All Battery Tests are Not Created Equal (Even when they have the same name!)

Examining the Test Parameters for the Most Common Lithium Ion Battery Standards

Compliance testing can be a very expensive endeavor and also a very confusing one. And when it comes to lithium ion cells and batteries, the expense is magnified by the number of samples needed for testing. Battery testing requires multiple samples per test and in most cases the testing can be degrading and so reuse of samples between tests is not permitted or at least not recommended.

However, when you look at the various battery test standards, many of the tests seem to overlap. One of the most common questions we get as a test lab is “why do we have to complete three short circuit tests using 28 batteries?”  The reason is that, even though the standards used all have a short circuit test in them, these tests are not the same.

In this article, we are going to review battery level testing in three of the most common small format lithium ion test standards, and look at the commonly named tests to explain how they are different.

Key Lithium Ion Battery Test Standards

First of all, we need to define which standards we are going to focus on. There are three standards that are commonly used in the evaluation of small format lithium ion battery packs for transportation and global compliance. These are: 1) the UN Manual of Tests, paragraph 38.3 for transportation: 2) UL 2054 Commercial and Household Batteries for the U.S. market; and 3) IEC 62133:2012 for the international market. The case for when each of these is applicable is an entirely separate topic, but these are the three most common standards for the evaluation of small lithium ion battery packs.

At the battery level, a full test program to each of these three standards would require over 80 battery samples. This can be an expensive and potentially difficult endeavor. Naturally, we want to check for areas to possibly reduce the number of samples and testing required. A quick look at the list of testing in each of these standards would identify only four tests that appear by name to be the same or similar. These would be: 1) external short circuit; 2) overcharge; 3) mold stress; and 4) free fall. However, the number of samples used across all three standards for these four tests would be at least 62 batteries. Only by taking a close look at the details of each test, can we determine if it is possible to reduce the samples and testing required based on similarity of these four tests across three standards.

UL 2054 is a test standard that requires a large number of samples but, unfortunately, the electrical testing in this standard does not overlap with any other battery test standards. This is because UL 2054 electrical testing has two unique requirements, faulting and no trip test limits. Single faulting of the samples under UL 2054 means that the batteries will be opened and any safety critical component that is not UL Recognized will be faulted in a manor to induce the worst-case performance of the component. Typically, this will include faulting of a FET or safety IC. Other common faults would include shorting a sense resistor. The faults will change based on the construction of the battery pack safety circuit. Examples of items that will not be faulted are Recognized PTCs or fuses.

In addition to this faulting of the safety circuit, the test parameters used in these tests are determined by finding a point of either loading or charging that allows the battery to operate while faulted for at least two minutes. This means that a short circuit test under UL 2054 would require multiple faulted units to be exposed to a load that allows the units to operate just below the limit of any secondary protection.

External Short Circuit Testing

Because of these differences, we can remove the UL 2054 electrical testing from the similarity list. This still leaves an overlap of two electrical tests across IEC 62133 and UN 38.3. Continuing to look at the electrical testing, we can do a comparison of the external short circuit and overcharge testing in these two standards.

Table 1 shows the similarities and differences between testing in these two standards. This shows that there is very limited similarity or overlap between the two tests even though they are both short circuit at elevated temperature tests. The circuit resistance used for the testing is one of the areas that could be used as an overlap. If the resistance used is within the limits of the IEC 62133:2012 parameters, it will also meet the UN criteria.

Table 1:  Short circuit test parameter comparison

However, the similarities end there. The IEC testing requires a larger number of samples for testing, and the samples have special charge parameters used to prepare them for testing. Additionally, the criteria used to terminate the testing, as well as determine if the units passed or failed the testing, are different. Finally, the UN testing requires use of four samples that have been conditioned through cycling and all eight of the samples used in the UN testing will have completed four other tests in a sequence prior to being exposed to the actual short circuit.

