314Ah & 587Ah Prismatic Energy Storage Cell Busbars: Integrated Thermal Runaway Venting & Overcurrent Capacity

A 314Ah prismatic cell delivers 314 A at 1C; a 587Ah cell delivers 587 A. At those levels, the interconnect stops behaving like a connector and starts behaving like a thermal and electromagnetic component. The production design targets that hold at 1500 V DC: conductor cross-section ≥ 400 mm² for 600 A continuous (1.5 A/mm² on C1100, 100 % IACS), busbar body ΔT < 30 K at 45 °C ambient, joint resistance < 10 µΩ per laser-welded interface, and ≥ 15 mm vertical clearance above the cell safety valve so vented electrolyte, carbonised copper and aluminium vapour cannot bridge two potential planes. Every 0.5 mm of conductor you remove to save copper mass is paid back as roughly 8–10 K of terminal temperature rise and a measurable drop in cycle life.

 

 

 

Capacity growth in ESS has not arrived through higher voltage - it arrived through bigger electrodes and thicker stacks. The electrical consequence is a step change in continuous current, and the mechanical consequence is a step change in terminal loading and stack growth.

 

Parameter	280 Ah (legacy)	314 Ah	587 Ah
1C continuous current	280 A	314 A	587 A
Typical envelope (W × D × H)	173 × 72 × 204 mm	174 × 72 × 207 mm	~220 × 90 × 270 mm (indicative)
Terminal pad/weld window	25 × 30 mm	30 × 38 mm	38 × 50 mm
Busbar section at 1.5 A/mm²	190 mm²	210 mm²	400 mm²
Thickness swell over 8,000 cycles	2–3 %	3–4 %	4–6 %
Vent jet temperature at valve	600–800 °C	700–900 °C	700–950 °C

 

Three failure modes follow directly from this table. First, the terminal weld window grows, so a weld designed for a 25 × 30 mm pad under-fills a 38 × 50 mm pad and leaves current crowding at the pad edge - the local resistance sits where the thermocouple is attached, and the cell BMS reports a terminal fault that is actually a weld-geometry fault. Second, thicker electrodes swell more, so a rigid busbar that survived 3,000 cycles at 280 Ah shears its own weld at 6,000 cycles at 314 Ah. Third, at 587 Ah the vent mass flow roughly doubles, and a busbar routed on the valve centreline becomes a plasma bridge during propagation testing under UL 9540A or GB/T 36276.

 

Apollo re-runs DFM on the cell maker's terminal drawing before quoting, because pad flatness (typically ≤ 0.15 mm), pad plating (Ni 2–5 µm, or bare Al for aluminium terminals), and the cell maker's stated terminal force limit (commonly ≤ 300 N and ≤ 2 N·m) all change the process window.

 

 

Current density is the only variable that simultaneously sets temperature rise, voltage drop, and copper cost. Design values below are for 600 A continuous, 200 mm unsupported span, 45 °C ambient, two joints at 10 µΩ, and an effective heat transfer coefficient of 12 W/m²K.

 

Design current density	Section for 600 A	Example profile	Body loss	ΔT incl. joints	Verdict
2.5 A/mm²	240 mm²	4 × 60 mm	5.2 W	40 K	Rejected at 1C duty
2.0 A/mm²	300 mm²	5 × 60 mm	4.1 W	36 K	Marginal, forced air only
1.5 A/mm²	400 mm²	5 × 80 mm	3.1 W	25 K	Production target
1.0 A/mm²	600 mm²	5 × 120 mm	2.1 W	16 K	Sealed enclosure, no airflow

 

Work the resistance directly, and the numbers are unforgiving. C1100 resistivity is 0.017241 Ω·mm²/m at 20 °C. A 400 mm², 200 mm long bar measures 8.62 µΩ; at 600 A that is 3.1 W. Two joints at 10 µΩ add 7.2 W - so the joints, not the conductor, dominate the thermal budget. Copper's temperature coefficient (α = 0.00393 K⁻¹) raises resistance by 33 % at 105 °C, which is why the 25 K figure must be validated at end-of-life contact condition rather than at T0.

