Battery Pack Prototype Testing: From First Sample to Final-Device Validation

Last technical review: July 2026

The first custom battery sample powered the product on the bench. Capacity looked acceptable, the connector fit and the protection board did not trip. Two weeks later, the integrated prototype reset during motor startup, stopped charging in a cold room and developed an intermittent fault after the housing was closed. The supplier said the battery had passed testing. The customer said the product did not work. Both statements were true because the two teams had tested different questions.

This is where many battery projects lose time. A first sample is treated as product approval, a short resistor-load test is treated as a real duty cycle, and a hand-built pack is treated as proof that mass production will be repeatable. The failure appears late, after tooling, firmware and certification planning have already moved forward.

Prototype Approval Is Not Product Approval

A battery prototype can prove several useful things: the proposed voltage is available, the cell arrangement can be assembled, the connector can be fitted and a protection circuit can operate. It cannot, by itself, prove that the complete product will meet runtime, charging, thermal, mechanical, life, safety and production requirements.

The distinction is easier to remember in three lines:

• A sample proves that one assembly can be built and tested.
• Design validation proves that the defined design meets the intended requirements in the product.
• Production validation proves that the manufacturing process can repeat that design without special handling or hand-selected parts.

NASA battery guidance separates engineering evaluation, qualification, lot acceptance and unit-level acceptance because each stage answers a different risk question [1,2]. Commercial OEM projects do not need to copy an aerospace test program, but the underlying lesson is directly applicable: evidence from one stage should not be used to approve a different stage.

In practice, a compact pouch-cell product, a high-current cylindrical-cell product and a backup-power module usually need different validation evidence. Early sample planning should therefore match the intended battery architecture, whether the project uses custom LiPo batteries, 18650 / 21700 battery packs or lithium-ion / LiFePO4 battery packs.

Why the First Battery Sample Often Creates False Confidence

The first sample is usually built under favorable conditions. It may use freshly selected cells, short wires, a laboratory charger, an open enclosure and a simple electronic load. The final product adds the conditions that expose the real weaknesses: startup pulses, constant-power behavior at low state of charge, charge-while-use, heat from adjacent electronics, cable routing, foam compression, firmware timing, cold charging limits and repeated duty cycles.

A bench test can be technically correct and still be irrelevant to the product. Common examples include:

• Testing average current while the device fails on a millisecond or second-level startup pulse.
• Measuring pack-terminal voltage while the product sees a lower voltage after the connector, wire and switch path.
• Charging the battery with a laboratory supply while the final charger uses different termination, restart or temperature logic.
• Testing the loose pack in free air while the final enclosure restricts cooling or compresses the pouch cell.
• Approving a hand-built unit without defining the cell lot, BMS firmware, welding process, harness or thermal materials that production must repeat.

Define What Each Prototype Stage Must Prove

Many hardware teams use the labels EVT, DVT and PVT. Those names are useful shorthand, but they are not universal battery standards. One company may call a hand-built fit sample DVT while another reserves DVT for a production-intent build. The approval record should therefore state the evidence required at each stage rather than relying on the label alone. The same stage definitions should be frozen for the selected battery product family.

A useful battery pack prototype approval process moves forward only when the open issues from the previous stage are visible. An engineering sample with temporary wiring can support an electrical decision, but the temporary wiring must be listed as a deviation. Otherwise it quietly becomes part of the evidence even though production will not use it.

Seven-Step Battery Pack Prototype Approval Workflow

The workflow below converts a successful first sample into a controlled approval. A project should not advance merely because the next build is scheduled; it advances when the evidence required at the current gate is complete or the remaining deviation has been formally accepted. This gate-based structure also aligns with how custom battery projects move from requirement review to production-intent validation.

How Many Battery Samples Should Be Reserved Before Mass Production?

Direct answer: there is no universal sample quantity. The sample plan should reserve separate units for the questions that can alter or damage a pack, while allowing controlled reuse for non-damaging integration tests. The quantity should reflect energy, product risk, third-party testing, failure-investigation needs and pilot-lot variation.

