Copper Busbar Ampacity Chart

Two things to read out of this table before using it. First, the AC and DC columns are close at 50/60 Hz because skin depth in copper at that frequency exceeds every bar thickness listed here; the columns only diverge as frequency rises. Second, these numbers assume a single bar in open air — stacked or packed bars derate quickly, per the factors below.

Cross-sections are reference values for planning, not release specifications: every released part is sized against your actual mounting and ambient. Our standard copper is C11000; when aluminum is chosen for weight or cost, it needs its own sizing — see copper vs aluminum busbar conductors.

One more reason the table uses a fixed 30 °C rise: temperature is what ultimately limits the part. Copper resistance rises about 0.4% per °C, so a conductor running at 100 °C dissipates roughly 40% more heat for the same current than one at 20 °C. A 30 °C rise above 40 °C ambient lands the conductor near 70 °C — comfortably inside a 105 °C dielectric system with margin left. If your enclosure already runs hot, you are buying margin you cannot use; size to the actual ambient instead. Every released part passes hipot, insulation-resistance and continuity checks under our in-house busbar testing process, with batch records supplied on request.

How the factors combine matters more than any single row. Derating multiplies: a part rated 450 A in free air at 40 °C, mounted in a sealed cabinet (about −20%) at a 60 °C ambient (about −20% from temperature), lands near 450 × 0.8 × 0.8 ≈ 290 A — not 450 minus 40. Treat the rows as independent multipliers and stack them.

Two further factors decide real designs less often but more heavily:

• Altitude above roughly 2000 m — thinner air cools worse and breaks down electrically more easily. Plan to derate both current and voltage class, and test parts with the corrected voltage, as explained in partial discharge in laminated busbars.
• Layer sharing — in a laminated stack the total copper area sets the heating, but each layer only sees its share of the current. Uneven sharing is a hidden derate; see the next section.

There is one ceiling the multiplier table cannot express: the dielectric temperature limit. The copper may survive 120 °C, but a 105 °C film system does not. In practice you cross-check the computed conductor temperature against the dielectric class — a 105 °C system wants the conductor near 70–80 °C, while a 130 °C high-temperature system lets the same copper carry roughly 20–30% more current. Choosing the dielectric class is therefore part of current sizing, not a separate step. Lifetime implications are covered in busbar thermal management.

The mechanism is symmetry. Where positive and negative terminals exit the stack side by side, their magnetic fields cancel and current distributes evenly across the parallel layers. Where one terminal sits offset, or layers enter from different edges, impedance differences — tiny at DC, significant at the switching frequency — push more current through one layer. That layer hits its temperature limit first, while the cooler layer's copper was bought for nothing.

Rules we apply in DFM:

• Route positive and negative exits at the same edge, with matched terminal counts per layer.
• Keep layer lengths within a few percent of each other; a long outer layer is a high-impedance layer.
• For the high-frequency ripple path from DC-link capacitor to module, verify sharing with a current-density simulation rather than arithmetic.

High frequency changes the picture again. Above roughly 1 kHz, skin and proximity effects push current to the surfaces of the conductor, so a single thick bar wastes its center. A laminated stack of thinner layers with overlapping planes behaves better electrically at these frequencies than the same cross-section in one plate — another reason the stacked geometry is preferred for DC-link paths, not just for terminal layout.

Ampacity tables hide the physics: a current rating is really a temperature-limit calculation. Treat heat flow like an electrical circuit — Ohm's law in thermal clothing:

• Heat flux Q = P_loss / A, in watts per square meter of surface.
• Temperature rise ΔT = Q × R_th, where total thermal resistance R_th = R_conduction + R_convection + R_radiation from the hot spot to ambient.

Conduction inside the stack is not the bottleneck. Copper spreads heat extremely well — thermal conductivity of roughly 370–400 W/m·K — and even the 0.2–0.4 mm dielectric film drops only about 0.5–1 °C across itself. The bar's own mass averages the heat over its whole surface.

Convection to ambient dominates natural cooling. A flat, thin laminate presents a very large surface for its volume — an inherent advantage over a round cable bundle. Reference heat-flux capability: natural convection about 5 W/dm²; forced air with fans about 50 W/dm². Orientation shifts the result too: vertical mounting typically improves heat removal by roughly 20% over horizontal.