 

Joint method	Joint resistance	Power at 600 A	Note
Laser weld Cu–Cu, 2.5 mm penetration	3–6 µΩ	1.1–2.2 W	No filler, no aged interface
Molecular diffusion welding Cu–Al	8–12 µΩ	2.9–4.3 W	IMC controlled ≤ 3 µm
Resistance silver brazing (BAg-8 / BCuP-5)	5–10 µΩ	1.8–3.6 W	Lap ≥ 3t
In-die riveted Ag contact rivet	15–25 µΩ	5.4–9.0 W	Field-serviceable
Bolted M8, Sn-plated, 12 N·m	20–40 µΩ	7.2–14.4 W	Torque relaxes 10–15 % after 500 h at 125 °C

 

Material selection follows the same arithmetic:

 

Material	Conductivity	Tensile (half-hard / H04)	Softening	Typical use
C1100 (Cu-ETP)	100 % IACS minimum (101 % typical)	300–380 MPa	180–220 °C	Main conductor, 3–5 mm
C2680 (CuZn30 brass)	28 % IACS	450–550 MPa	Not age-softened	Inserts, sensor brackets, stamped clips, nuts
C19400 (Cu-Fe-P)	60–65 % IACS	400–480 MPa	> 400 °C	Fuse links, spring terminals
A1060 aluminium	61 % IACS	80–120 MPa	-	Aluminium cell terminals, weight-critical DC links

 

C2680 carries roughly a quarter of copper's conductivity, so it is never used as a current path across a 400 mm² section. It appears where mechanical strength or thread wear matters - pressed-in nut plates, Hall-sensor brackets, and stamped retaining clips at 28 % IACS on non-load paths.

 

Request Free DFM Evaluation & Quote

 

 

During thermal runaway, a prismatic cell vents through the top safety valve at 700–950 °C with jet velocities above 100 m/s. The plume carries electrolyte aerosol, carbon black, and metal vapour. Deposited on an insulation surface, that residue cuts surface resistivity by orders of magnitude; deposited across two busbars at a 1500 V DC potential difference, it initiates an arc that ignites the next cell. Propagation test data under UL 9540A and IEC 62619 show most pack-level propagation events are electrical, not purely thermal.

 

Design rules applied at Apollo for 314Ah and 587Ah layouts:

Standoff:​ ≥ 15 mm vertical clearance between the busbar underside and the valve plane, with no conductor segment crossing the valve axis within a 30 mm radius.

Offset routing:​ dog-leg the busbar laterally so the conductor runs over the cell shoulder, not the valve.

Local shielding:​ mica plate (0.5–1.0 mm) or ceramic-filled silicone-coated glass cloth over any segment that cannot be moved, rated for 800 °C short-term exposure.

Edge conditioning:​ burr height ≤ 0.05 mm and edge radius R ≥ 1.0 mm to suppress field concentration and corona at 1500 V DC.

Sacrificial barrier:​ the insulation system must survive the plume, not just the operating voltage - epoxy powder coating at 150–400 µm alone is not a vent barrier.

Clearance and creepage are then set from the system voltage with a derating factor for electrolyte contamination:

 

System voltage	Clearance, IEC 60664-1 base	Applied creepage (1.5× derate)	Pack HI-POT verification
1000 V DC	5.0 mm	8.0 mm	3 kV AC / 60 s, ≤ 5 mA
1500 V DC	8.0 mm	12.0 mm	4 kV AC / 60 s, ≤ 5 mA
2000 V DC	10.0 mm	16.0 mm	5 kV AC / 60 s, ≤ 5 mA

 

Where the pack interfaces the PCS or inverter, a laminated busbar with stray inductance < 10 nH keeps switching overshoot bounded: at SiC di/dt of 10 A/ns, 10 nH produces 100 V overshoot, about 6.7 % on a 1500 V rail. Solid single-layer bars cannot reach that figure; the laminated stack with 100–200 µm dielectric between opposing Cu layers can.

 

 

 

Thick-section copper is hostile to conduction-mode laser welding. At 1070 nm, solid copper absorbs roughly 5 % of incident energy; absorption rises sharply once the melt pool forms, which is why keyhole welding on 3–5 mm C1100 needs power density above 10⁶ W/cm² and a wobble head (0.5–1.5 mm amplitude, 100–300 Hz) to stabilise the keyhole and let hydrogen escape. Without wobble, porosity exceeds 8 % of weld area, and the joint fails ISO 13919-1 Level B.

