Sizing High-Current Battery Cable: Conductor Area, Lug Crimp and Bolt Torque

High-current battery cable conductor sizing and lug crimping is the part of an energy-storage or charging build where a table lookup does the most damage. Somebody finds an ampacity chart, reads across to the amps they expect, writes "50 mm²" on the drawing, and considers the job done. That chart assumed a single cable in free air at 30 °C with a continuous load. The actual cable is one of eight in a bundle, inside a cabinet at 55 °C, landing on a terminal rated to 75 °C, carrying a duty cycle whose peak is three times its average. Every one of those four differences moves the answer, and three move it the same way.

What follows is how the number is actually derived, what to specify about the lug crimp so the joint is not the weak link, and why bolt torque is a safety-critical figure that has to come from somewhere other than a generic fastener table.

Sizing the conductor: temperature, drop and class

Ampacity is a temperature problem, not a current problem

Copper has no current limit. Insulation has a temperature limit, terminals have a temperature limit, and current is only the thing generating the heat. The real question is: what current, in this thermal environment, keeps the conductor below the lowest temperature ceiling in the chain? Four inputs change the answer.

Ambient. Published ampacity tables state their reference ambient — commonly 30 °C. A cabinet running at 55 °C is not a small correction; the derating factor is substantial and is published alongside the table. Use the ambient inside the enclosure at the worst operating case, not the room.

Grouping. Bundled cables heat each other. Standards publish a grouping factor by number of loaded conductors, falling quickly for the first few. A cable that is fine on its own can be well over its real rating once it is one of six in a tight bundle.

Duty cycle. The heating current is the RMS value over the thermal time constant of the cable, not the peak. A lead that sees 200 A for ten seconds every two minutes is not a 200 A lead. Give the supplier the duty profile, not a single number.

Termination temperature limit. This is the one that catches people. The insulation may be rated 105 °C, but if the lug lands on a device terminal rated 75 °C, the circuit is sized at the 75 °C column. The higher cable rating does not buy you anything; the joint is what is being protected.

And then voltage drop often governs anyway

On low-voltage DC — 48 V racks, 60 V and 72 V two-wheeler packs, charging-pile module wiring — the heat calculation frequently is not what sizes the cable. The drop budget is.

The working formula for a DC run, both conductors counted:

V_drop = 2 × L × I × ρ / A

with L in metres, I in amps, A in mm², and ρ the resistivity of copper — about 0.0172 Ω·mm²/m at 20 °C.

The detail that gets missed: copper's resistance rises roughly 0.39% per °C. A conductor sitting at 90 °C has about 1.27 times its 20 °C resistance. If you sized to a 2% drop using the cold figure, the hot cable delivers something closer to 2.5%, and the extra drop is extra heat, which raises the resistance further. Size the drop at operating temperature, not at bench temperature.

Conductor class is a spec line, not a detail

Two cables can both say 50 mm² and behave completely differently. IEC 60228 defines conductor classes:

• Class 2 — stranded, for fixed installation. Fewer, thicker strands. Stiff.
• Class 5 — flexible. Many finer strands.
• Class 6 — extra flexible. Finer still; welding-cable and battery-cable construction.

Three reasons this costs money if ignored.

First, bend life. A class 2 conductor in a lead moved at every service work-hardens and breaks strands at the crimp neck. Anything that flexes wants class 5 or 6.

Second, the crimp die is class-specific. A hexagonal die sized for a class 2 conductor in a given lug will under-compress a class 6 conductor of the same nominal area, because the fill and the strand behaviour differ. Shops that own one set of dies and one lug range will crimp whatever arrives in whatever they have. Specify the conductor class and the lug part number together, and ask which die is used.

Third, it is the honesty check. IEC 60228 publishes a maximum DC resistance in Ω/km at 20 °C for each size and class. That single table is the best acceptance test you have against undersized conductor and against copper-clad aluminium sold as copper: measure the finished lead's end-to-end DC resistance with a four-wire instrument, compare it against (Ω/km × length) plus a small allowance for the two joints, and a lead that is 30% high is not the cable you bought. Ask for this measurement per lot, in writing, with the instrument named.

