Battery Pack Configuration

How Does Battery Pack Configuration Work?

A battery pack is built by connecting individual cells in series (S) to increase voltage and in parallel (P) to increase capacity. The notation nSnP describes the configuration — for example, 96S3P means 96 cells in series with 3 parallel strings, totaling 288 cells.

Series connections add voltages while keeping capacity the same: 4 cells at 3.7V in series produce 14.8V at the same Ah rating. Parallel connections add capacities while keeping voltage the same: 3 cells of 3 Ah in parallel provide 9 Ah at the original voltage.

Cell matching is critical — mismatched cells in series can lead to overcharging/over-discharging of individual cells, causing safety hazards and accelerated degradation. Production cells are sorted by capacity and impedance before pack assembly.

The choice between high-voltage (more series) and high-capacity (more parallel) topologies has cascading effects on BMS complexity, contactor ratings, and fault tolerance. Higher series counts require more cell monitoring channels and active/passive balancing circuits, while higher parallel counts increase the challenge of detecting a single cell failure within a parallel group.

Formula: Series Cells (S) = ⌈Target Voltage / Cell Voltage⌉ Parallel Cells (P) = ⌈Target Capacity / Cell Capacity⌉ Total Cells = S × P Pack Voltage = S × Cell Voltage Pack Capacity = P × Cell Capacity

Example Calculation

Design a 48V 100Ah pack using 3.2V 50Ah LFP cells. Series cells = ⌈48/3.2⌉ = 15S. Parallel cells = ⌈100/50⌉ = 2P. Total = 15 × 2 = 30 cells. Actual voltage = 15 × 3.2 = 48.0V. Actual capacity = 2 × 50 = 100 Ah. Total energy = 4.8 kWh.

When to Use This Calculator

• Designing a new battery pack to meet specific voltage and capacity requirements from a motor controller or inverter specification
• Evaluating different cell formats (18650, 21700, prismatic, pouch) to compare total cell count and pack size for the same energy target
• Estimating bill-of-materials cost by calculating total cell count and then multiplying by per-cell pricing
• Planning BMS channel count and balancing architecture based on the series and parallel configuration

Common Mistakes to Avoid

• Using nominal voltage when maximum voltage determines BMS overvoltage thresholds — a 15S LFP pack has 48V nominal but 54.75V at full charge (15 × 3.65V), which must be within inverter input range
• Forgetting that actual capacity exceeds target when cell capacity does not divide evenly — ceiling function means you may get more capacity than specified, which affects charger settings
• Neglecting cell-to-cell impedance matching for parallel groups — cells with different impedances share current unequally, causing hot spots and premature degradation
• Ignoring the voltage range — a 15S LFP pack ranges from 37.5V (empty at 2.5V/cell) to 54.75V (full at 3.65V/cell); the connected load must tolerate this entire range

Frequently Asked Questions

What happens if cells in a pack are mismatched?

Mismatched cells cause the weakest cell to be over-stressed during both charging and discharging. In series, the lowest capacity cell reaches empty first (causing over-discharge) and full first (causing overcharge). Cell balancing circuits in the BMS mitigate this but cannot fully compensate for large mismatches.

How do I choose between more series or more parallel cells?

Higher series count (voltage) reduces current for the same power, decreasing I²R losses and enabling thinner wiring. Higher parallel count improves redundancy and thermal distribution. For high-power applications, prefer higher voltage; for high-energy applications, optimize parallel count for capacity.

What is the difference between nSnP and nPnS topology?

In nSnP (cells in series first, then parallel strings), each string is an independent series chain and fusing individual strings is straightforward. In nPnS (cells in parallel first, then series), parallel groups are formed first and then stacked in series. nPnS allows cells within a parallel group to self-balance but makes individual cell fault detection harder. Most EV packs use nSnP or hybrid topologies.