A copper trace is a patterned conductor that connects points on a printed circuit board. It is not a collection of fuses, insulation and components. Its resistance depends on copper resistivity, length and effective cross-section; its behavior at higher frequencies also depends on the surrounding electromagnetic structure. Good trace design therefore checks voltage drop, heating, signal return paths and manufacturing tolerances separately.
This guide explains the resistance calculation and how to turn it into a useful design check. For choosing a width from current and temperature-rise requirements, continue with the PCB trace-width guide. A resistance calculation alone does not establish a safe continuous-current rating.
Calculate copper trace resistance with consistent dimensions
What copper contributes to a PCB
Copper combines useful electrical conductivity with fabrication processes that can form pads, narrow interconnects and broad planes. It still has finite resistance. A long thin conductor can produce an unacceptable supply drop or measurement error even when it passes a simple continuity test. Increasing copper area can reduce that resistance, but the change must also fit spacing, etching, soldering and signal-integrity requirements.
Copper weight is an areal mass, commonly stated in ounces per square foot. One ounce per square foot corresponds to approximately 35 micrometres of copper, not one ounce spread over one square metre. Treat this as a nominal conversion, not a guarantee of the final conductor thickness. Outer-layer processing and plating differ from inner-layer construction. The copper-weight specification guide helps distinguish the material callout from the finished-board requirement.
Use R = rho 脳 L / A
For a uniform rectangular trace, A = w 脳 t and R = rho 脳 L / (w 脳 t). Here R is resistance in ohms, rho is resistivity in ohm-metres, L is length in metres, w is width in metres and t is thickness in metres. A convenient illustrative room-temperature copper resistivity is 1.72 脳 10鈭�8 ohm-metres. Actual material, temperature and geometry should be reflected in the design model.
Assume a 100 mm trace that is 1 mm wide and 35 micrometres thick. Convert the dimensions first: L = 0.1 m, w = 0.001 m and t = 0.000035 m. The calculated resistance is approximately 0.0491 ohm, or 49.1 milliohms. At a specified 2 A, the voltage drop I 脳 R is about 98.3 mV and the conductor loss I虏 脳 R is about 0.197 W. These are electrical estimates at the assumed temperature; they do not predict the resulting temperature rise.
For the same width, thickness and temperature, doubling the length doubles resistance. Doubling width halves resistance. Doubling thickness also halves the ideal resistance, but that does not automatically double a finished board’s allowable current. Heating and heat removal interact, and nearby copper and construction affect the temperature reached. Use the calculation to compare explicit geometries rather than making a universal ampere-per-millimetre rule.
| Illustrative geometry | Resistance | Drop at 2 A |
|---|---|---|
| 100 mm 脳 1 mm 脳 35 碌m | 49.1 m惟 | 98.3 mV |
| 100 mm 脳 2 mm 脳 35 碌m | 24.6 m惟 | 49.1 mV |
| 50 mm 脳 1 mm 脳 35 碌m | 24.6 m惟 | 49.1 mV |
Include temperature and the narrow sections
A first-order temperature model is R(T) = R(T0) 脳 [1 + alpha 脳 (T 鈭� T0)]. Using an illustrative alpha of 0.0039 per degree Celsius, a 60 degree increase raises the preceding 49.1 milliohm estimate to about 60.6 milliohms. At 2 A that corresponds to about 121 mV drop and 0.243 W loss. Temperature is an input to this calculation, not an output. If self-heating is significant, evaluate the electrical and thermal conditions together.
A real power path may include a wide pour, a short neck at a pad, several vias and a connector. Divide it into sections with suitable models and add series contributions. Do not assign the width of the largest pour to the complete path. For parallel routes, current sharing depends on the full path resistance and connection geometry; an equal split is an assumption to test, not a consequence of drawing two traces.
Use minimum or otherwise justified finished dimensions for the limiting case. Etching can produce a nonrectangular cross-section, and copper thickness has tolerances. A nominal CAD width is not a measured cross-section. State which value the fabricator should control and how it is verified. Keep the calculation record attached to the applicable PCB manufacturing specification.
Check the circuit consequence and verify the manufactured path
Resistance, current capacity and impedance answer different questions
DC resistance is useful for supply-drop and loss estimates. Current capacity adds a temperature-rise and reliability requirement under defined conditions. identify conductor sizing as a current, geometry and temperature problem and include board construction and environment considerations. Use the applicable complete guidance or a validated model for a design calculation; the public contents excerpt is not the full standard.
Controlled impedance addresses signal propagation and the relationship between the trace and its reference structure. A trace with low DC resistance can still be a poor transmission line if its return path is interrupted. Conversely, changing a signal width to reduce resistance can change its characteristic impedance. Keep the stackup, reference plane and routing context in the impedance-control review instead of treating ohmic resistance as the only trace property.
At higher frequencies, skin and proximity effects can make conductor loss differ from the DC estimate. Roughness, vias and transitions may also matter. Do not use the simple rectangular DC equation to claim an insertion-loss result. State the frequency range and use a model appropriate to the signal problem, then verify the relevant structure when the design requires measurement.
Keep load current out of precision pickup paths
A current-sense circuit can accidentally include copper resistance in the quantity it measures. For example, an intended 10 milliohm shunt with an additional 0.5 milliohm of load-carrying trace inside the sense pickup boundary produces a 5% scaling error under the simple series model. The remedy is to review the pickup locations and complete return path, not to insert an unexplained correction into firmware.
Use the Kelvin connection review when a shunt or other low-resistance element is measured. Separate current-carrying terminals from voltage-sense points according to the component design. Include connectors, test fixtures and board variants. A sense trace connected to the correct net but at the wrong physical location may still include an unwanted voltage drop.
Measure milliohms with a defined boundary
For low resistance, a two-wire meter reading can be dominated by test-lead and contact resistance. A four-wire measurement applies current through one pair of connections and senses voltage through another pair. Define the pickup points on the drawing, use an appropriate test current, and control self-heating and thermal offsets. Record instrument range, resolution and uncertainty rather than copying all displayed digits into an acceptance limit.
Compare the measured path with the calculation boundary. If the test includes a connector or solder joint that the calculation excludes, a discrepancy is not automatically a fabrication defect. Conversely, a continuity pass cannot establish a milliohm limit. Investigate geometry, temperature, fixture contact and assembly condition before concluding that copper resistivity has changed.
Release the requirement, not just a calculator screenshot
The useful output of a PCB design review is a traceable requirement: path endpoints, operating current, acceptable voltage drop, temperature conditions, finished copper dimensions and validation method. Include neckdowns and vias in the review. If the path changes during placement or routing, rerun the relevant check instead of retaining the original result as a generic approval.
For fabrication or assembly discussion, send the copper-path requirement with the board files. Identify whether the concern is voltage drop, heating, controlled impedance or precision sensing. These are related engineering tasks, but each requires its own assumptions and acceptance evidence.