What a LiFePO4 Capacity Test Can and Cannot Tell You
A distributor in Melbourne finishes a 0.2C discharge on two sample packs from a container shipment. Both results sit above the nameplate. The team signs acceptance. Months later a customer says the packs “feel like used cells.” The capacity test cannot answer that complaint. It never could.
This page exists because two live LJY Energy guides already point to it: the incoming inspection guide and how to test LiFePO4 capacity and internal resistance. Those pages cover the workflow and the bench method. This page covers the limits of the number you get, and how to write a method that can be used both at incoming acceptance and later in a warranty claim.
LJY Energy is a China-based LiFePO4 battery supplier and OEM/ODM solutions provider. Acceptance thresholds, sampling plans, and capacity-test methods are confirmed in the quotation and the order agreement. This page does not publish a fixed LJY test result, tolerance, warranty, price, MOQ, or lead time.
What a capacity test actually measures
A capacity test measures the charge or energy the unit delivered under a defined protocol: how it was charged, how long it rested, the discharge rate, the end-of-discharge limit, the temperature, and whether you recorded DC amp-hours at the battery terminals or energy at some other boundary.
It is not a hidden “true Ah” stored inside the product, independent of those conditions. Change the protocol, and the same pack can produce a different number.
A published cell example shows how tightly a serious capacity method is specified. The EVE LF280K product specification (RD-LF280K-S04-LF, Version A, April 2022) states a minimum capacity of 280 Ah at 0.5C, 25 ± 2 °C, between 2.5 V and 3.65 V. The same document’s calibration method charges at 140 A (0.5C) to 3.65 V, tapers to 14 A, rests 30 minutes, discharges at 140 A to 2.5 V, and averages repeated discharges after activation cycles. That is a cell method under a stated fixture and temperature. It is a useful illustration of completeness — not a pack acceptance protocol, and not an LJY Energy test result.
IEC 62620, the performance standard for industrial lithium cells and batteries, likewise treats capacity-related work as a set of named tests, including discharge performance at +25 °C, low-temperature discharge, and charge retention and recovery. IEC 62660-1:2018 does the same for vehicle-propulsion cells, including capacity among its basic performance tests. In both cases the standard is a method family, not a substitute for the buyer’s contract method.
What the number can tell you
Used with a written protocol, a capacity test can support four practical decisions.
- Nameplate comparison on the day of the test. It can show whether the measured units delivered at least the agreed share of rated capacity under that method.
- Batch consistency among the units you actually tested. A tight cluster and a clear outlier are useful receiving signals, especially when you also record voltage and internal resistance, as described in the incoming inspection guide.
- A baseline for later comparison. If the same method is repeated later, the difference is a residual-capacity measurement, which is what many warranties and performance guarantees ultimately need. DNV, writing about energy storage capacity warranties, emphasizes that reviewers look not only at the degradation curve but at how capacity is measured and calculated.
- A documented evidence pack. Serial numbers, date codes, temperature, instrument identity, and the raw charge/discharge curve are often more useful in a dispute than a single “pass” stamp.
What the number cannot tell you
It cannot prove the cells are new
A used or second-life cell can still deliver close to its present capacity. A fresh cell can sit slightly below nameplate under a conservative method. Capacity is not a provenance test. If new versus used or second-life matters to the order, you need traceability, date codes, batch documents, and a stated grade definition — see Grade A vs Grade B LiFePO4 cells — not a single discharge result.
It cannot predict cycle life
One capacity point does not show the shape of the fade curve, the presence of a knee, or how the pack will behave at a different depth of discharge, rate, or temperature. Cycle-life claims have to be read with their conditions, as set out in how to verify cycle-life claims and cycle life vs depth of discharge.
It cannot convert a cell result into pack life
A cell discharged to 2.5 V in a laboratory fixture is not the same object as a pack stopped by a BMS at a higher pack-level cut-off, with thermal gradients and interconnects in the path. IEC 62660-1:2018 describes cell tests that supply data for pack and system design. It does not make the cell number a pack result. See cell cycle life vs pack life.
It cannot separate calendar aging from cycle aging by itself
A low result on a pack that sat in a warm warehouse may be calendar fade, incomplete charge, a high resting state of charge, or a measurement-condition problem. The calendar aging guide covers the storage mechanism; a single capacity test only reports the capacity you measured.
It cannot speak for the units you did not test
Spot checks describe the sample. Unless the contract says how a failed sample is treated — retest, expanded sample, or lot rejection — a good result on two packs is not a statistical certificate for the container.
Conditions that move the number, even when the pack is unchanged
Buyers often treat a 3–5% gap as a quality defect when it is actually a method gap. The list below is not a complete laboratory standard; it is the set of conditions that should be identical, or explicitly converted, between incoming acceptance and a later claim.
