The care and feeding of a123-based packs...

slayer said:
Gary i agree with you about the lvc and most everything i know i learned from you especially with knucles setup with only 1p version but why do you think it would have made a difference with single cells charging
Click to expand...

Only because the cells might have been better balanced, and closer in voltage. In that case, more than one cell could have started dumping at the same time, That might have given him a better warning that the power was sudenly dropping, and he might have backed off the throttle before actually killing the cell(s).

The best option, of course, is to do LVC protection for each cell, but his current setup doesn't make that very easy to do.

-- Gary
 
GGoodrum said:
slayer said:
Gary i agree with you about the lvc and most everything i know i learned from you especially with knucles setup with only 1p version but why do you think it would have made a difference with single cells charging
Click to expand...

Only because the cells might have been better balanced, and closer in voltage. In that case, more than one cell could have started dumping at the same time, That might have given him a better warning that the power was sudenly dropping, and he might have backed off the throttle before actually killing the cell(s).

The best option, of course, is to do LVC protection for each cell, but his current setup doesn't make that very easy to do.

-- Gary
Click to expand...

Ok to bring them closer in voltage might give you the idea they are balance but if one is now deleviring 2.1 ah the other is 2.3 and 2.23 ah and so on bringing them to past 3.4 volts each capacities.... i don t think the voltage you bring them have anything to do with the (ah) they will deliver ...if we agree that past 3.4 volts is not real ah
 
i am not trying to say it is useless to charge them individally now , cause when you willl bring the bms out i will buy one ...it is just that small little point i wanted to say about the voltage and the ah ...of course if you start with brand new cells that are balanced in ah and in voltage then you assemble them with a good bms it is the best ...but most have already imbalanced cells in (ah) that no bms can bring back....
so like many have said before your pack is only as good as your worst cell suppose it is a 2 ah that is the worst and you have a 3p then 6 ah is the most you should spend
 
Yes, that's true. I was assuming that he started with reasonbly matched cells. It has been my experience that healthy a123 cells are all pretty close in capacity and "real" resting voltages after a charge. That means you check the cell voltages not right after charging, but after they've been put under some sort of load for 10-30 seconds. That will give you a more accurate indication of its "full" voltage. This will usually be a number somewhere around 3.5V. Fresh off the charger what you see is a surface charge that builds up on the electrodes. This number can vary quite a bit for various cells, even for cells that are closely matched in capacity. Putting a load on the cells briefly, will burn off this surface charge.

With a123s, abusing them will cause the cells to loose about 10% of their capacity. The easiest way to tell is that within a minute, or two, the surface charge will burn off by itself, and the voltage will only be around 3.35-3.40V. These cells still work, but like I said, the capacity will be reduced to about 2.0Ah. What I've found is that if a cell gets "stressed" to where it's resting voltage is around 1.5V, it can be recovered, but it will have this 10% reduced capacity condition. Cells that have been drained down below 1.0V, typically can't be recovered. Cells drained to about 1.8V, and above, seem to recover fine.

What is odd is that it appears that eithr the cells are healthy, and are matched in capacity within about 1%, or less, or they are stressed, and have lost roughly 10% of their capacity. These stressed cells are also fairly matched. In the 1500, or so, usable cells I have, there hasn't been a case that I'm aware of where a cell is not in one category or the other. I just don't have any that are 'tweeners. They are either healthy or they are stressed. In this case, I'm guessing that along the way, the one cell that died got stressed along the way, and then finally died when the whole pack was drained so for down. Personally, I can't see the practicality of just having a "single-P", 2.3Ah pack. It has no real usable range, and it is far too easy to run the cells down to where they can get damaged. Add to that no attempt to do any sort of balancing and/or individual charging, and absolutely zero low voltage protection and you have a recipe for failure. This was bound to happen, and frankly, I'm surprised it took this long.

