Battery Testing and Sorting
Currently I have a baseline target of 104 batteries at 2600mAh. That will provide the total capacity that I am targeting, but I also need to evaluate actual power draw and cell load. I plan to cover that later, just know what it's something that needs to be factored in. Right now though I am collecting cells, testing them, and sorting them. So what do we test for? In general terms we are looking for the battery capacity, ability to hold a charge, and internal resistance as measured by AC Impedance. If you start digging into this there seems to be a lot of opinions on how to test and sort batteries, what is most important, and the order you should do them. I'm of the opinion the AC Impedance test should be done first. As Ken pointed out though, to make any meaningful determinations from this test you need to know what the AC Impedance is supposed to be to begin with. So far LG, Samsung, and Sony have the best Product Data sheets that I have seen with LG providing really thorough data for the batteries I have looked into. A good place to start is the Cell Database
over at Second Life Storage. If they do not have your battery listed, and there is a number down the side of your battery type that into Google along with Product Data. If that doesn't turn up any results, I recommend looking for a battery you can test.
So, in my case I'm starting this build using LGABB41865
batteries. The LG Model number is ICR18650B4
. AC Impedance is 70 mOhms or less for my cells.
To test the AC Impedance, you really need to invest in a unit that uses 4 wire testing at 1kHz AC. We could start a whole new thread on this topic alone, but I'm going to leave it at this more or less. The YR1030 units seem to be the best bang for your buck in terms of cost vs decent measurement. For $50-$75 it's a relatively cheap way to add a degree of safety to this project. Having a unit to take measurements with can also help you down the road in troubleshooting if you start having a pack that is giving you fits.
After collecting battery packs, I need to remove the individual cells, clean the nickel strips from the ends, and inspect for any physical damage. WEAR GLOVES AND SAFETY GLASSES. Set aside any that have dents or dings. Those go to the recycler.(Overview video. I advise against the "whack it on the ground" method)
Now that I have a pile of cells, they need to be tested. By checking the AC Impedance now with the YR1030 I get cell voltage and a PRELIMINARY impedance value. I say preliminary because the product sheet specifies testing at full charge. Why not just charge the cells? Time. It can take a while to charge a cell, so I like to sort out anything I don't want to waste time on now. Yes, there will be more getting discarded at each level of testing, but his get's the bulk of them for me. In my first batch of 99 cells, 2 were DOA. Totally dead. 7 of them tested over 70 mOhms(over 85 mOhms at full charge). Of the remaining 90, they tested between 40.8 and 60.5 mOhms. I use a sharpie and write this preliminary value on the side of the battery.
Now that I have sorted out the batteries I'm mostly interested in, I need to charge them and test them for capacity. I use an Opus BT-C3100 v2.2
. There are others out there that in some respects are better, but I like the features of this one and it does well. There is a guy in the forums on Second Life Storage who has run numerous tests on multiple units of different makes and models. The Opus measures up well. I believe his review is in the FAQ now.
Four at a time, the batteries are loaded into the Opus and run through a test cycle. The cycle charges the battery up to full capacity, rests, then discharges it at a set rate until the voltage drops to a set voltage. It then calculates the capacity of the battery in mAh based on how long it took to discharge to said voltage. After it has completed the discharge test, it then charges the battery back to 4.2 volts and blinks between FULL and the mAh rating. I use a sharpie to write that value on the side of the battery below the mOhm value. The cells get set into a tray and the next 4 go in. I'm doing blocks of 36. When I get 36 batteries capacity tested and recharged I let them sit a few days, then run back through the AC Impedance test. If the value is +/-0.3 mOhms I call it good. If there is more that that I will remark the battery. I also check the voltage as the YR1030 provides that along with the mOhms. If any battery has dropped more than 0.5 volts since charging I pull it from the block and monitor it over the next few weeks to see if it really is self discharging or not. Anything that is self discharging gets pulled. Anything that tested over 65 mOhms during the full charge test may be usable in some projects, but I don't want to build it into the powerwall. Particularly when I have 70 battery cells testing 45.0 +/- 5.0 mOhms with capacities between 2650 and 2780 mAh.
So currently I have 70-75 batteries that are excellent and another 15-20 that are good to very good. If I can get another case or two in the few weeks to month I'll have enough to hit my target capacity.