Finished Building My A123 20ah Battery Pack

RayGo

10 W
Joined
Oct 24, 2012
Messages
66
I finished my build of a 12s 36v 20ah A123 battery pack back in February but I am still running some load tests. I’m happy so far with the results and had fun with the build process. Thanks again to all the posters on this forum for all their valuable contributions. I knew little about lithium batteries prior to coming across this forum. The NiMH batteries on my e-bike died a while ago and I needed a replacement battery system. Anyway, what follows is the recap of my battery pack build process.

I originally considered buying a Headway 36v15ah battery kit using 40152 cells including charger for $414 with free shipping. But then I came across this forum and all the talk about A123 20ah cells. I liked the idea that I could get 5ah more for about the same weight with the A123 cells. After viewing packs built by other ES members, I decided to go with the A123 cells. Then came the tough decision of where to get the A123 cells. I finally decided to take a chance with Victpower at the end of last year purchasing 13 cells at $14.70 each plus $114 shipping and $12.30 processing fee for a total of $317.30. I got “salvaged” short tab A123 cells from Victpower. I have since determined that one cell’s capacity is weak at 16,746ah when discharged down to 2.5v and another cell has developed a tiny black “leak”. The other cells appear to be within specs. However, I now wish I had ordered at least 2 spares instead of 1.

There still seems to be a lot of confusion over where to buy these cells. I consider myself lucky with my Victpower purchase. However, it now seems that Victpower is running low or out of the short tab cells and cell quality seems to be getting worse. If I had to get cells now, I would probably buy from OSN Power and get new full tab matched cells made in the USA or Korea since other ES members seem to have had reasonable results with them.

I spent about another $70 for parts from HobbyKing. That included 2 Celllog-8M voltage monitors and miscellaneous wiring and connectors. I also picked up a used Thunder 1220 smart charger for $50 and two 575 watt HP power supplies for $25. Add about another $30 in hardware from McMaster-Carr and Home Depot which included bolts, washers, nuts, ring terminals, vinyl tubing, aluminum strips and polycarbonate side pieces. So my total costs for the pack build ran about $500.

I tried to keep the pack costs down while building a safe setup that could easily be taken apart for repair. I also wanted to have compression on the cells to minimize cell puffing. This build is fine for my recreational e-bike usage drawing 30a max. After I began my build process, OSN Power came out with their 2 build kits which I would have considered using had I known about them earlier. However, they require the 12mm short tab Korean cells or full tab cells to work with their kits. They can be used with 10mm short tab A123 cells but only with “careful” cutting into the cell which I did not want to do.

Interestingly, OSN is now selling their 36v 20ah modular pack with BMS and 5a charger for $546 including shipping. So, it’s almost not worth doing your own build assuming you get good matched cells, their 1 year warranty is honored and you don’t get “killed” in return shipping costs. However, I prefer the use of voltage monitors for my pack rather than a BMS and OSN only builds the packs with a BMS. So if you don't want the BMS, it's easier to just order the cells and build kit parts rather than the assembled pack from them. I would stay away from the Victpower pre-assembled packs due to their soldered tab connections making for a tough time to repair the pack. Also, the quality of Victpower's assembled packs leave a lot to be desired.

I have included some photos of the build below. I'll post my load tests results and findings later.

pack 4.jpg
pack5.jpg
pack 1.jpg
pack7.JPG
View attachment 1
bike 2a.jpg
 
Great work!

Where did you got these JST connector holder grip? ( on teh balance connector).. The transparent plastic part that seem to clip on the white jst connectors?

Doc
 
Hello Zip!

I'm afraid I actually don't know that much about the motor specs. I bought this assembled "concept" bike back in 2004 from a company called Burro-ZVO Bikes here in SoCal and they have since gone out of business. The motor model is the X-Power X-500. It is a 36v, 1.5hp motor. Specs are 221 rpm, 81 NM torque, 60 amps peak, 17 amps input, 30 amps avg. I can go 20 - 40 mph with no pedaling. I'll be posting my load testing and findings with the new A123 pack later in a separate reply.

