Winding vs Stacking: Battery Cell Assembly Methods Compared — Cylindrical, Prismatic, and Pouch - GreenEngGuide - Industrial Engineering & Environmental Compliance Insights

Every lithium battery cell—whether it powers a power tool, an EV, or a grid storage system—starts with the same two materials: anode and cathode sheets with separator in between. But how you combine them into the final cell determines your production speed, energy density, thermal performance, and manufacturing cost.

There are only two ways to do it: winding (rolling the electrode sandwich into a cylinder or flattened spiral) and stacking (layering individual electrode sheets like a deck of cards). The industry has spent 20 years and billions in capital debating which is better. The answer, as usual, is: it depends on what you’re building.

This article compares the two methods across the dimensions that matter in production: speed, quality, cost, format compatibility, and what actually fails on the line.

The Processes in 30 Seconds

Winding

Anode sheet, separator, cathode sheet, and a second separator layer are fed from four unwind reels through tension control rollers to a winding mandrel. The mandrel rotates, winding the layers into a tight spiral. For cylindrical cells (18650, 21700, 4680), the winding is a true cylinder. For prismatic cells, the winding is flattened into a racetrack shape. For pouch cells using winding, the jelly roll is flattened and inserted into the pouch.

Speed: A modern cylindrical winder produces 15-25 cells per minute per spindle. Large-format prismatic winders run at 6-10 PPM.

Stacking

Individual anode and cathode sheets (die-cut or laser-cut from continuous coated foil) are picked up by a robotic arm or vacuum belt and placed alternately with separator sheets onto a stacking table. The separator can be a continuous Z-fold (zigzag folding) sheet or individual sheets.

Speed: A modern stacker produces 2-6 cells per minute for large format. Faster for small consumer cells, but fundamentally slower than winding because it’s a start-stop motion vs continuous rotation.

Comparison Matrix

Where Winding Wins

Speed and Throughput

Winding is fundamentally faster because it’s a continuous rotary process. The electrodes and separator move at constant linear speed through the tension control system onto a rotating mandrel. There are no start-stop motions, no pick-and-place, no waiting for a robot arm to return to home position.

For a 21700 cylindrical cell: winding time is roughly 3-5 seconds per cell per spindle. A 4-spindle winder produces roughly 40-60 cells per minute. A stacking machine for an equivalent-capacity pouch cell produces maybe 4-8 cells per minute.

This speed advantage is why every cylindrical cell ever made is wound. The 18650, 21700, and Tesla’s 4680 are all wound cells. There’s no practical way to stack sheets into a cylinder.

Lower Capital Cost Per Cell

A single-spindle cylindrical winder costs $100,000-200,000 and produces 15-25 PPM. A large-format automatic stacker costs $800,000-1,500,000 and produces 4-6 PPM. The capital cost per cell produced is roughly 5-10× higher for stacking. For consumer electronics where margins are thin and volumes are astronomical, this matters.

Proven Reliability

Winding has been the dominant cell assembly method for 30 years. The failure modes are well understood: tension control, edge misalignment, and tab welding. Maintenance technicians know how to fix these problems. Stacking is newer in mass production, and the robotic handling systems introduce failure modes the battery industry is still learning to manage.

Where Stacking Wins

Energy Density

When you wind electrodes, the curved sections at the ends of the jelly roll create inactive curved areas where the electrode isn’t perfectly flat against the separator. These curved sections don’t contribute to capacity but add volume and weight. In a large-format prismatic cell, this curvature penalty is 2-5% of active material area.

When you stack, every square centimeter of electrode is flat and active. No curvature penalty. This is why essentially every high-energy-density pouch cell uses stacking.

Rate Capability and Thermal Performance

A wound cell has a radial structure—current must flow through the spiral, and heat must conduct outward layer by layer. During high-rate discharge (3C+), the inner layers of the jelly roll heat up faster than the outer layers because the thermal path to the can is longer. This radial thermal gradient limits peak power and accelerates degradation in the hottest inner layers.

A stacked cell has a planar structure. Each electrode layer is one layer away from the cell surface. Current distribution is more uniform. Heat dissipation is more uniform. This matters increasingly as cells get larger—Tesla’s 4680 moved to a tabless design partly to address this radial current collection problem inherent to winding.

No Bending-Induced Stress

When you wind an electrode, the inner layers are compressed and the outer layers are tensioned. The minimum bending radius creates mechanical stress in the coating. The tighter the wind, the higher the stress. NMC cathode materials are more brittle than LFP—the bending-induced microcracking in NMC can contribute to capacity fade over cycle life, particularly at the innermost windings where the radius is smallest.

