.F (v4.3) 4/46

43 The contact person at the participating laboratory is: Masanobu Hirose National Metrology Institute of Japan Division of Electromagnetic Waves Tsukuba Central 3, Umezono 1-1-1, Tsukuba, Ibaraki Japan Tel: Fax: ) Measurement Parameters Based on CCEM.RF-K3.F, the following parameters will be measured. Gain For MI-1A-6 and Narda V637, the boresight gain will be measured at 6.5 GHz, 33.0 GHz, 40.0 GHz. The boresight line will be defined as the normal to the input flange of each antenna and it is not the normal to the aperture. Reflection Coefficients The real and imaginary parts of the reflection coefficients of the antennas will be measured at the same frequencies as the gain measurement. 5) Submission of results Each laboratory is expected to submit its report to the coordinating laboratory within 6 weeks after the end of its measuring period. When reporting results, participants should provide the following information. a) Antenna model and marking APMP.EM.RF-K3.F (v4.3) 43/46

44 b) Frequency [GHz] c) Gain [db] d) Gain uncertainty [db] f) Real and imaginary parts of reflection coefficient with their uncertainties In addition, a short description of the measurement method and set-up used, preferably with some schematic drawing, a traceability chain, and any theoretical corrections made should be included. 6) Reporting Uncertainties Participants should provide a comprehensive uncertainty budget which might include contributions from the following sources, among others: a) Direct connection measurement. b) System drift. c) Waveguide connection repeatability. d) Extrapolation or near-zone correction effects. e) Mismatch. f) Antenna alignment. g) Multiple reflections between antennas. h) Reflections from RAM. i) Receiver/power sensor linearity. Uncertainty estimates should be carried out in accordance with the recommendations of the ISO publication Guide to the Expression of Uncertainty in Measurement ISBN Uncertainties should be evaluated for a coverage factor k=1. 7) Link with other comparisons KRISS participated in the CCEM.RF-K3.F comparison, this allow us to link our APMP.EM.RF-K3.F (v4.3) 44/46

45 comparison to the CCEM.RF-K3.F. 8) Discussion of results It is expected that an open discussion will take place as quickly as possible after distributing a draft report containing a compilation of the results and a first attempt of interpretation. 9) Final report The draft version of the final report will be issued within two months after completion of the comparison. It will be sent to NMIJ for discussion and approval. The final report will be then submitted. Afterwards the final result can be published in an appropriate journal. 10) Problems during the exercise If technical and/or other problems arise, it is of the utmost importance to contact immediately each other to discuss the matter. It is assumed that each laboratory takes care of the travelling standards during the stay at the laboratory and the transportation to the next. 11) Transport and customs The travelling standards can be sent using regular package mail. The devices are stored in a cardboard container, which is provided by the coordinator. Additional packaging as protection is suggested. 1) Time schedule The travelling standards will be circulated in the following order. One of the traveling standards, MI-1A-6, to be provided by NMIJ will be sent to KRISS for the comparison. After the measurement at KRISS it will be sent back to NMIJ together with another traveling standard, V637, provided by KRISS for the measurement at NMIJ. Two traveling standards will be sent to KRISS again for the stability check at KRISS. The APMP.EM.RF-K3.F (v4.3) 45/46

46 more specific time schedule is to be determined after the discussion between the two institutes. Institute Measuring Period DUTs Contact person KRISS Feb. 007 ~ Mar. 007 MI-1A-6/V637 Jin Seob Kang NMIJ Apr. 007 ~ May 007 MI-1A-6/V637 Masanobu Hirose KRISS Jun. 007 ~ Jul. 007 MI-1A-6/V637 Jin Seob Kang APMP.EM.RF-K3.F (v4.3) 46/46