A useful battery datasheet is more than a list of impressive numbers. It describes what a particular cell can deliver, how that performance was measured, and the electrical, mechanical and environmental limits within which the result is meaningful. Reading it well means keeping those three things together: the value, its conditions and the identity of the product.
For a buyer comparing several manufacturers, the objective is not to find the largest capacity or current figure. It is to establish which candidate meets the project requirements and which questions still prevent a reliable quotation or sample evaluation.
1. Start with the product, not the headline capacity
Before comparing performance, record the manufacturer, full model designation, document revision and issue date. Check whether the document is a short product leaflet, a detailed specification or a report for particular samples. These documents have different purposes. A catalogue reference used by a distributor is useful for communication, but it is not automatically the original manufacturer's model number.
Then compare the document with the offered product. Do the capacity, dimensions, terminal arrangement and photographs identify the same item? Has the supplier confirmed the revision intended for samples and later orders? A similar sales name is not enough to resolve a mismatch.
A practical document-control record has four columns: offered product; applicable specification; outstanding questions; agreed acceptance basis. It keeps an attractive brochure from silently becoming an order specification. This is a procurement method, not a demand that every commercial discussion begin with a complete technical file.
2. Distinguish typical, minimum and rated capacity
Capacity in ampere-hours measures delivered electric charge. A capacity result depends on how the cell was charged, its temperature, discharge current and discharge endpoint. Different test methods can therefore produce different results without either measurement being fraudulent. Battery characterization manuals specify such conditions rather than treating Ah as a condition-free constant. S02
The label beside the number matters. A typical value describes representative performance under the manufacturer's stated basis. A minimum value may define a lower limit under specified conditions. A rated value is a declared rating; its exact relationship to measured capacity must be read in the applicable document. Do not assume these terms are interchangeable across suppliers.
For a concrete external reading example, Molicel's public P42A datasheet, version 1.7, lists 4,200 mAh typical and 4,000 mAh minimum capacity. It also separates 3.6 V nominal voltage from 4.2 V charging voltage. The point is the distinction between fields, not a recommendation or a CELL & PACK inventory claim. S03
When comparing two candidates, write the capacity method beside each value. If one is measured at a lower current or to a lower cutoff voltage, mark the comparison as incomplete until a common operating basis is available. Do not replace a supplier's declared value with your own preferred arithmetic.
3. Read the whole voltage window
A cell does not remain at its nominal voltage throughout discharge. Nominal voltage is a convenient rating for energy arithmetic and system description. Charge cutoff, discharge cutoff and the voltage under load answer different questions.
For procurement, ask what voltage range the equipment accepts, what charging profile is required, and where the pack's normal operating limits will be set. The BMS protection thresholds are not necessarily the desired routine operating endpoints. Protection is a final boundary, not a normal strategy for extracting every possible watt-hour. Modern monitor/protector devices illustrate that cell voltage, current and temperature protections are separate functions requiring configuration. S09
Illustrative arithmetic, not a pack design: twelve identical cells rated at 3.65 V give a nominal 43.8 V. If their specified upper voltage is 4.20 V, the corresponding series voltage is 50.4 V. An equipment label saying “44 V” does not, by itself, prove that the equipment accepts 50.4 V or the loaded voltage near the discharge endpoint.
This is why substituting a cell on the basis of nominal voltage alone can fail even when its dimensions and Ah rating look suitable.
4. Separate charging, continuous discharge and pulse discharge
Read the direction of current first. A 3C discharge reference does not establish a 3C charge limit. Charging may have a constant-current phase, a constant-voltage phase, termination-current conditions and temperature-dependent restrictions. The maximum setting of a charger IC is also not a permission to charge any attached cell at that setting. S30
For discharge, distinguish a permitted operating rating from a test condition and a short-duration result. A graph produced at a certain current demonstrates that test, not every repetition pattern at that current.
A meaningful pulse description includes duration, starting state of charge, temperature, minimum voltage and recovery or repetition. “100 A peak” is missing most of the information needed to compare it with a real load. Similarly, a continuous value should be read with the specified temperature or thermal cutoff and any cooling conditions.
For a 20 Ah reference capacity, 1C corresponds to 20 A and 3C to 60 A. That conversion is correct arithmetic; it does not establish that an unspecified 20 Ah cell permits either duty. The C-rate guide explains how to translate an actual load into a useful request.
5. Understand what an internal-resistance number represents
“Internal resistance” is not one universal measurement. AC impedance is measured with a varying signal at a specified frequency. A DC pulse method estimates a response from a current step and voltage change over a defined interval. Temperature, state of charge, rest time and connections affect comparisons. S04
The same public P42A document lists typical AC impedance of 10 mΩ at 1 kHz and DC impedance of 16 mΩ for a stated 10 A/1 s method. They are not conflicting quality grades: they are different measurements. These external figures must not be copied into another model's specification. S03
In a simplified pulse calculation, an increase of 50 A accompanied by a 0.25 V decrease corresponds to 5 mΩ for that observation interval:
R_apparent = |ΔV / ΔI| = 0.25 / 50 = 0.005 Ω
This assumed result is not automatically equivalent to a 1 kHz AC measurement or a long-duration resistance. A supplier comparison should preserve the measurement method instead of choosing whichever number looks smaller.
6. Read curves as evidence, not decoration
A discharge curve is useful only after its axes and conditions are understood. Check whether the horizontal axis is time, discharged Ah, state of charge or energy; confirm the current, temperature, charge preparation and cutoff. A curve of voltage against discharged Ah contains information about both capacity and energy: energy corresponds to the area under that voltage-versus-charge relationship.
