Impedance Spectroscopy Checks Battery Capacity in 15 Seconds

Battery testing can be guesswork and the lack of effective test methods results in batteries being replaced too soon or too late (mostly too late). Some batteries are swapped repeatedly without knowing the cause of failure, but most are left untouched until a failure occurs.

The market is saturated with battery testers. Most read only voltage and internal resistance and don’t work all that well. They do find a dead or dying battery; but so does the user. Batteries have improved and resistance-related failures are isolated. This makes single-dimensional impedance that only read resistance more or less obsolete. The solution lies in two-dimensional methods, but more on this later.

The leading health indicator of a battery is capacity, a value that is difficult to appraise. Many battery testers claim to measure capacity, but this is not possible with voltage and resistance alone, nor will a load test achieve this. Advertising features that lay outside the equipment’s capabilities is misleading and confuses the industry into believing that multifaceted results are attainable with basic test methods.

Battery diagnostics is complex. We don’t even have a reliable method to measure state-of-charge (SoC). Most rely on voltage and coulomb counting; assessing capacity lags far behind. Batteries cannot be measured per se; their health can only be predicted or estimated using indirect measurements. This is synonymous with the weatherman predicting the weather. All findings are estimations with various degrees of accuracy.

Battery testing is not new. The carbon pile introduced in the 1980s applies a load for a short duration. The voltage drop reveals the internal resistance and current flow assures cranking ability. The carbon pile cannot estimate capacity but a skilled mechanic can assess a battery on its kinetic behavior.

The AC conductance meter, also known as impedance tester, appeared in 1992 and was hailed as a breakthrough. The device injects an AC signal to measure the internal resistance. Today, these testers check the CCA of starter batteries and verify the resistive change on UPS batteries. While small and easier to use, these instruments cannot measure capacity.

Critical progress has been made with electrochemical impedance spectroscopy (EIS). EIS is not new. Size, high cost, long test times and required skills to interpret results have restricted their technology to lab environments. To facilitate capacity estimation, Cadex extended EIS and developed multi-model electrochemical impedance spectroscopy, or Spectro™. The heart of Spectro™ is a patented algorithm that performs 40 million transactions while scanning the battery with 20–2,000 Hz low amplitude signals. Figure 1 illustrates the Spectro CA-12 providing capacity, CCA, and SoC in 15 second test.

Configuration is by matrix selection. A matrix is a multi-dimensional lookup table against which the measured readings are compared. Text recognition, fingerprint identification and visual imaging operate on a similar principle. The Spectro™ system uses three types of matrix types: The battery-specific matrix serves a designated battery model and provides capacity readings in the form of a number, the generic matrix accommodates a battery group and gives a pass/fail on a capacity threshold of 40%, and the application-defined matrix inspects performance characteristics of new batteries.

Battery matrices for capacity measurements are created by scanning 10 or more batteries of the same model with different capacity levels. The SoC-matrix is made by reading batteries with different SoC levels, and the performance-matrix is derived by taking a snapshot of a good, marginal and poor battery from the assembly line. Batteries are scanned with the CA-12 and the data is sent to Cadex by the Internet for matrix assembly.

Spectro™ forms a platform that assesses most anomalies and characteristics of lead acid batteries with a single scan. Table 1 summarizes these functions.

Table 1: Spectro™ platform checks characteristics of lead acid batteries. The readings are checked against a matrix. Matrices can be created for various applications.

Spectro™ for Starter Batteries

Ever since Cadillac invented the starter motor in 1912, the SLI (starting, lighting and ignition) has taken on additional responsibilities. “Starting is easy, but can we steer and brake?” drivers ask. The main task of a starter battery was to crank the engine. Today, the duty is shifting towards assuring adequate energy reserve for auxiliary loads and start-stop function. Battery fades unnoticed and a periodic capacity check is vital. A battery behaves much like a horse that gallops until it drops dead.

Most testers for starter batteries are impedance-based and only read CCA (Cold Cranking Amp). CCA governs power delivery that impacts engine cranking and is related to internal resistance. CCA stays stable with age and the readings are ill-suited as a health indicator. Capacity, on the other hand, gradually decreases and serves as a dependable performance predictor. Figure 2 illustrates CCA and capacity of 20 aging starter batteries. All batteries were functional at time of testing.

Figure 2: Capacity and CCA of aging batteries. Batteries 1–9 have good CCA and high capacities; batteries 10–20 are at the end-of-life with capacity loss. All batteries crank well.

“Is there a connection between CCA and capacity?” many wonder. To find out, Cadex examined 175 aging starter batteries and found a correlation with r² = 0.55. Had the correlation been closer to 1, the diamond dots in Figure 3 would have gravitated towards the red reference line. This is not the case, meaning that CCA cannot be used reliably to estimate capacity or predict the end of battery life. A more in-depth assessment of the test results reveals that very few batteries fail due to low CCA; it is the low capacity that eventually pulls CCA down. Note the high number of batteries straddling the 40% capacity cut-off line. Capacity, not CCA, governs battery health and can be utilized to predict end-of-life.

