|Contents of this page
This page describes the construction of an ICR helicon ion source.
After the construction of the basic Fusor, which finally ended up as a cylindrical Fusor instead of the originally intended spherical Fusor, due to the unexpected availability of a suitable T-shaped cylindrical vacuum chamber, we realised that this cylindrical vacuum chamber was also a good and suitable starting point for integrating an ion source. The purpose of the ion source is to increase the ion density in the Fusor for obtaining a higher fusion rate observed by a higher neutron output.
A subproject was initiated to construct a helicon ion source that could be integrated into the Fusor project.
We discussed the characteristics of ECR/microwave sources and ICR helicon sources on our Theory webpage and concluded that a helicon source was a preferred choice.
Please note with respect to your health that by using the described device on this webpage the maximum permissible exposure levels averaged over 6 minutes for the FM and VHF range from 88 to 108 MHz are an RMS electric field strength of 63.2 V/m, an RMS magnetic field strength of 0.158 A/m and a power density of 1 mW/cm2.
Usually the preferred frequencies for operating ICR helicon plasma sources are chosen for ions with a charge of 1 at the natural frequency of 13.56 MHz in a magnetic field of 1.78T (and its harmonics) or at 2.45 GHz in a magnetic field of 86.6 mT, which is the natural frequency at which electrons gyrate and which is also a popular frequency due to the availability of cheap useable parts from microwave ovens. Most important fact is, however, that these frequencies are outside the RF frequencies that have been allocated to telecommunications.
Introduction to the Theoretical Background
The purpose of the helicon ion source is heating Deuterium gas to a plasma with a high density. The applied method of heating the plasma is by employing electromagnetic waves whose frequencies correspond to the cyclotron frequencies of ions and their harmonics. By launching electromagnetic waves at the ion cyclotron frequencies into the plasma a strong absorption of energy occurs because the ion cyclotron frequencies are one of the natural resonant frequencies of the plasma.
Helicon waves are electromagnetic waves, which travel along magnetic field lines from an external aplied magnetic field. The name "helicon" has been derived from the effect that the electric and magnetic field of the travelling waves rotate in a helix trajectory.
The advantage of an IRC helicon source as well as an ECR/microwave source is the fact that plasma heating occurs electrodeless. Both types of devices are low-pressure wave-heated plasma generators which produce a plasma density of 1010 - 1013 cm-3. The helicon ion generator is a device that produces electromagnetic waves propagating in magnetized plasma in the helicon mode (reference 1).
The helicon wave is a whisler wave (or R-wave) propagated in (a) magnetic field(s) where the applied RF frequency is in the ω hierarchy range:
ωci = ion cyclotron frequency
ωce = electron cyclotron frequency
The choice of RF frequency and the applied magnetic field strength are flexible parameters ranging from Mhz to GHz and from 30 G to >1 kG respectively because the plasma production does not strictly require a resonant condition.
The Lorentz force will cause an ion in a static magnet field to move in a circular pattern, a cyclotron motion. The cyclotron motion has an angular frequency ω equal to
For a given magnetic field strength B the angular frequency ω is given by
e is the elementay charge in Coulomb units (for a Deuteron = 1.60217733E-19 C)
z is the number of negative or positive charges of the ion, dimensionless (for a Deuteron = 1)
m is the mass of the ion in kg units (for a Deuteron = 3.3435860E-27 kg)
B is the magnetic field strength in Tesla units.
An electric exitation signal with a frequency f therefore will resonate ions with a mass-to-charge ratio of m/z given by
A uniform axial motion may be superimposed on the circular motion resulting in a circular curve in a three dimensional space, or a helix. When a uniform motion is superimposed perpendicular to the field (e.g. an electrical of gravitational field) a cycloid will result (reference 2).
When a neutral gas is entered into the plasma chamber and the electrons in the gas molecules are energized by the RF wave, the electrons have the change of ionizing or exciting neutrals upon collision. When the electron temperature is low (< 5 -7 eV) excitations followed by de-exitations or permanence as a meta-stable state are major loss mechanisms.
Only when high electron temperatures are achieved (> 10 eV) and when the ionization mean free path is smaller than the chamber length (λion << L) efficient ionization will take place. However, in addition the plasma chamber dimensions must be adequate for the desirable electromagnetic wave modes to propagate in it.
