Viewing blog entries in category: My Approach to Cell Testing

  • Mooch
    Do I have the qualifications to be doing stuff like this? Some say that I don't.

    You be the judge.

    Since 1992 I have been designing and building electronic devices for a large number of clients. My first products were camera remote-control systems for sports photographers. As these systems, and other devices, transitioned from wired to wireless the performance of the batteries they used was very important as there was no way to replace them during use. This led to me doing more and more battery testing and developing the electronics to charge, test, and protect them.

    As battery technology advanced, from NiCd to NiMH to Li-Ion and now ultracapacitor/Li-Ion hybrids, the devices I built and the battery testing I did advanced along with them. I started specializing in power management electronics; battery chargers, energy harvesters (for charging from heat, light, or vibration/movement), power supplies, battery analyzers, electronic loads for battery discharging, and battery management/protection systems. This is the work I still do today.

    My clients have included the US Army, National Geographic, Sports Illustrated, Eastman Kodak Company, and hundreds of other companies from large to small. Part of the work I've done for a lot of my clients has been battery testing. Sometimes they want to recommend a good battery to their customers. Other times they're batteries I have chosen to combine with my electronics to form a complete, protected power source for them to incorporate into their products.

    Depending on my client's requirements this testing can take up to several weeks to complete. I use a subset of these tests when evaluating the batteries we use when vaping. This includes continuous-current tests to establish the battery's true (and safe) ratings. They also help to determine if there's any risk of venting if one of our mods autofires or a mechanical mod's button is accidentally pressed. The pulse-current tests measure the degree of voltage sag we would see when vaping at different current levels. Both types of tests are done the same way for every set of batteries I test, over 110 different ones to date (almost 400 batteries total).

    This consistency in the testing allows for direct comparison of the performance of different batteries even if the pulse discharging I do doesn't match the way you vape.

    Safety is my number one priority. While I often test at discharge current levels that can result in unsafe battery temperatures, this is the only way to figure out what a battery's true and safe ratings are. No battery is totally safe but we can certainly avoid taking unnecessary risks.

    This is critical.

    There is a huge difference between a battery's rating and a capability of the battery. You might be able to vape with a battery at 40A but that doesn't make 40A the battery's rating. It's just something the battery can do without venting. You still don't how the battery performs compared to others, how much the battery is being damaged, or what the safe limits are.

    A rating is different from a capability because it uses a set of important criteria to establish the rating. Things like temperature, voltage when discharging, cycle life (how many times it can be charged/discharged) are defined and limits are set. This allows for direct comparison of the performance of different batteries and is how I test. The tests determine not only the safety limits of the battery but also the performance limits when vaping.

    In my blog at ECF I have listed the equipment I use and the steps I follow when testing. This allows anyone to replicate my tests if they want to:

    My cell testing equipment and setup | E-Cigarette Forum

    What's done for each cell test? | E-Cigarette Forum

    Does all this make me some kind of battery expert? Hell no. But I do feel I am qualified to do this testing. My results offer you a resource you can use when choosing a battery that will not only be safe for the way you vape but will also give you great performance.

    If there's something you don't like about the testing or the ratings/performance tables, let me know! I'd be happy to read what you have to say and discuss it with you. Over the past few months the feedback I've gotten has resulted in some good changes to the tables to make them less confusing and easier to read.

    Each of us has to decide which battery tester's results we will use. Different testers use different criteria when setting a rating or when comparing batteries. Find out how they test, compare their results, and pick the tester you trust the most with the batteries you use.

    Thanks for your time!
    Katya, Microzod, KS_Referee and 19 others like this.
  • Mooch
    Updated 10/24/15 to include pulsed discharge tests.

    For every cell I test, these are the steps I follow...

