Electronics Guide

Arc-Heated Processes

In atmospheric plasma spray, a direct-current arc burns inside a water-cooled torch between a conical cathode, usually thoriated or lanthanated tungsten, and a coaxial copper nozzle that serves as the anode. The arc does not leave the torch; the workpiece is electrically uninvolved. Plasma gas, typically argon with hydrogen, helium, or nitrogen added, flows through the arc, is heated and partly ionized, and exits the nozzle as a jet. Sources commonly quote plasma temperatures above 15,000 K near the arc core, with the working jet on the order of 10,000 K. Powder is injected into that jet, melts in flight, and impacts the substrate. Torch power in production equipment ranges from roughly twenty kilowatts to well over one hundred kilowatts.

Twin wire arc spray strikes its arc between two consumable wires that are continuously fed toward a meeting point, and a jet of compressed air atomizes the molten tips and carries the droplets to the workpiece. Plasma transferred arc runs a transferred arc from a torch electrode to the workpiece itself, so the part carries the full process current and forms a shallow molten pool that the injected powder joins.

Combustion and Gas-Dynamic Processes

High-velocity oxy-fuel burns kerosene, hydrogen, propylene, or a similar fuel with oxygen in a small chamber and expands the products through a converging-diverging nozzle. Published figures put combustion chamber pressure near 1 MPa and jet exit velocity above 1,000 m/s, with particles reaching as much as 800 m/s. High-velocity air-fuel substitutes compressed air for pure oxygen, which lowers the flame temperature, commonly cited between roughly 1,960 and 2,010 degrees Celsius, while still reaching particle velocities near 1,000 m/s. The detonation gun fires a discrete oxygen-acetylene detonation in a long barrel, purges with nitrogen, and repeats the cycle several times per second.

Cold spray heats a compressed carrier gas, nitrogen or helium, in an electrical resistance heater and expands it through a de Laval nozzle. Wikipedia and process vendors describe nitrogen heated to about 900 degrees Celsius in typical practice. Particles never melt; they bond by severe plastic deformation on impact. The heater is the only significant electrical load, and it is an ordinary resistive one.

Arc Voltage Is a Dependent Variable

The voltage across a plasma spray arc is set by the physics of the column, not by the supply. Arc voltage rises with the length of the column between the cathode tip and the point where the arc attaches to the nozzle wall, with the gas flow rate, and above all with gas composition. Argon alone sustains a comparatively low voltage. Adding hydrogen or helium raises the thermal conductivity of the gas, cools and constricts the column, and forces the voltage up substantially for the same current. A torch that runs at forty volts on pure argon may run at seventy or ninety volts on an argon-hydrogen mixture at the same current setting.

The consequence is that the operator selects a current and a gas recipe, and the torch reports back a voltage. Power is the product, and power is therefore only indirectly under control. A published condition-monitoring study of a three-cathode cascaded torch, the Oerlikon Metco TriplexPro-210, illustrates the numbers well: at a set current of 420 A with 46 standard liters per minute of argon and 4 of helium, the mean torch voltage was 89.3 V, giving 37.4 kW of electrical input, of which 22.0 kW appeared as net power in the gas. The remainder, close to forty percent, left through the cooling water.

Arc Root Motion and Voltage Fluctuation

In a conventional single-cathode torch the anode attachment point, called the arc root, does not stay put. Cold gas flowing along the nozzle wall pushes the root downstream, stretching the column and raising the voltage, until the accumulated field breaks down a shorter path upstream and the root jumps back. Researchers classify the resulting behavior into three modes. The steady mode shows essentially no voltage fluctuation. The takeover mode produces roughly sinusoidal fluctuation as one attachment hands off smoothly to another. The restrike mode produces a sawtooth: a slow voltage rise followed by an abrupt collapse. Published measurements on conventional torches report peak voltage drops on the order of 40 V with a mean restrike frequency near 7.5 kHz, and voltage spectra with a strong peak near 4 kHz are commonly reported for direct-current spray torches. Raising the arc current tends to increase the restrike frequency while reducing the amplitude of each jump, because the hotter, wider column resists being displaced by the cold boundary layer.

This matters to the coating, not merely to the oscilloscope. Because the supply holds current constant, every volt of fluctuation is a watt-for-watt fluctuation in delivered power, and therefore in gas enthalpy, particle temperature, and particle velocity. A forty-volt sawtooth on a sixty-volt mean is a power swing of more than half. The takeover mode is generally regarded as the most desirable of the three, because it spreads the heat load around the anode and produces well-defined, comparatively gentle fluctuations rather than violent ones.

