Your Very Own Altator: How to construct a tabletop, glass, toroidal plasma outreach device

Your Very Own Altator: How to construct a tabletop, glass, toroidal plasma outreach device Rutherford, G. & Saltzman, A. November 12, 2025 Contents 1 Introduction 2 2 Vacuum system 2 3 RF system 4 4 Implosion shield and Faraday cage 6 5 Torus supports and table base 7 6 Standard Operating Procedure 8 7 Potential Hazards 10 8 Cost Drivers 10 9 Future upgrades 11 9.1 Toroidal field coils ......................... 11 9.2 Diagnostics ............................. 12 9.3 RF system ............................. 12 9.4 Control system ........................... 12 9.5 Design Summary .......................... 14 9.6 Inductance ............................. 14 9.7 Length ................................ 15 9.8 Charging time ........................... 16 9.9 Dump Resistor ........................... 17 1 1 Introduction Altator is a device conceived of by nine graduate students from the Physics and Nuclear Science and Engineering departments during their first semester at MIT. It was motivated by the dual goals of providing hands-on experience and contributing to the outreach goals of MIT’s Plasma Science and Fusion Center (PSFC). For outreach, it was specifically designed to be included in tours as a bridge between the PSFC’s linear glow discharge demo and a full research tokamak. Ideally, Altator would allow the public to better understand the workings of a tokamak by seeing an RF-driven, toroidal plasma in action. In its current iteration, Altator consists of a glass toroidal vacuum chamber, a simple radio frequency (RF) antenna to break down the sustain the air plasma, and an implosion shield/Faraday cage. It has been included on dozens of tours of the PSFC and was featured in an episode of Building Stuff with NOVA ( https://www.youtube.com/watch?v=22b4x85UwPQ ). Altator cost approximately $17, 000, without counting labor costs and making use of many surplus parts from the rich MIT ecosystem. This document is aimed at individuals and institutions who are interested in building their own version of Altator and will provide details of the vacuum system (Section 2), the RF system (Section 3), the implosion shield and Faraday cage (Section 4) and the table and structure supporting the torus (Section 5). The standard operation produced (Section 6), safety considerations (Section 7), and a breakdown of key component costs (Section 8) are also discussed. 2 Vacuum system Being an outreach device, it is important that Altator’s plasma is readily visible and sufficiently large to be impressive. The vacuum vessel was therefore chosen to be made of glass and have a major radius of 10 inches (25.4 cm) and a minor radius of 2 in (5.08 cm). Creating such a shape from glass is no easy task, and Precision Glassblowing (https://precisionglassblowing.com/) in Denver, Colorado were the only glassblowers willing/capable to take on the project. The resulting torus is made up of four quadrants held together by ChemGlass CG-141-02 clamps. An example of the quadrants assembled is shown in Figure 1. On either end of each quadrant, the glass flares outwards to interface with these clamps. On one end, the face is flat; on the other 2 end, the face is grooved to fit an O-ring (which were provided by Precision Glassblowing). To interface with the vacuum pump and to provide space for diagnostics in the future, several 1 inch (2.54 cm) ports were attached to the quadrants. To connect to these ports, "NW40 to 1" quick coupling" connectors are used. As Altator does not currently feature a diagnostic that requires a port, all ports not needed by the vacuum system are blanked off. Figure 1: Altator’s vacuum vessel assembled from four quadrants. To pump down Altator, a surplus Edwards RV5 roughing pump is used. This pump sits underneath Altator and is connected to one of Altator’s ports via a corrugated tube. Between the tube and the pump is a valve, which starts out closed when the pump is turned on. Once a vacuum has been established in the region between the pump and the valve, the valve is opened. This prevents pump oil from being sucked into the vacuum chamber. To vent the vacuum vessel and bring Altator back up to atmospheric pressure, a valve is connected to another one of Altator’s ports. A Granville-Phillips 275 MiniConvectron vacuum gauge is used to monitor Altator’s vacuum. A diagram of Altator’s vacuum system is given in Figure 2. Unlike fusion devices, Altator’s working gas is air. In the future, a turbo pump may be installed to reach lower pressures, at which point backfilling with a different gas would be interesting. 3 Figure 2: Diagram of Altator’s vacuum system. 