A brief input-power peak cannot establish how long a power bank takes to recharge. It says little about the starting condition, interruptions or the point at which the test ended. For procurement, the useful deliverable is a traceable recharge record tied to an agreed start and completion rule—not a photograph of the largest wattage on a meter.
This guide proposes that record. It does not report ChargeKeku measurements, specify a universal recharge time or infer a product’s internal components. Keep supported input profiles, actual USB input, battery-side charging and subsequent output energy separate throughout the review.
Name the quantity before comparing the report
A source’s advertised capability is not the power bank’s supported input profile, and neither is a time-series measurement. Obtain the exact receiving-port specification. Where USB Power Delivery is supported, retain the observed contract and changes during the run. Do not assume that every Type-C port supports PD, accepts power or operates bidirectionally.
| Boundary | Evidence to request | Do not substitute |
|---|---|---|
| Supported input profile | Product input specification; source capabilities and observed contract where applicable | An output rating printed elsewhere on the enclosure |
| Measured USB input | Timestamped voltage and current at a declared measurement plane | The maximum rating of the source or contract |
| Battery-side charging | Supplier-supported internal telemetry or a separately authorized engineering method | USB input current or integrated USB input energy |
| Later output energy | A separate discharge test with defined load and cutoff | Input watt-hours from the recharge run |
Texas Instruments’ BQ25798 documentation separately identifies input-current limiting, power-path functions, thermal regulation and input/charging-current sensing. These functions illustrate why port-side conditions and battery charging are not interchangeable. They do not establish which component is inside a purchased power bank or how efficiently that product operates.
The BQ25895 documentation likewise lists power-path management and separate charge-current and input-current regulation. The purchasing implication is limited: ask which quantity a report actually measures. Component feature pages cannot establish a finished product’s capacity, recharge duration or termination behavior.
Freeze a reproducible starting condition
Define the starting preparation with the supplier before testing. Record the sample ID, hardware and firmware revision where available, prior use, rest interval and the agreed initial condition. Do not impose an uncontrolled deep discharge to make a nominally “empty” starting point. Follow the product’s approved operating limits and laboratory procedure.
A display reading may help identify the sample state, but it is not independent proof of a calibrated state of charge. If the agreement uses a displayed percentage, state that limitation and keep the preparation consistent. Different starting conditions can make two honest duration reports incomparable without either sample being defective.
Identify the source, cable, port and ambient conditions. Keep output ports unloaded unless simultaneous input/output operation is explicitly part of the test. In that separate use case, document the external load and supported operating mode; do not silently compare its recharge duration with an unloaded run. Record any powered display, accessory or other agreed operating state.
Define completion before pressing start
Agree on a non-destructive completion criterion that the supplier can explain and the test method can observe. It might use documented product status together with a specified observation interval, but the exact rule belongs to the product and purchase agreement. This article does not prescribe one termination-current threshold for every power bank.
Define when the clock starts, when it stops and how interruptions are handled. A display reaching 100% without an agreed endpoint is not enough to close an otherwise undefined claim. Keep elapsed wall-clock time and any separately reported connected charging time distinguishable; omitting an interruption can change the meaning of the result.
USB-C role negotiation and battery-charger control also need not be the same function. TI’s TPS25750 documentation describes a PD controller that can control external battery-charger ICs. TPS25750 is a historical example marked “not recommended for new designs.” It is not a component recommendation, a ChargeKeku bill of materials or evidence of a universal charging algorithm.

Capture the full run, not only its fastest moment
At the declared USB input measurement plane, instantaneous power is voltage multiplied by current. Integrating that input power over the recorded interval yields input energy in watt-hours when time is expressed in hours. State the logger’s sampling interval, integration method, gaps and uncertainty. This is an input-energy record, not measured cell capacity or usable output energy.
The source-side and power-bank-side ends of a cable are not identical measurement locations. Document meter placement and any added adapter. Assess whether the logger or fixture changes cable loss, negotiation or operation. An instrument that shows a peak but cannot preserve the interval and interruptions may be unsuitable for the agreed recharge claim.
Lower actual draw than a valid contract permits does not, by itself, identify a fault. Record the sample state, available thermal observations and protocol events before assigning a cause. Do not infer an internal cell current from an external USB reading or open the pack merely to obtain one. Cell-side investigation requires its own authorized, qualified method.
Use this six-step run sheet:
- Identify the claim: required recharge window, sample revision, receiving port and supported input mode.
- Prepare the sample: approved initial condition, prior history, rest interval and environmental record.
- Document the path: source, cable, logger, measurement plane, calibration and unloaded output-port state.
- Start the evidence window: agreed start event, synchronized timestamps and initial protocol observations.
- Retain the interval: voltage/current log, input-energy calculation, interruptions, status events and safe external temperature observations.
- Close the run: endpoint evidence, elapsed time, aborted-run reason if applicable, reviewer and disposition under the agreed rule.
Decide what the evidence actually supports
| Finding | Decision | Required next evidence |
|---|---|---|
| Peak screenshot with no start or endpoint | Hold the recharge-duration claim | Complete run with agreed boundaries |
| Valid contract but lower actual draw | Investigate; do not assign a cause from watts alone | Sample state, thermal context and event history |
| Matched start/end and agreed duration requirement met | Accept the tested recharge scope | Traceable logs, calculation and reviewer signoff |
| Display says 100%, endpoint rule absent | Hold completion determination | Supplier-defined observable completion criterion |
The procurement record should link sample ID, run ID and claim revision. Store raw-file names, clock basis, start/end evidence, source/cable IDs, input-energy calculation, deviations and the final disposition. Retain repeated and aborted runs rather than presenting only the best one. This is a proposed evidence structure, not a completed test form.
Stop testing for abnormal heat, swelling, damage or repeated protective faults, and follow the laboratory’s battery incident procedure. A safe stop is a recorded outcome, not a reason to conceal the run. Longer-term aging belongs in a separate power-bank cycle-life test plan; source setpoint performance belongs in the PPS charger acceptance scope.
For an OEM enquiry, send ChargeKeku the required recharge window and evidence fields. Request a record matched to the intended source, cable and starting condition. Keep later output-energy and capacity claims separate unless their own test evidence is supplied.