DIY stage 9: A123 "nano" battery pack [photos]

[daniel.weck]

Table of Contents:

http://www.pedelecs.co.uk/forum/electric-bicycles/4682-diy-stage-1-received-bafang-tongxin-kits-photos.html#post59183

http://www.pedelecs.co.uk/forum/electric-bicycles/4734-diy-stage-2-received-li-ping-lifepo4-battery-photos.html#post59988

http://www.pedelecs.co.uk/forum/electric-bicycles/4737-diy-stage-3-motor-testing-fork-photos.html#post60039

http://www.pedelecs.co.uk/forum/electric-bicycles/4747-diy-stage-4-16-wheel-lacing-building-photos.html#post60153

http://www.pedelecs.co.uk/forum/electric-bicycles/4769-diy-stage-5-prepping-fitting-electricals-photos.html#post60345

http://www.pedelecs.co.uk/forum/electric-bicycles/4778-diy-stage-6-fork-hub-torque-protection-photos.html#post60504

http://www.pedelecs.co.uk/forum/electric-bicycles/4786-diy-stage-7-first-ride.html#post60641

http://www.pedelecs.co.uk/forum/electric-bicycles/4825-diy-stage-8-continuous-improvements.html#post60991

http://www.pedelecs.co.uk/forum/electric-bicycles/6956-diy-stage-9-a123-nano-battery-pack-photos.html#post88165

------------------------------------------

Time for the A123 treatment !

I need my 10Ah LiFePO4 pack for many of my leisure rides, but for the shorter utilitarian trips, the much lighter and smaller 2.3Ah pack is ideal, especially as it fits perfectly inside the rear pocket of the Brompton C-Bag.

I will update this thread when I get more statistical data from my Cycle Analyst about the performance of the A123 battery pack. Meanwhile, here are some photos

I purchased two ready-made 6S1P packs from this eBay seller:

eBay My World - sz_ohmygod

ElectricWingman sells the same packs (if we trust the photos)...albeit at twice the price point:

6s A123 Battery Packs - 19.8V

They weigh 963g in total, including the charge/discharge Dean-terminated wires and the balancing tabs:

 

[daniel.weck]

I linked the 2 packs together using a pre-built series connector (also purchased from eBay) that uses Dean-terminated 14 AWG silicon wires.
I modified the series link by inserting an inline automotive blade 15A fuse holder (same as my other battery).
I taped the packs together (using strips of electrical black tape, easy to remove if needed). I organized the charge/discharge wires, inline fuse, and balancing tabs so that they don't catch on anything when moving the battery pack around.

I finally wrapped the whole thing into a plastic bag (the ones used to freeze food), so that the battery pack slides easily in and out of the Brompton C-Bag's rear pocket. The bag is transparent so I can check the connections if required.

The end-result is pretty neat:

You may notice that the Dean plugs in the picture above are of a slightly different shades of red. That's because I bought Dean-like plugs from eBay, as well as the real Dean brand. Now, I actually think that these "XT" Dean-like plugs are great, because they have a grippy grooved surface that makes them much easier to grab. They are good quality gold-plated high-ampere connectors, as far as I can tell, and of course they are totally compatible with real Deans.

 

[daniel.weck]

Eventually I will re-wire the controller connections to the various components (PAS / Pedal Assist Sensor, thumb throttle, brake cut-off switch, 3-way power switch). To cleanup the current mess, I will use these flat waterproof connectors (the picture shows the 6-way Superseal model):

My goal will be to make the controller totally separable, just like the battery. It's actually not trivial because the Cycle Analyst's shunt is wired into the battery <--> controller main 36V line, and because I need a system that is compatible with both my Tongxin and 8-Fun wheels (the 3-phase wires have different plugs). That's a DIY job for the dark winter days

 

[daniel.weck]

Dimensions comparison:

That's 2.3Ah versus 10Ah. The weight is the biggest differentiating factor. Size-wise, the A123 cylindrical cells are actually more voluminous per Ah than the LiFePO4 prismatic pouch cells.

 

[daniel.weck]

-- UPDATE -- I'm not using a CellLog now, instead I bought a proper charger/balancer. Scroll down to read about it.

The packs arrived from China with plenty of charge in them, showing 42V.
I intend to purchase a Cell-Log 8S to check the voltage of individual cells (I don't have a balancing charger), but in the meantime, I have verified via the built-in tabs using my basic voltmeter, one cell at a time. Most of them have 3.35V, some 3.34V. Sounds good.

The ideal charging voltage for A123 cells is 3.6V. My LiFePO4 charger delivers 45V 2.5A, which is a bit too much for the 12S A123 cells (45/12=3.75V). The safe max charge voltage is 4.2V per cell (50.4V for the whole 12S pack), so there is some margin ! Still, I'd rather use my other Lithium-Ion charger, which produces 42.3V 1.5A (42.3/12=3.52V per A123 cell...that'll do).