    Continuous Current Tests
    • I use two of each cell for testing.
    • Photograph the wrap from one cell and top if a button has been spot-welded on.
    • Remove the wrap and photograph the case, top, and bottom.
    • Attach the thermocouple (temperature sensor) halfway down the cell with Kapton tape, making sure to cover the tip of the thermocouple to prevent any flowing air from cooling it.
    • Clean the test rig contacts with a Scotch-Brite pad and then a 90% alcohol wipe.
    • Mount the cell in the low-resistance test rig.
    • Charge 18650's and 26650's to 4.20V at 2.5A until the current drops to 100mA. 18350's are charged at 0.5A unless the manufacturer specifies a higher rate.
    • Run three constant-current (CC) discharges down to 2.80V to check basic cell functionality, including capacity and temperature. 18650's and 26650's are discharged at 10A. 18350's-18500's are discharged at 5A.
    • If all three discharges are essentially identical then I continue. If something keeps changing for each discharge I keep running them until the cell's performance has stabilized. So far, every cell has stabilized within three discharges.
    • For each discharge I measure the actual current level using a 0.25% tolerance current shunt and a Fluke 8846A meter. This not only confirms the starting current level but by using the min/max/avg functions of the meter I can confirm that the current level has not drifted.
    • Run CC discharges, down to 2.80V, at every 5A increment above that until the cell reaches 100°C or the voltage just quickly collapses.
    • Note the maximum cell temperature reached for each discharge.
    • After each discharge let the cell cool to below 40°C before recharging.
    • Recharge each cell to 4.20V, stopping when the charge current has dropped to 100mA.
    • Determine the cell's continuous discharge rating (CDR) by noting the current level that brings the temperature closest to the 78°C average (74°C-82°C range) I measured for the Samsung, Sony, LG cells I tested at their CDR.
    • Run an additional two CC discharges at the cell's CDR to check for voltage sag, loss of capacity, or increasing temperature. These are all signs of cell damage and indicate that the cell's rating is too high.
    • Run an additional two CC discharges at 5A above the cell's CDR to check for voltage sag, loss of capacity, or increasing temperature. These are all signs of cell damage and indicate that it's being discharged at beyond its rating. It also gives us an idea of hard it can be abused.
    • Take the second cell, run the three initial discharges, and then discharge at 10A and at the CDR of the cell. If the results are within 2% of the first cell then the first cell's discharge graph is used. If the discharges of the second cell are different from the first I do not post any test results until I can source another set of cells to test and compare.

    Pulsed Current Tests
    • Discharge the second 18650 or 26650 cell at 30A, each pulse is 5 seconds on/30 seconds off, down to 2.50V. 18350-18500 cells are started at 10A.
    • A lower cutoff voltage is used for the pulse testing to give those cells that have a significant increase in voltage when hot (due to lowered internal resistance) a chance to warm up.
    • Run pulsed current discharges, down to 2.80V, at every 5A (for 18350-18500) or 10A (for 18650-26650) increment above that until the cell reaches 100°C or the voltage drops to 2.50V for the first pulse.
    • After each discharge let the cell cool to below 40°C before recharging.
    • Recharge each cell to 4.20V, stopping when the charge current has dropped to 100mA.
    • Note the maximum cell temperature reached for each discharge.

    I don't have a standard yet for determining the pulse rating for a cell. When I have enough pulsed current discharge data I will give each cell I test a pulse rating. In the mean time you can view the discharge graphs to see what the voltage drop is for the cells I have been testing recently. All of the Samsung, Sony, and LG cells are being retested to add this pulsed current data.

    Important Notice!
    Testing batteries at their limits is dangerous and should never, ever be attempted by anyone who has not thoroughly studied the dangers involved and how to minimize them. My safety precautions are the ones I have selected to take and you should not assume they will protect you if you attempt to do any testing. Do the research and create your own testing methods and safety precautions.
    Ariel_MX, CG138, dcfluegel and 3 others like this.
  • Mooch
    As an example, suppose one cell had a 30A continuous discharge rating (CDR) and another cell had a CDR of 20A. The 30A cell sounds like it would be a better choice for use in a device that draws lots of current. But if it rises to a temperature of 120°C delivering that current, that cell is overrated and will have a very short life...if it doesn't vent or burst first. The 20A cell might actually be a better choice depending on how hot it gets. Something you could determine if you had the temperature data.

