MASS DRIVER-TWO RESEARCH FINAL REPORT National Aeronautics and Space Administration Grant No. NSG-3176 August 1982 Submitted to National Aeronautics and Space Administration Lewis Research Center Cleveland, Ohio, 44135 Prepared by William R. Snow U.S. Army Armament Research and Development Command Dover, New Jersey, 07801
FOREWARD This final report, submitted per the requirement of Grant No. NSG3176, documents the results of theoretical and experimental research on Mass Drivers. A forty nine stage, two phase push-pull, SCR switched accelerator of 1.25 m length was constructed and tested. Analysis of Mass Driver performance in orbit to orbit transfer and lunar launcher missions are also presented in this report. W.R. Snow Dover, N.J. August 1982
ACKNOWLEDGEMENTS The author wishes to express his gratitude to Professor Gerard K. O'Neill and all the students and staff who contributed to the effort of constructing and testing Mass Driver Two (see APPENDIX B). The following members of the staff of the Physics Department; Bart Gibbs, Joe Horvat, Ann David, Marius Isalia, and John Gomany are thanked for their assistance. Special thanks to Roger Miles, Ruth Miles, and Willie Werosta are due for their contributions. This effort has been done in collaboration with M.I.T. which was responsible for the superconducting bucket that was to be used in mass driver two testing. The members of the M.I.T. group include: Dr. Henry H. Kolm, Kevin Fine, Peter Mongeau, Fred Williams, and Peter Graneau. This effort was supported by the National Aeronautics and Space Administration under Grant NSG 3176 and in part by the Space Studies Institute. Finally, special thanks to Suzanne Bozell, Lois Smith, and Carolyn Dorko are due for their assistance in preparation of this report.
TABLE OF CONTENTS Chapter 1. INTRODUCTION 1.1 Mass Driver History 1.2 Mass Driver Applications 1.3 Mass Driver Two Chapter 2. ACCELERATOR 2.1 Drive Circuit 2.2 SCR and Optical Trigger Circuits Chapter 3. OHMIC BUCKET 3.1 Ohmic Bucket Requirements 3.2 Ohmic Bucket Circuit Analysis 3.3 Ohmic Bucket Design 3.4 Ohmic Bucket Capacitor Bank and Controller Chapter 4. RECHARGING SYSTEM 4.1 Recharging Requirements 4.2 Recharging Circuit Analysis 4.3 Recharging Capacitor Bank 4.4 Recharging Controller and SCR Trigger Circuitry 4.5 Recharging System Performance Chapter 5. INSTRUMENTATION AND CONTROL 5.1 Capacitor Bank Voltage Monitoring 5.2 Capacitor Bank Current Monitoring 5.3 SCR Trigger Pulse Recording
Chapter 6. ACCELERATOR TEST FIRING RESULTS 6.1 Theoretical Performance 6.2 Test Firing Results Chapter 7. SUMMARY 7.1 Mass Driver System Performance 7.2 Mass Driver Two Results 7.3 Suggestions for Future Work REFERENCES
APPENDIX A Mass Driver Bibliography APPENDIX B Personnel APPENDIX C Budget Summary APPENDIX D Mass Driver Two Blueprints (Listing) APPENDIX E Mass Driver Reaction Engine Blueprints APPENDIX F Mass Driver Two Overall Layout (Blueprint) APPENDIX G Mass Driver Two Electrical Block Diagrams (Blueprints) APPENDIX H Mass Driver Two Operating Procedure H.l Operating Procedure Block Diagram H.2 Operating Procedures H.3 Rack Panel Layout (Blueprint) APPENDIX I Construction and Testing of the 2.5 m mass driver APPENDIX J Mass Driver Two: A Status Report APPENDIX K Mass Driver Reaction Engine Characteristics and Performance in Earth Orbital Transfer Missions APPENDIX L High Performance Mass Drivers APPENDIX M A Small Scale Lunar Launcher for Early Lunar Material Utilization APPENDIX N The Effects of Lateral Deflections of a Mass Driver Bucket APPENDIX 0 Effective Inductance Measurements for Full-Size Mass Driver Coils
CHAPTER 1 INTRODUCTION 1.1 Mass Driver History A Mass Driver is an electrical synchronous motor used to accelerate payloads of any material to a high velocity. A bucket containing superconducting coils carry the payloads. These buckets are accelerated by pulsed magnetic fields to a predetermined velocity. Timing is done by optical position sensors and magnetic levitation forces from guidestrips is used to prevent contact of the bucket to the guideway. These buckets then release their payload and then are slowed down (empty) by a decelerator for recirculation to be reused. The first linear synchronous accelerator was built by Edwin Northrup in the 1930's at Princeton University . Later Arthur C. Clarke formally proposed a lunar launcher and used it in one of his . novels . Robert Heinlein used them in two of his novels as a lunar and earth launcher Other efforts in the 1940's were conducted in Germany during World War II for launching guided missiles and by Westinghouse for launching aircraft(B 25's). The German effort failed due to the induction drive melting the skin of the missiles before they reached any appreciable velocity. The Westinghouse "electropult” was successful but did not compete with the steam catapults then being used. These efforts were premature in that the required technology for energy storage and switching were not available. By the mid 1970's, technology became available for electromagnetic accelerators to be considered again as efficient and competitive launchers.