Freeze the Requirement Before the First Test

A test cannot prove a requirement that was never defined. Phrases such as “long runtime,” “high current,” “safe charging” and “works in cold weather” sound clear in a meeting but lead to different test conditions in the laboratory.

The battery validation requirement should define at least:

• Exact device operating states, startup and stall events, continuous load, peak load, pulse duration, repetition and recovery time.
• Voltage range at the battery and at the device input, including the minimum voltage at which the product must remain functional.
• Required runtime or delivered energy under a documented duty cycle, not only nominal capacity in amp-hours.
• Charger, input source, charging current, termination behavior, low-temperature charging rule and whether the product operates while charging.
• Operating, charging, storage and transport temperature ranges, including warm restart and cold start where relevant.
• Final enclosure, mounting orientation, cable route, connector retention, sealing, vibration, drop or impact conditions and allowable pouch-cell swelling space.
• PCM/BMS thresholds, recovery behavior, communication, state-of-charge reporting and fault logging.
• Target markets, product standards, transport requirements and the evidence that must be supplied to the customer or test laboratory.
• Beginning-of-life, end-of-life and production-tolerance conditions, with pass/fail rules for each test.

Use a Prototype Validation Checklist Before Approving the Sample

Download the Battery Pack Prototype Validation Checklist before approving a sample or requesting a revised build. It helps the OEM and battery supplier record the sample stage, tested configuration, acceptance criteria, raw evidence and changes that require revalidation.

• Define the sample stage: fit sample, engineering validation, design validation, production-intent build or pilot lot.

• Record sample identity: cell lot, BMS hardware and firmware, connector, wire, NTC, harness route, labels and deviations.

• Tie each test to a requirement: load waveform, charger state, ambient, enclosure condition, pass/fail rule and owner.

• Record raw evidence: synchronized current, voltage, temperature, BMS state, charger behavior, photos and failure logs.

• List changes that require revalidation before the design can move to mass production.

Make Every Battery Sample Traceable

When a sample fails, the first question should be “What exactly was tested?” A surprising number of projects cannot answer it. The sample may have the same external label as the previous unit but use another cell lot, revised BMS firmware, a different NTC position or a hand-reworked connector.

A golden sample is useful as a physical reference, but it is not a substitute for drawings, a BOM, firmware identification and acceptance criteria. A sample can be lost, damaged or quietly reworked. The controlled specification must be able to recreate it.

Validate the Production-Intent Battery in the Final Product

Final approval belongs to the production-intent battery in the actual host product. Electrical performance, charging, mechanical installation, thermal behavior and fault recovery are interconnected; each must be tested with a common sample identity and a common requirement set.

Validate the Electrical System Under the Real Load

Battery sample testing should combine the pack data with the product data. Voltage, current, temperature, BMS state and device behavior need a common time base. A final number such as “10.8 V after the test” rarely explains why a reset, cutoff or charger fault occurred.

Open-Circuit Voltage Is Only a Starting Point

Open-circuit voltage can confirm gross state of charge, polarity and obvious imbalance. It does not prove loaded voltage, capacity, internal resistance, connector integrity or protection behavior. A pack may show the correct voltage with no load and collapse when the product starts.

Capacity and Runtime Must Follow the Device Duty Cycle

A low-rate capacity test is useful for incoming comparison, but it does not automatically predict product runtime. The correct validation uses the actual or a justified equivalent load profile, including active, peak, standby and sleep states. The Battery Runtime Calculation Guide explains why nominal amp-hours alone are insufficient when converter efficiency, cutoff voltage, pulse load and end-of-life margin matter.

For a constant-power load, current rises as battery voltage falls: I = P / V. A 30 W product draws 2.5 A at 12 V but 3.0 A at 10 V before conversion losses are added. The highest current and voltage sag may therefore occur near the end of discharge, not at the beginning.

