Battery Technology

Two batteries both labeled 3000 mAh may have different energy capacities, may not drive the same load, and cannot share the same charging parameters. A battery is not a fixed voltage source: the chemistry determines the operating window, structure and materials affect power capability, temperature and aging alter the currently available capacity, and protection circuits only handle a portion of the boundaries.

First, Separate the Three Classification Dimensions

Li-ion refers to the family of lithium-ion batteries and does not equal LCO (Lithium Cobalt Oxide). LCO, NMC, and LFP describe different cathode chemistries; "Li-Po" in consumer products often refers to lithium-ion batteries using polymer-related electrolytes or pouch structures, and you cannot infer the cathode formulation, charging upper limit, or discharge C-rate from this alone. Cylindrical, prismatic, and pouch are packaging dimensions: a pouch cell does not automatically imply the highest energy density, nor does it mean the shape is unconstrained by manufacturing and expansion limits.

The nominal voltages in the table are for classification, not for charging settings. NiMH chargers determine termination based on voltage changes, temperature, and time conditions as designed; lead-acid float voltage is also not a universal "fully charged open-circuit voltage". Panasonic Ni-MH Battery Handbook, Yuasa: VRLA Charging Conditions

Names like 18650, 21700 primarily indicate cylindrical cell size series and do not guarantee capacity, chemistry, or rated current. Actual dimensions may change after adding a protection board, raised terminals, or outer casing; flat tops cannot be reliably judged as "no protection". 14500 cells are similar in shape to AA batteries but may have incompatible voltages; replacement requires checking device specifications.

Ah is Charge, Wh is Energy

Capacity Q is the charge extracted under specified temperature, current, and cutoff conditions, where 1 Ah = 3600 C. Energy requires integration over the discharge process: E = ∫V(t)I(t)dt. Multiplying nominal voltage by Ah to get Wh is a practical approximation and does not claim that voltage remains constant throughout.

A 3.6 V, 3 Ah cell is approximately 10.8 Wh. Assuming the conversion efficiency over the entire operating range is approximately 90%, and the system average consumption is 2 W, the energy budget gives 10.8 × 0.9 / 2 = 4.86 hours. In reality, you must also consider remaining energy at cutoff, standby power consumption, temperature, and load variations. Dividing 3000 mAh directly by the current on the regulated 5 V side mixes charge from different voltage sides and leads to errors.

C-rate normalizes current by rated capacity: for a 3 Ah cell, 1C is 3 A, and 0.5C is 1.5 A. It is not a default allowed value; charging upper limits, continuous discharge limits, and short-term pulse limits must be viewed separately regarding duration, temperature, cutoff, and heat dissipation conditions. The appearance of a specific test current on a discharge curve does not mean that current is allowed for long-term use in any environment.

Rated capacity answers "how much charge can be extracted under specified conditions"; power capability answers "whether it can provide the required current in the current state." This explains why high-capacity models may not be suitable for large pulse loads: they may trigger voltage drop or temperature rise limits before discharging the full nominal capacity.

Why Loaded Voltage Drops and Recovers After Shutdown

Use a simplified model to understand the discharge terminal voltage: V_terminal ≈ OCV(SOC, Temperature) − I × R_ohmic − V_polarization. OCV is the open-circuit voltage approached after sufficient rest; ohmic voltage drop responds almost immediately to current, while polarization has a process of building up and recovering over time. Actual cells also exhibit hysteresis, so a static resistance model cannot be treated as a complete chemical model.

Assume at a certain moment OCV is 3.6 V, short-term equivalent resistance is 80 mΩ, and the load changes from very small to 5 A; I×R alone causes a voltage drop of about 0.4 V; considering connection resistance and polarization, the terminal voltage may hit the system undervoltage threshold. The recovery after load removal indicates that the voltage drop has decreased, not that SOC has automatically increased.

Therefore, "low battery shutdown" may be governed by energy depletion, insufficient power, or protection actions. Observing rest voltage, minimum loaded voltage, temperature, and current changes is necessary to distinguish the cause. When estimating internal resistance via ΔV/ΔI, record the pulse length and sampling moment; millisecond and second-level results contain different polarization components and cannot be compared without conditions to 1 kHz AC internal resistance.

Voltage rebounds—does charge return?

Move across 15 s: current becomes zero and ohmic drop disappears, while polarization remains. Later terminal voltage recovers but discharged charge stays fixed. No charge is restored; voltage components change.

SOC is an Estimated State, Not Another Scale on a Voltmeter

SOC usually represents the ratio of remaining charge relative to the current full-charge capacity baseline; the percentage displayed by manufacturers may also include reserved areas and smoothing strategies. Nominal 3.7 V does not mean "37% remaining," nor is 4.0 V a unified 80% for all lithium-ion batteries. Charge/discharge direction, load, rest time, temperature, and aging all affect the mapping; the flat plateau of LFP makes it harder to convert small voltage errors in the middle section into reliable percentages.