The first serious Long term study of coaxial accelerators (originally called TLA - transport linear accelerator, later called a mass driver) began with the efforts of Prof. Gerard K. O'Neill and Dr. Henry H. Kolm for use in space colonization and industrialization studies 1974-1976. Since then a large number of papers have been written and published by this Princeton - M.I.T. group on mass drivers and their applications, see Appendix A. Originally, mass drivers were of a planar geometry for use as a lunar launcher. Later it was found that a coaxial geometry gave much better coupling (factor of 2) and thereby better performance. However, this coaxial geometry reduced the flexibility of the bucket to contain and release payloads. Since 1974, the mass driver has gone through a number of changes in its geometry, number of phases, number of bucket coils, kind of bucket coils, guidestrips, and overall configuration and this process is still continuing. The planar geometry lunar launcher contained a three phase drive circuit and a bucket with two superconducting bucket coils. This was very close to the design of the M.I.T. Magneplane so that calculations on the design could be checked with their experimental results. The first coaxial mass driver design also contained three phases in its push-pull drive circuit but it had four superconducting bucket coils. At this time a simplified mass driver was constructed to demonstrate some of their features. Mass Driver One was constructed during the winter and spring of 1977 at M.I.T. It was a quarter cycle push only accelerator with 20 drive coils individually fed by an electrolytic capacitor. The bucket was ohmic (not superconducting) and had sliding brush contacts to obtain power. After the successful testing of Mass Driver One, further work on mass drivers continued at the 1977 Ames Summer Study. Improvements in
the designs were made with emphasis on the reaction engine used for asteroid retreival misssions. The drive circuit was changed from three to two phases but remained push-pull and the bucket now contained only two superconducting coils. At this time it was felt that very little improvement could be made and that it was time to build a working model to test this out with as many features in a real mass driver as was possible. This resulted in the Mass Driver Two effort at Princeton and M.I.T. (this report). This was the first full time effort into the design, construction, and further research on mass drivers. During this time (1978-1981) research was done on alternative mass driver designs by the groups at M.I.T. and Princeton. The M.I.T. group developed a new family of arc-commutated mass drivers that does away with SCR's and superconducting coils (in favor of ohmic coils) and which could obtain < 250,000 g's of acceleration in principle. A pulsed induction reaction engine for orbit to orbit transfer missions was developed. This "mass driver" consists of only one drive coil and an aluminum ring (< 1gm) "bucket coil" which is accelerated (> lO^6g's) and vaporized. This new mass driver can be scaled down to relatively small (few tons) payloads and should be competitive with MPD thrusters. Finally, the M.I.T. group developed another new accelerator termed the helical rail accelerator. This accelerator is equivalent to taking the rails and armature of a rail gun and twisting them into coaxial helices. At Princeton a new Pull-Only Mass Driver was developed, see Appendices L, N, and 0. This mass driver abandoned the two phase push-pull drive circuit of Mass Driver Two in favor of a multi phase (~10) drive circuit with a single superconducting bucket coil with small trailing guidance coil. If acceleration were sufficiently high, an ohmic coil
could be used instead of a superconducting coil. This new drive circuit allowed for less demanding requirements on the SCR's for large caliber accelerrators (> 40cm). And finally, a combination of the pull-only mass driver and arc- commutated mass driver using only an accelerator (no decelerator and return line) was developed for use as a lunar launcher, see Appendix M. Research is continuing on coaxial accelerators at: M.I.T. and EML by Dr. Henry H. Kolm, Dr. Peter Mongeau, and their associates (induction accelerator, helical rail launcher, etc.); ARRADCOM by William R. Snow and his associates (solenoid accelerators); Sandia National Laboratories by Maynard Cowan and his associates (theta gun), internal research by General Electric and General Dynamics, and finally at Princeton by Professor Gerard K. O'Neill.
1.2 Mass Driver Applications Mass drivers used for space applications come in the form of either a stationary launcher or as a reaction engine. As a stationary launcher, either the Earth or the Moon are suitable locations. The earth launcher could be used in launching lOOOKg payloads into earth orbit or on an escape trajectory from the solar system. Launching into orbit (from every few minutes to once a day) would require an apogee kick motor to insert it into orbit. The escape trajectory mission would be used for nuclear waste disposal (one launch per day) if that were considered desirable. In addition a mass driver could be used in a hybrid chemica1/electrica1 scheme where the mass driver replaces the first stage of a chemical (solid) rocket. This reduces the required launch velocity from the mass driver and simplifies its design. The lunar launcher could be used to support large construction efforts in space such as space manufacturing, solar power satellites, and eventually space colonization. In the near term however, it could be scaled down to a project delivering liquid oxygen (LOX) to low earth orbit quantities of 500-5000 tons per year. This LOX could be used to refuel chemical (hydrogen/oxygen) OTV's used in LEO to GEO orbit transfer missions, see Appendices A,M. This mass driver would be operated in an open loop mode with only an accelerator^ see Appendix M. It would be mostly reuseable and self sufficient but would result in a vastly simplified system for near term use. Or it could be operated in a closed loop (classical mass driver) system with recirculating buckets launching payloads to a catcher located at L2. This results in a much more complicated but nearly 100% reuseable and self sufficient system. Reaction engines (MORE) fall into two types of missions, either orbit to orbit transfer (LEO to GEO) or asteroid retreival missions.
The LEO to GEO mission MDRE would have an Isp of about 1000 sec and carry a payload of more than 4000 tons, see APPENDICES A,E, and K. The asteroid retreival MDRE would bring back to high earth orbit Apollo- Amor or Earth Trojan (if discovered) asteroids weighing up to 10^6 to 10^7 tons for large scale space manufacturing efforts . Finally, the pulsed induction reaction engine could be used for the LEO to GEO transfer mission where a few tons of payload are involved, which are clearly too small for the classical MDRE.