The lug crimp, and the number that proves it

What actually makes a lug crimp sound

A large lug crimp is not a scaled-up terminal crimp. Different tooling, different acceptance.

Hexagonal compression is the mainstream method: a hex die closes around the barrel and deforms it into the conductor. For a given lug and conductor, the die number and the number of compressions along the barrel are both specified by the lug maker — small sizes may take one or two crimps, large barrels three, four or more spaced along the length. A barrel that has been crimped once in the middle and left alone at the ends is a joint with a fraction of the contact it was designed for.

Deep-indent crimping puts a single deep punch into the barrel — easier tooling, common on cheaper work, and it leaves more of the barrel loose. Hydraulic crimping is the same hexagonal geometry with the force from a hydraulic head rather than a mechanical press, the practical route above the size a bench press can close.

What a buyer should specify and check:

• Lug part number and its maker's crimp specification (die, crimp count, positions).
• Conductor fill: the barrel is designed for a specific area, and shimming a smaller conductor into a bigger barrel is not acceptable.
• Strip length and inspection window: many lugs have a sight hole, and the conductor should be visible at it — the fastest visual reject there is.
• Cross-section evidence: a sectioned crimp shows whether the strands have been deformed together into a solid mass or whether there are voids. This is the same discipline applied to signal-size crimps, described in crimp height, pull-force and cross-section, scaled up.
• Pull-force: a stated minimum tensile figure per size, tested to destruction on a sample.

And one thing to prohibit: do not solder a high-current lug. Solder wicks up between the strands past the barrel and creates a stiff transition where the cable will fatigue-crack under vibration, and it creeps under sustained bolt pressure, which quietly unloads the joint you torqued. Crimp it, or have it welded by a process the supplier can describe.

Joint resistance: the number that ends the argument

Everything above is a proxy for one measurable thing: how much resistance the joint adds. IEC 61238-1 tests compression connectors for power cables by comparing the joint's resistance against that of an equal length of the conductor itself, then heat-cycling the assembly to see whether the ratio stays stable. That acceptance concept — the joint should be no worse than the conductor it replaces, and should not drift upward over cycling — is what belongs in a purchase specification.

Practically, for an incoming-goods check:

Note what the last row is not. Withstand-voltage and insulation-resistance tests say nothing about whether the crimp is any good. They are worth having, and most harness shops list the equipment for them — but a supplier who answers "we hipot every lead" to a question about joint resistance has not answered it.

Bolt torque: where a good cable is destroyed at the last centimetre

The lug goes on a stud and a nut goes on top. That is where high-current assemblies fail, and the reasons are mechanical rather than electrical.

• Torque is a proxy for preload, and a lossy one. Most of the applied torque goes into friction under the nut face and in the threads. Lubricate a thread specified dry and you over-tension badly at the same reading; leave dry a thread specified lubricated and you under-load it. State the condition with the number.
• The number comes from the terminal, not the bolt. A battery post insert, a busbar tapped in soft copper, a device terminal with a moulded body — each has its own limit, usually well below what an equivalent steel fastener could take. Get the figure from the component manufacturer's documentation and put it on the drawing with its source cited.
• Copper creeps. A joint torqued cold and then cycled to temperature relaxes. A conical spring washer under the nut maintains load through that relaxation; a plain washer does not. On accessible installations a re-torque after commissioning is standard; on sealed assemblies, design the spring washer in.
• Contact geometry. The lug palm should sit fully on the pad, not overhang it. Where two lugs on one stud are unavoidable, the convention is the larger palm underneath — and two is the limit.
• Plating pairs. Tin on tin is straightforward. A tin lug on bare aluminium busbar is a galvanic couple that corrodes at the interface in a humid cabinet; bimetallic lugs exist for this and cost more for a reason.
• Torque-stripe every joint. A paint line across nut, palm and stud makes joint movement visible in a photograph, which is the only practical way an inspector who does not carry a calibrated wrench can verify anything after the fact.

Termination choice at the other end of the lead — bolted versus pluggable — is a separate decision with its own trade-offs, covered in ring terminal or quick-disconnect for two-wheeler EV battery leads.