Cascaded and Multi-Cathode Torches

The modern answer to arc root wandering is geometric rather than electrical. A cascaded torch inserts a stack of electrically floating neutral rings, separated by insulators, between the cathode and the anode. The arc must traverse the stack, so its length is fixed by the hardware rather than by the balance of gas drag and breakdown. Voltage rises, because the column is longer, and stability improves dramatically. In the TriplexPro measurements cited above, the standard deviation of torch voltage was 0.31 V on a mean of 89.3 V, with the mean drifting only between about 89.0 and 90.0 V across the test period. That is a level of steadiness a conventional torch cannot approach.

Manufacturers pair the cascaded principle with single or multiple cathodes. Oerlikon Metco specifies its single-cathode cascaded SinplexPro gun at a maximum of 60 kW, and markets it on the claim that cascading decouples the arc from the process gas flows and mixtures. The three-cathode TriplexPro distributes the current among three arcs, which reduces the current each electrode must carry and smooths the jet further. These are vendor design claims, but the published voltage statistics support the general contention that cascaded torches fluctuate far less than conventional ones.

Electrode Wear and the Drifting Operating Point

Torch consumables erode. The cathode tip blunts and loses emissive material; the anode bore widens where the arc root attaches. Both change the arc column and therefore the voltage at a given current. Arc voltage trend is consequently the cheapest and most widely used health indicator for a torch, and many spray controllers log it precisely for that purpose. The condition-monitoring study noted above cautions that wear does not always announce itself cleanly in a single electrical parameter, and found that substrate surface temperature, measured by infrared pyrometry in the eight to fourteen micrometer band, tracked deposited coating mass in a nearly linear way and served as a better indicator of torch status than voltage alone.

Why Current and Not Voltage

An arc has a low, and over parts of its range negative, dynamic resistance. Connect it to a stiff voltage source and the current runs away until something limits it. Regulate the current instead, presenting a high output impedance, and the arc settles wherever its own physics puts it. Current regulation also gives the process engineer the parameter that most directly governs electrode heating and erosion, and it makes the recipe portable: a current setpoint and a gas recipe describe an operating point in a way that a voltage setpoint cannot, because voltage depends on hardware wear.

Some controllers offer a constant-power mode layered on top, trimming the current setpoint slowly so that the product of current and measured voltage holds a target as the electrodes wear. This is a supervisory loop, not a change of topology; the inner loop remains a current loop.

Thyristor Phase-Controlled Rectifiers

The traditional plasma spray supply is a line-frequency transformer feeding a phase-controlled thyristor bridge, in six-pulse or twelve-pulse configuration, with a large series output inductor. It is rugged, tolerant of the short circuits and open circuits that torch faults produce, and cheap per kilowatt at the powers involved. Its weaknesses follow from phase control. The output ripple appears at six or twelve times the line frequency, which is 360 or 720 hertz on a sixty-hertz supply and 300 or 600 hertz on a fifty-hertz supply. The control loop can only act once per firing interval, so the achievable regulation bandwidth is a few hundred hertz at best. Phase control also draws a lagging, harmonic-rich current from the line, which matters when a spray cell shares a substation with sensitive equipment.

Switch-Mode and Chopper Supplies

Newer supplies rectify the line, correct the power factor, and then use an insulated-gate bipolar transistor chopper or a full-bridge inverter with a high-frequency transformer to produce the output current. Switching in the tens of kilohertz shrinks the magnetics, lowers the ripple frequency requirement on the output filter, and raises the current-loop bandwidth by an order of magnitude or more. The practical benefit is that the loop bandwidth begins to approach the kilohertz band where arc root fluctuation lives, so the supply can at least partially reject the voltage disturbance rather than passing it straight through as a power disturbance. Efficiency improves as well, and an active front end can present a near-unity power factor with low harmonic distortion.

The tradeoff is fault tolerance. A thyristor bridge shrugs off a hard short at the torch; a transistor bridge must detect it and act within microseconds. Spray supplies therefore carry fast desaturation and overcurrent protection, and their designers pay careful attention to the energy stored in the output inductor and in the long torch cables, which has to go somewhere when the arc extinguishes.

Output Filtering and the Series Inductor

The output inductor does three jobs at once. It smooths rectifier ripple. It limits the rate of current change during arc transients, which helps the arc survive a momentary extinction. And it stores the energy that reignites the arc after a restrike event. Sizing it is a compromise: more inductance means smoother current and a more forgiving arc, but slower response to a deliberate current change and more stored energy to dissipate at shutdown. In supplies where the cable run to the torch is long, the cable inductance is not negligible and must be counted as part of the filter.

The High-Frequency Starter

The usual solution is a high-voltage, high-frequency starter of the same family used for touch-free arc initiation in tungsten inert gas welding. A step-up transformer charges a capacitor, a spark gap or a solid-state switch discharges it into a resonant circuit, and the resulting damped oscillation, at kilovolt amplitude and frequencies from the hundreds of kilohertz into the megahertz range, is coupled onto the torch cable through a series coupling inductor. The spark it produces ionizes a conducting path, the direct-current supply takes over, and a sensing circuit removes the high-frequency drive within milliseconds. Some designs instead use a capacitive discharge pulse or a low-current pilot arc, and torches designed for automated cells sometimes retract the cathode to make contact and then withdraw, avoiding high-frequency emission entirely at the cost of mechanical complexity.