1.3 Mass Driver Two Mass Driver Two went through substantial changes from what was originally planned to what was finally constructed and tested. Original plans were that MD 2 would be a 1000g's, 10m closed loop synchrotron model with a recirculating bucket. It would of had 5m acceleration and 5m deceleration sections with an injection leg and possibly a free flight test section inbetween the accelerator and decelerator for observation of the superconducting bucket. This model would have had all the features of a full scale mass driver except for payload release. Since there was insufficient space at Princeton to house this device and a feeling that the return line and recirculating buckets need not be demonstrated at this time as an integral part of the mass driver, the synchrotron model was abandoned. Instead a 10m (5m acceleration, 5m deceleration) single shot linear mass driver with 1000g's acceleration and a superconducting bucket became the proposed device at the beginning of this grant. Work on the superconducting bucket was the responsibility of M.I.T. lead by Dr. Henry H. Kolm. While all other aspects and construction location was to be at Princeton. The caliber was chosen to be 13.1cm which permitted the use of four inch ID glass pipe (later plexiglass was substituted) as the vacuum jacket for the superconducting bucket and allowing for a 0.1D build on the drive coils. Energy transfer calculations were made to determine the voltages and currents required by the SCR's. Results showed that 1000g's would require a substantially larger and more expensive SCR than was desired. So, 500g's became the nominal design acceleration allowing for safety margins in the SCR specs, and repetitive operation. It was decided upon to make Mass Driver Two (10m version) into six modular units for mobility and ease of assembly. This resulted in the mass
driver becoming 2.5m, 6.4m, 10m (acceleration and deceleration sections) devices as sections were added. Work began in 1978 to design and construct (SSI paid for materials) the drive coils, SCR feeders, and I beam mounts for the entire 10m version, see Chapter 2. During that summer large quantities of IR 151RA100 SCR's, capacitors, high power diodes, and other material became available from the PPA accelerator located at the Princeton Plasma Physics Laboratory. These parts were then used to construct Mass Driver Two (10m version). Optical trigger and SCR trigger circuits were breadboarded at this time and used to fire a group of four drive coils from sector four up to 750 volts by the end of the summer (1978), see Chapter 2 and Appendix I. The fall of 1978 was spent designing the printed circuit boards for the optical and SCR trigger circuits as well as designing the electronics housings to hold these circuits and the optical sensors. Early in 1979 these electronics housings were constructed and assembled for the 2.5m section. The printed circuit boards were assembled and installed in the electronics housings in late spring. Overall assembly of the accelerator proceeded in June and is shown in Fig. 1.1
Fig. 1.1 Mass Driver Two Assembly. At this time the superconducting bucket wasn't ready for firing in the mass driver. An ohmic bucket was then built, see Chapter 3, to simulate the superconducting bucket without the need for liquid helium, vacuum system, and guide strips. The ohmic bucket became a major enterprise of its own in addition to the accelerator. In order to get the ohmic bucket moving into the accelerator while its current remained near its peak value, it required and injection system. The recharge capacitor banks for phases one and two were also constructed at this time and can be seen in Fig. 1.2 along with the accelerator, injector coils, and an ohmic bucket. The control circuitry used to operate the mass driver during these early tests is shown in Fig. 1.3. At this time a decision was made to complete only the 1.25m accelerator section. The added systems needed to operate the mass driver and the growing complexity was pushing the schedule far behind what was planned.
Fig. 1.2 Mass Driver Two during early testing with ohmic bucket. Fig. 1.3 Control circuitry during early testing with ohmic bucket.
During the fall of 1979 circuitry was constructed to simulate the firing of a bucket to test the SCR trigger and optical trigger circuitry, see Chapter 5. During this period instrumentation to measure voltages, currents, and periods were designed and constructed. Further refinements and additions to this instrumentation circuitry was continued through 1980 and is shown in Fig. 1.4. Fig. 1.4 Mass Driver Two instrumentation and control circuitry. A simplified recharge system was assembled in early 1980 to test the simplified switching circuit proposed then, see Chapter 4. After being proved feasible for the full scale system, work proceeded on completing the feeders and fusing for the recharge capacitor bank. Design and construction of the circuitry to automatically control the recharging was begun during the summer of 1980, see Chapter 4.
The Mass Driver Two accelerator section as it appeared during the summer of 1980 is shown in Fig. 1.5. Shortly afterwards the PPA capacitors Fig. 1.5 Mass Driver Two accelerator. shown in the figure had to be disposed of because they contained PCB's. New capacitors were ordered but a nine month delay occurred before these new non PCB capacitors were received and installed. Testing of the complete two phase accelerator could not be done during this period. After the new capacitors were installed, testing resumed of the full two phase accelerator with recharging with an ohmic bucket. Test results during this period (spring 1981) are presented in Chapter 6 before the accelerator was disassembled in the author's absence. A summary of these test results of the accelerator, system analysis of mass drivers, and future work are presented in chapter 7.
CHAPTER 2 ACCELERATOR 2.1 Drive Circuit The two phase push-pull drive circuit used in Mass Driver Two is shown in Fig. 2.1. Details of the circuit, operating conditions, and drive coil designs can be found in APPENDIX I. Construction of drive coils used in Mass Driver Two are shown in Fig. 2.2 to Fig. 2.5. Fig. 2.1 Six drive coils of phase 1, shown at the time of peak positive current; at this time there is zero current in phase 2. Four coils are excited, one to pull and one to push each of the two bucket coils. Spacing is chosen equal to the separation between a drive and bucket coil at peak gradient dM/dx. SCR's are triggered through pulse transformers for electrical isolation between the drive circuit and the electronic triggering circuitry.
Fig. 2.2 Drive coil winding Fig. 2.3 Attachment of flexible lead to drive coil
Fig. 2.4 Attachment of neutral bus for phase one of sector four.
Fig. 2.5 Construction of top drive coil clamp.