 

Cu–Al joints carry a second constraint: intermetallic compounds CuAl₂ and Cu₉Al₄ are brittle and 5–10× more resistive than either parent metal. Holding the IMC layer at ≤ 3–5 µm is the difference between a joint with 60 MPa lap shear and one that cracks during module stacking.

 

Process	Thickness range	Heat input / HAZ	Joint strength	Cycle time	Reference
Fibre laser welding, 4–6 kW, wobble	0.5–5.0 mm Cu; 1.0–3.0 mm Al	HAZ 0.5–1.5 mm	≥ 150 MPa Cu–Cu; ≥ 60 MPa Cu–Al	0.5–3 s/joint	ISO 13919-1 Level B, ISO 6520-1
Resistance silver brazing, BAg-8 (780 °C) or BCuP-5 (645–800 °C, self-fluxing on Cu)	1.0–6.0 mm	Whole part annealed	≥ 200 MPa shear	8–20 s	AWS A5.8, ISO 17672
Molecular diffusion welding Cu–Al	0.2–3.0 mm	400–500 °C, 20–40 MPa, no melting	IMC ≤ 3 µm, lap shear ≥ 70 MPa	60–180 s	Internal weld spec
In-die riveting, Ag/Cu solid rivet	1.0–4.0 mm	Cold	Push-out ≥ 500 N	40–80 spm in press	Press-force monitored
Ultrasonic metal welding	0.2–1.5 mm foil stacks	Cold	Foil peel per cell spec	0.3–1.0 s	Not applicable at 4 mm

 

Resistance silver brazing has one consequence that is routinely missed in RFQs: the whole part reaches brazing temperature, so cold-worked C1100 anneals. Yield strength drops from roughly 300 MPa (H04) to 60–80 MPa (annealed), and any bend or fold formed before brazing relaxes. Either form after brazing, or specify a stiffening rib and accept the softened condition.

 

Fault current is the last check on the joint, not on the conductor. Adiabatic heating from a cleared fault is negligible - a fuse with I²t of 8.0 × 10⁴ A²s (gR per IEC 60269) dissipates under 1 J in a 400 mm² bar. The real limits are electromagnetic: two parallel bars 30 mm apart carrying 10 kA generate about 133 N of repulsion per 200 mm span, rising to roughly 533 N at 20 kA. Support spacing and fold geometry, not section area, decide whether the busbar survives a short circuit.

 

Download Busbar Design Specification

 

A 314Ah cell grows 3–4 % in thickness across its cycle life. On a 174 mm-wide cell, that is 5–7 mm of stack movement; across a 20-cell module the end-plate displacement reaches double digits, and every cell-to-cell busbar sees the differential. A rigid bar converts that displacement into terminal shear.

 

Fold type	Axial displacement absorbed	Reaction force at max travel	Bend radius	Application
U-fold, single	± 1.5 mm	35 N	R ≥ 1.5t	Cell-to-cell within a module
Ω-fold, double	± 3.0 mm	55 N	R ≥ 1.5t	Module-to-module, end plates
S-offset / dog-leg	± 0.8 mm	90 N	R ≥ 2t	Vent-clearance routing
Braided copper flexible	± 6.0 mm	20 N	n/a	HV junction, service loops (resistance +30 % vs solid)

 

Fold design details that decide fatigue life: use soft-temper C1100 (R220) in the fold zone so the strain distributes instead of concentrating; orient the rolling grain perpendicular to the bend line; hold bend radius ≥ 1.5t to avoid grain-boundary cracking; and verify grain size to ASTM E112 after any annealing step. The target is 3,000 displacement cycles with resistance drift ≤ 10 %, then vibration per UN 38.3 T3 and SAE J2380 random profiles (10–2000 Hz, three axes), with the busbar first natural frequency either above 100 Hz or deliberately decoupled from the cell stack mode.

