How to Read a Cell Datasheet
A nine-step checklist, why the cycle-life line needs five conditions attached, and why the omissions tell you more than the numbers.
Battery Fundamentals · Part 4 — Specification and Sourcing · Chapter 10 · 13 min read
A cell datasheet is a sales document that happens to contain engineering. Every number on it is true under some condition, and the skill is working out which condition — because the conditions are where the difference between two apparently identical cells actually lives.
9
Checks, in order
3
Temperature ranges
5
Conditions a cycle claim needs
4
Omissions that matter
10.1 — Why the order matters
Work down the checklist in sequence. It is arranged so that the cheap disqualifying checks come before the expensive ones: there is no point evaluating a cycle-life claim on a cell whose voltage window does not match the BMS you already have.
In plain English
Read a datasheet the way you would read a loan agreement. The headline number is designed to be memorable. The conditions attached to it are designed to be skipped, and they are where the actual commitment sits.
10.2 — The nine-step checklist
- 1Identity. Model number, chemistry, form factor, dimensions with tolerances, weight — and the manufacturer and factory. Large manufacturers run multiple plants with genuinely different quality records, so the brand alone is not an identity.
- 2Nominal ratings, with their test conditions. “3,200 mAh at 0.2C, 25°C, to 2.5 V” is a specification. “3,200 mAh” on its own is marketing. If the conditions are absent, the number cannot be compared with anything.
- 3Voltage window. Nominal, charge cut-off, discharge cut-off. Check these against your BMS thresholds and your charger settings now, not after delivery. A mismatch here is the most common cause of a pack that overcharges or never fills.
- 4Current limits. Standard charge, maximum charge, standard discharge, maximum continuous discharge, and peak discharge with a duration. A peak figure without a duration is not a specification, and sizing a pack against it is how packs overheat.
- 5Internal resistance. ACIR at 1 kHz and DCIR with the pulse duration stated — and crucially the maximum acceptance limit, not just the typical value. Typical tells you about a good cell; the limit tells you what you are contractually allowed to receive.
- 6Cycle life. See the next section. This line deserves its own scrutiny.
- 7Temperature ranges. Three of them, separately. See below.
- 8Certification. UN 38.3 as a minimum, since it is legally required for transport, plus IEC 62133 or IS 16046 and whatever the application demands. Check the certificate names the exact model, not a family.
- 9What is absent. The most informative section of any datasheet. See 10.5.
10.3 — The cycle life line, read properly
Important
“≥2,000 cycles” means nothing on its own. A cycle-life claim needs five conditions attached before it is a claim at all: charge rate, discharge rate, depth of discharge, temperature, and the state-of-health threshold being counted to.
The same cell, honestly described two ways
| Statement | Status |
|---|---|
| “≥2,000 cycles” | Not a claim. There is no condition under which it can be tested or disputed. |
| “2,000 cycles at 1C/1C, 100 % DOD, 25 °C, to 80 % SOH” | A claim you can hold someone to, design a warranty against, and verify on a cycler. |
Each missing condition moves the number by a lot. A cell rated 3,000 cycles at 0.5C may deliver 1,200 at 2C. The same cell at 50 per cent DOD instead of 100 per cent may deliver two to three times more cycles. At 45°C instead of 25°C, expect roughly a quarter of the calendar life. And a figure quoted to 70 per cent SOH will always look better than one quoted to 80 per cent.
The relationship between DOD and cycle count, and why shallower cycling wins, is worked through in Chapter 5.
In plain English
The commercial version of this: if your warranty says five years and your customers cycle daily, you have promised roughly 1,800 cycles at whatever DOD they actually use, at whatever temperature your packs actually live at. Match the cycle-life conditions to that reality, not the cycle-life number to your marketing.
10.4 — Three temperature ranges, not one
These are separate specifications and they are routinely conflated.
| Range | Typical | Why it differs |
|---|---|---|
| Charge temperature | 0 to 45 °C | The narrowest, because charging a cold cell plates lithium. This is the one that must be enforced by the BMS. |
| Discharge temperature | −20 to 60 °C | Wider — cold discharge reduces capability but does not damage the cell. |
| Storage temperature | 0 to 35 °C ideal | Quoted with a recommended storage SOC, usually 30 to 50 per cent. Both halves matter and the SOC half is usually ignored. |
Important
Compare the charge range against your BMS default. Many boards ship permitting charge from −10°C, which is looser than any cell manufacturer allows. The board will authorise damage the cell datasheet prohibits, and closing that gap is the integrator’s job. The mechanism is in Chapter 6.