Based on drive coil design parameters given in APPENDIX I theoretical values of the half cycle ringing period for phases one and two are given in Tables 2.1 and 2.2. Measured ringing periods are also given in Tables 2.1 and 2.2. For phase one, circuit resistance and inductance were calculated from actual capacitor voltage and ringing period measurements. where Construction and assembly of the SCR feeder strips are shown in Fig.'s 2.6 and 2.7. Details of the SCR feeder strips can be found in APPENDIX I. circuit resistance circuit inductance sector capacitance ringing period initial capacitor voltage final capacitor voltage
Table 2.1 Phase One Circuit Parameters Firing Number Ringing Period (theory) (us) Ringing Period (meas.) (ps) Inductance (uh) Resistance (m^ ) 1 995 999 49.1 35.5 3 960 944 43.8 33.3 5 870 868 43.8 33.3 7 765 781 30.0 27.2 9 745 745 27.3 25.4 11 745 731 26.3 24.6 13 700 707 24.6 23.1 15 588 605 18.0 20.1 17 516 528 13.7 17.4 19 516 524 13.5 18.4 21 516 528 13.7 17.1 23 475 485 11.6 16.3 25 451 445 — — 27 451 444 9.71 15.2 29 417 415 9.72 14.5 31 383 378 7.04 12.9 33 383 378 7.04 12.3 35 383 376 6.97 12.5 37 360 357 6.28 11.9 39 338 334 5.49 11.8 41 338 335 5.52 11.7 43 338 335 5.52 11.7 45 338 335 5.52 12.1 47 338 334 5.48 12.5 49 338 337 5.59 11.8
Table 2.2 Phase Two Ringing Periods Firing Number Ringing Period (theory) (us) Ringing Period (meas.) (us) Firing Number Ringing Period (theory) (Us) Ringing Period (meas.) (us) 2 960 960 26 451 439 4 849 871 28 417 410 6 765 760 30 383 374 8 745 742 32 383 373 10 745 737 34 383 373 12 700 708 36 360 355 14 588 600 38 338 332 16 516 — 40 338 332 18 516 516 42 338 — 20 516 519 44 338 — 22 475 479 46 338 — 24 451 — 48 338 334
Fig. 2.6 Construction of SCR feeder strips.
Fig. 2.7 Assembly of SCR feeder strips.
2.2 SCR and Optical Trigger Circuits The SCR and optical trigger circuits are described in detail in APPENDIX I. Preliminary testing of the trigger circuits is shown in Fig. 2.8. After suitable circuits were obtained, prototype circuit boards were made and tested with phase one sector four drive coils in their actual locations at full power. This was done to test for false triggering and pickup problems before freezing the circuit board designs. Testing of these circuit boards is shown in Fig. 2.9. Fig. 2.8 Preliminary testing of SCR and optical trigger circuits.
Fig. 2.9 Test firing of drive circuit with prototype optical and SCR trigger circuit boards.
Final circuit boards were assembled and installed into the electronics housings. The alignment and testing of the installed circuit boards is shown in Fig. 2.10. Fig. 2.10 Alignment and testing of optical and SCR trigger circuit boards in phase two.
The actual firing delays were measured in the optical trigger circuit boards and are given in Table 2.3. Table 2.3 Firing Delays in Optical Trigger Circuit Boards PHASE 1 PHASE 2 1 A4 1840 2 Bl 808 3 Bl 470 4 B2 369 5 B2 225 6 B3 151 7 B3 158 8 B4 230 9 B4 120 10 Cl 143 11 Cl 76 12 C2 114 13 C2 90 14 C3 160 15 C3 116 16 C4 129 17 C4 114 18 DI 119 19 DI 86 20 D2 100 21 D2 94 22 D3 123 23 D3 94 24 D4 104 25 D4 90 26 El 81 27 El 102 28 E2 99 29 E2 60 30 E3 78 31 E3 88 32 E4 105 33 E4 84 34 Fl 70 35 Fl 78 36 F2 84 37 F2 78 38 F3 71 39 F3 84 40 F4 83 41 F4 70 42 G1 60 43 G1 54 44 G2 72 45 G2 66 46 G3 42 47 G3 60 48 G4 53 49 G4 44
CHAPTER 3 OHMIC BUCKET 3.1 Ohmic Bucket Requirements The ohmic bucket originated from the desire to be able to test mass driver two before the superconducting bucket was ready from M.I.T. This permitted routine testing of mass driver two without the need for vacuum and liquid helium. The ohmic bucket consists of two aluminum wire coils mounted on a G-10 tube as shown in Fig. 3.1. The capacitor bank that powered the ohmic bucket through a tether is shown in Fig. 3.2. To obtain current densities comparable to the of a superconductor it was necessary to cool the ohmic bucket to liquid nitrogen temperature. This reduced its resistance by a factor of more than nine. Cooling the ohmic bucket with liquid nitrogen is shown in Fig. 3.3. A hose with warm air is positioned near the resistor-diode network at the back of the ohmic bucket to prevent frost buildup. During a test firing the ohmic bucket is removed from the liquid nitrogen and placed into the injector. Within 15 seconds (to prevent it from warming up too much and increasing its resistance) the ohmic bucket is fired. It was caught with a backstop consisting of a box of styrofoam backed with lead bricks as shown in Fig. 3.4. Based on energy transfer calculations for different buckets (mass and coupling) currents of 24,000 to 45,300 amp turns are required to obtain 500 g's of acceleration. Previous design details of the ohmic bucket and its associated injector can be found in APPENDIX I.
Fig. 3.1 Modular ohmic bucket (a = 0.514) Fig. 3.2 Ohmic bucket capacitor bank.
Fig. 3.3 Ohmic bucket chilldown with liquid nitrogen.
Fig. 3.4 Ohmic bucket caught by backstop after a test firing.
3.2 Ohmic Bucket Circuit Analysis The basic ohmic bucket circuit is shown in Fig. 3.5 Fig. 3.5 Basic ohmic bucket circuit.
After the SCR is turned on the following circuit shown in Fig. 3.6 is used for analysis. Fig. 3.6 Ohmic bucket circuit. The solution of this circuit for bucket coil current is:
The shape of this bucket coil current is shown in Fig. 3.7. The main feature is a relatively flat current after reaching its peak (<5% droop) during the acceleration period. To prevent unnecessary heating of the bucket coils a 20 resistor is connected across the capacitor to quickly discharge it at time A bucket with an = 0.514 was tested at liquid nitrogen temperature to obtain its total ( ) coil current waveform. The circuit parameters and operating point were used then to calculate the total coil current using the previously derived equation. The following values were used: where initial capacitor voltage bucket coil inductance bucket coil resistance (at 77°K) capacitor bank capacitance 176 Q (meas) 2.68 h (theory) 6000 turns (meas) 10,000 uf (label) 100 volts (meas)
Fig. 3.7 Ohmic bucket current waveform.
