999zip999
100 TW
My new 1220 shows I.R. for ea. Cell.
Arkmundi's new A123 26650 battery build
- Thread starter arkmundi
- Start date
wineboyrider
1 MW
Hyperion 1420i gives you a reading of the internal resistance. I just keep an eye out for any cells that are starting to show ir. And in the hyperion does up to 14s so when I balance and charge a pack in series it's not a parallel pack with a wonky cell. I think this methodology actually, gives us rc charging/balancing types a better idea of what the cells are doing. It just made more logical sense and less wiring to series first, then parallel.
tomtom123
100 W
What bms do you use for this? or none at all?
999zip999
100 TW
Winerboy your 1420 justs shows pack I.R. ? 1220 shows pre. Cell.
arkmundi
10 MW
Custom made box to fit the 12S8P string set, with protection against moisture, heat and shock. Trick was to do that, have it not slide off the back rack and yet provide easy access for deconstruction, and achievable with tools and materials on hand. Stared at it for days before coming up with a good solution. Watts' Up meter to monitor the ride., at 39.5V initially. And ready to go...
Attachments
999zip999
100 TW
Yes Ark I like a low end voltage as I find 3.45- 3.5 to be full. And get very little at 3.6 or 3.8v to worry about. But my pack is 20ah or 17ah if I push it out of the A123 20ah pouches. I due believe in A123...
Jason27
10 kW
How much did this battery building experiment cost you arkmundi ?
arkmundi
10 MW
Interesting. I'm pushing them to 3.3V, 39.6V for the pack on the premised that 80%<SOC<20% will extend the life of the pack. Certainly never above 3.6V for sure.999zip999 said:
Just the cells 100 x $8.75 per. Everything else was from materials on hand. Oh, and the digital soldering station and flat tip for $60. My first battery build was an experiment. This build is hardly that and expect 5 years. So maybe $0.60 a day roughly including electricity. What does a car take these days?Jason27 said:
999zip999
100 TW
I find 3.0-3.9v is big difference. Remember l have 470cycles on two 12s so 940 charges on my 12s charger. Yes a lot watching and triing to understand things.
arkmundi
10 MW
for 2013, $9,122 a year on average according the AAA study, or $25 a day. http://newsroom.aaa.com/2013/04/cos...ording-to-aaas-2013-your-driving-costs-study/arkmundi said:
wineboyrider
1 MW
No, just pack resistance, initially, but if you watch which cells it takes longer to balance, and immediately after charge for all cells that is with the 5.82 firmware update999zip999 said:
Architectonic
10 kW
arkmundi said:
I was actually considering this, but another $300 for inernational shipping, so probably not.
Still, it is a good deal for those in the USA.
arkmundi
10 MW
arkmundi
10 MW
The project is now fully road-tested, tweaked and all is well with the battery build, a 110% success (ie., better than expected). The real test was today, hauling that load of groceries up the hill, after a trip to the YMCA for a swim and other errands. So a 5ah trip and the battery is delivering just fine! Yea! 
999zip999
100 TW
Good job let us know over time. A123. Yea.
arkmundi
10 MW
In the process of doing a periodic check of the individual cells in my 12S8P cylindrical pack, so here's a new angle on that:
Its so far been a relatively easy to accomplish build, therefore within the ability and budget of the common eBike maker. The process can be applied for any of the cylindrical format cells, though I recommend the 26650, and for instance the 18650 format which is more commonly available. Currently working on spec'ng out an arc spot welder for the Technocpia hackerspace, so its even more accessible.