10.5 — What is missing tells you most
A datasheet is a positive document — it lists what the manufacturer is willing to stand behind. Four common omissions each carry a specific meaning.
- •No K-value or self-discharge specification. They are not screening for self-discharge. That is the screen that catches internal micro-shorts, and it is the defect that presents months later as one permanently low cell. This is the most serious of the four.
- •No DCIR. They may not measure it. You then have no basis for predicting voltage sag or heat generation under your actual load, and you will find out on the road.
- •No calendar life data. You cannot predict shelf ageing, which matters enormously for inventory and for any pack that spends time parked at high state of charge.
- •No cycle life test conditions. Treat the number as fiction. Not as optimistic — as fiction, because there is nothing there to be optimistic or pessimistic about.
Ask for each of them explicitly. A manufacturer who measures these will send them; one who does not will change the subject, and that answer is worth more than the document. What to ask for alongside the datasheet — batch test reports, UN 38.3 summary, date codes — is covered in Chapter 11.
Quick check: test yourself
1.A datasheet says “Max discharge current: 300 A”. What is your next question?
Show answer
2.Cell A claims 2,000 cycles, Cell B claims 3,500. Which lasts longer?
Show answer
3.The datasheet lists a charge temperature range of 0 to 45 °C. Your BMS default allows charging from −10 °C. What happens if you ship it?
Show answer
4.A datasheet has no self-discharge or K-value line. How much does that matter?
Show answer
Chapter summary
- ✓Every number on a datasheet is true under some condition. The conditions are where the difference between two apparently identical cells lives.
- ✓Work the checklist in order: identity and factory, ratings with conditions, voltage window against your BMS, current limits with durations, resistance with acceptance limits, cycle life, temperatures, certification, omissions.
- ✓A cycle-life claim needs five conditions — charge rate, discharge rate, DOD, temperature and SOH threshold. Without them it is not a weak claim, it is not a claim.
- ✓There are three temperature ranges, not one. Charge is narrowest because cold charging plates lithium, and BMS defaults are routinely looser than the cell allows.
- ✓A peak current without a duration is not a specification. Size packs against continuous ratings.
- ✓Ask for the maximum acceptance limit on internal resistance, not the typical value — typical describes a good cell, the limit describes what you may legally be sent.
- ✓Four omissions each mean something specific: no K-value means no self-discharge screening; no DCIR means no sag prediction; no calendar life means no shelf-ageing model; no cycle conditions means the cycle number is fiction.
Frequently asked questions
What does “≥2,000 cycles” actually tell you?+
Nothing on its own. A cycle-life claim needs five conditions before it is a claim at all: charge rate, discharge rate, depth of discharge, temperature, and the state-of-health threshold being counted to. “2,000 cycles at 1C/1C, 100 per cent DOD, 25°C, to 80 per cent SOH” is something you can hold a supplier to, design a warranty against and verify on a cycler.
A datasheet lists a maximum discharge current of 300A. Can I design to that?+
Not without a duration, temperature and state of charge attached. A peak current with no duration is not a specification — it might be a three-second pulse at 25°C and full charge. What you need for pack sizing is the maximum continuous discharge rating, which is often a fraction of the peak. Sizing against a peak figure is one of the most reliable ways to overheat a pack.
How many temperature ranges does a cell have?+
Three, and they are separate specifications. Charge temperature is narrowest, typically 0 to 45°C, because charging a cold cell plates lithium. Discharge is wider, typically −20 to 60°C. Storage has its own range, usually 0 to 35°C ideal, quoted alongside a recommended storage state of charge of 30 to 50 per cent — and that second half is routinely ignored.
What should I worry about if a datasheet has no self-discharge or K-value line?+
That the manufacturer is not screening for self-discharge at all. That screen is what catches cells with internal micro-shorts from particle contamination, and without it defective cells ship and present three to six months later as one permanently low cell in a customer’s pack. Of the four common omissions — no K-value, no DCIR, no calendar life, no cycle test conditions — this is the most serious.
Should I ask for typical or maximum internal resistance?+
Maximum. Typical tells you about a good cell; the maximum acceptance limit tells you what you may contractually be sent. Ask for ACIR at 1kHz and DCIR with the pulse duration stated, both with limits rather than typical values.
Reviewed by
Battery Fundamentals is an original educational series on lithium battery technology. Threshold tables, cycle-life curves and worked examples use representative values drawn from published product specifications and widely-observed industry patterns, not measured data from a specific product. Always consult the current manufacturer datasheet before making design, purchasing, warranty or certification decisions.