-- Gary
 
My simple rules of thumb, based on my experience and the suggestions of others (I recognize that different folks will have different ways of solving problems, or differing priorities):

Before starting out on a trip, be sure the pack is fully charged, and have a good idea how much of the pack's capacity the trip will use. (I use a 20S3P pack most of the time; each of the 6 10S subpacks has a 30 amp fuse.)

On a trip, never use more than 80% of the pack's capacity. Monitor both pack voltage and Ah (or Wh) used.

Don't leave home without a cell phone.

Carry a spare pack. (Mine is an unfused 20S1P pack, which can get me home from 5 or more miles out.)

Keep records for each ride. Every so often, measure individual subpack voltages. As warranted, measure all cells in a subpack.

Larry
 
How can I easily detect a bad cell in a 5p pack? Assuming the pack has been fully charged and balanced. Packs are mounted in its case in the bike frame and cannot be easily removed.
In this drawing, can the yellow wires work to monitor and measure voltage for each cell, and can the same wires be used for charging one cell?
*tabs are spot-welded
 
Yikes! what happened to your avatar? Trying to keep up with Tyler I guess. :shock:

In a 5p welded pack, it's going to be very difficult to detect a bad cell other than it's one of the bunch.

You might be able to put a high current through the pack and measure the voltage drop in the tabs somehow, but there's not much voltage drop in the first place. Basically you're going to see the same voltage everywhere you measure.

If one of the 5 went bad, you have to cut the pack apart to fix it anyway. Take a guess. If you're lucky, you'll start tearing at the right end. Test each cell as they come off until you find the bad one.
 
As Richard points out, this won't work. The cells will self-equalize as soon as they are connected in parallel. You can usually detect a weak block of cells, though, by comparing voltages after charging thells individually, and then putting a slight load on the pack for 20-30 seconds, to burn off the surface charge.
 
Thank you guys! so there is really no easy way, especially when the cells are permanently soldered/welded. It's best I think to use bigger capacity cells(Ex. Headway, Lifebutt) with non-permanent connections so it can easily be separated for individual cell charging and voltage testing.
Fech, I thought you like exotic beauties. :mrgreen:
 
shinyballs said:
Thank you guys! so there is really no easy way, especially when the cells are permanently soldered/welded. It's best I think to use bigger capacity cells(Ex. Headway, Lifebutt) with non-permanent connections so it can easily be separated for individual cell charging and voltage testing.
Fech, I thought you like exotic beauties. :mrgreen:
Click to expand...

When you tie cells together in parallel they become a larger capacity cell.

The main advantage to starting with larger cells is reduced assembly time labor.

Smaller cells gain you flexibility in battery pack geometry, both physical and electrical. You can more easily shape the pack to suit the bike. And get just the desired amount of capacity too.
 
GGoodrum said:
I'm sorry, but unless you figure out how to either individually charge each cell, or figure out a way to balance them, you will likely continue to have this problem where a low cell can die. Having an LVC capability would at least keep the cells from being damaged.

-- Gary
Click to expand...
No need to be sorry. You are dead on right and I had always expected this to happen. Amazing it took 2 months of regular 30 MPH riding before killing my first a123 cell.

I certainly appreciate the excellent information on this thread. My initial goal was achieved by installing the cells in the forks without soldering them. I also use the chopper on dirt trails more than on-road because I don't want hassles. The chopper looks to "unnatural" for NY "locals" not to "freak out". And the chopper handles so well that I ride it for enjoyment and not really for commuting or shopping (I have diamond frame 'normal' bikes for that).

With that said ... I agree with all you say. If I can somehow get "taps" to measure each cell voltage (in the tubes) they would be very thin wires at best. Not enough for anything other than a warning system to let me know if one cell is going flat before the others. So I had a thought on some kind of visual display on the handle bars.

I could use LEDs to monitor each cell voltage for a low V condition. A red LED from radioshack needs about 1.60 volts to barely illuminate (about 5 mA). If I use 2 cheap diodes in series with the LED, the red LED needs about 2.70 volts to barely illuminate (anything less than 2.7 volts and the LED is basically off). At 3.30 volts the red LED is very illuminated (even in bright daylight). At 4.00 volts it is super bright. The voltage across the 2 cheap diodes rises from about 1.1 volts to about 1.9 volts (a123 cell voltage 2.7V to 4.0V).