Typically, battery packs based on A123 cells can be charged using basic cheap SLA (lead acid) chargers. Pb charging normally comprises of 3 phases:

(1) CC (Constant Current) until cut-off voltage is met (A123: 3.6V/cell in normal slow charge 1A to 3A).

(2) CV (Constant Voltage) until current drops below threshold (A123: 50mA per paralleled cell...or 30 minutes max).

(3) Float (trickle charge) (3.45V/cell)

I'll see what happens with my Lithium-Ion charger.

Regarding using the pack without a BMS: the safe discharge cut-off voltage (LVC) per A123 cell is 2.0V, and what normally happens is that the pack suddenly stops working. The Cycle Analyst will tell me when the pack reaches 24V (including voltage sag under load), at which point I'll know I should stop ! Apparently A123 cells are almost empty at 2.5V, so 30V is probably a safer threshold.

At any rate, experience shows that these A123 cells are very robust, and will usually recover from abuse. Wonderful stuff.

The CellLog detects HVC (charge cut-off) and LVC (discharge cut-off), and warns by beeping. I want to get one mainly to check long-term health of individual cells. It's small enough to be left in the toolkit, and I'm sure it will be handy at times.

 

[NRG]

Neat as ever Dan, good job. Why not use the Cellog alarm output and drive the brake cutout cct when you hit LVC?

 

[jerrysimon]

Glad it all turned up ok. That ebay seller is very reliable.

You would need two cellogs for 12s. However IMHO and from experience over the last 8 months of daily use (200 charge cycles approx), these cells really don't seem to need any form of BMS or protection. I have continually checked the balance voltage every few weeks and they remain in balance.

As you state the power cuts off and then returns briefly for short bursts, after resting when the pack becomes discharged.

I am thinking of getting another second pack myself. The beauty of this DIY route, is you know that if one cell does die it is pretty easy to identify it, remove it and replace it with a good cell.

However I would also agree that the solution is not the best one in all applications, being more suited for short trips. I still think though, that there is no point in carrying extra battery weight unless you need it.

Regards

Jerry

 

[Pedalo]

Interesting (OK maybe not that interesting) that yours came in the funky looking red wrap with A123 written all over it.

Mine (from the same supplier) was wrapped up in a far more dull black shrinkwrap:

http://www.pedelecs.co.uk/forum/electric-bicycles/6859-my-diy-electric-brompton-project.html#post87008

I've put mine through about 20 charge/discharge cycles so far and they seem to be holding up O.K.

 

[daniel.weck]

-- UPDATE --

Ok, in the end I didn't purchase the Cell-Log monitor, instead I bought a Turnigy Accucell 6 to properly charge and balance each individual A123 6S pack, one after the other. The voltage discrepancies between the cells was considerable, so I only used my 36V A123 battery twice since I received the parts...I prefer to get everything balanced properly before making proper use of it.

PDF manual:

http://www.sparkfun.com/datasheets/Prototyping/Batteries/ACCUCEL_manual(2).pdf

It took me a little while to figure out how to switch the Accucell to LiFe (it's in the same Lithium "category" as LiPo...but via the "user settings" menu, see PDF manual download link). I then chose 2.5A as my charge current, and witnessed the charger moving from the CC to the CV phase...very exiting !

EDIT: silly me ! (well, the PDF manual is not always as clear as I'd wish, to be fair) There is a normal charging mode, *and* a balanced one !! I initially thought that there was a charging mode, followed by a separate balancing function...anyway, back to square one, the charger is now charging+balancing the cells at the same time. Let's see how the individual cell voltage stabilizes ... yep, 3.6V ... it takes more than an hour to get to this stage though (Constant Voltage, near-zero current).

At the moment I use a Lenovo laptop charger (16V 4.5A), which happens to use the connector required by the Turnigy device. I'm awaiting delivery for a cheap (7 GBP) generic charger from eBay (12V 5A).

 

[daniel.weck]

I rode 13km (50% moderate hills) so it gave me the opportunity to charge the A123 pack for the first time. I am not leaving the charging station unattended because I'm not sure whether this cheap Lithium-Ion charger is doing the right thing for the A123 pack (see my analysis in a previous post in this thread).

I have been monitoring each individual cell's voltage via the balancing tabs (using my voltmeter), and it is showing great discrepancies. I guess this is an artifact of the charging process, the cells will eventually self-balance.

I've also been checking the voltage at the charger's spliced output, in order to figure out whether it is in the CC (Constant Current) or CV (Constant Voltage) phase. It's actually hard to tell so far, as there are fluctuations.

Anyway, after a while the light on the charger went from red (which I imagine means "charging...") to blank (or actually, some kind of *really* faint green). At this stage the charger still produces 43V. I've checked the pack's standalone voltage, now 43V too (rounded value, as it fluctuates a little). So my deduction is that the charger is off, and I am measuring the potential of the battery itself.