The two waveforms are tabulated in Table 3.1. Given the uncertainty in the values of inductance and capacitance the two waveforms show close agreement after reaching peak current. The real bucket coil reached its peak current approximately 10 ms earlier than expected and is probably due to a smaller inductance or capacitance. Table 3.1 Total Ohmic Bucket Current for \ = 0.514 Bucket Time (ms) Total Current (meas.) (Amp) Total Current (theory) (Amp) Capacitor Voltage (theory) (V) 0 0.000 0.000 100.0 5 0.468 0.318 99.9 10 0.679 0.546 99.7 15 0.818 0.710 99.4 20 0.911 0.827 99.0 25 0.975 0.910 98.6 30 1.018 0.968 98.1 35 1.045 1.009 97.6 40 1.060 1.036 97.1 45 1.072 1.054 96.6 50 1.077 1.066 96.0 56 1.078 1.073 95.4 70 1.072 1.074 93.9 80 1.064 1.068 92.8 90 1.053 1.059 91.7 100 1.043 1.048 90.7 120 1.023 1.026 88.6 150 0.990 0.991 85.6 170 0.969 0.969 83.6 189 0.950 0.947 81.8
3.3 Ohmic Bucket Design Three different sizes of ohmic bucket coils were designed, built and tested. The first one used had an = 0.514 and both coils were wound on a GIO tube (see APPENDIX I). This was of integral construction to minimize mass. This proved to be troublesome. When a coil burned out the other would be still good but the entire bucket would be useless. A modular design then was made so that all parts could be interchanged and mass produced for rapid testing (see Fig. 3.1). Later ohmic buckets were built with 's = 0.63 and 0.65 to improve coupling and a reduced mass was attempted in these designs. The basic parameter of these ohmic buckets are given in Table 3.2. Table 3.2 Ohmic Bucket Parameters Tests were performed to measure the bucket coil resistance after removal from the liquid nitrogen and are given in Table 3.3. This shows that a test firing should happen within 15 seconds from removal of the ohmic bucket from liquid nitrogen and placed in the injector.
Other tests were performed to measure the bucket coil resistance before and after a test firing to observe heating effects at different currents. These tests are summarized in Tables 3.4 and 3.5 for the two different buckets. Table 3.3 Ohmic bucket resistance as a function of time after removal from liquid nitrogen. time (sec) = 0.514 bucket resistance () = 0.63 bucket resistance () q= 0.65 bucket resistance () 0 96.0 62.4 92.8 5 97.3 63.5 93.7 10 99.9 66.6 98.9 15 104.9 71.8 107.3 20 111.2 77.7 117.8 25 119.0 83.8 128.2 30 126.5 90.6 138.9 40 143.3 103.1 161.0 50 159.4 116.5 183.5 60 176.1 128.6 205 80 210.2 151.5 242 100 237 173.2 279 120 264 195 309 150 302 221 351 180 335 246 387 210 367 269 420 240 395 289 450 270 421 306 476 300 445 323 500 00 905 577 849
Table 3.4 Ohmic bucket resistance before and after firing at different voltages (a = 0.514). Table 3.5 Ohmic bucket resistance before and after firing at different voltages (a = 0.63). capacitor voltage = 795V 411V 113V initial bucket coil resistance (ft) 124.8 124.8 124.8 final bucket coil resistance (ft) 276.6 225.0 198.8 peak bucket coil current (amps) 4.03 2.22 0.64 time of current peak (ms) 24.1 28.5 23.3 capacitor voltage = 794V 392V 108V initial bucket coil resistance (ft) 192.0 192.0 192.0 final bucket coil resistance (ft) 320.2 277.8 256.0 peak bucket coil current (amps) 3.28 1.63 0.48 time of current peak (ms) 40.8 46.4 47.5
3.4 Ohmic Bucket Capacitor Bank and Controller Two ohmic bucket capacitor banks were designed and constructed for mass driver two testing. The first one was a 1 KV, 10,000 unit which is described in detail in APPENDIX I. Modifications to the circuit since that report was written include: elimination of over voltage protection circuit, disconnect on tether, single 60A fuse (replaced with 20A fuses on individual capacitors), addition of mechanical shorting bar, addition of current shunt on negative side of capacitor, elimination of case ground, a DPDT was substituted in place of the SPST relay with a 50 clamping resistor on the relay input, and an auxiliary SCR trigger circuit was added to insure that the "turn on" SCR latched properly. This circuit is shown in Fig. 3.8. The previous SCR trigger circuit had a 3 ps pulse width (used in drive circuit) which was not on long enough to have the SCR latch due to a very small di/dt. The controller circuit that was used to sequentially fire the ohmic bucket, fire the injector after some delay, and finally turn off the charging is shown in Fig. 3.9. The first ohmic bucket capacitor bank was normally operated at 800 volts. Later energy transfer calculations showed that a higher current was required by the heavier buckets. The calculations also showed that 1500 volts would be required for the = 0.514 bucket in order to obtain 500 g's acceleration. Therefore a new ohmic bucket capacitor bank was designed to meet this requirement with a large reserve margin in voltage if higher acceleration tests were desired. It had a maximum voltage of 2 KV, a
capacitance of 8300 uf and was comprised of the same electrolytic capacitors used in the first capacitor bank (2500 , 350 volt). The circuit diagram of this capacitor bank is shown in Fig. 3.10. Each "capacitor” in the diagram consists of 8 parallel connected capacitors as they were obtained from PPA surplus. The new ohmic bucket circuit is shown in Fig. 3.11 and charging circuit in Fig. 3.12. Features of these circuits are that the "turn on" SCR's had to be stacked for voltage which required voltage dividers and snubbers as well as auxiliary SCR trigger circuits with isolation. A clamping resistor is connected on the DPDT relay input to prevent the voltage across the SCR's from floating around. The contacts of the heavy duty relay required RC snubbers across them to reduce contact bounce effects. The ohmic bucket capacitor bank circuit shows that all operations are done by remote control with relays for safety.
Fig. 3.8 Auxiliary SCR trigger circuit.
Fig. 3.9 Ohmic bucket controller logic.
Fig. 3.10 2 KV ohmic bucket capacitor bank.
Fig. 3.11 2 KV ohmic bucket circuit
Fig. 3.12 2 KV ohmic bucket capacitor bank charging circuit.