In order to prevent plasma losses to the plasma chamber wall good electron magnetization will be required. This can be achieved when the electron Larmor radius is smaller than the tube radius (le << R) and when the Hall parameter is larger than 1 (ΧH >> 1).
Larmor radius: the radius of gyration;
Hall ratio: the ratio between the electron gyrofrequency and the electron-heavy particles collision frequency.
The ion cyclotron frequency for a given magnetric field strength therefore can be found by re-arranging the terms of equation IV to
For a more comprehensive overview of the helicon theory and plasma dynamics we refer to reference 3 and reference 4.
For the plasma chamber we will use a long glass tube with a small diameter. The preferred material is quartz glass as it is non-conductive and heat resistent up to 1450ºC and it has a low thermal expansion but sometimes borosilicate glass (max. temperature 600ºC) is also used, which is possible because the plasma is heated under low-pressure (vacuum) conditions and the applied magnetic field keeps the plasma away from the glass walls by means of confining the plasma because a plasma is a diamagnetic medium. Nevertheless, some heat is exposed to the tube walls coming from ions and electrons reaching the wall, depositing their energy and recombining.
For our purposes we have acquired a quartz tube with dimensions 1500 mm length, 16 mm I.D. and 12 mm O.D (60" LG x 5/8" OD x 1/2" ID) (image 1):
Image 1: Quartz tube (source: Supplier)
Our quartz glass tube was a lucky find as the dimensions of an outer diameter of 16 mm and an inner diameter of 12 mm make it possible to apply EU standard brass fittings and standard vacuum fittings for creating the inlet port for feeding of Deuterium gas as well as a vacuum connector. See the next paragraph about the vacuum connections.
The acquired tube is far too long for our purposes and therefore we have cut off a section with a length of 150 mm by using an electric tile cutting machine. The tile cutting machine is the only instrument provided with water cooling and a diamond saw blade, which can make a decent cut in quartz glass. After the cutting job and after carefully drying the tube from adhering water, the quartz glass endings are heated in a propane/oxygen torch flame for melting the quartz glass cut surfaces to a smooth surface, which increases the strength of the tube endings and prevents from spontaneous cracking.
The length of the tube and the diameter of the tube constitute an aspect ratio (diameter to length) of 0.11, which is within the typical range used for helicon plasma sources (reference 5)
The vacuum tight Deuterium inlet into the quartz tube is made by means of a Henco transition coupling 16 mm x 1/2" with O-ring (image 2):
Image 3: Henco transition coupling 16x½"with O-ring (© Henco)
The Henco transition coupling consists of nickel plated brass and has been designed for coupling copper tubes, but can also be used (with some care) on quartz glass tubes too. Copper tubing will permit some compression for a tight fit but quartz glass doesn't. This means that tightening of the large swivel nut (at the left in the image, with an opening diameter of just over 16 mm) should not be done too heavily as the quartz may crack. Additionally, the compression ring (internal diameter just over 16 mm) should not be permitted to squeeze directly onto the quartz glass surface. A double layer of PTFE-tape under the compression ring is recommended. The fact that the coupler has an O-ring part (i.e. two O-rings at a small distance of each other) with an outer diameter of 12 mm that fits tightly into the inside of the quartz tube is essential for preventing cracking of the tube and to ensure a vacuum tight fitting. Though we are not in favour of using grease in vacuum systems it is advised here to grease the O-ring slightly with a good quality vacuum grease to faciliate easy sliding into the quartz glass tube.
The ½" threaded stub at the end of the coupling (at the right of the image) will be closed with a ½" brass end-cap (identical to image 6) by using PTFE-tape. In the center of the end-cap a 1 or 1.6 mm diameter hole has been drilled to accommodate either a stainless steel O.D. 1 mm injection needle or a length of 1.6 mm OD stainless steel HPLC tubing, which has been brazed into the drilled hole. The needle or HPLC tubing acts as the Deuterium gas inlet and protrudes into the quartz glass tube beyond the metal parts of the coupling.
