Design data for the control of multipactor discharge in spacecraft microwave and RF systems.
Multipactor, a discharge occurring within satellite RF and microwave payloads, causes system degradation and failure. During testing of the European remote sensing satellite (ERS-1), problems were encountered in multipactor control. A study was initiated to examine the subject in greater detail. As part of this study, multipactor parallel plate susceptibility curves were constructed from measurements and a computer model. The effects of contamination, venting of the components, the use of dielectrics, types of glues and handling and storage on the multipactor breakdown performance also were examined. Methods of enhancing the multipactor performance were reviewed. This paper presents some of the study's findings and shows that susceptibility curves can be used directly for the design of spaceborne waveguide components. They also have applications in other transmission systems provided corrections for geometry and field intensity are made.
Multipactor discharges can occur within spaceborne high or medium power RF and microwave payloads if proper design and handling procedures are not followed. With lower power satellites this anomalous effect has not been a major problem, but difficulties were encountered when carrying out multipactor testing for the ERS-1. One of the payload's instruments has an RF power of 5 kW at 5.3 GHz. In trying to resolve these difficulties, it became evident that good multipactor susceptibility design information was not readily available. Also, in the available previously published data,[1-4] there are large discrepencies. The control of multipactor by choice of surface finish, component venting, component cleanliness and handling has been investigated. The results from the study have given a clearer insight into the mechanism of the multipactor effect and those parameters that are important for its eradication or reduction.
Multipactor susceptibility design curves for several different materials, which can be used directly for the design of waveguide components, are presented. They also have applications in other transmission media provided corrections are made for the different geometries and field distribution. The effects of different types of contamination on the multipactor threshold, the inclusion of dielectrics and other means of multipactor reduction also are presented. While component venting is a subject that has not been addressed before, it has been found to be important and design information has been provided.
The Multipactor Effect
Multipactor is a discharge produced in vacuum when an RF field exists between two surfaces and where the mean free path of the electrons is greater than the gap spacing. A discharge occurs when free electrons within the gap are accelerated by the field, strike the surfaces and initiate an electron avalanche by the release of secondary electrons, as shown in Figure 1. The discharge is resonant in nature and is dependent on the RF field, the spacing between the two surfaces and the surface secondary electron emission properties (SEE). The electron transit time must be an odd number of half cycles of the RF field period, given by (2n - 1) [Pi], where n = 1,2,3 ..., the mode index, and can occur to an n of > 16. Multipactor occurs when the primary electron energy is greater than [W.sub.f1] and less than [W.sub.f2], as shown in Figure 2. Between these energy values, the number of secondary electrons produced is greater than the number of incident primary electrons ([Sigma] = > 1), so sustaining a discharge. The classic presentation of the multipactor occurrence boundary has been used to facilitate comparison. In this format, the X axis is the product of the frequency and gap spacing in the GHz per mm and is a constant independent of frequency. The Y axis is the peak voltage occurring across the gap. When a discharge is initiated, it occurs at the lowest energy level for the particular f X d product, regardless of mode. In other components besides parallel plate, this resonance can be changed by the geometry of the discharge area, nonuniform field distributions and other second order effects.
The effects of a multipactor breakdown are well known and have been presented in many publications.[2-5] It can happen within waveguide systems, coaxial, barline and stripline components. Multipactor is most destructive when occurring within a poorly vented component. Outgassing caused by the multipactor discharge may lead to a gas discharge and component failure.
The ESTEC facility used for the multipactor test program was a pulse system operating at 5.3 GHz with a ring resonator used for power enhancement. The maximum peak power in the ring can exceed 30 kW, as shown in Figure 3.
Multipactor Design Margins
A measurement program using waveguide (parallel plate) test pieces was undertaken in conjunction with computer modeling of the susceptibility zones using simple theory. Appendix A discusses the generation of the computer model. This measurement program used 10 test pieces with (f X d) products from 1.05 to 58 GHz/mm, manufactured in five different materials, as shown in Figure 4. Table 1 lists the measurements.
The test pieces were manufactured of aluminum and copper, with extra sets plated gold, silver or with alodine treatment for aluminum. The plated coatings were 7 [Mu] m thick and contained no brighteners nor other additives. Each test piece consisted of a reduced height waveguide section matched by two quarter wave step transformers. With the low impedance test samples, stainless steel waveguide shims were necessary to maintain a good impedance match, which did not effect the threshold measurement in any way.