Its so far been a relatively easy to accomplish build, therefore within the ability and budget of the common eBike maker. The process can be applied for any of the cylindrical format cells, though I recommend the 26650, and for instance the 18650 format which is more commonly available. Currently working on spec'ng out an arc spot welder for the Technocpia hackerspace, so its even more accessible.
arkmundi
10 MW
Completed the next step of the project, a homemade analyst for the 12S string as built. Uses two celllogs. This way I can monitor each cell in the string as its charged. Alarm set to 3.62V so if any cell enters an over-charge condition, I pull the plug. My 5 amp smart charger cycles up in five steps, turning on one of the LED's on the front panel as it completes. The final step is special, as it will trickle charge the pack and cycle up & down several times to determine if the pack is fully charged or not. My experience has been that when its almost done, it'll force a cell charge higher than I want. So rather than let the charger decide if its finished, I let the celllogs decide, by alarm.
arkmundi
10 MW
I'm OK, but it was an un-expected test of the battery pack I knew would come, to be slammed yet again to the pavement by a ICEr driver with a blind spot for bikes. Ye'gods I'm on the sidewalk and entering a marked crosswalk and the driver right turns me. Got out and asked if I was OK and gave me a hug for driving off, so a bit more considerate than the usual shmuck on the road. But the joy of seeing my hard-won battery pack survive unscathed!
999zip999
100 TW
I try to ride in a bubble. 2400 watts helps.
arkmundi
10 MW
Two additional tests passed today! First, I had never fully charged the pack and did that today. By "fully charged" I mean by my Smart-charger's decision. Its got some internal logic chip that does charging in 5 steps, throwing an LED on the front when entering the next phase. At the end, all 5 LEDs are green and the charging automatically stops. The last, final step is unique in that it trickle charges at less than the 5 amps of the other 4 steps. It also does minute charge/discharge incremental steps. I believe its testing internal resistance and has presets for what a LiFePO4 36V battery pack should register for IR, Voltage and Amps in/out. Anyway, it went through the full cycle, lit all 5 LEDs and stopped as expected. At the end, voltage was at 41.49V. That's on average 3.4575V per each of the 96 cells in the pack. And that's a very happy result, well below the the max of 3.62V that I'm comfortable with, and above the 3.3V which is the cell's native mid-range.
I've been doing some close monitoring of the pack since its build and use. The cells are amazingly all extremely well balanced, so performing above expectation with regards to being robustly self-balancing. I have no BMS on the pack and have not as yet had to use my Thunder 1220 balance charger. That may change as the pack ages.
Second, on the road today in heavy rain. So the battery pack passed the moisture test, as no leakage. Remember, from earlier in this thread, that I built a custom box for the pack. So this also was good news.
I've been doing some close monitoring of the pack since its build and use. The cells are amazingly all extremely well balanced, so performing above expectation with regards to being robustly self-balancing. I have no BMS on the pack and have not as yet had to use my Thunder 1220 balance charger. That may change as the pack ages.
Second, on the road today in heavy rain. So the battery pack passed the moisture test, as no leakage. Remember, from earlier in this thread, that I built a custom box for the pack. So this also was good news.
docnjoj
1 GW
Perhaps (no doubt) those 26650 cells are higher quality than the AMP 20 pouches I got a year ago. However I need to use the 1220 balance charger after each 20 mile ride. I only use about 12 AH for these rides. but the cells really need balancing each time. With balance they are within 1/100. Good enough for me.
otherDoc
otherDoc
arkmundi
10 MW
Yea, I don't know. I don't think it has to do with the nanophosphate core, as its all the same whatever the format. I think it may be that the cylindricals are pressure constant. Those AMP20s have to be built into a pack with the right pressure packaging, and I never got that part right. But yea, my AMP20 pack needed periodic balancing.docnjoj said:
999zip999
100 TW
I don't see how you can balance charge with a 1220 or put a bms on your pack build as 12s8p. I can't see how a cellog type meter would work.
If it was built as 8p 12s you could balance charge and or bms and cellog. Easier to monitor.
If it was built as 8p 12s you could balance charge and or bms and cellog. Easier to monitor.
arkmundi
10 MW
In total. Its 8 strings of 12S. The celllog monitoring and Thunder balance charge is for one 12S string. Repeated 8 times.999zip999 said:
999zip999
100 TW
So 16 6s balance tabs ? For balance charging ?