Peak Current and Voltage Sag Need Fast Measurement

A handheld multimeter can miss the transient that resets a processor or trips a BMS. Use an oscilloscope or a logger with sufficient bandwidth and sampling rate to capture the battery terminals, device input and current waveform during startup, motor stall, radio transmit, heater activation or inverter surge. The measurement fixture must not add enough resistance to change the event being measured.

Milliohm-level connection problems should be checked with a four-wire Kelvin method or an equivalent fixture. Two-wire resistance measurements can include the leads and probe contacts, obscuring the actual crimp, weld or connector loss.

Protection and Recovery Are Part of Function

A pack does not pass merely because the PCM/BMS disconnects under an abnormal condition. The project must also verify which threshold was reached, how quickly the event was detected, whether the switching path interrupted the correct direction and how the product recovers. The PCM/BMS Selection Guide should be used to coordinate overcurrent, undervoltage, overvoltage, temperature and recovery behavior with the charger and device firmware.

Validate Charging With the Actual Product

Charging is often validated last, even though it involves the cell limits, BMS, charger, device firmware, connector and thermal environment at the same time. A battery that charges correctly on a laboratory supply may behave differently with the final USB-C source, adapter, solar controller, docking station or embedded charger.

The Charger, BMS and Firmware Must Agree

Verify the complete charge profile at the battery terminals: precharge where applicable, constant-current operation, constant-voltage transition, termination, restart and post-charge standby. Confirm that the charger voltage matches the exact series count and chemistry, and that the BMS does not become the normal charge controller. Protection should remain a boundary for abnormal conditions, not the routine way to end every charge.

Charge-While-Use Can Break Normal Termination

When the product remains active during charging, load current can hide the battery current from the charger. The charger may fail to terminate, restart repeatedly or report a full battery while the cells are not balanced. Test the real operating states, including display on, radio transmit, motor idle, standby and power-off charging. Record charger input current, battery current and load current separately where the architecture allows.

Low-Temperature Charging Needs a Defined Strategy

Do not solve a cold-charge complaint by raising or disabling the low-temperature threshold. Confirm the exact cell charging limits, NTC location, sensor tolerance, thermal lag and whether the product requires delayed charging, reduced current or a controlled preheat function. The battery, charger and firmware must enter and leave the restricted state predictably.

Test the Final Mechanical and Thermal Environment

A battery that fits the CAD envelope may still fail after assembly. The wire can be pinched by a rib, the connector can be loaded sideways, foam can block a heat path, a pouch cell can be compressed without swelling allowance, or an NTC can lift from the surface after vibration. These are product-integration failures, not cell-capacity problems.

Fit Is Not Integration

Check installation with the real housing, fasteners, gasket, display, PCB, speaker, motor, converter and cable routing. Verify insertion and removal, polarity protection, connector access, strain relief, bend radius, sharp edges, abrasion, service procedure and the worst assembly tolerance. A pack should not require an experienced technician to “massage” the wires into one safe position. Use the Connector, Wire Gauge and NTC Selection Guide when finalizing the electrical and sensing interface.

Mechanical Tests Should Surround Electrical Tests

Where vibration, drop, shock or repeated insertion is relevant, record electrical performance before and after the exposure. Inspect loaded voltage, insulation, connector retention, welds, cell movement, sensor position and enclosure damage. IEC 60068-2-6 and IEC 60068-2-27 provide generic vibration and shock test methods [9,10], but the product requirement must define the severity, mounting and acceptance criteria. A generic method number is not a complete test plan.

Thermal Approval Belongs to the Installed Product

Repeat the actual duty cycle in the final enclosure and operating orientation until the required sequence is complete or temperature approaches the defined steady condition. Measure the cells, BMS power path, connector and relevant enclosure locations. The Battery Pack Thermal Management Guide explains how to separate cell heat, current-path hotspots, sensor lag and enclosure heat accumulation.

