The corona effect is the localized ionization of air surrounding a high-voltage conductor that occurs when the electrical field gradient exceeds the dielectric breakdown strength of the surrounding gas, resulting in a faint glow, hissing noise, and power loss. In a real circuit or installation, corona changes the system by introducing parasitic capacitance, generating ozone (which rapidly degrades nearby silicone and rubber insulation), causing severe radio-frequency interference (EMI), and bleeding off real power as heat and light. Hobbyists and junior engineers most commonly confuse corona with arcing (a complete, low-impedance plasma channel bridging two potentials) or tracking (carbonized conductive paths forming along the surface of a solid insulator). Corona is strictly a localized, non-bridging partial discharge.
The Physics of Air Breakdown (With Worked Numbers)
To understand corona, you have to look at the dielectric strength of air. At standard temperature and pressure (STP), dry air breaks down and becomes conductive at an electric field gradient of approximately 30 kV/cm (3 MV/m). When the voltage on a conductor is high enough, and the conductor's radius is small enough, the electric field at the surface of the metal exceeds this 30 kV/cm threshold. The air molecules immediately adjacent to the metal strip their electrons, creating a localized plasma. Because the field strength drops off rapidly as you move away from the curved surface of the wire, the ionization doesn't bridge the gap to ground—it just hovers around the conductor as a glowing, hissing cloud.
Imagine you are building a 15 kV DC power supply for an electrostatic precipitator or a custom X-ray tube. You route the high-voltage output using a bare, solid copper wire with a 2 mm radius (0.2 cm).
For a simplified bench calculation of a wire far from ground planes, the peak electric field (E) at the conductor surface approximates to Voltage divided by radius (E ≈ V/r).
- Applied Voltage (V): 15,000 V
- Wire Radius (r): 0.2 cm
- Surface Field (E): 15,000 / 0.2 = 75,000 V/cm (75 kV/cm)
Because 75 kV/cm is more than double the 30 kV/cm threshold of air, this wire will generate heavy corona discharge. You will hear a loud hiss, smell ozone, and watch your 15 kV supply sag under the parasitic load.
The Fix: To eliminate corona in free air, you must increase the radius until the surface field drops below 30 kV/cm. r = V / 30,000. Therefore, r = 15,000 / 30,000 = 0.5 cm. You need a conductor with at least a 5 mm radius (10 mm diameter), or you must alter the dielectric environment.
For a deeper dive into the fundamental physics of dielectric breakdown thresholds across different gases, the Georgia State University HyperPhysics database provides excellent baseline constants for bench calculations.
Where You Meet Corona in Practice
You will rarely design utility-scale transmission lines, but the corona effect shows up frequently in bench-level high-voltage and high-frequency projects:
- Flyback Transformers and SMPS: In switch-mode power supplies generating 5 kV to 30 kV (like neon sign drivers or CRT flybacks), sharp solder joints or thin enamel magnet wire at the peak of the winding will corona. This eats away the enamel, eventually causing a dead short.
- Tesla Coils and High-Frequency HV: The fast dV/dt (rate of voltage change) in Tesla coil top loads exacerbates corona. This is why builders use smooth aluminum toroids (top loads) rather than sharp spheres or bare wires—to increase the radius and keep the surface gradient below the breakdown threshold.
- Humid Utility Lines: If you live near high-voltage transmission lines, you can hear corona on humid or foggy days. Water droplets on the cable distort the local electric field, creating microscopic sharp points that trigger localized corona, resulting in an audible crackling or hissing sound.
- Ozone Generators: Corona is intentionally harnessed in ozone generators. By forcing a high-frequency, high-voltage field across a sharp mesh or wire inside a ceramic tube, the localized corona splits O2 molecules, which recombine as O3 (ozone) for water purification or air treatment.
Corona vs. Arcing vs. Tracking
Misdiagnosing a high-voltage failure mode leads to the wrong fix. Here is how to tell them apart on the bench:
| Phenomenon | Visual / Audio Signature | Physical Mechanism | Primary Damage |
|---|---|---|---|
| Corona | Faint purple/blue glow, steady hissing, strong bleach/ozone smell. | Localized air ionization; does not bridge the gap to ground. | Ozone degrades silicone/rubber; EMI disrupts nearby microcontrollers; gradual power loss. |
| Arcing | Bright, blinding white/blue flash, loud sharp crack or bang. | Complete dielectric failure; a low-impedance plasma channel bridges the gap. | Instantaneous catastrophic failure, melted copper, tripped breakers, destroyed MOSFETs. |
| Tracking | No glow, silent, leaves a permanent black, charred line on an insulator. | Surface leakage currents carbonize organic insulation, creating a conductive path. | Permanent degradation of PCB FR4 or plastic bobbins, leading to eventual surface arcing. |
If your high-voltage PCB is failing silently over a few weeks and leaving black scorch marks across the FR4 fiberglass between high-side and low-side traces, you are dealing with tracking, not corona. If your microcontroller keeps resetting when the HV supply turns on, but there are no scorch marks, you are likely dealing with the massive EMI generated by corona.