The opposite side of the quartz tube, where it enters into the Fusor, has a vacuum tight connector which holds the quartz tube by means of an O-ring in a vacuum tight closure as shown in image 4:
Image 4: Glass tube to KF25 NW vacuum connector (source: Supplier)
The connector in image 4 is unfortunately only available for KF25 NW, which is too large for our quartz tube. Therefore we have made a similar connector by using a KF16 NW to ½"BSP thread connector (image 5):
Image 5: Dimensions KF16 NW to ½" BSP connector
The dimensions in image 5 are: A = 30mm / 1.2 inch, B = 20mm / 0.8 inch, C = 16mm / 0.7 inch, D = 40mm / 1.6 inch; the thread is "G" BSPP.
The connector is completed with a suitable Viton O-ring and a ½" BSP "G" thread chrome plated brass nut cap (image 6), which gets a hole drilled in the center with a diameter of almost 16 mm, which is reamed to an almost tight fit around the quartz tube. Actually, we have drilled a hole of 10 mm diameter into the cap nut and than reamed it in a lathe up to 16+ mm. The reason for that action is that a conventional steel drill bit does not make perfectly round holes, which is better achieved by rotating the workpiece in a lathe and reaming the hole with a cutting chisel.
Image 6: ½" BSP Cap Nut
By carefully tightening the end cap onto the KF connector, the O-ring enclosed inside the cap and around the quartz glass tube is compressed and expands radially towards the quarz glass tube and towards the brass cap wall thus ensuring a vacuum tight closure. A brass flat ring, fitting into the cap, may be needed in order to prevent rotational forces onto the O-ring when tightened, causing deformation and consequently vacuum leaks.
An axial magnetic field is applied by means of a series of permanent magents.
The magnetic field has a number of functions:
For helicon discharges a magnetic field strength of 10 -100 mT is mandatory, which equals 100 - 1000 G.
For our device we have chosen to use 20 pieces of annular magnets with dimensions of 27 mm outer diameter (OD), 5 mm thickness and 16 mm inner diameter (ID); all dimensions with a tolerance of ±0.1 mm. These magnets are made of sintered Neodynium-Iron-Boron with an outer coating of Nickel-Copper-Nickel. They have the following characteristics:
Initially the idea was to use spacers made of 3-D printed ABS plastic but a calculation of the expected wall temperature made it clear that this material with a melting point around 200°C was not suitable:
Ions as well as electrons will strike the wall with energies for ions of ±5 KBTe (the sheath drop) and for electrons of ±2 KBTe, i.e. a total of ±7 KBTe. The Bohm flux ½nKBTe is approximately the flux of ion-electron pairs striking the walls. When we neglect convective cooling and radiation from the inside surface we can make a conservative estimate by inserting this Bohm flux in the stefan-Boltzmann law. For KBTe = 3 eV and n = 2E13 cm-3, the wall temperature is approximately 860°C, well below the maximum temperature valid for quartz glass but far above the melting point of ABS.
Conclusion: For placing the magnets concentrically around the quartz glass tube we will need to make use of rings cut from a non-magnetic metal, e.g. an alumina or copper tube with an ID just over 16 mm or by using plastic spacers at the outer circumference of the magnets.
Measuring the Magnetic Field
The magnets should be placed at specific fixed distances from each other to obtain the best possible almost homogenous magnetic field. Also we would like to know in an absolute value the strength of the magnetic field inside the quartz tube. For this purpose sophisticated magnetometers exist, but as we intend to do the measurements only once and because we are not interested in a highly accurate (calibrated) result we have considered constructing a simple magnetometer in our workshop.
Different methods and schematics how to build a magnetometer can be found in the internet. As an example we have picked a very simple one (reference 6).
However, we could not withstand the challenge and decided to build a slightly different model that uses an Arduino Micro-Pro to read out and process the signal (any Arduino version can be used).
The main component in the schematic is the SS495B Honeywell Hall sensor, a miniature ratiometric bipolar Hall sensor with a linear range from -640 to +640 gauss (image 7):
Image 7: Honeywell SS495B Hall Sensor
With an Arduino Mini-Pro connect the +5V pin of the SS495B to VCC of the Arduino, the GND pin to GND and the Output pin to Analog input 0 of the Arduino. The Arduino source code can be found in reference 7. The readout of the Hall sensor can now be done in the serial print mode.