Data Matching and Results
Using the computer program discussed in Appendix A, the measured data was fitted to models of the different materials. Results obtained for aluminum are given in Figure 5. One important difference from the classic measurements is the knee between the n = 1 and n > 1 zone boundaries. Only alodine and copper follow a common mode boundary. This knee is thought due to second-order effects from the small gap dimensions used (0.2 mm) for an f X d of 1.06 GHz/mm. This would not have been observed in the classic work because the frequency was changed and not the gap dimensions. Another difference was the zone boundary slope change in copper and aluminum when n was greater than 3.
The values of primary electron energy at the multipactor threshold ([W.sub.f1]), for the primary mode (n = 1) and higher modes (n > 1) were extracted from the computer model. As a comparison, secondary electron emission (SEE) measurements were made on the surface of the different materials by electron spectroscopy chemical analysis (ESCA). These compared well with the n = 1 extracted values, considering experimental error.
Susceptibility Design Curves for Waveguide
Based on the results obtained for the different materials, a set of multipactor susceptibility design curves has been generated that is based on Equation 3 in Appendix A, in which all constants are included in the slope constant (C). [V.sub.peak] = [C.sub.(b)] X [(f X d).sup.2] for region (b) [V.sub.peak] = [C.sub.(a)] X (f X d) for region (a) The fXd at the changeover between n = 1 to n > 1 has been included. However, for copper it would have been prudent to use the Au/Ag boundary, but in this region not enough data points were available.
Dielectrics within a Discharge
Waveguide test samples were constructed that included commonly used dielectric on one or both of the multipacting surfaces. These were examined for their multipactor performance. Table 2 lists the multipactor thresholds of various dielectric materials. Teflon was found to be as good as alodine treated aluminum, the best metal. Alodine treated aluminum was applied as a spray to the metal surface. Kapton gave equivalent performance to silver and gold when applied as an adhesive film.
In another test, alumina strips 6 X 0.6 mm were set into the central broad walls of a copper waveguide test sample. The multipactor breakdown was similar to copper. The test piece showed erratic generation of second harmonics below the normal multipactor threshold, due to passive intermodulation within the bulk of the alumina.
Effects of Contamination
From the many multipactor tests done for ERS-1, it had been observed that contamination reduced the multipactor discharge threshold greatly. Excessive contamination caused a surface discharge to occur, which led to large temperature rises and, in some cases, to component failure.
A measurement program was undertaken using those contaminants present from normal laboratory handling, such as dust, fingerprints, glues and potting compounds. The contaminants were applied to previously tested clean waveguide test pieces. Table 3 lists the effects of various contaminants on the multipactor threshold.
One of the most destructive contaminants was the lubricant (oleamide) contained within plastic bags. It caused up to 4.5 dB of degradation. Another finding was that by using standard contaminant measurement techniques, no contamination was registered, demonstrating that a molecular layer was sufficient to reduce the measured multipactor threshold significantly.
Effects of Venting
Good venting is a necessary precaution against discharges. Multipactor starts at pressures below 2 X [10.sup.-1] mb. Above this pressure, a more destructive gas ionization discharge will occur. Different ERS-1 components of composite construction discharges other than multipactor were seen to start at pressures significantly below 2 X [10.sup.-1]. The internal pressure should always be below 2 X [10.sup.-5] mb to avoid discharges. Venting holes should be included to achieve this pressure before switch-on of the high power payload in space.
Venting Hole Calculation
Appendix B lists the calculation of venting hole size in order to avoid multipactor and gas discharge within an RF component. The ultimate pressure P [infinity] is the final achievable pressure within the component after launch, which will only reduce after many months in orbit. It can be considered as the balancing pressure achieved between pumping conductance of the venting hole and the outgassing rate of the inner surface of the component.
Need for Multipactor Testing
Although the susceptibility zones are useful for predicting multipactor in RF components, the actual multipactor threshold cannot be absolutely identified. Therefore, it is always necessary to test a representative example of the actual hardware. The flight hardware itself also should be tested. For the case of single carrier operation, the test margin should be 6 dB, which allows for degradation in the multipactor threshold due to long term contamination through handling and storage, migration of contamination in the payload after launch and SWR degradation when in orbit.
For multiple carriers the situation is more complex; the margin should ideally be 6 dB above a test power calculated from [n.sup.2] X [P.sub.i] where n = the number of carriers [P.sub.i] = the individual carrier power
Multipactor is dependent on the voltage across the gap. The peak voltage is the vector sum of the individual peak voltages for all carriers. With a large number of carriers, the probability of these carriers all being in phase is small. There is some justification for reducing the 6 dB margin.