CHAPTER 4 RECHARGING SYSTEM 4.1 Recharging Requirements After each firing of a group of drive coils, energy must be supplied to the sector capacitor in order for it to be fully charged for the next firing. Energy loss from the sector capacitor is due to energy transfer to the bucket and to resistive heating in the drive coils, feeders, and sector capacitor. Measurements of sector capacitor voltage before and after firing were made in order to determine total resistive heating loss. The initial stored energy in the sector capacitor is: The voltage on the capacitor after a half cycle is given by:
Final stored energy is given by: one-half cycle period circuit resistance circuit inductance Therefore the circuit efficiency is: The results of measurements made of phase on of the accelerator are given in Table 4.1.
Table 4.1 Phase One Accelerator Drive Coil Efficiency with no Energy Transfer to Bucket Firing Number Initial Voltage (Volts) Final Voltage (Volts) Efficiency 1 396 -293 0.548 3 -265 191 0.520 5 371 -273 0.542 7 -282 204 0.523 9 363 -272 0.561 11 — — — 13 374 -283 0.573 15 -301 225 0.560 17 354 -270 0.582 19 — — — 21 359 -275 0.587 23 -297 223 0.564 25 351 -270 0.591 27 -302 225 0.555 29 356 -272 0.584 31 -292 216 0.548 33 356 -272 0.584 35 -262 193 0.543 37 340 -259 0.581 39 -295 215 0.531
Table 4.1 Continued: - = No Data Available Due to SCR Failure 41 348 -262 0.567 43 -286 210 0.539 45 353 -266 0.569 47 -274 197 0.517 49 381 -286 0.564
The amount of energy required during a recharge cycle is given by where Two types of buckets were intended to be used in testing. One was an ohmic bucket with a maximum mass of 0.6 KG and mechanical guidance. The other would be a superconducting bucket with a mass of 1.0 KG and copper guide strips which were assumed to reduce the drive force by 20%. The ohmic bucket results in a low energy transfer requirement of . The superconducting bucket with guide strips has a high energy transfer requirement of . The final voltage on the sector capacitor is determined to be:
Required recharge energy and sector capacitor voltage are given in Table 4.2 for the phase one accelerator with the low and high energy transfer demands. The initial voltage on the sector capacitor was assumed to be 750 volts (the nominal design) and efficiencies from Table 4.1 were used. Several other requirements placed on the recharge system were a) that the time of recharging be approximately ) switching be done as to prevent voltage spikes (from forced commutation) so that the SCR's in the drive circuit would not be destroyed, c) polarity reversing switching be used to enable recharge on both polarities of the sector capacitor, and d) that the control circuitry be as simple as possible. Table 4.2 Phase One Accelerator Recharge Requirements for Low and High Demand () Low High Firing Number Recharge Energy (J) Final Cap. Voltage (volts) Recharge Energy (J) Final Cap. Voltage (volts) 1 335 483 417 390 3 351 467 433 369 5 339 479 421 385 7 350 468 432 370 9 328 490 410 398 11 — — — — 13 321 497 403 407 15 329 489 411 397 17 316 502 398 413 19 — — — —
Table 4.2 Continued: 21 313 505 395 417 23 326 492 408 401 25 311 507 398 419 27 331 487 413 395 29 315 503 397 414 31 336 482 418 388 33 315 503 397 414 35 338 480 420 386 37 317 501 399 412 39 345 473 427 377 41 325 493 407 402 43 341 477 423 382 45 324 494 406 403 47 353 464 435 366 49 326 492 408 401 - = no data available due to SCR failure
4.2 Recharging Circuit Analysis The recharge system went through a number of changes before a final design was made to work. In APPENDICES I and J, earlier work on the recharge system is described. It was first thought that a forced commutation circuit would be necessary to turn off the recharging pulse. While developing that circuit it was soon realized that if sufficient series inductance were present in the circuit and the Q high enough it would oscillate with a period close to the desired value. This would permit the use of an SCR as the "turn on" switching device which later would commutate off at the end of a half cycle of recharging current. A simplified circuit diagram for one polarity of the recharge system is shown in Fig. 4.1. Fig. 4.1 Simplified recharge circuit.
A requirement of the circuit is that the recharge capacitor be much greater than the sector capacitor. This is to prevent significant droop of the recharge capacitor voltage and makes it look like a relatively constant voltage source. Before completing the full system for mass driver two, a small capacitor bank was assembled and used for charging the sector capacitor as shown in Fig. 4.2. Fig. 4.2 Recharging system test circuit.
Test results showed that the basic circuit would work. These results will be presented later in this chapter. Since the recharge capacitor bank is much larger than the sector capacitor it can be treated as a constant voltage source (battery) during a recharging period. The circuit shown in Fig. 4.1 can be redrawn using Laplace Transforms shown in Fig. 4.3. Fig. 4.3 Simplified recharge circuit.
It has been assumed that there is no initial stored energy in the inductance but there is initial stored energy in the sector capacitor. The voltage across the sector capacitor can be expressed as: Solving for the sector capacitor voltage in the time domain gives: where At the end of one-half cycle of recharging the final voltage across the sector capacitor is: where = period of recharging
The recharging current can now be determined from: which gives: At one-quarter cycle the recharging current is near its maximum giving: After each recharge period the recharge capacitor voltage will be lowered by an amount equivalent to the energy lost by the recharge bank to charging the sector capacitor and ohmic heating losses. The energy stored initially in the recharge capacitor bank on the nth is given by: On the next firing
where = energy lost by the recharge capacitor on the nth firing. Therefore the new recharge capacitor voltage for the (n + 1) firing is: Some of the demands placed on the SCR used to do the switching (actually two will be used) are: 1. Voltage rating be at least two times greater than the recharge so as to be able to withstand the voltage reversals on the sector capacitor. 2. The di/dt rating be: 3.
4.3 Recharging Capacitor Bank The completed recharge capacitor bank for phases one and two is shown in Fig. 4.4. Fig* 4.4 Recharge capacitor bank for phanse one and two of the accelerator.