The Electromagnetic Interference Problem

A megahertz spark of several kilovolts, injected onto a cable running twenty or thirty meters through a factory, is an efficient radiator and an efficient conducted-noise source. Spray cells have a long history of unexplained faults at torch ignition: robot encoders losing counts, fieldbus segments dropping frames, powder feeder controllers resetting, and mass flow controllers glitching. The starter is almost always the culprit, and the failures cluster at the instant of ignition rather than during the spray pass, which makes them easy to diagnose once the pattern is recognized.

The countermeasures are conventional but must actually be applied. The torch cable bundle is shielded and the shield bonded at both ends. The high-frequency energy is confined to the loop between starter and torch by radio-frequency bypass capacitors and common-mode chokes at the supply output. Signal cabling is routed away from the power umbilical rather than tied to it. Sensitive controllers are fed from a separately derived supply with line filtering. Under the CISPR 11 classification, spray equipment that deliberately generates radio-frequency energy for material treatment falls in Group 2, the same category as industrial radio-frequency heating and plasma processing equipment, and arc-based spray power sources are commonly built and tested to the arc welding equipment standards, IEC 60974-1 for safety and IEC 60974-10 for electromagnetic compatibility.

Constant Voltage Rather Than Constant Current

Because the electrodes are consumed, arc length is not fixed by geometry; it is the difference between how fast the wire is fed and how fast it melts. That makes a constant-voltage supply the natural choice, exactly as it is for gas metal arc welding. If the arc grows too long, its voltage rises, and on a constant-voltage output the current falls, which slows melting and lets the feed catch up. If the arc grows too short, voltage falls, current rises, melting accelerates, and the gap opens. The arc length regulates itself without any control loop at all, through the intersection of the supply's flat output characteristic with the arc's own voltage-current relationship. This is the opposite of the plasma spray case, and it is the single most important electrical distinction inside the thermal spray family.

Vendor literature for commercial arc spray systems describes constant-voltage direct-current rectifier supplies with arc voltages adjustable across roughly 18 to 40 volts and continuous output in the range of 100 to 400 amperes at a full duty cycle, with 450-ampere machines offered for higher throughput. Power is therefore modest, generally under twenty kilowatts, which is one reason wire arc spray is the cheapest process in the family to buy and to run.

Deposition Rate Follows Current

Melting rate in a consumable-electrode arc scales closely with current, and industry figures reflect this directly. Vendor and trade sources commonly quote copper- and iron-based alloys spraying at about 4.5 kilograms per hundred amperes per hour, and zinc at roughly 11 kilograms per hundred amperes per hour, the difference following from the far lower melting point and heat of fusion of zinc. Current setting is therefore the throughput control, and voltage setting is the quality control: higher voltage lengthens the arc, produces larger and hotter droplets, and generally coarsens the coating, while lower voltage gives finer droplets and denser deposits until the arc becomes unstable.

Transients and Modern Inverter Machines

A consumable arc extinguishes and reignites frequently, especially with cored wires and at low voltage settings. Each event is a short-circuit or open-circuit transient that the supply must ride through without tripping. Traditional machines used a transformer-rectifier with enough leakage inductance to soften the transitions. Inverter-based arc spray supplies now dominate new installations, because a fast output stage can shape the current through a short-circuit event rather than merely surviving it, and because the same converter can be reprogrammed for different wire chemistries. The wire feeder is part of the control problem: feed speed must track current setpoint, and a stalled or slipping feeder is a common cause of arc instability that looks, on the supply's front panel, like a power problem.

The Conventional Parameter Set

A production spray controller regulates and logs a familiar list: arc current, plasma or atomizing gas flows by mass flow controller, carrier gas flow, powder feed rate, cooling water flow and temperature rise, standoff distance and traverse speed from the robot, and part rotation speed. It logs arc voltage as a measured rather than commanded value. In a qualified process, every one of these is recorded for every part, because the qualification specification requires it. That parameter set is sufficient to reproduce a process on the same hardware. It is not sufficient to detect that the hardware has changed.

In-Flight Particle Diagnostics

Commercial diagnostics close part of that gap by measuring the plume itself. Single-particle instruments, of which the Tecnar DPV family is the best known, image a small measurement volume onto a detector through a mask with two parallel slits: the time between the two pulses from one particle gives velocity, and the ratio of emission in two wavelength bands gives temperature by two-color pyrometry, which is insensitive to the unknown emissivity of the particle. Ensemble instruments such as Accuraspray trade single-particle resolution for a fast average temperature, average velocity, and plume position, which suits production monitoring better than research. Imaging systems such as Oseir SprayWatch photograph the plume to extract position, width, and angle. These are vendor products with vendor performance claims, but the underlying measurement principles, two-color pyrometry and time-of-flight, are standard optical techniques.

