BB-730: Diagnosing Starter Battery Capacity Loss with EIS Testing

A new starter battery begins service with its expected rated capability, but usable performance declines with age, operating conditions, and deterioration. The important failure point is not always the moment the engine stops cranking. A battery may still deliver a brief high-current burst while having too little stored energy for repeated starts, accessory loads, or prolonged operation.

Battery diagnostics therefore need to distinguish stored energy capacity from short-duration cranking performance. Reserve Capacity (RC) is traditionally expressed as the number of minutes a battery can deliver a specified load; the referenced automotive procedure uses a 25-amp discharge. Capacity can also be expressed in ampere-hours using an IEC-style discharge procedure, commonly described as a 0.2C, or five-hour, rate. These measurements answer different questions from Cold Cranking Amps (CCA), which evaluates the battery’s ability to deliver high current during a cold-temperature cranking test.

The distinction matters because conventional cranking tests can indicate that a battery is capable of starting an engine without showing how much reserve energy remains. Capacity-oriented diagnostics, including non-invasive electrochemical impedance spectroscopy (EIS), attempt to estimate that remaining reserve without a full charge-discharge test. Their usefulness depends on test conditions, model quality, and validation for the battery being measured.

Diagnostic Data Connectivity and Service Records

A connected battery diagnostic system normally has two functional layers. The first is the local test operation: a host device receives the selected test settings, controls or coordinates the measurement, and displays numerical and graphical results. The operator may see a pass/fail decision together with values such as estimated state of charge, internal resistance, temperature, and battery-health indicators.

The second layer is optional connectivity. Where a suitable platform is available, results can be uploaded to a cloud service and shared with technicians, supervisors, or fleet managers. A service record may associate the test with a vehicle identification number, work order, location, operator, and time of test. Historical readings can then help identify deterioration between visits rather than treating every test as an isolated event.

Typical records may include:

• pass/fail status;
• state of charge;
• measured or estimated internal resistance;
• battery temperature;
• vehicle or equipment identification;
• work-order and service notes; and
• numerical or graphical test results.

Some systems can also print reports or labels when compatible printers and label hardware are available. This can be useful when batteries are removed from service, tagged for charging, or tracked through a fleet workshop.

Connectivity should be treated as a workflow feature, not as proof that the underlying battery-health measurement is accurate. A cloud database can preserve a result, but it does not independently validate the test method, the selected battery model, or the interpretation of the measurement. The accuracy question remains dependent on the instrument, connections, test conditions, reference data, and diagnostic algorithm.

Organizations using cloud-connected testing also need to define who can access records, how long records are retained, how vehicle and operator information is protected, and whether the platform can be integrated with existing maintenance software. Local-only operation may be preferable where connectivity is unreliable or where service data must remain inside a controlled environment.

Internal-resistance values also require caution when comparing historical readings. Different instruments can use different frequencies, test currents, filtering methods, temperature corrections, and calculation conventions. A resistance value from one tester should not automatically be treated as directly comparable with a value from another tester unless the methods and conditions are known.

Source: Battery University

Why Capacity Fades Before Cranking Fails

Starter batteries have at least two diagnostically important characteristics: the ability to deliver high current for a short time and the ability to store and deliver energy over a longer period. CCA primarily addresses the first characteristic. Capacity, Reserve Capacity, and ampere-hour measurements address the second.

CCA is associated with the battery’s ability to supply a specified high-current load during a cold-temperature test. The source describes the SAE J537 procedure as applying a load related to the CCA rating while the battery is cooled to -18°C (0°F). Reserve Capacity uses a timed discharge, while an IEC capacity procedure can use a 0.2C, five-hour discharge. The exact rating and acceptance procedure depend on the applicable standard and battery application.

A battery can therefore continue to crank an engine even after its reserve energy has declined substantially. Such a battery may start the vehicle once in favorable conditions but fail after several start attempts, during extended accessory operation, or when used in a start-stop system. Trucks, buses, and vehicles with hotel or auxiliary loads generally need more reserve margin than an ordinary passenger car.