Doctorbass:

You can also get the balance plug savers at the site below. They are in Ohio so shipping is cheap and fast but you need to check what is in stock. I got mine from HobbyKing and they appear to be in stock in the USA warehouse so shipping should also be cheap and fast.

http://www.buddyrc.com/6s-ab-clips-5-sets.html
 
So I finally got around to running some load tests on my pack. My Thunder 1220 smart charger allows me to make a 5a discharge setting, but for whatever reason will only discharge the pack at 1a. Not wanting to wait almost 20 hours for a discharge test and not having a separate wattmeter to make a discharge setup, I decided to ride my bike with its in-line wattmeter for the load test. I was averaging about 1ah per mile with no pedal assist over fairly hilly terrain. I gave up after about 12 miles due to the cold and windy weather. I wound up making 3 runs over 5 days due to the weather. I kept track of the ah used and to my surprise, I got over 23ah out of the pack when the battery started sagging and the voltage finally dropped down to 28v. I knew the voltage would recover after resting but I assumed the ah capacity of the pack would be “fixed” since the voltage drops off quickly after restarting each time. So much for that theory and test results. However, it did help me in identifying a weak cell at 1.76v when the rest were all around 2.5v. There is no BMS in my pack and I did not hear the Cell-log alarm with the LVC set at 2.1v.

So I balance charged the pack to 3.65 per cell and ran the discharge test at 1C and low voltage cutoff set to 2.5v per cell. After 17 hours, the test ended with 17.359ah with the weak cell cutting off the test at 2.5v with all the other cells at 2.9v. So I took the pack apart and replaced the weak cell with the spare and ran the load tests again. I finally completed the discharge test and got 18.699ah with the LVC set at 2.5v. My Thunder 1220 charger will only discharge down to a minimum of 2.5v. So I'm pretty happy with the Victpower A123 cells.

Following are some of my other findings. My cells charge up to 3.65v but then drop off overnight to about 3.567 - 3.579v. They tend to stay in the 3.2 to 3.3v range the bulk of the time and seem to stay in balance after use. Even a bulk charge rather than balance charge keeps the cells pretty much in balance. My Victpower short tab cells pretty much perform the same as described in two other topics in this forum: How important is balancing a battery pack? and How long does a a123 20ah hold 3.65v?

http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=48258#p710240
http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=46939

My Internal Resistance value for the pack was 21.8 micro-ohms at the start of the charge cycle with individual cell IRs ranging from .4 to 3.9mo. Then the pack IR value dropped to 18.3mo at the end of the charge. IR values get lower/better as the cells reach their HVC. Any idea what are good IR values for individual cells and when should the IR checks be made?

So as I stated earlier, this build works well for my e-bike recreational use to 30a and using A123 cells with short tabs. I also like the drutledge or 999zip999 type builds if you have full tabs and don’t mind drilling holes in the tabs. However, I prefer the OSN modular build kit based on price and its clean approach to separating and clamping down the tabs. Based on what I’ve read in other topics in this forum, if you have higher amperage needs, use the agniusm kit or modify the modular OSN kit by adding a nut and bolt to the middle of the tab clamps to even out the clamping pressure. You can find info on these other builds and kits in "A123 20AHr Pouch Cell Battery Build & Info Thread".

http://www.endless-sphere.com/forums/viewtopic.php?f=14&t=38761

My pack weighs in at 16 pounds. I also wanted Cell-logs rather than a BMS since I believe you still need to manually monitor the pack. You cannot assume the BMS works correctly or worse, it might fail at some point with no warning. I agree that periodic human monitoring is necessary to protect this expensive investment.

So, hopefully this helps others in building their own pack.

I do have a question about prolonging the life of the battey pack. My rides are usually short runs about 12 miles and use up only about 5ah. Is it better to immediately charge the battery back to 43.8v or 3.65v per cell or is it "easier" on the battery if I make 1 or 2 more trips and run the pack down to about 75 - 80% DOD and then recharge the battery?
 
I really don't like my cells below 3.1 v after a ride and time to check with my hyperion sentry monitors. I bet under load it hit's 2.9v or a a little bite more. Maybe 2.8v or so just a guess. I only will take out 17ah. I have my 9 fet only draws 30amp. I had aproblem for 4 months and couldn't charge but check and all stayed at 3.4- 3.35v. after 4 months no bms. Victpower full tabs gotten early. I charge to 3.59v. and they float down to 3.48v or so if left over night. I follow the durtlegde build.
P.s. Thanks Raygo for helping me Win the lunch date with Nancy 5 months ago. Still saving the money for the plane ticket.
 
Back
Top