The Arduino model type as used has a 10-bit analog to digital converter, which therefore maps input voltages 0 and 5 volt into integer values between 0 and 1023. The resolution between readings of the Arduino A/D-converter is therefore 5 volt divided by 1024 units or 4.9 mV per unit. When a higher resolution is required, than an Arduino Due or Zero can be used with the 12-bit ADC or a TI ADS1110 16-bit ADC can be integrated separately into the schematic.
The resolution of the output of the Hall sensor in our setup is in a range from -640G to +640G, or a total of 1280 G units in a range from 0.5 V to 4.5 V, which is a range covering 4000 mV, where 4.9 mV is one ADC unit. Or: 1280 G units in 4000 mV divided by 4.9 mV, which equals about 1.6 G per ADC unit of 4.9 mV. With other words, our gauss readings are shown in resolution steps of 1.6 G. For our purposes, i.e. determining the magnetic field strength of our setup where the required magnetic field strength is between 100 to 1000G, this is sufficiently accurate and there is no need to change to a higher bit rate ADC.
The interface for delivering RF power to the plasma is formed by the antenna. The RF signal on the antenna causes waves to be emitted, which can be
Where possibility number 2 represents the desired situation, possibility number 3 can be quite disturbing to our neighbourhood especially when the applied helicon frequency is in the 27 Mhz or the 80 - 110 Mhz range, which are respectively the CB and the FM radioband. Radiation into free space therefore should be prevented by encapsulating the helicon plasma source in a (grounded) cage of Faraday.
Possibility number 4 can occur with an unmatched impedance between power generator and antenna coupling causing the forward wave to be reflected back in the power line resulting in standing waves in the transmission line. Further about this issue can be found in the matching circuit paragraph.
The simplest for of an antenna is the m = 0 type antenna, i.e. an axisymmetric single loop antenna, which is known to have less wall losses compared to longer helical or Nagoya antennas.
The m = 0 antenna is bidirectional (azimuthally symmetric) whereas the m = +1 helical antenna exites the wave in one direction.
A calculator for calculating a classical helix antenna can be found in reference 8.
Once the antenna type and dimensions have been chosen, it must be matched to the transmission line from the RF power source. With "matching" is meant impedance matching between the RF transmission line impedance, usually 50 Ohm, and the antenna impedance which acts as a load impedance for the transmission line that carries the power signal to the antenna. The matching can only be done with the antenna fully functional and in place in the complete setup because everything around the antenna affects its performance. The radiated electromagnetic fields from an antenna interact with nearby materials and may change its frequency of operation and energy transfer into the plasma. Nearby objects to our antenna are the magnets, the spacers that hold the magnets at a certain distance, the quartz envelope, etc.
The main purpose of the matching network is to maximise the forward power to the plasma and to minimise the power reflected back to the RF generator. The ratio between the forward power and the reflected power can be expressed as the standing wave ratio and it is a measure of the efficiency of the RF power transmission.
For the design of a matching network a choice can e.g. be made from different possible types of LC-matching network, like the L-network, the Inversed L-network, the Pi-network and the Split-capacitor network.
We shall take a look at the L-network and the Inversed L-network as the simplest forms of LC-matching networks.
The L-network can be used when R1 > R2, where R1 is the impedance of the amplifier output (usually 50Ω) and where R2 is the impedance of the antenna (to be measured), as shown in image 8:
Image 8: L-network (© FRS 2017)
The inductive reactance XL of this network can be calculated as:
XL = √(R1R2-R22)
The capacitative reactance XC of this network can be calculated as:
XC = -R1R2/XL
For calculating the values of the inductor L1 and the capacitor C1 we make use of the general formula (below), but first we will consider the matching network when R2 > R1. In that case we will make use of the inverted L-network as shown in image 9:
Image 9: Inversed L-network (© FRS 2017)
For the Inverted L-network, the inductive reactance can be calculated as:
XL = R2√(R1/(R2-R1))
and the capacitative reactance XC of this network can be calculated as:
XC = -R1R2/XL
For both networks the values for inductor L1 can be calculated as:
L = XL / (2*π*f )
where f is the operating frequency.