| Condition | Why it changes the result |
|---|---|
| Discharge rate (C-rate) | A 0.2C discharge and a 0.5C or 1C discharge are different tests. Higher rate usually reduces delivered capacity and increases voltage drop. |
| Temperature | Capacity and resistance are temperature-sensitive. A test at 15 °C is not interchangeable with a test at 25 °C. Say where temperature is measured. |
| Rest time after charge | The EVE LF280K method, for example, inserts a 30-minute rest after charge and after discharge. A test that starts discharging immediately is a different method. |
| Charge completion rule | Constant-current only, versus constant-current/constant-voltage to a stated cut-off current, changes how full the starting point is. |
| End-of-discharge limit | Cell voltage (for example 2.5 V per cell) versus a higher BMS cut-off changes usable capacity. Always state whether the BMS remains in the circuit. |
| Boundary | DC amp-hours at the pack terminals is not the same as AC energy at an inverter or point of connection. Project warranties often care about the latter; incoming warehouse tests often measure the former. |
| Instrument and calibration | An uncalibrated electronic load and a calibrated cycler can disagree by more than a typical commercial tolerance. Record the instrument. |
The how-to-test guide describes a practical buyer protocol (charge to 3.65 V per cell with taper, rest 30+ minutes, discharge at 0.2C or the datasheet rate to 2.5 V). That is a workable receiving method when the contract adopts it. It is not a universal industry law, and it should not be mixed silently with a 0.5C datasheet method.
Incoming acceptance and warranty claims need the same method
This is the procurement mistake that turns a good laboratory habit into an unenforceable claim.
- Incoming acceptance asks: “Did these units meet the receiving threshold today?”
- A residual-capacity warranty asks: “Has remaining capacity fallen below the promised floor under the agreed method?”
If the two questions use different rates, temperatures, rests, or boundaries, neither side can compare the results fairly. DNV’s capacity-warranty review makes the same practical point for storage projects: operational restrictions and the measurement method decide whether a warranty can actually be used.
The companion page warranty vs incoming acceptance vs cycle-life claims explains why the three promises are not interchangeable. The drafting rule for this page is narrower:
- Write one capacity protocol in the purchase order.
- Use it for factory or incoming acceptance.
- Use the same protocol, or a documented conversion, for any later residual-capacity claim.
- Record the raw data, not only a pass/fail label.
A common commercial reference on the incoming inspection page is a 5% tolerance against rated capacity. That is a reference often used in trade, not an IEC-mandated limit. If you use it, write it down as your contract threshold, and say which method it applies to.
What to lock into the purchase order
Copy this list into the RFQ and fill the values per order. The numbers are confirmed per quotation.
- Test object: finished pack with BMS, or cell, or module
- Charge rule, including voltage and current taper
- Rest time after charge and after discharge
- Discharge rate and end-of-discharge limit
- Temperature setpoint and measurement location
- Result reported as DC Ah, DC Wh, or AC energy at a named boundary
- Incoming pass threshold and sample plan (including what happens if a sample fails)
- Residual-capacity threshold, if any, and the term (years, cycles, or throughput)
- Who may witness the test and how raw data are stored
- Instrument accuracy or calibration requirement
For storage projects, include the same method in the wider commercial and industrial BESS RFQ checklist. For the claim route after a failed test, use the warranty-terms guide.
What LJY Energy can confirm in writing
On request, we can discuss a capacity-test method for your model, an incoming sample plan, and whether that same method can be reused for a later residual-capacity check. Those items are confirmed in the quotation and the order agreement, not as a published standard on this page.
To specify a capacity-test method for an order, use the contact form: https://ljyenergy.com/contact/.
FAQ
Can a pack fail a capacity test and still be within a normal commercial tolerance?
Yes, if the contract tolerance and the test method were never aligned. A result 4% below nameplate may be a fail under a 2% clause and a pass under a 5% clause, and both can be consistent with the same physical pack if the discharge rate or temperature also changed.
Why do suppliers quote capacity at 0.5C when incoming tests often use 0.2C?
Those are different methods. 0.2C is a common receiving choice because it is gentler and often closer to a “available capacity” figure; many cell datasheets use 0.5C as the rated point. Neither is wrong. Mixing them without a conversion is wrong.
Does a capacity test show state of health?
Only relative to a defined beginning-of-life or nameplate baseline, under the same method. Without that baseline and method, “SOH” is a label, not a measurement.
Should I disable the BMS during a pack capacity test?
Only if the contract says so, and only with a safe, documented procedure. For a finished-pack purchase, the commercially relevant capacity is usually the capacity the pack delivers with the BMS in its normal protection and cut-off role. State the choice in writing.
Is IEC 62620 or IEC 62660-1 enough, by itself, as my acceptance method?
No. Those standards define families of performance tests. Your order still has to name the test object, the exact procedure, the pass threshold, and whether the same method will be used later for a warranty claim.
Notes and sources
General measurement and procurement guidance as of November 2026. Nothing on this page is an LJY Energy test result, certificate, warranty term, price, MOQ, or lead time. Manufacturer figures are cited from the named public specification and are not LJY guarantees.
- EVE Power Co., Ltd, LF280K Product Specification, RD-LF280K-S04-LF, Version A, April 2022 (public distributor copy used for verification: https://www.boetronics.nl/files/EVE_LF280K.pdf)
- DNV, “Energy Storage Capacity Warranties: Beyond the Fine Print,” https://www.dnv.com/article/energy-storage-capacity-warranties-beyond-the-fine-print-200339/
- IEC 62620:2014+AMD1:2023, official abstract, https://webstore.iec.ch/en/publication/85493; test list also summarized by DEKRA, https://www.dekra.nl/en/iec-62620/
- IEC 62660-1:2018, official abstract, https://webstore.iec.ch/en/publication/28965
- Related LJY Energy guides: how long LiFePO4 batteries last, how to test capacity and internal resistance, incoming inspection
To agree a capacity-test method before shipment, contact LJY Energy: https://ljyenergy.com/contact/.