3 RF system To break down and sustain the plasma, a 13.56 MHz radiofrequency system is used. This frequency was chosen as 13.56 MHz sources are plentiful, and it lies within an industrial, scientific, medical (ISM) radio band, ensuring there are no legal difficulties in operating the source. Altator’s RF source is an ENI Power Systems ACG-5 and was bought off of eBay. The RF source can output up to 600 W, but breakdown is typically achieved with powers as low as ∼ 10 W. A single, inductively coupled antenna is currently in use (see Figure 3). This antenna was constructed by wrapping a 14 gauge copper magnet wire around a quadrant of the torus 20 times prior to assembly of the torus. The ends of the antenna are soldered to a female type N connector, allowing connection to the rest of the RF system. The impedance mismatch between the transmission lines and the antenna causes power to be reflected back towards the source. To prevent this power from reaching the source, a matching network is placed after the source. This matching network is a modified MFJ Versa Tuner III and consists of two variable air core capacitors and a variable inductor. These tuning elements are wired in a nontypical manner, which can be seen in Altator’s RF schematic shown in Figure 4. This wiring scheme was chosen due to the limited range of the tuning element values and limited access to additional components beyond those that came with the Versa Tuner III. A more typical scheme, such 4 Figure 3: Altator’s inductively coupled antenna. as a Pi-type matching network is recommended instead of the circuit shown here. A picture of the internals of the matching network is given in Figure 5. Figure 4: Schematic of Altator’s RF system. To match the system, the RF source is disconnected, and a vector network analyzer (VNA) is connected in its place. The VNA is useful for many purposes, but here it is used to measure S 11 , which in this case corresponds to the ratio of the power reflected back towards the source divided by the source output power. By varying the capacitor and inductor values in the matching network, S 11 can be made quite small. A target value of S 11 <− 20 dB was chosen empirically as this was found to prevent the source from overheating at typical power levels. Routinely, S11 <−25 dB is achieved. 5 Figure 5: Modified internals of the MFJ Vera Tuner III matching network. 4 Implosion shield and Faraday cage Though the risk of implosion as the glass torus is pumped down is quite low, an implosion shield was installed out of caution. This consists of 1/4 inch (6.35 mm) sheets of polycarbonate surrounding the sides and top of Altator. These sheets are held in place by vertical pieces of 15 series aluminum extrusion from 80/20. To ensure the RF power densities incident on members of the public are below the legal limits, a Faraday cage was added underneath the implosion shield. Using a frame made from 1 inch (2.54 cm) wide, 1/8 inch (3.17 mm) thick aluminum bars, wire mesh was attached to each polycarbonate panel. The wire mesh and aluminum bars were purchased from McMaster-Carr. The aluminum frame comes into contact with the vertical 80/20 supports (which had their anodized coating stripped off) to form an electrical connection. The top of the Faraday cage uses right-angle aluminum bars to connect to the other aluminum frames. Underneath Altator, the final wire mesh is attached to the plywood table top. Using right-angle aluminum bars and standoffs, this mesh is connected to the side meshes. To ground the Faraday cage, a wire was run to a nearby Earth ground connection point. A picture of the Faraday cage and implosion shield is given in Figure 6. 6 Figure 6: Altator’s implosion shield and Faraday cage. 5 Torus supports and table base Altator sits on a table largely constructed from 15 series 80/20 aluminum extrusion and measures 42"x62" (106.68 cm x 157.48 cm). Two plywood sheets (painted black) are used the tabletop and a lower level used for storage and housing the vacuum pump. The RF source and matching network are set on the upper level as operators need easy access. The table sits on four large castors, allowing Altator to be relocated with some ease (though its width requires double doors). The torus is held in place by laser cut acrylic supports. However, due to the difficulty in manufacturing the torus, its minor radius varies somewhat, leading to some of the supports having little to no contact with the torus. In the future, these supports will be replaced by new supports featuring a foam contact surface that will conform to the shape of the torus. A picture of Altator with the supports visible is given in Figure 8. 7 Figure 7: Table constructed out of 80-20 and plywood. 