Validate Fault Handling Without Improvised Abuse Tests

A good validation plan includes foreseeable faults and recovery states, but hazardous abuse testing should not be improvised on a normal engineering bench. External short circuit, overcharge, forced discharge, crush, impact or thermal-runaway work requires the correct standard, facility, instrumentation, containment and trained personnel.

System-level fault injection can still answer many practical questions safely when it is planned and risk-assessed:

Common OEM Mistakes During Battery Prototype Approval

The following mistakes repeatedly move failures from the engineering bench into tooling, certification or pilot production. Each one can make a technically valid test irrelevant to the final product. A controlled custom battery project workflow keeps these approval gates visible before release.

Approving the pack because capacity passed

A low-rate amp-hour result does not prove startup voltage, high-current behavior, thermal performance or device runtime. Use the actual duty cycle and the final cutoff.

Testing with a laboratory charger instead of the final charger

Termination, restart, source limits, USB-C negotiation and temperature logic can change the result. Validate the exact charger and device power path.

Testing only a fresh sample at room temperature

Low temperature, high ambient, low state of charge, aging and repeated duty can expose different limits. Define the worst credible operating boundary before testing.

Writing acceptance criteria after seeing the result

Moving a sensor or shortening a test after failure converts evidence into negotiation. Agree limits, stabilization and allowed control actions before the run.

Using the golden sample as the only specification

A physical reference can be reworked, damaged or lost. The controlled BOM, drawings, firmware and acceptance criteria must recreate it.

Reusing one unit for unrelated or damaging tests

A pack exposed to vibration, overcurrent or teardown no longer represents a fresh approval sample. Maintain test history and allocate samples by purpose.

Ignoring pilot-lot variation

An engineer-built unit can hide normal variation in crimp height, weld energy, NTC pressure, wire routing or cell matching. Sample across normal production.

Accepting an “equivalent” cell or BMS without revalidation

A similar rating is not proof of equivalent voltage sag, charge behavior, thresholds, thermal performance, transport scope or life.

Approving the battery but not the host product

The enclosure, charger, converter, firmware and user operation determine the system result. Final approval is shared between supplier and OEM.

Keeping only a one-page PASS report

Without raw data, sample identity and test conditions, another engineer cannot reproduce the decision or investigate a future failure.

Certification, Transport Testing and Product Validation Answer Different Questions

Certification planning should start early, but a certificate should not be used as a substitute for product engineering. The battery, host product and transport chain can each have separate requirements.

The current UN Manual of Tests and Criteria requires lithium cells and batteries to be of a tested design type for transport, and a change that materially affects the test result can create a new type requiring the relevant tests [4]. That change-control principle matters even when the project is not yet shipping: cell substitution, series-parallel changes, protective-device changes or major mechanical revisions should not be treated as paperwork-only updates.

The finished product may also fall under a host-equipment standard. IEC 62368-1:2023, for example, applies to audio/video and information and communication technology equipment and uses a safeguard-based safety approach [11]. Household appliances, medical equipment and other products have different standards. The customer should define the target market and product category before the battery test plan is frozen.

Move From a Successful Prototype to Repeatable Production

The production question is simple to ask and difficult to prove: can ordinary production build the approved battery without heroic manual adjustment, hand-selected cells or an engineer standing beside the line? A perfect golden sample is not enough if the process cannot reproduce it.

The Golden Sample Is a Reference, Not the Specification

Keep an approved reference sample, but freeze the information needed to recreate it: drawings, BOM, approved manufacturer part numbers, firmware hash or version, connector pinout, wire length and gauge, NTC position, weld or busbar design, fuse, labels, materials, work instructions and inspection criteria. Photograph hidden construction before the pack is closed.

Pilot Production Must Test Variation, Not Only the Best Unit

A pilot lot should represent normal tools, operators, fixtures and material flow. Select samples across the run, not only the first unit or the unit prepared for a customer visit. Check critical-to-quality characteristics such as cell matching, weld quality, harness length, connector retention, loaded voltage drop, BMS firmware, temperature sensing, insulation and final functional test.

