What this calculator does
At DC, current spreads evenly over a conductor's cross-section. At high frequency, the changing magnetic field inside the metal induces eddy currents that oppose the current in the interior, and the current crowds towards the surface. The current density falls off exponentially with depth, and the skin depth δ is the depth at which it has dropped to (37 %) of its surface value:
Skin depth sets the RF resistance of every conductor. If the metal is several skin depths thick, it behaves like a sheet of thickness δ, and its resistance per square is the surface resistance
goes straight into the conductor-loss equations of transmission lines, cavities and antennas. Because it grows as , conductor loss in dB per metre also grows as , until dielectric loss takes over.
The calculator also reports the frequency at which δ equals a given thickness,
and the ratio . Above this frequency the conductor no longer behaves like a DC conductor. Below it, the thickness, not the skin effect, limits the resistance.
Materials
The presets use room-temperature (about 20 °C) bulk values.
| Material | σ (MS/m) | Source |
|---|---|---|
| Copper, annealed | 58.0 | IEC 60028 annealed-copper standard (100 % IACS) |
| Silver | 63.0 | CRC Handbook, ρ = 1.587 µΩ·cm at 293 K |
| Gold | 45.2 | CRC Handbook, ρ = 2.214 µΩ·cm |
| Aluminium | 37.7 | CRC Handbook, ρ = 2.650 µΩ·cm |
| Nickel | 14.4 | CRC Handbook, ρ = 6.93 µΩ·cm (μr entered separately) |
| Brass, cartridge (C26000) | 16.2 | 28 % IACS (typical alloy value) |
| Tin | 8.70 | ρ = 11.5 µΩ·cm |
| Stainless steel 304 | 1.39 | ρ = 72 µΩ·cm (typical alloy value) |
Pure copper is slightly better than the IACS standard: CRC gives 59.6 MS/m. The 58.0 MS/m standard is the usual design value for commercial annealed copper and PCB foil. Choose Custom to enter any conductivity and permeability.
Nickel is the exception. It is ferromagnetic, and its permeability depends strongly on frequency: it can be a few hundred at low frequency and falls towards 1 above its ferromagnetic resonance, in the low-GHz range. It also depends on the plating process and on the field strength. When you choose nickel, the calculator shows the μr input. Enter a value that applies at your frequency, and it warns you if you leave μr at 1. This matters for ENIG finishes, where a few microns of nickel sit under a thin gold layer and can add noticeable loss below a few GHz.
How to use solve-for
- Solve Frequency for Skin depth to find where δ reaches a target. For example, enter 1 µm to see where a 1 µm plating starts to carry most of the current.
- Frequency where δ = t is already a direct output. Solving Conductor thickness for it gives the thickness that equals one skin depth at a frequency.
- Solve Conductor thickness for t/δ = 3 to find the thickness that is three skin depths at your frequency. In a conductor that thick, 95 % of the current () flows within the conductor.
Worked example
Copper at 1 GHz:
- skin depth δ = 2.090 µm;
- surface resistance = 8.250 mΩ per square;
- 1 oz foil (m) is = 16.75 skin depths thick;
- δ equals 35 µm at 3.565 MHz.
At 1 GHz, the other presets give δ = 2.005 µm for silver, 2.368 µm for gold, 2.591 µm for aluminium, 5.397 µm for tin and 13.50 µm for 304 stainless.
As a cross-check against the literature: at 10 GHz, with σ = 5.8×10⁷ S/m, textbooks quote δ = 6.6×10⁻⁵ cm (0.66 µm), and this calculator gives 0.6609 µm.
Accuracy and validity
The formula is exact for a flat, smooth, homogeneous conductor much thicker than δ, with a frequency-independent conductivity. It also holds for curved conductors whose radius of curvature is much larger than δ. Its limitations are in the inputs and in a few physical effects:
| Effect | Impact |
|---|---|
| Temperature | Copper's σ falls about 0.39 %/°C, so δ grows about 0.2 %/°C. At 85 °C, is about 12 % higher than at 20 °C |
| Surface roughness | A real foil with roughness comparable to δ has 1.2–2× (and more) the smooth-surface loss |
| Thin conductors () | The current also flows on the back surface; the resistance does not follow |
| Alloys and platings | Conductivity varies with composition and process, often ±10 % or more |
| Above ~1 THz or at cryogenic temperature | The anomalous skin effect and carrier relaxation break the classical model |
Practical notes
- Platings. A finish thinner than about one skin depth is largely transparent: the current flows mostly in the metal underneath. Thick silver plating on brass gives silver-like loss above a few hundred MHz.
- Thicker copper doesn't help at RF. Above a few tens of MHz, extra copper weight barely reduces conductor loss, because the current lives within a few microns of the surface. On a microstrip, most of it flows on the underside of the trace, next to the dielectric, where the foil is roughest.
- Litz wire works only at low frequency. Stranded insulated wire reduces skin and proximity losses only while each strand is thinner than about δ, which in practice limits it to below a few MHz.
- Shielding. The absorption loss of a shield is 8.69 dB per skin depth of thickness, which is why thin copper shields work well at RF but poorly at power-line frequencies, where δ in copper is about 8.5 mm at 60 Hz.