6 Standard Operating Procedure This section should not be viewed as a formal standard operating procedure. Readers should instead use it as a means to inform the development of their own standard operating procedure. To turn the device on: • Unlock the plug lockouts on the RF supply and vacuum pumps, and connect these devices to the outlet. •Ensure the valve connecting the vacuum pump to the torus is closed. •Turn on the vacuum pump. • After about 10 seconds, open the valve connecting the vacuum pump to the torus. • While the torus is pumping down, use the VNA to check the RF matching and ensure S11 <−20 dB. •Ensure the RF source output power is set to 0 W. 8 Figure 12: CAD sketch of preliminary cryostat design. Not shown: pressure relief valve 9.7 Length Initially, we considered installing a helicon antenna on Altator, which would have set the requirement for at least 150 Gauss on-axis throughout the torus, requiring 5,000 Amp turns for a 4.75" inner radius magnet. This was used as the basis of this planning document, but the following methodology could be adjusted for any target field. Both the critical current and number of turns are now questions at hand. We calculate the critical current using a COMSOL with the Magneto-Angular Anisotropy Model (Zhang et al., 2016 and Robert et al., 2019), the basis of which was set up by Nicolò Riva. In this model, Jc0= 3.5 ×1010A/m2 and the critical current then is given by JcB =Jc0 (1 + ε(θ)(B/B0)β)α The dependence on the angle the field makes to the tape enters the model through ε(θ) = qsin2(θ)/γ +cos2(θ) 15 The three numerical constants are α= 1.2 β= 1 γ= 8.02 We assume a tape width of 4mm. The results of this model can be found in Fig. 13. Through iteration we estimate needing 152 turns and perform our simulation on this geometry. We find a critical current of 65.256 A. To prevent quenching, we want to operate at no more than 45% of the critical current. Therefore, we expect to be able to operate at a current of about 29.25 A. To reach the required 150 G field strength at the minimum field point on the toroidal axis, we require 5,000 amp-turns. Dividing this number by 66.58 A gives 75.1 turns of the two-in-hand tape. The required length of tape can be found by estimating the total length of tape required as Ltotal = 2π(Nr1+ N−1 X k=0 kLsw) Here Nis the number of turns, r 1 is the inner bore of the magnet, which is 4.75 inches, and L sw is the thickness of a two-in-hand piece of tape, which from Figure 11 can be seen to be 0.554 mm. The total length of tape required is then L total = 66.5 meters of two-in-hand tape (133 meters if only singlelayer HTS is available), with a coil thickness of 1.646 inches, found by multiplying Nby Lsw. 9.8 Charging time The charging time is constrained by the supply voltage that would cause arcing between the current leads as well as the dump resistor providing resistance in parallel to the superconducting coil. As we plan to use a supply on order of 0.066 V, a separation of at least 0.1mm will be sufficient to prevent arcing. The charging time is given by the L/R time, which for the circuit shown in Figure 14, is L/R = 0.008 H/0.001 Ω = 8 seconds. Note the resistor that plays a role in the charging time is R2, which represents resistances in the wiring from the power supply to the magnet. This is confirmed by the LTspice current plot (green curve in Figure 14) through the HTS magnet, which is modeled as the inductor in Figure 14. Once the system reaches steady state, the power draw required to keep the magnet energized is 66.58 A * 0.06658 V = 4.43 Watts. 16 (a) 76 turn design. (b) Magnetic field generated by HTS coil. Figure 13: COMSOL modeling was performed to determine the magnetic field generated by our coil. 9.9 Dump Resistor The dump resistor circuit diagram is shown in Figure 14. The dump resistor is placed in parallel with with the superconducting coil. Using an air breakdown voltage of 3 × 10 6 V/m and a Teflon break down voltage of ∼ 7.5 × 10 8 V/m, we find the dump resistor needs to be between 0.8 Ω< R < 300 Ω . However, we plan to be nowhere near these voltages, and instead set a target of keeping system voltages below 100 V, as wire insulation and feedthroughs are typically rated to 300 V. Consider a quench in the superconducting coil, so that the path in Figure 14) is now resistive. In this event, current now flows through the dump resistor. At t = 0 s after the quench occurs, the current flowing will be 66.58 A. So, the required dump resister resistance which keeps the voltage across it limited to 100 V is 100 V / 66.58 A = 1.5 Ω . In the model in Figure 14, the dump resistor is broken up into two resistors in series, R1 and R3, in order to give the modeler the option to ground the dump resistor at its center. 17 Figure 14: Circuit diagram for dump resistor (bottom). HTS magnet charge time (top) 18