Figure 3: Relationship between CCA and capacity on 175 starter batteries. The correlation is only 0.55. Passing starter batteries dwell in the green acceptance field bound by the 40% capacity and 50% CCA threshold.

The test accuracy of the Spectro CA-12 depends on battery anomaly and the quality of matrix used. A reputable test lab in Germany tested the CA-12 against a competitive impedance tester on a large number of starter batteries. Using the battery-specific matrix, the CA-12 attained correct capacity prediction of 87%; CCA was 97%. In comparison, the CCA accuracy of the impedance tester was only 51% (with no capacity reading). Service personnel may be unaware of the low accuracies impedance testers provide. As a result, many faulty batteries pass as good, only to fail on the road, while good batteries are replaced by error, causing undue expenses for customers.

Spectro™ for Deep-cycle Batteries

Traditional battery testers measure capacity by a discharge and the batteries must first be fully charged to get a correct reading. This is not the case with Spectro CA-12 DC, and capacity testing at partial charge is possible. If below 60% SoC, the unit advises to charge and retest. The 15-second test time does not stress the battery as a full discharge would. The instrument uses a battery-specific matrix and provides numeric readings.

In Cart Mode, the capacity of each monoblock is displayed numerically and graphically. The graphic display allows verifying capacity-match of all monoblocks at a glance. Matching is important for good performance and long service life; weak monoblocks can be spotted and replaced. With the ability to measure capacity on-the-fly, strong monoblocks can be regrouped for a second life.

State-of-charge by Impedance

Troubleshooting electrical systems during fabrication, as well as transportation and showroom activities drains the vehicle battery. A battery should leave the factory with a state-of-charge of about 90%, but verifying this by voltage alone can prove inaccurate. Applying a load distorts the battery voltage, and neutralizing the effect takes up to 24 hours. Voltage based SoC measurements are inherently inaccurate.

The impedance integrity of a battery remains steady with load. This opens the door to measure SoC by impedance and Spectro CA-12 SoC makes use of this. The test takes 15 seconds and can be done with a parasitic load, as well as immediately after charge (but not during charge). Surface charge and voltage agitations have minimal effect on the readings. The instrument uses an application-defined matrix and displays SoC in percentage. Figure 4 shows the relationship of battery voltage and “Spectro SoC” when applying a 50A load. Note the stable Spectro SoC compared to the decreasing voltage.

Checking Quality in Battery Manufacturing

Battery manufacturers are often at a loss why some batteries have low capacity. Checking each battery on a high-volume production line is prohibitive, and yet the buyer judges quality on the lowest performing battery. The CA-12 QA makes performance checking possible by verification against a “golden sample.”

New batteries improve with use, but weak ones from the crib tend to have a shorter service life than the strong ones. Spectro CA-12 QA identifies underperforming batteries and prevents them from entering distribution channels. Performance testing leads to improved manufacturing practices and higher yields. Furthermore, a performance evaluation at the point of fabrication allows for the efficient sorting of batteries into different price classes. Classification and elimination of under-performing battery packs ensures consistent quality and reduces warranty returns.

What does Impedance look like?

Impedance is resistance in an alternating current (AC). Resistance depends on the frequency and produces a real and an imaginary part. The impedance of a battery can be presented by the Nyquist plot in which the imaginary impedance is shown on the vertical axis and the real impedance on the horizontal axis.

Figure 5 illustrates three batteries with different characteristics in a Nyquist plot. The scanning frequency is 20–2,000 Hertz; the upper end represents 20 Hertz and contains the most information about the battery. A healthy battery has low real ohmic resistance and gravitates to the left of the graph. As the battery ages, its spectrum moves to the right an upwards. Good batteries show straight tails and only display a small bend above the neutral line; aging ones shift to the top right and bend over.

Nyquist plots can be used to compare test signatures against a “golden sample” that is taken from a battery with solid performance characteristics. Battery A in Figure 5 represents one. To prepare a golden sample, the user scans one or several good batteries with Spectro CA-12 NQ and downloads the data to PC-Companion software for display. Nyquist also assists in manufacturing and incoming inspection, as well as warranty claims through fault analysis and field service.

Summary

There is no ideal battery test instrument, but scientists predict that the battery industry is moving towards electrochemical impedance spectroscopy. The Spectro CA-12 is one of the first battery testers using this technology outside research laboratories estimating capacity. This significantly improves the accuracy of state-of-health prediction. Cadex is working on expanding Spectro™ to include lithium-ion batteries.

Advanced instruments often use matrix systems but the development of a matrix can be a challenge if no naturally aged samples are available. Stress-aging batteries may be possible but this does not reflect a true life situation. The release of the generic matrix is welcome news, especially in the service sector. The generic matrix for the automotive market comes at a time when the industry recognizes the importance of capacity measurement. The same matrix can test most starter batteries.

Additional information on the Spectro technology can be found on the Cadex website.