For both networks the values for capacitor C1 can be calculated as:
C = 1 / (2*π*f *XC)where f is the operating frequency.
As mentioned earlier for solving the above equations we will need to know the impedance of the antenna system and measuring the antenna impedance requires expensive equipment, a network vector analyzer.
A less costly investment can, however, be done in a VSWR meter, which will enable us to match the antenna impedance empirically to the transmission line. For that we will need a tunable matching circuit for which we introduce a variable capacitor and a variable inductor in the LC-network as shown in the schematics of image 10.
Image 10: Variable Matching Network (© FRS 2017)
Details about this circuit can be found in reference 9.
For the capacitor Cp in the matching circuit schematic we intend to make use of a Ducati air spaced variable capacitor with a capacity of 6 - 250 pF at a maximum voltage rating of 1500V (image 11).
Image 11: Ducati Variable Capacitor (Source: Supplier)
As second capacitor Cs we will use a Jackson air spaced trimmer capacitor with a capacity of 4 - 30 pF at a maximum voltage rating of 750 V (image 12).
Image 12: Jackson Trimmer Capacitor (Source: Supplier)
For the variable inductor Ls we will use an in-house built coil with an adjustable iron core. The iron core exists of the tip of a micrometer (image 13).
Image 13: Micrometer (Source: Supplier)
The coil will exist of about six turns of 1 mm diameter copper wire. Inside the coil an iron core of 6.4 mm diameter can very precisely be positioned, fully or partially, over a distance of 28 mm. Insertion of the iron core, whitch has a high permeability compared to air, results in a greater magnetic field flux causing a higher inductance.
The approximate inductance can be calculated, assuming that we have a coil with a radius r and a lenght l as
L = (N2µA)/land
µ = µrµowhere
L = Inductance of the coil in Henrys
N = Number of turns in wire coil, straight wire is 1
µ = Permeability of core material (absolute, not relative, 1.25663753x10-6 for air, 6.3x10-3 for iron)
µr = Relative permeability (dimensionless, 1.00000037 for air, 5000 for iron)
µo = 1.26 x 10-6 T-m/At (permeability of free space)
A = Surface area of coil in square meters or πr2
l = Average length of the coil in meters
Suppose, we have a coil of 6 turns copper of 1 mm diameter with 3 mm between the windings and an inside diameter of 8 mm, than our coil has an outer diameter of (8 + 1 + 1) = 10 mm and a length of (5x(1+3)+1) = 21 mm.
With no iron core we will have an open air core and by applying the formulas we find an approximate inductance L of: (62 x 1.25663753x10-6 x π x (5x10-3)2) / 21x10-3 Henry, which equals 169 nH.
With an iron core fully inserted and applying the formulas we find an approximate inductance L of: (62 x 6.x10-3 x π x (5x10-3)2) / 21x10-3 Henry, which equals 84.8 mH.
Hence, by moving the coil in our example out and in the core we can change the inductance approximately between a value of 169 nH to 85 mH. This calculated result is more or less in the same range as the setup described in reference 9, where the variable coil has a measured variable inductance range of 100 - 160 nH, taking into account that in the article the core can only be inserted about halfway into the coil.
In RF transmission systems the standing wave ratio is a measure for how efficient RF power is transmitted from the power source, through the transmission line, into the load, the antenna. SWR is the ratio between the transmitted and the reflected waves. A high SWR indicates a poor power transmission into the antenna and a high reflected power, which can damage the equipment. SWR refers to a voltage ratio and is therefore expressed as a dimensionless VSWR number.
In an ideal system voltage does not vary and in such a system a VSWR of 1.0 means that the ratio is 1:1, or with other words reflection is absent. This is rarely the case.
An example of a suitable VSWR meter is shown in image 14:
Image 14: VSWR meter (Source: Supplier China)
This VSWR meter has an operational range between 100 and 520 MHz and an input power of 120W and measures the VSWR from 1.00 to 19.9 at an impedance of 50Ω.
RF power source
The RF Power Source has a modular setup and consists of a number of rather inexpensive commercially available modules. The total amount of modules, however, will affect our budget quite intensively. Only the matching circuit and the antenna are made in our workshop and for these items separate paragraphs can be found on this page.