The sizing worksheet

Fill this in before you send a drawing, and the quotes you get back will be comparable — one of the more common reasons quotes for the same spec differ is that the spec was not the same spec.

1. Continuous current (RMS over the duty cycle) and peak current with its duration.
2. Worst-case ambient at the cable, inside the enclosure.
3. Number of loaded conductors bundled with it, and the routing.
4. Run length, one way, and the allowed voltage drop as a percentage of system voltage.
5. System voltage and insulation voltage rating required.
6. Insulation material and temperature rating; the temperature limit of every terminal the cable lands on.
7. Conductor class per IEC 60228 (2, 5 or 6) and whether the lead flexes in service.
8. Lug part numbers at both ends, stud size, and the crimp specification from the lug maker.
9. Bolt torque per end, with the source document and the dry/lubricated condition.
10. Acceptance evidence required: DC resistance per lot, joint resistance per joint, pull-force sample, cross-section photograph, torque-stripe photograph.
11. Marking: polarity, part number, and length, printed or on a durable label.

Common questions

Can I just give the supplier the amps and let them size it?

You can, and you will get a cable sized against their assumptions about ambient and bundling — optimistic assumptions, and certainly not your installation. Give the four thermal inputs and the drop budget, then take the sizing decision yourself or review theirs.

Is bigger cable automatically safer?

Bigger conductor is thermally safer, but not free: bend radius grows, the lug and die range changes, the terminal may not accept the barrel, and the mass hanging off a battery post rises — a vibration problem. Oversizing past two steps usually trades a thermal problem for a mechanical one.

What is a reasonable joint resistance limit?

It is set against the conductor, not against a universal number: the joint should add no more than roughly the resistance of an equal short length of the same conductor, and should not drift upward through heat cycling. Derive your limit from the conductor's published Ω/km and put the figure, in µΩ or mΩ, in the purchase specification.

What to ask the supplier next

For high-current battery leads and cabinet cabling, these six questions separate a shop that builds them from a shop that will learn on your order:

• What is the biggest conductor cross-section you can crimp in-house, and on what press, at what tonnage?
• Do you use hexagonal compression dies matched to the lug maker's specification, and can you show the die chart you work from?
• Is hydraulic crimping available above your press range, and is ultrasonic welding an option?
• Do you measure DC resistance of finished leads, and do you have a four-wire micro-ohmmeter for joint resistance? What is recorded, per lot or per piece?
• Can you supply a cross-section photograph and a pull-force result per size on the first article?
• Who supplies the bolt torque figure you build to, and will you torque-stripe and photograph the joints?

A listed cross-section analyser and a set of testers is the right bench for crimp and insulation checks, and a good reason to book a video walkthrough — but it does not cover low-resistance joint measurement, which needs a four-wire micro-ohmmeter. Ask for that instrument by name and to see a recorded result on a lead like yours; the rest of the walkthrough follows an ordinary factory audit checklist.

As a worked example: Huisheng Electronics, a Shenzhen wire harness and cable assembly maker incorporated in 2016 whose registration was checked against the Chinese corporate registry in August 2026, lists energy-storage and charging cabling as one of its six product lines — high-current battery leads and cabinet wiring for energy-storage units and charging-pile modules, with an Energy-Storage Power Cable and a Battery Terminal Lead Set in the published catalog, a range the company's own history dates to a 2022 catalog extension. The supplier-published equipment roster names automatic cut-strip-crimp machines, servo crimping presses and benchtop crimpers, plus a terminal cross-section analyser and tensile, withstand-voltage, insulation and continuity testers. What it does not name is the ceiling: the biggest cross-section crimpable, the press tonnage, whether hydraulic or hexagonal-die tooling is on hand for big lugs, and whether any low-resistance joint measurement is done. Those are the six questions above, worth answering before a purchase order rather than after — as are the stated trade terms (MOQ 1,000 pieces, samples in 7–10 days, volume in 15–25 days), which are supplier figures to confirm at quote time. The adjacent electronics side of the same rack build is covered in the guide to energy-storage BMS manufacturing in China.