The supplied source reports a study of 175 trunk-mounted AGM starter batteries tested in a moderate climate. Capacity and CCA were measured according to DIN and IEC-related procedures in the described dataset. Its plotted results showed capacity moving toward an approximately 40 percent end-of-life gate while many CCA readings remained acceptable. This is a source-specific dataset, not a universal failure curve for every AGM battery or vehicle.

The same source reports the claim that approximately 48 percent of starter-battery failures are associated with capacity fade and about 12 percent with low CCA. Those percentages should be treated as findings from the cited source and study context rather than as a general statistic that applies to every market, battery design, climate, or vehicle population.

The practical implication is that CCA, internal resistance, capacity, and state of health are related but not interchangeable:

• CCA indicates short-duration high-current capability under a defined cranking condition.
• Internal resistance provides information about power delivery, but its interpretation depends strongly on the measurement method and battery condition.
• Capacity or Reserve Capacity indicates how much usable energy remains under a specified discharge procedure.
• State of health is a broader assessment that may combine capacity, power capability, age, operating history, and abnormal behavior.

A minimum operational reserve concept can make the service decision more useful. A pass means that the estimated reserve remains above the selected operating threshold. A fail may justify replacement even if the engine still starts, because the battery no longer has enough margin for the expected duty cycle.

The source gives approximately 40 percent as an example capacity threshold for passenger vehicles. A higher value, such as 50 percent or more, may be selected for trucks, buses, or equipment with hotel and auxiliary loads. These are application decisions, not universal limits. Climate, electronic load, start-stop use, battery age, duty cycle, and manufacturer requirements can all change the appropriate reserve margin.

A diagnostic system may use selectable or downloadable matrices for different applications. A generic automotive matrix may be suitable for ordinary vehicle batteries, while a separate reference set may be needed for commercial vehicles, marine equipment, machinery, or deep-cycle systems. The source also describes matrices for lead-acid and lithium-based deep-cycle batteries. Such claims must be separated from independently validated capability: a matrix developed for one chemistry, construction, or operating environment should not automatically be applied to another.

Capacity-oriented testing complements, rather than replaces, other diagnostic methods. A conventional load test directly observes behavior under a defined load but can be time-consuming and may not reveal all aging mechanisms. Conductance and basic electronic testers are fast, but their results can be influenced by the reference model and may not directly measure remaining capacity. Vehicle battery-management systems can monitor charging behavior, current, voltage, temperature, and starting events, but they may not perform a complete capacity test. Fixed replacement schedules based only on the number of starts are also indirect estimates.

A full charge-discharge test remains an important reference method when time and equipment permit. It is often impractical for a large installed starter battery, however, which is why rapid non-invasive methods are of interest. EIS measures the battery’s response to electrical signals over a range of frequencies and attempts to relate that response to known battery conditions.

How Rapid EIS Battery Testing Works

Electrochemical impedance spectroscopy applies or observes a small alternating electrical signal and measures the resulting voltage and current response over multiple frequencies. The instrument can use the resulting frequency response to create a Nyquist plot, in which the real and imaginary components of impedance are represented graphically. The shape and position of the response contain information about electrochemical and electrical behavior.

In a multi-model EIS approach, the measured response is compared with computational models or a reference matrix containing batteries tested under known conditions. The analysis may estimate capacity, CCA, state of charge, and operational reserve energy. Some systems also describe machine-learning or data-fusion methods, but those terms do not by themselves establish accuracy. Validation against an appropriate reference test is still required.

The quality of the reference matrix is central to the result. The source uses Matrix Integrity Level (MIL) to describe how well a matrix represents the battery population and operating conditions being tested. Adding batteries with different states of health can improve the model’s coverage, especially when the additional samples represent the same general battery type but include healthy, aged, partially charged, and degraded units. A large matrix is not automatically a good matrix if it contains batteries that are irrelevant to the application.

For that reason, matrix selection should reflect the battery’s:

• chemistry and construction;
• nominal capacity and physical size;
• vehicle or equipment application;
• temperature and climate exposure;
• typical state-of-charge range;
• age and operating history; and
• expected failure mechanisms.