Radiation contributes more than people expect. Bare polished copper radiates almost nothing (emissivity about 0.03); a painted or film-laminated outer surface reaches about 0.9. For a naturally cooled part, radiation accounts for roughly 40% of total heat removal — which is why a black-painted busbar can run about 25 °C cooler than the same bar left bare, counter-intuitive as it sounds.

Busbar conduction losses are typically 10–100 W, so natural air cooling almost always suffices; you do not need to fan the bar itself, though neighboring fan cooling adds margin. The full mechanism, including measurement practice, is in busbar thermal management.

A quick worked example makes the surface budget tangible. With a 30 °C maximum rise and an overall heat-transfer coefficient near 20 W/m²·K on both faces, a naturally cooled laminate sheds about 6 W per square decimeter of surface. A 25 × 40 cm bar therefore dissipates roughly 60 W without any fan — comfortably above the 10–100 W loss range typical of these parts. Forced air raises the heat-transfer coefficient to 50–100 W/m²·K, roughly 5–10 times the natural value, and allows 2–3 times the current on the same copper. In practice the airflow around a power stack already adds this margin, which is why busbar cooling calculations usually start from natural convection and stop there.

Continuous rating asks what the bar can carry forever. Pulsed duty asks something else: how much heat the copper can absorb before its temperature limit is reached. Copper mass is a heat reservoir — short, high-current pulses dump energy into the body faster than the surface can shed it, and the rise depends on pulse duration and repetition, not just RMS current.

Reference practice for sizing pulses:

• Very short pulses, milliseconds: the bar can absorb several times its continuous rating, because surface cooling barely has time to act. Energy (I² × t), not steady current, is the currency.
• Long pulses, seconds to minutes: the reservoir fills. Average current over the duty cycle eventually sets the temperature, and you fall back to continuous sizing.
• Between the two regimes we simulate the transient rather than guess — conductor thickness, not just area, decides how quickly the heat spreads.

Pulse scenarios in real designs include inrush into DC-link capacitors, motor starting currents, and charge-discharge peaks in storage systems. If your current profile is pulsed, send us the pulse width, peak and repetition rate; the thermal-mass calculation is documented with the sizing sheet.

Contrast the two regimes on the same part: a 30 × 5 mm copper bar rated about 450 A continuously in free air may comfortably carry 900–1000 A for a fraction of a second during a motor start, because the heat has no time to reach the surface. Hold that same 900 A for a minute, and the same bar overheats — the reservoir has filled and continuous rules take over. Knowing which regime you are in is half the sizing problem. Thermal mass also decides fault behavior: a short-circuit peak of several kiloamperes for a few cycles does not overheat a well-sized bar, which is one more reason conductor cross-section, not just continuous current, is documented in the release drawing.

The chart becomes useful once anchored to a duty cycle. Three reference configurations — typical planning values, to be verified per drawing:

The rail row maps to our rail traction busbar page, where the environmental case — wide temperature swings, humidity, service continuity — drives a more conservative derate than any of the other applications.

These map to the field roles described in EV busbars, energy storage busbars and inverter busbars — each application page lists the thermal and electrical environment that drives the derate. The same sizing pattern recurs in EV charging station busbars — continuous rectifier-to-outlet distribution with low ripple — and in data center power distribution, where flat, low-profile planes carry rack-level DC with high reliability. In every case the continuous number from the chart is the starting line, not the final specification.

• Start with your continuous current and voltage class.
• Apply the derating factors for your actual mounting and ambient.
• Select conductor cross-section and layer count with margin.
• Confirm the thermal class of the dielectric against the resulting temperature.

Then confirm the loop inductance of the selected geometry. Current capacity and inductance are optimized together, not separately — see low-inductance busbar design.

To run this process for you we need four things: the continuous current and any peak or pulse profile, the voltage class, the enclosure ambient and cooling mode, and the terminal types. With those, we return a recommended cross-section, layer count and dielectric choice, with the calculation documented alongside the quotation — typically within one business day. Values in the chart are planning references; every released part is computed against your drawing, and the quoted part always carries its sizing sheet.