The capacitor bank for one phase consists of three shelves of electrolytic capacitors (2500 pf, 350V) wired in series. Each shelf consists of sixteen groups of eight capacitors all wired in parallel and connected to a common feeder as shown in Fig. 4.5. Fig. 4.5 Interior of recharge capacitor bank showing shelves of capacitors and SCR trigger circuit.
Each individual electrolytic capacitor was fused with a ceramic ABC 20A slow blow fuse rated at 250 volts. This was done for safety reasons since each recharge capacitor bank could contain as much as 50 KJ of energy which might discharge through an individual capacitor if it failed. The circuit diagram of the recharge capacitor bank is shown on Fig. 4.6. Fig. 4.6 Recharging capacitor bank circuit.
The circuit diagram showing how the SCR1s are connected for polarity reversing are shown in Fig. 4.7. This enables the recharge capacitor bank to charge the sector capacitor after every half-cycle of oscillation of the drive circuit. Fig. 4.7 Polarity reversing switch for recharging system.
The interconnecting feeder that wires the three shelves of capacitors in series, and contains charging chokes and SCR's is shown in Fig. 4.8. Fig. 4.8 Interconnecting feeder for recharge capacitor bank.
The feeders that run from the recharge capacitor bank to the sector capacitors of phases one and two are shown in Fig. 4.9. They are in a protective housing for safety. Fig. 4.9 Side view of accelerator showing feeder strips (in protective housing) connecting recharge capacitor bank to sector capacitor (not shown).
The sector capacitor that is being recharged is shown in Fig. 4.10. It is also in a protective housing for safety reasons. The sector capacitor consists of five () energy storage capacitors (made by Maxwell Laboratories) wired in parallel. The remote charge/discharge panel for the sector capacitor can be seen on the side of the housing. Fig. 4.10 Phase one sector capacitor (in protective housing).
The charging circuit for the recharge and the sector capacitor banks is shown in Fig. 4.11. The power supplies and control circuitry for phases one and two can be seen the fourth rack in Fig. 1.4.
Fig. 4.11 Recharging capacitor bank charging circuit.
4.4 Recharging Controller and SCR Trigger Circuitry The recharge controller unit is used to determine the polarity and magnitude of both the voltage and current of the sector capacitor. This permits the proper triggering of SCR's after a drive coil firing to recharge the sector capacitor. The sequence of events before and during the recharge period is shown in Fig. 4.12. A block diagram showing how the proper SCR trigger signals are generated from sensing sector capacitor voltage and current is shown in Fig. 4.13. The circuit diagrams for the different circuits contained in Fig. 4.13 are shown in Figs. 4.14-4.17.
Fig. 4.12 Recharging system event sequence.
Fig. 4.13 Recharging controller block diagram.
Fig. 4.14 Switching circuit.
Fig. 4.15 Current clipping circuit.
Fig. 4.16 Capacitor voltage amplifier circuit.
Fig. 4.17 Scope trigger circuit.
Fig. 4.18 Construction of recharging controller. The recharging controller which contains the circuit cards for phases one and two is shown in Fig. 4.18. It can be seen in Fig. 1.4 in the third rack with its associated power supplies.
The SCR trigger circuit used to trigger the proper SCR* s for recharging is derived from the one used for the drive circuit. Two SCR's must be turned on simmultaneously for recharging to occur for a single polarity. The block diagram of the SCR trigger circuit is shown in Fig. 4.19 and the individual trigger circuit diagram is shown in Fig. 4.20.
Fig. 4.19 SCR trigger circuit block diagram for recharging system.
Fig. 4.20 SCR trigger circuit for recharging system.
Output pulse transformers were used to provide electrical isolation between the two SCR's being triggered. The pulse transformer consisted of a stack of four ferrite cores used in the drive circuit SCR trigger circuit and wound with fifteen turns of wire on both primary and secondary. These were driven with a Darlington pair which in turn was driven with a Darlington pair with isolated outputs. Inverters with open collectors were used to get the proper wave form to drive the primary Darlington. Since the power supplies operating the recharging controller and SCR trigger circuits were differnt and had different cable lengths, optical couplers were tried so as to prevent ground loops. They were not satisfactory and a futher pulse transformer was used on the input to provide isolation and noise immunity. 4.5 Recharge System Performance Results from testing the recharging system test circuit shown in Fig. 4.2 are now presented. The recharge capacitor bank in that circuit consisted of 16 parallel electrolytic capacitors. The sector capacitor was comprised of 11 surplus PPA capacitors wired in parallel with a total capacitance of 2033 pf. Since the recharge capacitor bank has a total capacitance of compared to of the sector capacitor it can be treated as a voltage source. Measured values of voltage, current, and recharging period are given in Table 4.3. The results of these tests are given in Table 4.4. The results showed that this simplified recharge circuit (over that of a forced commutation circuit) could be used to recharge the sector capacitor.
Table 4.3 Recharge system test circuit parameters Table 4.4 Recharge system test circuit results However, the recharge period was 50 Us longer than desired. This was due to the larger than expected circuit inductance in the recharge capacitor and its associated wiring. The larger resistance is due to losses in the electrolytic capacitors and skin effect in the wiring. After these tests confirmed the basic operation of the recharge system, the full scale recharge capacitor bank and control circuitry were then completed. Results from testing the full scale recharge capacitor bank and control circuitry are now presented. Tests were conducted by precharging the recharge and sector capacitors and manually triggering a drive coil firing. After the drive coil has oscillated for one half cycle the recharge controller circuitry then senses the proper polarity and time to trigger the recharge SCR's. Typical waveforms for the sector capacitor voltage and recharge current are shown in Figure 4.21 for the tests performed. Voltages measured for the recharge and sector capacitors are given in Table 4.5. Tests no. 1, 3, 5 and 6 were done with drive coil firing no. 1 and tests no. 2, 4, 7 and 8 were done with drive coil firing no. 48.