Multipactor Reduction Techniques
As part of the study, various methods were assessed for increasing the multipactor threshold. One method involved coating by plastic or other low SEE surface. For example, Teflon spray gave the greatest increase in compactor threshold over untreated aluminum for the tested plastic materials.
Another method involved magnetic or DC biasing of the critical area, in which a fixed magnetic or electric bias was applied to suppress multipactor. With a fixed bias, electrons can be deflected away from the multipactor sensitive surfaces, which reduced the onset of a discharge. Experiments were carried out with fixed magnetic fields. Improvements were on the order of up to 1.5 dB. It was possible to instigate a multipactor discharge at a lower threshold than the nominal one if the bias voltage was incorrect.
A third method involved ion cleaning of the critical surfaces. The values of primary electron energy are much lower than those measured under high vacuum. Experiments were conducted to determine the multipactor threshold after ion cleaning; in all cases except aluminum, large increases (3 to 5 dB) in the multipactor threshold were observed. On re-exposure to air, this improvement quickly degraded to its original value. For such a technique to be used as a multipactor suppressant, the component would be required to be kept under vacuum at all times prior to launch.
Several methods for multipactor control that will allow optimum high power microwave and RF designs to be achieved have been presented. The sets of multipactor susceptibility curves can be used directly for the design of waveguide components and for other RF and microwave transmission systems.
Multipactor control should be exercised for all steps in component construction, such as material choice, component venting, surface treatment, testing, handling and storage.
SPECIFICATIONS FOR THE STANDARD WAVEGUIDE TEST SAMPLE
Gap (mm) f X d (GHz/mm) 0.20 1.06 0.28 1.48 0.40 2.12 0.70 3.70 1.00 5.30 1.50 7.95 2.00 10.60 4.00 21.20 7.00 37.10 11.00 58.30
[Tabular Data 2 to 3 Omitted]
PHOTO : Fig. 1 Multipactor electron avalanche process within parallel plates.
PHOTO : Fig. 2 Secondary emission coefficient to electron energy.
PHOTO : Fig. 3 A C-band multipactor test facility.
PHOTO : Fig. 4 A standard waveguide multipactor test sample.
PHOTO : Fig. 5 Gap volts to f X d for aluminium waveguide test samples with fitted theoretical susceptibility zones.
A.J. Hatch and H.B. Williams, "Multipacting Modes of High Frequency Gaseous Breakdown," Journal
of Applied Physics, Vol. 25, April 1954. "The Study of Multipactor Breakdown in Space Electronic Systems," NASA Report CR-448, Hughes
Aircraft Co., July 1966. R. Woo, "Final Report on RF Breakdown in Coax Transmission Lines," NASA Report 32-1500, October
1970. P. Clancy, "Multipactor Control in Microwave Space Systems," Microwave Journal, March 1978. G. August, "Multipactor Breakdown: Lessons Unlearned," (SRI) AIAA 10th Communications Satellite
Conference, 1984. A. Woode and J. Petit, "Investigations Into Multipactor Breakdown in Satellite Microwave
Payloads," ESA Journal, Vol. 14, No. 4, 1990. G. Matthaei, L. Young and E. Jones, Microwave Filters, Impedance-Matching Networks and Coupling
Structures, McGraw-Hill, New York, 1964.
Alan Woode received his electrical engineering degree from Hatfield Polytechnic in 1962. Since 1976, he has been with the European Space Agency Technology Centre in the Netherlands, where currently he is principal engineer in charge of a laboratory dealing with the technology of microwave remote sensing instrumentation. Many activities of the laboratory have been associated with technology for the successful European remote sensing satellite (ERS-1). Previously, Woode was with the Standard Telecommunications Laboratories, where he worked on early Gunn device technology. He also worked for Microwave Electronic Systems Ltd., where he developed a bistatic perimeter protection radar. Woode is a fellow of the IEE, London.
John Petit received his BSc honors in physics from Sussex University in 1980. From 1980 to 1984, he worked with Marconi Defence Systems, Stanmore, where he was an antenna engineer. Then Petit transferred to Marconi Command and Control Systems for a year. Since 1985, he has been working with the European Space Agency's Research and Technology Centre in the Microwave Instrumentation Section, where he is responsible for the development of the multipactor test facilities. Petit has been involved in diagnostic research of the multipactor phenomenon and in providing test support in the field to various satellite projects.
|Printer friendly Cite/link Email Feedback|
|Author:||Woode, A.D.; Petit, J.|
|Date:||Jan 1, 1992|
|Previous Article:||Small and large signal analysis of MESFETs as switches.|
|Next Article:||Designing iris-coupled waveguide filters using the mode-matching technique.|