NASA programs use qualification, lot and acceptance testing as separate controls because one approved design does not eliminate incoming-lot or assembly variation [1-3]. Commercial projects should scale the sample size and destructive testing to the product risk, energy, volume and regulatory obligations; there is no universal number of samples that is correct for every battery pack.

Changes Need Defined Revalidation

The change-control plan should state who reviews a change, which documents are updated and which tests are repeated. “Equivalent component” is not a validation result.

A Practical Battery Pack Validation Matrix

The matrix below is a starting structure, not a universal list. Each project should add the application-specific hazards, standards and user scenarios that matter to the finished product.

Final Production Approval Gate

A project is ready for production release only when the approved requirement, tested configuration and manufacturing controls point to the same design. Open issues may remain, but each must have a documented owner, risk decision and revalidation plan.

What the Buyer Should Send to a Battery Supplier

• Final or current enclosure drawing, battery space, mounting, cable route, connector access and nearby heat sources.
• Measured battery-side current and voltage waveform for startup, continuous load, peaks, charging, standby and shutdown.
• Device minimum operating voltage, converter behavior and any reset or cutoff logs.
• Required runtime, duty cycle, ambient range, charge source and charge-while-use behavior.
• Mechanical and environmental conditions, including drop, vibration, sealing, service and pouch-cell swelling constraints.
• Communication, SOC, alarm, wake, sleep and fault-recovery requirements for a smart BMS project.
• Target markets, product standards, certification plan, transport route and required documentation.
• Planned production volume, expected service life, field-replacement strategy and allowed component alternatives.

The most useful input is measured data from the product. A spreadsheet containing only “3.7 V, 5000 mAh, 10 A” leaves the supplier to guess the events that will decide whether the battery actually works. This evidence is the right starting point for a custom battery engineering review.

What a Useful Supplier Validation Package Should Contain

• Sample identity and as-built BOM, including exact cell, lot, topology, BMS hardware/firmware, connector, wire, fuse and sensors.
• Approved requirement matrix showing the test condition, instrument, sensor location, pass/fail rule and result.
• Raw or exportable voltage, current, temperature, charging and fault data, not only a one-page “PASS” certificate.
• Clear list of deviations between the tested sample and the intended production design.
• Failure log, root-cause evidence, corrective actions and the tests repeated after each change.
• Production drawings, critical-to-quality controls, pilot-lot plan and final functional-test coverage.
• Change-control list identifying substitutions or revisions that require customer approval and revalidation.
• Applicable compliance reports and UN 38.3 test summary matched to the exact shipped design type.

How THOR Power Supports Battery Prototype and Final-Device Validation

THOR Power supports custom LiPo pouch batteries, lithium-ion battery packs, 18650 and 21700 packs, LiFePO4 backup and energy-storage batteries, and custom special battery packs. Prototype work begins with the device requirement and sample purpose, not with a generic request to “send one battery for testing.”

For concept-stage comparison, teams can also review LiPo battery models, 18650 / 21700 models and energy storage models before moving into a custom validation plan.

Depending on the project, the engineering scope can include cell and topology review, load-profile analysis, PCM/BMS coordination, connector and wire review, NTC planning, sample traceability, charging tests, voltage-sag capture, thermal instrumentation, enclosure integration, production-intent builds, pilot-lot review and change control. Final product approval remains a shared responsibility: the battery supplier controls the pack design and process, while the OEM controls the host product, firmware, charger, use conditions and market requirements.

For a difficult prototype, send the enclosure drawing, load waveform, charger details, ambient range, existing sample data and the failure condition. The first task is to identify which question the current sample has not yet answered.