Decision Tree: Mitigating Corona in Your HV Design
When you detect corona (via ozone smell, AM radio static, or UV camera), you must alter the geometry or the dielectric. Use this decision path to select the exact mitigation strategy and part for your build.
| IF your design scenario is... | THEN apply this mitigation... | CONCRETE PICK / PART NUMBER |
|---|---|---|
| Scenario A: Bare high-voltage busbars or sharp solder joints in an enclosed, low-vibration power supply (e.g., SMPS, HV multiplier). | Pot the assembly. Submerging the sharp points in a high-dielectric solid resin eliminates the air gap entirely, raising the breakdown threshold from 30 kV/cm to >200 kV/cm. | MG Chemicals 832C (Thermally conductive epoxy potting compound. Dielectric strength: 450 V/mil. Mix 1:1 by volume). |
| Scenario B: High-voltage RF outputs in free air where heat dissipation is critical and potting is impossible (e.g., Tesla coil top loads, antenna feeds). | Increase the surface radius. Attach a smooth, hollow toroid to the sharp termination to distribute the electric field gradient over a larger surface area. | 50mm OD Aluminum Toroid (Spun aluminum, mirror-polished. For >100kV, step up to a 100mm OD toroid to maintain the E < 30 kV/cm rule). |
| Scenario C: High-voltage PCB traces operating between 2 kV and 10 kV where creepage distance is limited by board size. | Conformal coat and route properly. Apply a dielectric barrier to prevent surface tracking and localized air ionization at trace edges. Ensure traces have no 90-degree corners (use teardrop pads). | MG Chemicals 419D (Acrylic conformal coating. Apply two brushed coats, allowing 30 min cure between. Dielectric strength: 1500 V/mil). |
| Scenario D: High-voltage cable terminations entering a grounded metal chassis (e.g., 30kV DC coax entering a metal enclosure). | Use a stress cone or grading ring. The sharp transition from insulated cable to bare terminal creates a massive field gradient. A geometric stress cone smoothly transitions the field lines. | Silicone Rubber Stress Cone (Rated for your specific cable OD. Alternatively, build a DIY grading ring using 10 AWG bare copper wire wrapped in 3M Scotch 23 high-voltage splicing tape). |
FAQ: Bench and Field Diagnostics
How can I detect corona on my bench without an expensive UV camera?
Tune a cheap analog AM pocket radio to the bottom of the dial (around 540 kHz) where there are no local stations. Turn the volume up and bring it within 12 inches of your high-voltage circuit. Corona discharge generates massive broadband RF noise. If you hear a loud, steady static hiss that syncs with your HV supply turning on, you have corona. If you hear sharp, intermittent pops, you are likely seeing micro-arcing.
Does corona only happen at high voltages, or can it happen at high frequencies?
Corona is primarily a function of peak voltage and conductor geometry, not frequency. However, high frequencies (or fast dV/dt square waves from SiC/GaN MOSFETs) make the effects of corona much worse. Rapid voltage transitions cause the localized capacitance of the corona cloud to charge and discharge violently, massively amplifying the EMI and accelerating the ozone generation rate. A 5 kV peak sine wave might barely hiss, but a 5 kV peak, 100 kHz square wave will generate aggressive corona and rapid insulation failure.
Can I just use hot glue or standard silicone RTV to stop corona on a PCB?
No. Standard hardware-store silicone RTV (like GE Silicone II) releases acetic acid while curing, which corrodes copper traces. More importantly, standard hot glue and cheap silicones are highly susceptible to tracking and have poor dielectric consistency. For high-voltage work, always use purpose-built electrical grade silicones (like Dow Dowsil 3140) or rigid epoxies. As noted by Fluke's diagnostic guidelines on partial discharge, using the wrong dielectric material can inadvertently create voids that trap air, making internal partial discharge worse than the original surface corona.
Why does my silicone high-voltage wire turn brittle and crack after a few months?
This is the hallmark of ozone damage caused by nearby corona. Ozone (O3) is a highly reactive oxidizer. While silicone is generally heat resistant, it is highly vulnerable to ozone attack, which breaks the polymer chains, causing the jacket to turn chalky, brittle, and eventually crack. If you see this, you have a corona source nearby. You must either eliminate the corona source using the decision tree above, or switch to an ozone-resistant wire jacket material like PTFE (Teflon) or EPDM rubber.