An earlier endpoint at a higher load can arise because voltage falls to the test cutoff sooner. It does not necessarily mean that the same amount of active material permanently disappeared. A temperature curve may likewise show a performance limitation under that test rather than permanent aging. Determining recoverable performance versus damage requires the appropriate follow-up test, not visual guesswork.
Manufacturer curve sets, such as the separate charge, rate-discharge, temperature and cycling charts in the P42A reference, show why the caption matters as much as the line. Read the original document rather than extracting an unlabelled curve image. S03
When data are missing, request the curve most relevant to the project. A buyer needing a repeated peak near the end of discharge gains little from another low-current room-temperature capacity graph.
7. Treat cycle life as a test record
A cycle claim should identify the charge and discharge methods, temperature, voltage or SOC window, depth of discharge, rest periods, measurement method and end-of-test criterion. “2,000 cycles” is not a complete specification.
Also distinguish retained capacity from retained power. A cell may still store adequate charge but experience sufficient resistance growth that the equipment reaches a low-voltage limit during a peak. Public aging datasets measure both capacity and DC internal resistance because one cannot fully describe the other. S14
For example, 80% retention can mean 80% of initial measured capacity or 80% of another declared reference. Those denominators can change the interpretation. Ask which one the document uses. Do not convert a laboratory cycle count directly into a warranty in years; the lifetime guide explains the missing variables.
8. Check dimensions, mass and operating modes
Dimensions need an agreed boundary. Does the drawing include tabs or terminals? Are values maximum, nominal or tolerance-banded? For a soft-pack cell, what measurement state and mechanical condition are specified? The cell-body rectangle alone is not the complete installation envelope.
Likewise, a mass value can be typical or a maximum. When calculating Wh/kg, state whether you used a typical, maximum or measured mass. A maximum mass and typical energy are not necessarily measurements from one sample.
Keep charging, discharging and storage temperature ranges separate. An allowed storage range does not authorize operation throughout that range, and permitted discharge at a cold temperature does not authorize charging there. Low-temperature charging can involve lithium-plating mechanisms that differ from a temporary discharge-capacity reduction. S15
Finally, operating limits do not mean equal performance everywhere within them. A cell may be permitted to operate at a temperature while delivering less energy or power than at the reference test condition.
9. Turn two datasheets into a procurement comparison
The following is a proposed comparison worksheet, not a set of universal pass/fail thresholds.
Scroll horizontally to read every column.
| Review item | Record for each candidate | What resolves a mismatch? |
|---|---|---|
| Product identity | Manufacturer, full model, offered revision | Supplier confirmation tied to the quotation |
| Capacity and energy | Value, typical/minimum status, test basis | Comparable discharge conditions |
| Voltage | Nominal, charge limit, discharge limit | Equipment and charger review |
| Load | Continuous and pulse conditions | Duty-profile assessment and relevant tests |
| Resistance | Method, SOC, temperature, interval | Like-for-like measurement |
| Lifetime | Protocol, retention reference, tested sample scope | Application-relevant evidence |
| Mechanical fit | Drawing, tolerances, terminals, installation conditions | Pack-level engineering review |
| Documents | Exact product coverage and sharing status | Matched specification and applicable evidence |
| Commercial terms | Product state, quantity, price basis and delivery | A written offer with unresolved items identified |
A useful conclusion is not always “Candidate A wins.” It may be “A is ready for the agreed sample evaluation; B needs a pulse test at the required temperature.” That is already a more actionable result than a spreadsheet ranked by headline capacity.
Common questions
Can two cells with the same Ah and size replace each other?
Not automatically. Their voltage boundaries, charge behavior, load capability, terminals, thermal conditions and documentation can differ. Treat them as candidates for review, not drop-in substitutes.
Is a lower resistance always better?
It may help a particular load, but only comparable methods support a comparison. Resistance alone does not establish capacity, safety, authenticity or lifetime.
Does a catalogue summary replace the full specification?
No. A summary helps initial selection. The offered product and acceptance requirements should be confirmed against the applicable specification and project documents before an order.
Should every missing field stop an initial enquiry?
No. Submit the known requirements and identify the unknowns. The important distinction is between something still to be clarified and something already represented as a confirmed product fact.
Put the datasheet to work
CELL & PACK supports multi-brand battery sourcing beyond the featured online catalogue. Send the preferred brand and model—or the requirements you need a candidate to meet—together with quantity, destination and project timing. We coordinate product information, quotations and samples; the manufacturer or customer's technical team confirms design suitability and testing.
Discuss a cell or pack sourcing request. For the next step in selection, read LFP vs NMC and cylindrical, prismatic and pouch cells.
Sources and further reading
- S02 Idaho National Laboratory / U.S. DOE. Battery Test Manual for Electric Vehicles, Revision 3 (June 2015), INL/EXT-15-34184.
- S03 Molicel. INR-21700-P42A Product Data Sheet, Version 1.7.
- S09 Texas Instruments. BQ76952 battery monitor and protector.
- S30 Texas Instruments. BQ25170 800-mA linear battery charger for 1-cell Li-ion and LiFePO4.
- S04 HIOKI. Hioki Releases White Paper on More Effective Solution to Lithium-ion Battery Testing.
- S14 Scientific Data. A multi-stage lithium-ion battery aging dataset using various experimental design methodologies (2024).
- S15 National Laboratory of the Rockies / Journal of The Electrochemical Society. Investigation of Lithium Plating-Stripping Process in Li-Ion Batteries at Low Temperature Using an Electrochemical Model.