    I feel strongly that temperature must be a part of any cell testing or otherwise the tests can't be used to compare cells. These tests are my first step in trying to get a handle on how we can give these cells true CDR's. Based not only on capacity and voltage-under-load but also based on how safe they are to use at different discharge current levels.

    Manufacturers rate their cells for use at temperatures up to 60°C, maximum. At temperatures exceeding about 45°C a cell's aging accelerates, shortening its life. At 70°C-80°C a cell starts increasing its self-heating due to additional exothermic chemical reactions. If this self-heating is not stopped, or the heat pulled away by cooling, it will eventually lead to venting, bursting, and possibly thermal runaway. At approximately 120°C an important component in a cell starts melting (the separator), leading to short-circuiting and more self-heating. This is a point where the cell starts to be in big, big trouble. And so is anyone using that cell.

    Different chemistries have different temperature thresholds for thermal runaway but all suffer similarly at temperatures below this (accelerated aging, exothermic reactions creating gas and increased internal pressure, separator melting, etc) which can lead to venting and/or bursting. It's why I did not differentiate between the chemistries when setting the maximum temperature I would let a cell reach before stopping a test.

    Testing ICR cells (LiCoO2, "lithium-cobalt-oxide", "LCO") is riskier than testing the IMR (LiMn2O4, "lithium-manganese-oxide", "LMO") cells we normally recommend for use in a vaping device. This is due to the lower thermal runaway temperature of lithium-cobalt. It makes measuring of the cell temperature during a discharge all the more important.

    I have set a safety limit of 100°C for all of my tests, which is a ridiculously high temperature to operate a cell at! But I know that vapers will always want to reduce device size by reducing the number of cells so we'll go as hot as we can without getting too close to thermal runaway. Know that operating at over about 45°C reduces cell life though. If the cell exceeds 100°C before completing a discharge at its continuous discharge rating (CDR), then the cell is definitely overrated. It's just to dangerous to use continuously at that discharge current level. Under certain circumstances I'll let the discharge continue even if the cell temperature is above 100°C. But this is guaranteed to damage the cell and might lead to venting or thermal runaway.

    For reasonable cell life, I have set a limit of 75°C. While this is high enough to speed up the aging of the cell, it will still allow using the cell for a reasonable amount of time before needing to replace it. Beware of using any cell at higher temperatures than this. Not only can the damage become quite severe very quickly but it also takes you closer to the temperature at which the cell could vent.

    I realize that vaping does not discharge the cell continuously and that it will run cooler when used in a device, even if each time the device is fired it draws current equal to the cell's CDR. But we must have a safety margin when using these cells! If a device autofires then knowing that the cell you have picked will not vent, or worse, is very important. And a cell that is short-circuited might not destroy itself, and your device, if we can pick the one that runs cooler at high discharge current levels. This can only be done if we know how hot these cells get.

    I would love to see the ECF community come together to create a set of standardized test requirements to use when comparing cells and determining their safety at different discharge current levels. Using these tests we could set an accurate and safe current limit for each cell. Not just for continuous current, but also for "pulse" current testing that better simulates what happens when cells are actually used in a device. Additional tests could include cell leakage rate (good for estimating degree of damage to a cell), internal resistance, total joules delivered for each discharge current level (not a test, just some math), and cycle life testing.
  • Mooch
    This is the equipment I use for cell testing:

    • West Mountain Radio CBA IV Pro battery analyzer, modified for low voltage drop . Accuracy = +/-1% up to 20A.
    • Custom constant-current electronic load, rated 150A/400W. Accuracy = +/-0.6% up to 150A. A second load is available, if not being used for other testing, if I need to discharge at over 150A (up to 300A).
    • Adjustable 5A/30V CC/CV power supply for charging the cell.
    • Omegaette HH308 dual type-K thermocouple thermometer. Accuracy = +/-0.3% + 1C.
    • 20A, 100A, and 200A current measuring shunts. Accuracy = +/-0.25%.
    • Fluke 8846A 6-1/2 digit DMM. Accuracy better than +/-0.01%.
    • Low-resistance cell clamping rig.
    • Safety glasses, fire-resistant apron, fire-resistant gloves. I wear all when doing destructive testing or if I think the cell temperature will rise much above 100°C. Otherwise just the safety glasses.

    • The CBA battery analyzer handles 10% of the discharge current and creates the graphs. The custom load handles 90% of the current.
    • The cell has its wrapper removed so the thermocouple can be placed directly against the metal can to sense temperature.
    • The thermocouple, plus 0.5" of its cable (to prevent ambient air from cooling it and affecting the thermocouple), are tightly Kapton-taped to the cell with the thermocouple positioned halfway down the cell.
    • All discharges, unless noted, are constant-current to within +/-1% of the stated value. Confirmed via +/-0.25% tolerance current shunt.
    • All measured temperatures were rounded to the nearest degree-C. Only the highest temperature for the discharge is recorded.
    • The cell is placed in an insulated fireproof container with a lid loosely placed on top. The container is left open at the top to allow heat to escape during discharge. If a cell vents the cover is closed and a fan is turned on to evacuate the gas/vapor outside via a flexible metal hose.
    • The cell holder is a non-conductive c-clamp using a 1" x 1", 0.040" thick copper plate to connect to the bottom of the cell. A 1/4" diameter copper rod is used to connect to the top of the cell. Both the top and bottom connections have 10AWG and 12AWG silicone-insulated wires soldered to them. The 10AWG wires go to the 400W load and the 12AWG wires go to the CBA, directly soldered to the CBA circuit board. I do not use the CBA's PowerPole connectors as their resistance is too high.
    • Why only a 1/4" copper rod for the top of the cell? One reason is that's the equivalent of 3AWG wire...plenty beefy enough. The other reason is that I don't want to conduct any more heat from the cell than I have to.
  • Mooch
    The performance of a cell over months of use in our devices doesn't just depend on that how that cell does for one or two discharges when new. It depends on how fast the cell ages over time.

    What affects this aging? Assuming you're not overcharging or overdischarging, temperature has the biggest effect on your battery's health. The hotter you run them, the quicker they will start to lose capacity and run at a lower voltage.

    To better determine how a cell will perform over time I run multiple discharges at the cell's specified or accepted continuous discharge rating (CDR). If the cell gets too hot, much over 75°C, then I know that the CDR rating is too high for that cell. I then run another lower-current discharge to measure the capacity loss and damage that the high temperatures might have caused. If there was a capacity loss then I have confirmation that running at the cell's specified CDR is harmful. This means that the CDR needs to be lowered.

    When picking a CDR I'll select the discharge current level that results in the cell temperature rising no higher than approximately 75°C. While this is an extraordinarily high temperature to operate a cell at, it does allow for powering the small devices we use now while still maintaining decent battery life and safety margins.

    If you see test results elsewhere that rate a cell a lot higher, or lower...ask questions! What temperatures did the cell reach during testing? Did the tester do multiple discharges to see if the cell started to show signs of damage at its CDR? What method did the tester use to measure temperature? Some popular methods are very poorly suited for measuring cell temperatures.

    All these questions may seem overly geeky but the answers will make you a better, safer vaper and help you to pick the tester(s) you trust to review the cells you're putting up to your face.

    Another question might be, why do I say "cell" instead of "battery"?
    To prevent confusion with the portion of a vaping device that powers the coil, often (confusingly) called a "battery", I use the term "cell". It's also the term used by cell manufacturers and my clients so I'm just used to it. :)