RayGo said:
wb9k - Thank you very much for all your valuable input to this forum.
Now I have a follow-up question for you about the A123 20ah cell's Internal Resistance values. I was always wondering what were considered "good" IR values for these cells. I have the Thunder 1220 which measures the IR values of my 12s pack during the charge cycle. Recently, I got a total pack IR value of 43.6mO at the start of the charge cycle and then the value dropped to 21.8mO at the completion of the cycle at 42.97v. I did not record the individual cell IR values. The pack had been discharged at that time by about 13ah. I set the individual cell voltage to charge to 3.65v but have dropped that to 3.60v based on your advice. I don't know how to interpret these IR values given that your A123 cell spec is 0.3mO.
Back on April 3rd, I recorded these values at the end of my charge cycle: 43.7v & 30.5mO pack values; avg 3.637v. & actual cell IR at 2.9, 2.6, 3.3, 2.3, 2.3, 1.9, 1.6, 2.9, 2.9, 1.9, 3.6, 2.3. The IR values seem to always be changing from charge to charge. On June 19th, I got pack values of 43.61v & 40.1mO.
What do you make of these values? Thanks again.
Hi RayGo,
The answer to many of your questions is in my earlier post (end of second page), particularly this bit: "Impedance is dynamic not only with regard to temperature, but also with C rate and SOC. Meaningful apples-to-apples comparisons dictate that all measurements be taken at the same temperature (ambient and actual cell temp), SOC--nominally 50%, and C-rate--nominally 1C." So, if you're not taking a good deal of care to make sure you're measuring under the same conditions every time, you should expect variation in your numbers. I see nothing terribly amiss in any of the numbers you report here. Do re-read the post I quoted from here, because you seem to have misunderstood a few other things. For starters, A123's spec is 3 mOhms, not 0.3. That's a typical measurement for a brand new 20 Ah cell USING AN AC mILLIOHMMETER, which is a totally different test from what we're talking about here. So let's clarify the test we are talking about now.
I'll describe a test we used for warranty service on modules from a large vehicle pack. I like this test better than other methods that are more common because I think it gives a more meaningful indication of what's going on with pack and cell health, which is what your concern should be. The modules were 26650 cells in 12S8P, for a nominal rating of 40V and 18.2 Ah. The total test included a charge and balance, capacity test, and simulated drive cycle test. But before all that comes an impedance test. All of the following is important:
1) The module must have been at rest in an evironment held at room temperature for a minimum of 24 hours prior to testing. No charging or discharging allowed AT ALL immediately prior to the IR test. If stored below ~68F or above 80F, the results of the test are not valid "hard" numbers (though, significantly, the rest of the test
is valid under those conditions. In other words, actual module performance trumps the single number IR result at the beginning of the test. We get our "hard" IR number at a later test anyway; this measurement I like so well is really just for our reference--but it's very useful.)
2) The module must be between roughly 30 and 70% SOC. If SOC is too low or too high, the Voltage moves too quickly wrt to SOC, invalidating the results of the test. Notice that this concept basically invalidates the numbers your Thunder test is giving you--if I understand your description of the test above correctly.
The test itself:
3) Starting Voltages of the module and each individual cell group are recorded.
4) Discharge at 2C for 10 seconds and record minimum Voltages reached.
5) Ten seconds of rest.
6) Charge at 2C for 10 seconds and record maximum Voltages reached.
7) Use Ohm's law to calculate resistance values from these measurements.
8 ) Average the charge and discharge results together. These are your figures for impedance.
The thing to do is to characterize your pack when it is new and you have high confidence that things are working as they should. Keep these numbers as a baseline for comparing future tests against. Usually when things go wrong, it's a cell here or there that is straying from the others. You want to watch out for cell groups with IR numbers that stray well away from the mean of the other cell groups,
or all cell groups experiencing radical (inevitably upward) leaps in value together. You can use the impedance numbers to do this, but I found it actually more useful to stare at the Voltage curves during these measurements. Vmax-Vmin is the number I liked best. A cell that's wrecked may show 3.2 Volts OCV, but you put a 2C load on it, and voltage falls to 2.0. Put a 2C charge on it, and it rockets to 3.9 Volts. With this particular type of module, the best cells have a Vmax-Vmin of <300 mV. Cells in the 300 to 350 mV range were a "grade" lower if you will, 350 to 400 mV another grade lower yet. These numbers are for groups of 8 cells in parallel. Above 400 mV, the cells would have difficulty passing the remainder of the test, failing either the minimum capacity test or drive cycle test. Even many of those cells have life left in them, but not enough for this application.
There's more. Pack or module-level impedances are not just the sum of the impedances of the cells. The impedance of the connections to and between the cells are also added, and a pack impedance test is a good way to find connections that have become unstable. That's really the main value in a pack level impedance test.
That's quite a bit of info. Hopefully you'll find it useful.
dh
