Playing with A123 20Ah lithium cells

[Daox]

I bought my cells from Eric Powers (user name ericbecky). He did a group buy on the cells and I reserved a few before they were all gone. He is currently out of stock but was considering doing another group buy when I talked to him last weekend.

[Ryland]

There are a few people selling them on Ebay too a123 20ah | eBay

You can also buy them directly from China http://www.alibaba.com/showroom/a123-20ah.html

From what I've heard how these came to be is that the factory that is making them had a contract to make them, tooled up to make them then something changed and A123 systems was no longer going to use that factory, no idea who's choice that was, but the factory decided to make them anyway and sell them on their own. so as I understand it, it's a little shady, but they are still good cells and people are using them for good projects.
The same kind of thing happened with the E-tek motors, I did some work for one of the guys who helped design that motor and they tooled up in China to build it but before it was supposed to be released they found out that the factory had opened early and was selling it on their own, law suits fallowed and the improved version of the E-tek motor was never released.

[oil pan 4]

Thanks.
Next starting battery I build will be based off something like these.

[Daox]

The vendor Eric gets his cells from, Victpower (min order of 300 cells), has a pretty good reputation. I would be very cautious buying from elsewhere. As Ryland said there are a lot of grey market cells floating around, some without the tabs on them, others may not even be A123 cells.

[oil pan 4]

I may have gotten ahold of those grey/black market cells for my 28 cell LiFePO4 battery. The 28 cell battery seems to have about as much power as its 20 cell counterpart.

I will be looking the next group buy, don't need them now but when I do I will want between 4 and 12 cells.

[Daox]

Originally Posted by Ryland

I forgot all about the dissimilar metals like copper and aluminum, they should not be in contact with each other! aluminum oxide is also not conductive, so I'd get something like a scouring pad, to clean up your contacts then grease them with dielectric grease or conductive aluminum never seize on the aluminum side and copper never seize on the copper side.

I posed this question to the local makerspace and got some interesting answers.

1) Aluminum oxidizes pretty much immediately.

2) The battery manufacturer knows this and likely plated the aluminum with something that doesn't oxidize. This doesn't help my spacers though. So I'm still trying to figure out what to do there. I'm still not even sure that the very thin layer of aluminum oxide will create that much resistance.


Also, I think I mentioned it, but the copper lead is plated with something. I'm not sure what. My first guess would be tin, but it might also be zinc. However, zinc is much farther up on the galvanic series than tin...

[Daox]

Just got some more info.

Apparently the copper terminal is plated or is made of 'aluminized copper', so attachment to an aluminum spacer should not be an issue.

Also, a makerspace member recommended "noalox". It claims to:

•Anti-oxidant
•Reduces galling and seizing on aluminum conduit joints
•Suspended zinc particles penetrate and cut aluminum oxide
•Carrier material excludes air to prevent further oxidation
•Improves service life of aluminum electrical applications
•Costs a lot!

Ok, last one I added. But, with that new info and some noalox I think I should be in pretty good shape.

[Cobb]

Oh, ok. Yeah, I am sure those will do better in your gasser vs your hybrid. Id even add a solar panel too if you were me.

Ive wanted to replace the lead acid battery in my Insight with something like a few prius sticks, IMA sticks or a lifepo4 battery for some time now. Price has been the main hold up. From what I see the Honda charges the battery for the first few minutes after start up, then when the voltage falls before 12.6 volts. The only exception to this is regen braking. When ever you regen it also activates the charging system. As we know lead acid batteries need a long continue low amp charge vs a short high amp, whish is what the Honda does by way of the ELD and a dc to dc converter.

Although I replaced the motorcycle battery the car came with with a group 51, I did find 2 more mpg from disconnecting the power steering and see an effect with use of the lights, drls, hids vs halogen, etc.

Good luck.

[Daox]

I've pretty much finished up the first round of testing on the first cell. Capacity wise it isn't quite up to the 20Ah spec (I don't have the 1C discharge Ah number handy), but its fairly close. I also found that with my mottcell cells that sometimes multiple cycles tended to increase the capacity almost like the cells needed to be exercised. However, this wasn't true for all the cells I tested.

Anyway, here are a few charts.


The first test was to charge/discharge at .5C (10A). My laptop went to sleep during the charge so I only have the discharge which is thankfully the more useful (to see voltage sag).







I have a few more charts from my 1C (20A) tests. The first shows the voltage over time for the discharge cycle.





This shows the volts over time for the charging cycle.




Below you'll see the Ah put into the cell over time.

With these two charging charts, I can fairly accurately estimate the state of charge (SOC) that these cells will sit at when they are connected to my alternator on the Paseo. For example, lets say my alternator puts out 14V. 14V gives us 3.5V per cell (well below the 3.65V max charge voltage, so no need for a BMS). By looking at these charts we know the cells hit 3.5V around the .9 hour mark. By looking at the capacity chart at the .9 hour mark it looks like I'll have close to 17Ah stuffed into the batteries. This is actualy quite a bit more than I thought which is nice. Though, I do have to verify my alternator output. If the alternator puts out 13.5V (3.375V/cell) we're looking at a capacity of around 10Ah. So, a little voltage variance goes a long way.









Next up is the 2C (40A) tests. You can see the voltage is starting to sag a tiny bit more, but its very reasonable. I don't have a mottcell chart handy right now, but I'm pretty sure there is more sag at 1C on my mottcell cells than there is at 2C with these cells.

Here is the voltage over time for the discharge cycle.




Here is the voltage over time for the charge cycle.




And here is the capacity over time for the charge cycle.

[Daox]

I will also mention I monitored cell temperatures here and there. The cell was simply sitting on a board, no insulation or end plates squishing things together.

At .5C there was almost no noticably increase in temperature. I shall recheck this when I test the other cells.

At 1C dis/charge I saw a pretty constant 4F above ambient in cell temperature. The cell heated up very uniformly. The terminals I thought would heat up as well, but they always remained at a lower temperature than the cell. This may be due to the large aluminum blocks on them.

At 2C there was a good temperature increase. Probably more like 10-12F. I'll also have to verify this because I'm not 100% what ambient was.