RF Voltage Breakdown: Case Studies and Prevention

Transcription

1 WMB-5 RF Voltage Breakdown: Case Studies and Prevention H. Clark Bell HF Plus

2 References [1] R. Woo, Final Report on RF Voltage Breakdown in Coaxial Transmission Lines, Jet Propulsion Lab., Pasadena, Calif., Tech. Rept , Oct. 1, [2] G. August, Multipactor breakdown-lessons unlearned, AIAA 10th Comm. Satellite Conf. Proc., Mar [3] S. B. Cohn, Rounded corners in microwave highpower filters and other components, IEEE Trans. Microwave Theory Tech., Vol. MTT-9, pp , Sep

3 Outline Ionization Multipaction Breakdown Thresholds Resonator Voltage Filter Tuning 3

4 RF Voltage Breakdown Breakdown between conductors Ionization: electron mean free path < gap Multipaction: electron mean free path > gap Transition: both types may occur Characterized by Vrms = rms voltage across gap fd = fd product = frequency x gap distance [MHz-cm] 4

5 Ionization Mechanism E-field accelerates free electrons impact gas molecules avalanching ionization Dependencies temperature pressure type of gas 5

6 Ionization (cont.) Arcing occurs at fd < 100 MHz-cm current flow between conductors Corona occurs at fd > 100 MHz-cm buildup of electrons and ions in gap 6

7 Critical altitude Ionization (cont.) Corresponds to pressure with lowest breakdown threshold voltage for given fd Must consider for vented operation during ascent/descent May consider for thermal-vacuum (T-V) test (unintended pressurization) 7

8 Multipaction Mechanism E-field accelerates free electrons in gap impact conductor, more electrons emitted E-field reverses avalanching emission Dependencies availability of free electrons surface properties 8

9 Practical problems Multipaction (cont.) Barely detectable in test Changing surface conditions [2] Not repeatable 9

10 RF Voltage Breakdown Thresholds Uniform Electric Field [1] 200 Breakdown Threshold 150 Vrms Ionization at critical pressure (Woo Fig. 21) Multipaction (Woo Fig. 5) MHz-cm 10

11 RF Voltage Breakdown Thresholds Uniform Electric Field [1] 200 Breakdown Threshold 150 Vrms Ionization at critical pressure (Woo Fig. 21) Multipaction (Woo Fig. 5) MHz-cm 11

12 Breakdown Threshold Factors Temperature local uncertainty reduces ionization threshold probably reduces multipaction threshold Voltage [2] accuracy of model field enhancement [3] standing waves 12

13 Resonator Voltage TEM Resonator: 90 deg at center frequency Equate Im(Y)/Re(Y) at lower band edge (1): t = G g1 tan θ1 Y 0 13

14 Resonator Voltage (cont.) For 2% bandwidth ( 0.99 π 2) tan θ1 = tan = Let R = 1 G = 50Ω, Z0 = 1 Y 0 = 75Ω, g1 = t = 8.00 For 2 W input power tip of resonator Vrms = t P R = 80 V ½ way down (45 ) = 56 V 14

15 Resonator Voltage (cont.) 200 Multipaction at 56 Vrms 150 Multipaction region 74 < fd < 156 MHz-cm Vrms MHz-cm 15

16 Resonator Voltage (cont.) For 2% BW and 2 W Multipaction when 74 < fd < 156 MHz-cm Usually operate to right of multipaction region 500 MHz gaps > cm (0.123 inch) 2000 MHz gaps > cm (0.031 inch) 16

17 Resonator Voltage (cont.) In practice, get even higher voltages Band edge (higher delay, stored energy) Internal resonator (proportional to g value) Mismatch (SWR > 1) 17

18 Resonator Voltage (cont.) What is a safe margin? Find worst-case voltages take into account sharp edges temperature mismatch manufacturing variations qualification limits Design for 2:1 voltage ratio (6 db margin) 18

19 Best practice Resonator Voltage (cont.) Calculate worst-case threshold matched Design for 4:1 voltage ratio (12 db margin) Gives 6 db margin for T-V tests of component with unintended short/open system with other component failure 19

20 Filter Tuning Combline filter: n = 3 resonators 0 S21 (db vs MHz) 0 S21 (db vs MHz) Passband loss: Stopband rejection: 0.16 db max 53.2 db min 20

21 Filter Tuning (cont.) Filter tuned down in frequency by 5% 0 S21 (db vs MHz) 0 S21 (db vs MHz) Passband loss: Stopband rejection: 0.16 db max 59.1 db min 21

22 Filter Tuning (cont.) Effect on loss and rejection Stopband margin went from 3 db to 9 db No change in passband loss Effect on breakdown threshold Tuning screw C initially was about 10% of total C Total C went up about 10% (to tune down 5%) Tuning screw C increased nearly 100% Tuning screw gap reduced to ½ initial gap Moved into multipaction region 22

23 Filter Tuning (cont.) Actual scenario Filter space qualified at initial production Years later, another production order Technician tuned filter lower in frequency no tuning procedure for guidance wanted lots of stopband margin original design engineer not around Multipaction at 17 th T-V cycle of system test increase in system noise detected no damage to filter 23

24 Risky method: Filter Tuning (cont.) capacitive loading tuning screw with variable minimum gap combline resonator 24

25 Safer method: Filter Tuning (cont.) tuning screw with fixed minimum gap 25

26 Diplexer Transmit-Receive Diplexer TX filter and CJ designed for high power RX filter assumed to be low power What about RX filter resonator next to CJ? 26

27 Diplexer (cont.) Example RX filter CF = 2000 MHz, BW = 40 MHz TX power 40 W at 2200 MHz 43 Vrms on 1 st resonator from CJ Multipaction is possible 27

28 Matching Circuit Input: Vrms = 59 V fd = 129 MHz-cm Capacitor: Vrms = 87 V fd = 103 MHz-cm Inductor: Vrms = 50 V fd = 714 MHz-cm Isolator: Vrms = 46 V fd = 114 MHz-cm 28

29 Matching Circuit (cont.) 200 Matching Circuit 150 Vrms MHz-cm 29

30 Matching Circuit (cont.) Tested at vacuum no apparent breakdown Pressurized chamber with power breakdown was noticed Damage burn marks vaporized metal TFE melted 30

31 Breakdown Prevention Careful design Dielectric filling Clean, no contaminants Avoid critical altitude 31

32 Conclusion Multipaction is often not predictable not detectable Prevention requires careful design careful fabrication careful verification 32