Stacking imposes no bending stress on the electrodes. Each sheet lays flat with its entire area contributing uniformly to capacity and aging uniformly over cycle life. For NMC and high-nickel cathodes that are inherently more mechanically fragile than LFP, this is a meaningful advantage.

Superior for Large-Format Prismatic and Pouch

As cells get larger (the industry trend is toward bigger cells: 4680, blade cells, 300+ Ah prismatic), the thickness of a wound jelly roll becomes a problem. The more layers you wind, the larger the diameter, and the more pronounced the radial thermal gradient and inner-layer stress become. Stacking scales linearly—add more sheets, the cell gets thicker but each layer is still one layer away from the surface.

The Format Lock-In

Real Production Issues

Winding: Tension Control Is Everything

The number one cause of winding yield loss is tension variation. Too loose, and layers delaminate during electrolyte filling or cycling. Too tight, and the innermost layers are crushed, separator pores collapse, and lithium plating occurs at the compression points.

A good winder has closed-loop tension control with feedback from dancer rollers or load cells at each unwind spindle. Tension variation should stay within ±5% of setpoint. The setpoint itself is material-dependent—copper foil stretches differently than aluminum, NMC coating tolerates less tension than LFP.

The number two cause: edge misalignment. If the anode overhangs the cathode edge by more than 0.5 mm, lithium dendrites grow at the overhang during cycling, eventually causing an internal short. Modern winders use CCD cameras with real-time edge position feedback and automatic alignment correction.

Stacking: Burr Control Is Everything

When electrode sheets are die-cut or laser-cut for stacking, every cut edge is exposed. Winding buries most edges inside the jelly roll, so a small burr on the cut edge of a wound electrode is less likely to cause a short. But in a stacked cell, every burr on every sheet is a potential short site because the sheets are stacked face-to-face with only 12-25 μm of separator between them.

Burr standards for stacked cells are extremely tight: <15 μm burr height for cathode and <10 μm for anode. Achieving this at production speed requires:
– Precision-ground die sets replaced on a strict cycle (typically every 200,000-500,000 cuts)
– For laser cutting: optimized pulse parameters to minimize recast layer and dross
– 100% automated optical inspection (AOI) of cut edges—a single missed burr can cause an internal short that passes formation but fails after 50 cycles in the field

The burr problem is why stacking yields are inherently 2-4% lower than winding in production. Every sheet is inspected. Every burr is a reject. Multiply by 40-100 sheets per cell and the cumulative probability of a defect per cell is nontrivial.

Z-Folding vs Individual Separator Sheets

For pouch cells using stacking, the separator can be applied two ways:

1. Z-folding (continuous separator): A single separator sheet is zigzag-folded between electrode layers as they’re stacked. Advantage: one separator sheet, no alignment of individual separator pieces, faster, mechanically simpler. Limitation: tension must be perfectly uniform across the fold; any wrinkle is a defect that propagates.

2. Individual separator sheets: Pre-cut separator pieces placed between each electrode layer. Advantage: no tension issues; can use different separator materials for anode and cathode facing. Limitation: slower, more robotic picks, electrostatic handling challenges (separator is light, static-prone).

Z-folding dominates in mass production because it’s faster and the tension control problem is easier than coordinating individual separator sheets.

The BYD Blade Cell Example

BYD’s blade cell (used in their LFP battery packs) is perhaps the most famous example of stacking winning on format constraints. The blade cell is roughly 960 mm long × 90 mm tall × 14 mm thick. There is no practical way to wind an electrode sandwich that long and that thin into a jelly roll. The aspect ratio forces stacking.

BYD uses a proprietary stacking process that achieves 0.3 seconds per electrode sheet placement, which is extremely fast for stacking. Combined with the format advantage (blade cells pack into a pack with much higher volumetric efficiency than cylindrical cells), the stacking vs winding economics shift when the cell format itself is impossible to wind.

Summary: How to Choose

For most new production lines in 2026, the decision tree looks like:

“`
Cylindrical cell? → Winding. Period.
Pouch cell? → Stacking. Period.
Prismatic cell:
Under 100 Ah and cost-driven? → Winding.
Over 100 Ah or energy-density-driven? → Stacking.
LFP chemistry? → Either (LFP is forgiving of winding stress).
High-nickel NMC? → Stacking (bending-induced microcracking is a real degradation mode).
“`

Neither method is going away. Winding owns the high-volume cylindrical market. Stacking owns the high-energy pouch market. Prismatic cells are the battleground where the two compete, and stacking is slowly winning as cells get larger and energy density targets get more aggressive.

📖 Related Reading

📊 China Lithium Battery Supply Chain Reports

Equipment supplier intelligence, material pricing, and policy analysis — built from factory-floor experience, not desk research.

View Reports → Work With Me →