Cells are clearly unbalanced at this stage (from 3.3V to 3.8V), but once again, I think it takes a while for the 12 cells to self-balance across the pack. I will check again later...

 

[jerrysimon]

Dan some self balancing cleary does take place both after charge and discharge when the packs are left to rest. The problem may be that the pack was only partly charged to start with.

When I enquired, ohmygod told me they charge the single cells to about 80% capacity. What was the voltage of each 6s pack when you first got them ?

I think its important to have them fully charged and balanced before you start cycling them.

Regards

Jerry

 

[daniel.weck]

When I enquired, ohmygod told me they charge the single cells to about 80% capacity. What was the voltage of each 6s pack when you first got them ?

It was pretty close to 41V for the whole pack. A fully-charged, run-in pack produces close to 42-43 volts, right ?

Nothing to worry about, once I've got the Cell-Log plugged-in, I will be able to figure out whether I need a proper balancing charge to optimize the pack.

I'm not going to be cycling for a few days anyway so the A123 cells will have plenty of time to self-balance.

 

[Pedalo]

After reading this thread, I started looking at how the balancing is shaping up in my two packs. I hadn't really bothered before and just assumed they would all be self balancing but this was not really the case. I found that the cells in each pack were all holding about the same voltage but the difference between the two packs was significant. For example, off the cheap SLA charger, the cells in one pack were all at about 3.45V whereas the cells in the other were all at about 3.75V. I wonder if my two packs were built from different batches of cells...

I've run everything through a balancing charge on my Turnigy charger (all back to 3.6V) and will see how they go over the next week or so.

 

[jerrysimon]

I was careful to balance each of my 6s1p packs after building them but have not done so since.

As I said no cell is more that 50mv out and usually more like 25mv.

I must admit I have not checked the seperate overall voltage of each 6s1p pack. I will do that tonight and report back.

As I said it may be that it is important to fully charge/balance them before the first use, to maintain the balance as you then cycle them.

I also recall reading somewhere about not discharging them below 30% or so in the first 20-30 cycles. Whether that is true or not I am uncertain.

Regards

Jerry

 

[Pedalo]

I thought I'd balanced them all before first use - but it's such a long time since I first obtained the batteries and then put them to one side while waiting for the Tongxin motors to arrive I now can't be sure.

I guess it would be a very good explanation if I actually initially charged/balanced just one pack and then put them to one side. When I subsequently attached the SLA charger to both packs in series I could have been repeatedly overcharging the pack that was previously charged and not charging the other enough.

Now I've balanced it all up I'll keep an eye on them over the next week or so. If things still drift I'll just do a balancing charge every week.

 

[jerrysimon]

Yes that could be the cause.

I just rode home having charged them at work before I left. When I got home I measured each pack,

Pack one

Max voltage difference between each cell 8mv
Total Voltage 19.79v

Pack two

Max voltage difference between each cell 4mv
Total Voltage 19.79

So pretty much perfectly in balance. Thats after 8 months and over 150 charge cycles and balanced charged only once when they were new as I said.

I tend to always charge the at work as well as in the evening when I get home, even though I could probably get away with one charge. Maybe that helps as I doubt the packs ever get below 1/3 of remaining charge.

I will post data after I have charged them tonight. Maybe I should try one charge and run them down almost flat ?

Regards

Jerry

 

[jerrysimon]

Ok this morning when I got up I checked the pack after they had been resting, following a full charge the night before.

Pack one

Max voltage difference between each cell 8mv
Total Voltage 20.83v

Pack two

Max voltage difference between each cell 40mv
Total Voltage 20.61

I noticed one cell was down very slightly at 3.39v the rest being around 3.44v

It seems my pack is very slightly better balanced after 2/3 discharge.

Regards

Jerry

 

[jerrysimon]

Having ridden to work and taken 1.6Ahrs out the battery (I was lazy this morning) all the cells are either 3.26v or 3.27v, again confirming that they seem to be better balanced (pretty much perfectly balanced) when the pack is 2/3 discharged.

First time I have used the Turnigy meter in months

Regards

Jerry

 

[Pedalo]

I've also now observed that when the battery has been used to half or 1/3 capacity, the cells are almost perfectly balanced. I had a similar small imbalance as you saw when the cells were charged.

I'll keep an eye on it over the next week or so. I'm pretty sure now I must have started from an imbalanced state as suggested in my last post.

Now I've got them balanced, hopefully they will stay that way.

 

[daniel.weck]

A123 cells now balanced !

I edited a previous post, see here:

http://www.pedelecs.co.uk/forum/electric-bicycles/6956-diy-stage-9-a123-nano-battery-pack-photos.html#post88203