Overcharge Testing

Overcharge is another test that is common to both the UN and IEC standards. Table 2 summarizes the important parameters for this test between the two standards. With this test, there are no overlapping test parameters. The charge voltage and currents used are different in both cases as well as the criteria used to terminate the testing and determine pass or fail. Finally, the number and condition of the samples to be tested is different.

Table 2:  Overcharge test parameter comparison

Mechanical Testing

In the area of mechanical testing, we have similar testing across only two standards, UL 2054 and IEC 62133. The mechanical testing under UN 38.3 does not overlap with any of the UL 2054 or IEC 62133 testing at the battery level. Within the UL and IEC standards, the two mechanical tests that potentially overlap are mold stress and free fall (drop impact).

Table 3 shows a comparison of these tests. To start, both tests require the same number of samples for testing; however, the mold stress testing in the two standards are evaluating different concerns. UL 2054 is looking for exposure concerns, while IEC 62133 is looking for fire or explosion hazards with the battery. As a result, the UL test is performed on a battery that is fully discharged, while the IEC test is done on a set of fully charged batteries. 

Table 3:  Mechanical test comparison

The final test we will examine for the battery level is the free fall or drop impact test. Again, the testing used the same number of samples for the testing and, in this case, they are both in the fully charged condition. The drop height and number of drops per samples is also the same for both tests. Even though the pass/fail criteria for this test is different for these two tests, the testing could be done and different verification done for each standard. The only difference that would prevent this would be an additional preparation step in UL 2054 that requires batteries rated for use below 0^oC be soaked at the lowest operating temperature for three hours and then dropped cold. If you have a battery that does not have an operating range for either charge or discharge that is below 0^o, then these tests can be considered to be the same. If the battery is intended to be used below 0^oC, then the cold soak and drop could be considered as a worst case as well. However, some engineering justification may be needed to support this.

Potential for Sample Re-Use

From this review, it becomes apparent that there is very little overlap between these three common lithium ion battery standards. There is only one test that could potentially be considered to overlap at the battery level, representing a savings of just three samples from the over 80 needed to complete all three test programs. This leaves re-use as a potential for saving on samples and this is limited as well since many of these tests can be destructive or degrading in nature.

UN 38.3 does specifically state that you can reduce by four the number of samples used if you reuse cycled batteries from T1-T5 for T7. This will reduce the samples required, but it will increase the time to complete since T1-T5 are sequential, and T7 requires a 7 day wait. Looking at IEC 62133, there is nothing in the standard that excludes reuse; however, the charge parameters for the short circuit are extreme and different from the other tests in the standard. This leaves a potential for reuse of samples between overcharge and the electrical testing. This would potentially reduce the number of samples for testing by five.

Finally, looking at UL 2054, the standard dose not explicitly state that you cannot reuse samples from test to test. However, most samples are not easily reused after being exposed to a faulted and no trip condition. You could potentially reuse the mechanical test samples for other mechanical testing, but this only allows for a reduction of six samples. If a manufacturer decides to reuse any samples, there should be a verification of proper function of the samples before they are reused. Typically, this would be running one or two charge/discharge cycles to show they are working correctly and have appropriate available capacity.

Conclusion

In conclusion, battery testing under the three most common lithium ion battery testing standards requires a large number of samples, and there are few avenues available to reduce the required number samples based on similarity or even reuse. Therefore, understanding when these standards are required is very important, as well as efforts to support the global harmonization of these standards. In the past year, there has been some movement on harmonization, with UL releasing UL 62133, which is harmonized with IEC 62133 as well as the Canadian version of the standard. If a manufacturer is able and willing to move to the use of UL 62133 in place of UL 2054, the samples required under the current revisions of the standards would be around 37 instead of over 80.

Cindy Millsaps is President and Chief Executive Officer (CEO) of Energy Assurance, LLC. Prior to establishing Energy Assurance, Cindy worked in global regulatory approvals, quality systems management, product safety and product qualification testing with emphasis on information technology equipment, power/energy and batteries. Cindy can be reached at cindymillsaps@energy-assurance.com.