Plume position deserves particular attention, because it drifts as the anode bore erodes and the arc root migrates, and because it is invisible in the electrical parameters. A plume that has shifted by a few degrees puts the powder stream through a different part of the jet, changing melting and velocity without changing a single number on the controller display.

Substrate Temperature and Deposition Feedback

Infrared pyrometry aimed at the part measures the variable that governs residual stress and adhesion, and it responds to the total energy actually delivered. The condition-monitoring study cited earlier found substrate surface temperature, measured in the eight to fourteen micrometer band, to correlate almost linearly with deposited coating mass, and recommended it as an indicator of torch status. Closed-loop control on substrate temperature, adjusting traverse speed or cooling air rather than torch power, is now common on thick or thermally sensitive coatings. Gravimetric powder feeders, which regulate mass flow by weighing the hopper rather than counting disk revolutions, close another loop that used to run open.

The Cooling Loop

Torch cooling is a closed loop through a chiller, and the coolant is deionized water. The reason is electrical rather than thermal: the nozzle and the cathode holder sit at process potential, and the coolant passages are in direct contact with both, so ordinary water would carry leakage current between the electrodes and to the machine frame. Plant practice is to run the loop through a mixed-bed deionizer cartridge and to monitor conductivity continuously, interlocking the supply if it rises. Flow rate and inlet-to-outlet temperature rise are monitored as well, and the pair provides a bonus measurement: the calorimetric heat removed, subtracted from the electrical input, gives the net power in the gas without any optical instrument. Many spray controllers compute and log exactly that.

The Umbilical

A plasma spray torch on a robot arm receives everything through one flexible bundle: the direct-current power conductors, cooling water supply and return, plasma and carrier gas lines, and the powder hose. In many designs the power conductors are themselves water-cooled, which allows a smaller and more flexible cable for a given current. The bundle must survive continuous articulation, and cable failure is a routine maintenance item rather than an anomaly. Electrically, the run length matters twice over: resistive drop wastes power and shifts the supply's sense point away from the torch, and cable inductance adds to the output filter and to the loop the high-frequency starter must drive. Long runs also make the bundle a better antenna, which is the reason shielding and bonding practice deserves the care described earlier.

The Booth

Spray happens inside an acoustically treated, ventilated booth for good reasons. Plasma spray is extraordinarily loud, well above one hundred A-weighted decibels at the operator position, which places the cell squarely inside hearing conservation requirements. The plume emits intense ultraviolet radiation, which produces ozone and nitrogen oxides in the surrounding air and burns unprotected skin and eyes as an arc welding arc does. Overspray produces fine particulate that must be captured. Ventilation and dust collection are therefore part of the installation rather than accessories, and their fans and controls belong in the electrical load schedule alongside the supply.

Electrical and Physical Hazards

Plasma spray supplies present an open-circuit voltage of a few hundred volts direct current at the torch terminals, which is lethal and, unlike an alternating-current shock, produces a sustained muscular contraction that makes release difficult. The high-frequency starter adds a second hazard of a different character, since kilovolt radio-frequency energy causes surface burns and can couple into anything nearby. Interlocked booth doors, torch-mounted covers, and a discharge path for the output inductor energy are all standard. Arc-based spray power sources are commonly designed and certified to IEC 60974-1, the safety standard for arc welding power sources, with electromagnetic compatibility addressed under IEC 60974-10. Fine metal powders introduce a combustible dust hazard, and aluminum, magnesium, and titanium feedstocks in particular fall under NFPA 484, the Standard for Combustible Metals, which drives dust collector selection, deflagration protection, and wet-collection practice. Robot cells add the motion hazards governed by ISO 10218.

Process Qualification

In aerospace and medical work, the coating is a controlled characteristic and the process that produces it is qualified rather than merely specified. SAE AMS 2437, Coating, Plasma Spray Deposition, currently in revision E, sets engineering requirements for applying plasma spray coatings and for the properties of the resulting deposits. Suppliers to the aerospace primes are additionally audited under the Nadcap coatings program, whose audit criteria are published as AC7109 with a set of slash sheets covering particular process families. Engine manufacturers layer their own specifications on top. The practical effect on the electrical designer is that every process parameter, including current, voltage, gas flows, and powder feed rate, must be recorded with traceable calibration and retained, and that a change to the power supply or its firmware is a change to a qualified process, requiring requalification rather than a service note. This is why spray equipment is unusually conservative in its control electronics, and why installations frequently run supplies far older than any other converter in the plant.