A reference set developed for conventional automotive lead-acid batteries may not represent a marine deep-cycle battery, a commercial AGM battery, or a lithium-based system accurately. The same caution applies when a generic automotive matrix is used on trucks, machinery, or batteries exposed to unusual temperature and duty-cycle conditions.

The supplied material attributes rapid EIS testing to a Spectro system and describes an EIS Direct Drive approach for larger batteries. It reports testing of batteries up to approximately 300 Ah, with a typical test time of about 30 seconds in that range. Larger batteries or scans extending below 4 Hz may require more time. The same material reports measurement down to approximately 300 micro-ohms for the later large-battery approach. These figures are product-specific claims and should not be generalized to every EIS instrument or treated as universal industry specifications.

The reference material supplied here does not establish a current universal frequency range, voltage range, or independently verified accuracy specification for all successor devices. Those details should be checked against the current instrument documentation before being used for equipment selection or service procedures.

Measurement performance can also change with temperature, state of charge, terminal connection quality, battery size, chemistry, construction, age, and internal damage. A poor connection, surface charge, recent charging, severe sulfation, electrolyte imbalance, or an operating condition outside the reference matrix can affect the result. A rapid test is only as reliable as the measurement conditions and the model used to interpret them.

EIS can be particularly valuable where a complete discharge test is impractical and where reserve capacity is more important than a simple starting decision. It may also identify unusual responses that warrant further investigation. However, a pass result should not override visible damage, leakage, swelling, overheating, charging-system faults, loose connections, or a known vehicle electrical problem.

A practical diagnostic sequence may include:

1. Inspect the battery, terminals, cables, mounting, and signs of leakage or physical damage.
2. Confirm state of charge and allow for the effects of recent charging or surface charge.
3. Check the vehicle or equipment charging system and parasitic loads.
4. Select a reference model appropriate to the battery chemistry, application, and size.
5. Perform the rapid test with clean, secure connections and record the test conditions.
6. Use a conventional load, capacity, or controlled discharge test when the result is uncertain or consequential.
7. Compare the result with the operating reserve required by the vehicle or equipment, rather than relying on CCA alone.

Rapid impedance testing is therefore best used as one part of a battery-health program. It can improve screening and support predictive replacement when its models are properly validated, but it does not eliminate the need for inspection, charging checks, conventional testing, or diagnosis of the wider vehicle electrical system.

References

1. Battery University | BU-922: What Causes Starter Batteries to Fail?. (n.d.). http://www.batteryuniversity.com/article/bu-922-what-causes-starter-batteries-to-fail
2. Battery University | BU-905b: Knowing when to Replace a Starter…. (n.d.). http://www.batteryuniversity.com/article/bu-905b-knowing-when-to-replace-a-starter-battery
3. Battery University | BU-919: Building a Matrix to test Batteries. (n.d.). http://www.batteryuniversity.com/article/bu-919-building-a-matrix-to-test-batteries
4. Battery University | BU-806: Tracking Battery Capacity and Resistance…. (n.d.). https://www.batteryuniversity.com/article/bu-806-tracking-battery-capacity-and-resistance-as-part-of-aging
5. Battery University | BU-904: How to Measure Capacity. (n.d.). http://www.batteryuniversity.com/article/bu-904-how-to-measure-capacity
6. Battery University | BU-925: EIS Direct Drive, non-invasive Capacity…. (n.d.). http://www.batteryuniversity.com/article/bu-925-eis-direct-drive-non-invasive-capacity-test
7. Spectroscopy checks Battery Capacity in 15 Seconds | Cadex Electronics. (n.d.). https://shop.cadex.com/batteries/spectroscopy-checks-battery-capacity-in-15-seconds
8. Battery University | BU-905b: Knowing when to Replace a Starter…. (n.d.). https://www.batteryuniversity.com/article/bu-905b-knowing-when-to-replace-a-starter-battery
9. Cyber-Physical Cloud Battery Management Systems: Review of Security Aspects. (n.d.). https://www.mdpi.com/2313-0105/9/7/382
10. Battery Failure Analysis. (n.d.). https://www.intertek.com/blog/2021/01-15-battery-failure