The layout of the RF Power Source is shown in the block schematic in image 15:
Image 15: Block Schematic of the RF Power Source (© FRS 2016)
The inputs and or outputs in the block schematic are given in Watt units (W), whereas the power gain is given in decibel (dB). Impedances of all modules up to the matching circuit is 50Ω.
RF Function Signal Generator
Any RF function signal generator can be chosen when it has a sine output ranging from 1 - 100 MHz or higher. The signal from such a function generator is quite weak and therefore several amplification steps are required before we can dump 100 W of energy into the plasma.
A good example for a sofware controlled function signal generator is shown in image 16:
Image 16: RF function signal generator
The function signal generator is based on a AD9854 chip, a digital synthesizer, which uses direct digital synthesis (DDS). The module requires an input of 5V at 1A and provides a maximum outpuit of 100 MHz sine wave and 10 MHz square wave in steps of 0.1 Hz with an adjustable amplitude with a 12 bit precision. A total of four output ports are present and these can be fed with signals with different phase angles.
Images 17 and 18 show the pc screen GUI's of the included software:
Image 17: GUI of the signal generator software
Image 18: GUI of the signal generator software
However, we found a cheap old circulator, working in the range of 0.14 to 112 MHz, a 1979 model Prüfgenerator "EP 105 BS a" by the manufacturing company of Dipl.-Ing. H.-G. Neuwirth, from Hannover, Germany (image 19).
Image 19: Neuwirth HF Signal Generator (Source: Supplier)
This signal generator has 7 Frequency ranges:
The weak signal from the function signal generator needs to be amplified prior to further processing. For this purpose we have chosen a low noise broadband amplifier working in the range from 2 MHz to 1 GHz with an ampification gain of +22 dB. A signal as low as -110 dB can be amplified. The pre-amp runs at 5 VDC / 80 mA (Image 20):
Image 20: RF preamp +22 dB gain
RF Step Attenuator
The output power level is controlled by a step attenuator between preamp and the broadband booster amplifier (image 21):
Image 21: RF Step Attenuator
The step attenuator offers a bandwith from 1 MHz to 4 GHz with an attenuation of max. 31.5 dB in 0.5 dB steps. The power supply required is 5 V DC.
RF Broadband Booster
The broadband booster amplifier will amplify the signal up to a level suitable for the power amplifier (image 22):
Image 22: broadband booster amplifier
This amplifier has a bandwith of 1 MHz to 500 MHz and with an input signal of 1 mW a gain can be achieved of 32 dB to an output of 1.6 W.
The output of the broadband booster amplifier suits the RF Power Meter that is the following module in the chain of modules.
RF Power Meter
The RF Power Meter has a measuring range from 1 - 500 MHz at 1 nW - 2 W power input, with a measuring power range from -8- - +10 dB with a resolution of 0.1 dB (image 23):
Image 23: Power Meter (power setting)
The power meter requires an input of 7 - 12 VDC at 50 mA. It will be applied between the broadband amplifier and the power booster. Additionally, RF power attenuation can be applied by adding or subtracting attenuation power in dB units (image 24):
Image 24: Power Meter (attenuation mode)
Between the circulator and the matching circuit we will apply a combined Power/SWR meter to measure the power output and the standing wave ratio for optimal antenna tuning.
RF Power Amplifier
The RF power amplifier (image 25) is used to convert a low power RF signal into a higher power, suitable to drive the antenna.
Image 25: RF Power Amplifier
The RF circulator or an RF isolator is used for blocking reflected power from the antenna circuit by dumping this power into the dummy load. It is rather difficult to find a suitable (used and cheap) circulator in the required frequency range of 80 - 120 MHz VHF bandwith and for the applied power of 100 W; moreover they are somewhat bulky due to the required low frequency range. Most used RF circulators on the market appear to be designed for the GHz UHF bandwith and therefore can do with smaller dimensions.
The principle of operation of a circulator, a three port ferro-magnetic device with the ports arranged in a triangle, is that it has directional properties from port to port. RF energy is directed from port 1 to port 2, but not to port 3, caused by ferromagnet coupling inside the circulator. This occurs on all ports, thus obtaining a directivity travelling in one direction around the triangle.