 

 

Method	Thickness	Dielectric strength	Continuous rating	Flammability	Dominant failure mode
Epoxy powder coating, electrostatic/fluidised bed	150–400 µm	> 15 kV/mm (typ. 20–30 kV/mm)	130 °C RTI	UL 94 V-0	Pinhole at a burr, edge chipping
PET heat shrink sleeve	200–500 µm	15–20 kV/mm	125 °C	UL 94 V-0 / VW-1	Sliding, trapped air pockets
PVC dip coating	300–800 µm	10–15 kV/mm	90 °C	UL 94 V-0	Plasticiser migration, cracking
PA66 / PPS overmolding	0.8–2.0 mm	15–25 kV/mm	140 °C	UL 94 V-0	Tooling cost, sink marks
Mica + ceramic silicone shield (vent zone)	0.5–1.5 mm	> 20 kV/mm	800 °C short-term	Non-combustible	Mechanical damage

 

At 300 µm, an epoxy coating rated at 20 kV/mm gives 6 kV of dielectric headroom against a 4 kV HI-POT - a 1.5× margin that collapses to 1.1× at 150 µm, which is why Apollo specifies 150 µm as a floor and 400 µm as the vent-zone target. Partial discharge inception voltage is checked on sample parts; EN 45545-2 (R22/R23) applies where the pack enters rail-adjacent installations, and ISO 14001 governs the coating line's VOC and powder-recovery handling.

 

 

Stamping 4 mm C1100 in a progressive die is a tonnage and burr-control exercise, not a forming exercise. Blanking force for a 5 × 80 mm profile at 200 MPa shear strength is about 170 kN before stripping and drawing loads, so production runs sit on 250–400 t servo presses at 40–60 spm with 8–14 stations, in-die tapping and in-die riveting where the design calls for contact rivets.

 

Characteristic	Method/gauge	Frequency	Acceptance
Terminal pitch, hole diameter, slot width	CMM, 2.4 µm + L/300 µm	First/last + 5 pcs/lot	± 0.01 mm, Cpk ≥ 1.67
Burr height	Digital microscope / CCD vision	100 %	≤ 0.05 mm
Weld penetration and porosity	Metallographic cross-section	1/500 + every setup	≥ 1.5 mm penetration on 4 mm stack, ≤ 5 % porosity
Joint resistance	4-wire micro-ohmmeter at 100 A	100 %	≤ 10 µΩ
Coating thickness	Eddy-current gauge, 5 points/part	AQL 0.65 per ISO 2859-1	150–400 µm, Cpk ≥ 1.33
Dielectric	AC HI-POT	100 %	4 kV AC / 60 s, ≤ 5 mA
Burst pressure (liquid-cooled variants)	Hydrostatic rig	1/500	8 bar / 60 s, no rupture
Thermal cycling	Chamber −40 / +85 °C	3 pcs/lot	300 cycles, ΔR ≤ 10 %
Salt spray	ISO 9227 NSS	Quarterly	480 h, no base-metal corrosion
Material verification	Spectrometer + mill certificate	Per heat lot	Cu ≥ 99.90 %, ≥ 100 % IACS

 

Documentation ships under IATF 16949 with PPAP Level 3, IMDS entry, RoHS and REACH declarations, and full heat-lot traceability from mill certificate to finished pallet.

 

 

 

What is the T1 sample lead time for a custom 314Ah prismatic cell busbar, and is PPAP Level 3 included?
T1 stamped samples ship 15–20 working days after DFM sign-off on soft tooling; a hard progressive die takes 35–45 days. PPAP Level 3 ships with the first production lot - dimensional reports, CMM data, weld cross-sections, HI-POT and salt spray records.

How do you verify coating thickness and dielectric strength on epoxy powder-coated ESS busbars?
Thickness is measured at five points per part by eddy-current gauge: 150–400 µm, Cpk ≥ 1.33. Every part passes 4 kV AC HI-POT for 60 s, leakage under 5 mA, plus per-lot 6 kV DC surge and 480 h ISO 9227 salt spray.

What tooling life can we expect on a 4 mm C1100 progressive die, and who owns the tool?
A 4 mm C1100 progressive die with powder-metallurgy inserts is rated for 1.5–2 million strokes, with regrinds every 150,000. Tooling ownership transfers to you on final payment; Apollo stores and maintains it free for five years.