Fig. 4.21 Sector capacitor voltage and recharge current waveforms
Table 4.5 Recharge system circuit voltages The amount of energy delivered to the sector capacitor and energy lost by the recharge capacitor is given in Table 4.6. The recharge system efficiency is defined as the ratio of SCB energy gain to RCB energy lost. Test 5 showed a 98% efficiency which is not consistant with the other values. A one volt difference in the recharge capacitor voltage would reduce the efficiency to 73% which is more in line with the other values. Similar remarks can be made for tests no. 7 and 8 since tests 5-8 had a - 1 volt measurement accuracy for the recharge capacitor voltage. Therefore the average recharge system efficiency is 72%. Other parameters of the recharge system are given in Table 4.7. The recharge period (average) is which is reduced from the value in the early test circuit for the recharge system. The goal was to have a period of but stray inductance in the electrolytic capacitors prevented this from being obtained. Allowing for about for reverse recovery on the SCR's and a delay before SCR turn on brings the total recharge time to . This period is a dead spot in the drive current waveform which reduces the duty cycle of that phase. The peak current and average i^2t values (when scaled up) seem reasonable for the SCR's.
The di/dt values (when scaled up) are close to the maximum design specs for the IR 15/RA100 SCR's used. This is a problem which was later confirmed when accelerator tests were done at higher voltages. A number of failures occurred where the SCR's fused in a dead short across the recharge capacitor bank. This was due to excessive di/dt and thermal shock within the silicon wafer. A higher voltage (2200V) and higher amperage flat pac SCR would have made the recharge system useble at full voltage (full power). Table 4.6 Recharge system circuit energies Table 4.7 Recharge system circuit parameters
CHAPTER 5 INSTRUMENTATION AND CONTROL 5.1 Capacitor Bank Voltage Monitoring Sector capacitor bank and recharge capacitor bank voltages were monitored by two means. One was with digital panel meters (NLS RM-350) having three and one-half digits and a fixed range (1000 volt maximum) for DC voltage measurements during charging and discharging. These can be seen in Fig. 1.4 in the third rack from the left for the accelerator, decelerator, and recharge capacitor banks. The two digital panel meters seen in the first rack are for the ohmic bucket and injector capacitor banks. The drive circuits and recharge circuits containing the sector and recharge capacitor banks are electrically floating and balanced. To monitor the AC waveforms by an oscilloscope required that the voltage be divided to be displayed. This had to be done to eliminate a potential shock hazard and the introduction of unwanted ground loops. A balanced voltage divider circuit was employed and is shown in Fig. 5.1.
The voltage observed by the oscilloscope with balanced inputs is given by: where also Typical resistor values are:
These values give a high input impedance to reduce loading on the capacitor while giving a low output impedance to reduce loading effects by the oscilloscope. Voltage divider circuit card calibrations are given in Table 5.1 using the previous typical resistor values. Table 5.1 Voltage Divider Card Calibration
At a later date several voltage divider circuit cards were modified to give a larger voltage output to operate the recharge controller circuitry. The values of and were increased to . The new voltage divider circuit card calibrations are given in Table 5.2.
5.2 Capacitor Bank Current Monitoring Current measurements were made with Rogowski coils (with integrators). They were surplus units from the Princeton-Pennsylvania Accelerator. Calibration of these current transformers gave 5.00 mV/amp over a frequency range of 100 Hz to 20 KHz with a flat response. These are well suited for the 1 KHz frequencies found in the drive circuit. 5.3 SCR Trigger Pulse Recording As the bucket travels through the accelerator it interrupts evenly spaced beams of light to sense bucket position. This signal is then delayed before being used to fire the group of four SCR's (see APPENDIX I). These pulses (width) from the optical sensors of each phase are fed to a fan in circuit which is a "massive OR”. This results in all of the pulses appearing on a single channel that can be easily recorded. A typical output for the accelerator is shown in Fig. 5.2. Velocity and acceleration of the bucket can be obtained from this information. In addition, a pulse counter can be used to quickly check that all optical sensor circuits are properly functioning by counting the number of pulses. The fan in circuit can be seen in Fig. 1.5 at the end of the accelerator in the aluminum box. Along with the pulse recording system there is an associated "firing simulator" circuit. It is used to generate a sequence of pulses that electrically simulate the motion of a bucket through the accelerator. This permits easy and routine testing of all the SCR and
optical trigger circuits without a bucket. It can be fired in a single shot or repetitive mode. The firing simulator circuit can be seen in Fig. 1.5 at the beginning of the accelerator in the large aluminum box. Fig. 5.2 Fan in circuit output.
CHAPTER 6 ACCELERATOR TEST FIRING RESULTS 6.1 Theoretical Performance The waveform for the drive current of one phase of the two phase push-pull drive circuit of mass Driver Two is shown in Fig. 6.1. Fig. 6.1 Current and voltage waveforms during the passage of one bucket coil. A computer program was written to calculate the energy transfer of the drive circuitry using actual values of firing delay, inductance, resistance, and capacitance for the different drive coil firings and including recharging. The values of inductance and resistance used can be found in Chapter 2. The circuit used for modelling the energy transfer is shown in Fig. 6.2. Details of the circuit analysis and difference equations used can be found in Appendix J.
Fig. 6.2 Basic drive coil configuration for Mass Driver Two Performance calculations were made with four different bucket designs. They had 's of 0.514 (the original ohmic bucket), 0.60, 0.65, and 0.70. The total ohmic bucket mass used was 400 gms (the original one had a mass of 530 gms) and was a value that was thought could be achieved, but never was. Current in an individual bucket coil was chosen to be 20,000 amp turns (20,000 to 25,000 amp turns was being achieved in the original bucket). The injection velocity into the accelerator was chosen to be 10 m/sec (10-15 m/sec was the range used in testing). And the sector capacitor voltage was varied from 600 volts to 800 volts (750 volts was considered nominal in the design). The results of these performance calculations are presented in Fig. 6.3. These results lead to the construction of higher ohmic buckets of reduced mass. An ohmic bucket with an = 0.73 was constructed and
RkJQdWJsaXNoZXIy MTU5NjU0Mg==