The directivity (or "isolation") is measured in dB, e.g. a port to port isolation of 20 dB means that it conducts RF in one direction 100 times better than in reverse. Isolation, however, is only provided when the circulators are terminated, typically at 50Ω.
As a passive device circulators can be designed and constructed as a waveguide (cavity), coaxial or stripline circulator.
A stripline circulator usually is constructed with the coupling between the ports consisting of flat insulated overlapping terminated copper sandwiched striplines in a 120 degree angle position. The flat copper striplines are sandwiched between two discs of optimised ferrite. A permanent magnet imposes a strong magnetic field across the discs for generating the directivity.
In some circulators internal matching networks may be present as L/C circuits or variable trimmer capacitors. When chosing a circulator near to the required frequency, such a matching network may permit to shift the operating frequency of the device and change the degree of port to port isolation.
For tuning, always with the circulater connected in "reverse", it will be imperative to terminate all ports with 50 Ohms and then tune for maximum rejection of the signal at the required frequency, rotating through all three ports for tuning to the same. Never tune the circulator in the forward direction as this will not work!
Our search for a suitable circulator made us find a used high power circulator, which came from army surplus and its working range is in the 200 to 400 MHz bands (image 26).
Image 26: Temex Circulator (Source: Supplier)
It is a Temex brand device, though it is also marketed by Cobham Microwave, a model BB3007 with the following specifications:
Table 1: Circulator Specifications
This three-port coaxial circulator is a so-called Y-junction circulator which uses ferrites in the presence of a magnetic bias field, in order to provide a non-reciprocal effect. If a signal is applied at port 1 it will emerge from port 2 with a loss characteristic, called the insertion loss (I.L), which in our circulator equals 0.70 to 0.80 dB. In the reverse direction there will be a leakage at port 3 from the incoming signal at port 1. This leakage, called isolation, is 14 to 17 dB below incoming power at port 1 in our ciculator.
The lucky find of this circulator necessitates us to feed our ion source plasma with a frequency in the range of 200 to 400 MHz.
RF Dummy Load
The dummy load is a 50Ω Non-Nichrome resistive RF termination impedance resistor with a 150W power rating. It can be used up to a frequency of 3 GHz. The resistor will need to be attached to a large metal cooling block with a thin layer of thermal paste applied between the contacting metal surfaces (image 27):
Image 27: RF Dummy Load 150 W
Image 28 shows a standard LED heatsink with a capacity of 100W that can be applied with the dummy load:
Image 28: Heatsink for Dummy Load
Tuning and Tests
Ref.1: Very Efficient Plasma Generation by Whistler Waves near the Lower Hybrid Frequency: http://people.physics.anu.edu.au/~web112/publications/papers/boswell_1984_very_efficient_plasma_generation_by_whistler_waves_near_the_lower_hybrid_frequency.pdf
Ref. 2: Ion cyclotron resonance: https://en.wikipedia.org/wiki/Ion_cyclotron_resonance
Ref. 3: Plasma Dynamics in a Helicon Thruster: http://www.eucass-proceedings.eu/articles/eucass/pdf/2013/01/eucass4p337.pdf
Ref. 4: Design and Development of a 1 kW-class Helicon Antenna Thruster: http://aero.uc3m.es/ep2/docs/publicaciones/meri15c.pdf
Ref. 5: Neutral Gas Expansion in a Cylindrical Helicon Discharge Chamber: http://mwalker.gatech.edu/papers/JPP_V29_No3_Gianelli_2013.pdf
Ref. 6: Build your own Gaussmeter: http://www.coolmagnetman.com/magmeter.htm
Ref. 7: Gaussmeter: https://arduining.com/2012/07/17/arduino-hall-effect-sensor-gaussmeter/
Ref. 8: Sophisticated Helix Antenna Design (Calculator): http://www.changpuak.ch/electronics/calc_12a.php
Ref. 9: Development of a compact permanent magnet helicon plasma source for ion beam bioengineering: http://aip.scitation.org/doi/abs/10.1063/1.3646467?journalCode=rsi
|Last Updated on: Wed
Jun 7 23:49:34 2017