Key Takeaways for OEM and ODM Buyers

• A first battery sample is an engineering input, not automatic approval for the final product or mass production.
• Test the production-intent pack in the real device with synchronized current, voltage, temperature, charger, BMS and firmware data.
• Define pass/fail criteria before testing and keep every sample traceable to its cell lot, BOM, firmware and test history.
• Certification, UN 38.3 and supplier bench tests answer different questions; none replaces final-device validation.
• Freeze the design and prove that normal production can repeat it, then revalidate changes that can affect performance, safety or compliance.

Conclusion

Battery prototype approval is a chain of evidence, not a single test result. The first sample may prove fit or basic function, but final approval requires the production-intent configuration to work with the real load, charger, firmware, enclosure and environmental conditions.

The final production decision should link a written requirement to a traceable sample, raw test evidence, resolved failures, pilot-lot controls and a change-management plan. When one of these links is missing, the project may still move forward, but it is moving forward with an unowned risk.

For OEM and ODM teams, the practical goal is not to produce the largest possible test report. It is to create enough clear evidence that another engineer can understand what was approved, reproduce the decision and recognize when a later change requires revalidation.

FAQ: Battery Pack Prototype Testing and Validation

Author, Technical Review and Evidence Standard

This article is written for OEM and ODM battery-pack development, sample approval and supplier validation. It does not replace cell datasheets, product risk analysis, accredited laboratory testing, transport requirements, safety certification, quality-system procedures or market-specific compliance assessment.

Technical review scope: prototype-stage definition, requirement traceability, sample identity, electrical and charging validation, enclosure integration, fault and recovery behavior, certification boundaries, production-intent evidence, pilot-lot control and engineering change review.

Evidence standard: engineering claims should be supported by exact component specifications, controlled sample records, calibrated measurements, raw test data and final-device evidence under defined conditions. The composite cases explain recurring diagnostic patterns and are not presented as named customer projects or first-party performance claims.

Review this page at least annually, and sooner when referenced standards, internal product pages, sample-development methods, validation equipment, production controls or compliance claims change.

Technical References

1. NASA – Guidelines on Lithium-Ion Battery Use in Space Applications.

Guidance on battery design, charge control, qualification, lot acceptance and unit acceptance. Used here for the general distinction between engineering, qualification and acceptance evidence; commercial OEM projects should scale the approach to their own risk.

2. NASA – Requirements or Specifications, Certification, Testing.

Technical presentation covering sample testing, lot testing and acceptance approaches for commercial and custom lithium-ion batteries.

3. NASA – A Cell Selection Method and Validation Process for the Aerospace Battery.

Example of documented cell selection, pre-acceptance testing, lot identification, screening and validation controls.

4. UNECE – Manual of Tests and Criteria, Revision 8, Part III, Subsection 38.3.

Official transport test framework for lithium cells and batteries, including design-type testing, test-summary requirements and treatment of material design changes.

5. IEC 62133-2:2017+AMD1:2021 CSV.

Safety requirements and tests for portable sealed secondary lithium cells and batteries under intended use and reasonably foreseeable misuse.

6. IEC 61960-3:2017.

Performance tests, designations, markings, dimensions and requirements for secondary lithium cells and batteries for portable applications.

7. IEC 62619:2022.

Safety requirements and tests for secondary lithium cells and batteries used in industrial applications, including stationary applications.

8. IEC 62620:2014+AMD1:2023 CSV.

Marking, performance tests and requirements for secondary lithium cells and batteries used in industrial and stationary applications.

9. IEC 60068-2-6:2007.

Generic environmental test method for sinusoidal vibration. Product-specific severity, mounting and acceptance criteria still need to be defined.

10. IEC 60068-2-27:2008.

Generic environmental test method and guidance for shock. It is a method reference, not a complete battery or product validation requirement.

11. IEC 62368-1:2023.

Safety standard for audio/video, information and communication technology equipment, illustrating that the host product can impose safeguards beyond battery-level evidence.

12. UL Solutions – Preparing for Your UL Mark Evaluation (U.S. and Canada).

Official guidance emphasizing early design review, prototype information, pre-production conformance, critical components, manufacturing details and factory test equipment.