Coulomb counters update charge by integrating current. If discharge is defined as positive, calculating with hours and amperes, Q_remaining(t) = Q_initial − ∫I dt; but initial state, charging efficiency, unmeasured leakage current, and capacity baseline all introduce errors. For example, a constant bias of 1 mA accumulates 24 mAh over a day, which is 0.8 percentage points for a 3 Ah battery; infinite integration without correction is not possible.

Practical fuel gauges combine current integration with voltage, temperature, chemical models, and capacity learning. An incorrectly configured cell model may still report incorrect SOC even if voltage measurement is accurate; fuel gauges cannot replace hardware overcharge/overdischarge protection. TI BQ27Z561-R1: Current, Voltage, Temperature Measurement and Impedance Track Fuel Gauge

Series/Parallel: Calculate Ideal Model First, Then Check Limitations

Four consistent 3.6 V, 3 Ah cells, first connected in pairs in parallel, then the two groups in series, result in 2S2P: nominal 7.2 V, 6 Ah, total energy 43.2 Wh. If each cell's equivalent internal resistance is 30 mΩ, ignoring wires and protection devices, each parallel group is 15 mΩ, and the two groups in series remain 30 mΩ. When the pack outputs 10 A, each cell ideally shares 5 A, the pack voltage drop is 0.3 V, and internal loss is 3 W. This is a topology calculation, not a conclusion that allows outputting 10 A for selection.

Normal total series voltage does not mean each string is normal. For example, in a two-string system measuring 8.4 V, it could still be 4.3 V + 4.1 V; you cannot judge that no string has exceeded limits based solely on total voltage. Balancing can reduce state differences, but it cannot repair capacity degradation or allow a string to exceed its specifications; different systems have different balancing start intervals and methods.

Parallel makes terminal voltages converge, but does not guarantee that SOC, health state, or current are immediately identical. If two sources differ by 0.2 V and the total resistance of the connection loop is 40 mΩ, the simplified estimate of the initial balancing current is 5 A; therefore, there is no rule that "it is safe as long as the difference is less than 0.1 V" independent of loop impedance. These models are used to understand battery pack design; actual grouping requires manufacturer-specified matching, connection, insulation, and fault isolation schemes.

Lifespan Requires Looking at Calendar Aging and Cycle Aging Simultaneously

Batteries undergo side reactions over time even without cycling; temperature and storage state affect calendar aging. Cycling adds effects such as electrode stress and interface changes, with results varying by chemistry, SOC window, current, and temperature. You cannot write "20% loss in one year" from a specific set of experiments as a universal law for all batteries, nor can you rank damage from high temperature, fast charging, and deep discharge as always applicable.

Reading "2000 cycles" requires knowing at least the definition of a cycle, depth of discharge, charge/discharge C-rate, temperature, and end criteria. When counting discharge throughput based on initial rated capacity, discharging 50% of rated capacity twice is approximately one equivalent full cycle (EFC); this only normalizes throughput and does not mean it has the same aging result as one deep cycle.

Capacity retention rate is not the entirety of SOH: increased internal resistance may prevent high-power devices from starting first, while low-power devices may still be usable for a long time. Cross-product comparisons must place capacity, power, temperature, and termination criteria under the same set of conditions. Charging temperature limits should especially be executed according to cell specifications; "can discharge at low temperature" cannot be extrapolated to "can also charge at the same low temperature."

What Do Cells, Protection, and Device Strategies Respectively Handle?

Cell datasheets specify allowed windows; chargers regulate the daily charging process; protection devices handle anomalies such as overvoltage, undervoltage, and overcurrent; the main controller formulates strategies based on temperature and operational needs. Bare cells do not necessarily come with protection boards, and having a protection board does not imply complete temperature monitoring, balancing, or fuel gauge estimation. Protection action thresholds are not normal charging target values.

LFP is still a lithium-ion system storing large amounts of energy and cannot be described as "almost never catching fire." Significant swelling, abnormal heat, odor, or damage should lead to stopping use and contacting the manufacturer or qualified service provider; smoke or fire should lead to evacuation and contacting fire services. Do not confuse lithium-ion batteries with metal lithium fires. NFPA Lithium-ion Battery Safety Tips

Recalculate Energy and Series/Parallel Accounts

from math import isclose
v, ah, r = 3.6, 3.0, 0.030
s, p = 2, 2
pack_v, pack_ah, pack_r = s*v, p*ah, s*r/p
assert isclose(pack_v*pack_ah, 43.2)
assert isclose(pack_r, 0.030)
assert isclose(10*pack_r, 0.3)
assert isclose(10**2*pack_r, 3.0)
assert isclose(v*ah*0.9/2, 4.86)
assert isclose(0.001*24/ah*100, 0.8)
assert isclose(0.2/0.040, 5.0)
assert isclose(5*0.080, 0.4)
print('Energy, internal resistance, voltage drop, balancing current, and integration bias calculations passed')

The example only verifies the model and units; it does not perform charge/discharge, assembly, or lifespan testing on physical cells.

Continue Reading

• Charging Protocols and Fast Charging: Entry protocols, CC/CV, conversion, and system power distribution.
• Protection Circuits: Protection thresholds, action times, and abnormal energy paths.