A yellow LED from radioshack with 3 cheap diodes in series with the LED is very dim when the cell voltage is 3.3 volts but is very bright (in daylight) at 3.7 volts and super bright at 4.00 volts. This could be used as a 'full charge' indicator for each cell in the 2 - 12-cell sticks.

So basically I could have 24 red LEDs (for a visual LV cell warning) and 24 yellow LEDs (for a visual HV cell indicator). Each cell gets a red and yellow LED. Both LEDs are 'off' below 2.70V. The glowing LEDs would look kinda cool (maybe?) on the chopper while also serving a practical (visual) cell monitoring system.

How to install tap wires to each cell is tough because the cells do fit very snugly (with original paper wrap) in the tube forks. Still working on that. All without any soldering to the individual cells of course.

btw. I charge the bike forks with 2 36V-SLA chargers (one for each fork) and I use a 1-hour light timer with a 24-hour clock. So when not in use, the chopper sits in the garage off the charger for 23-hours and floats on the chargers at 42.5V per fork for 1-hour on a daily basis. Max voltage during charge is about 44.5V (if more than 0.5 amp current draw). Idiotically simplistic I know. But the chopper is so much fun to ride that I am still happy using this system for now.
 
Knuckles said:
So basically I could have 24 red LEDs (for a visual LV cell warning) and 24 yellow LEDs (for a visual HV cell indicator). Each cell gets a red and yellow LED. Both LEDs are 'off' below 2.70V. The glowing LEDs would look kinda cool (maybe?) on the chopper while also serving a practical (visual) cell monitoring system.
Click to expand...


You need a couple of Gary's LVC boards.

http://www.tppacks.com/products.asp?cat=26

(second item on the page)

Add some flashing lights if you must, but the LVC functionality is a must.


Richard
 
I hear you. These boards would take out the guess work and shut off the controller before I do damage to any more cells.
I really need a 24-Cell Low Voltage Cutoff Protection Circuit. My lame-ass visual LV indicator is just a goofy mod that is very simple to construct.

But at least (I think) I have a way to tap into each cell in the series sticks (forks).
Trying now to see if the wires fit into the fork tubes (still without any soldering of those a123s).

Toa Chie for the link. Much appreciated.

K
 
Knuckles said:
I hear you. These boards would take out the guess work and shut off the controller before I do damage to any more cells.
I really need a 24-Cell Low Voltage Cutoff Protection Circuit. My lame-ass visual LV indicator is just a goofy mod that is very simple to construct.

But at least (I think) I have a way to tap into each cell in the series sticks (forks).
Trying now to see if the wires fit into the fork tubes (still without any soldering of those a123s).

Toa Chie for the link. Much appreciated.

K
Click to expand...


I'm using two of Gary's LVC boards for my 20-cell pack. Ride to work and back every day for a couple months now -- no problem. Cooked a few cells before I got them.

Gary can probably set some up for your particular pack.


Richard
 
Excellent testimonial. I agree. I endorse Garys's logic (and product). Maybe I will send him detailed pics of my chopper. It is a prototype after all.

Yes ... Care and Feeding. Man how I do love those a123 amps!

Cheers Mate.

Toa Chie!

(And thanks to all whom contributed to developement)
TOP NOTCH SOLUTION!
 
The TC54 voltage detection chip, which is the heart of the LVC circuit, only draws a few microamps, so the wires going to each cell junction can be as thin as a hair. Just find the thinnest wire you can, and maybe it will fit, along with the cells. :)
 
Gary et al,

Great design work on the BMS! :D Your hard work has provided us all with the "missing piece" in the A123 solution.

Questions for long time owners of A123 cells:

Excluding cells damaged by discharge below 2V, have any of your A123 cells lost signifigant capacity due to:
a) age?
b) cycles?
c) high amperage draw?

Will these cells last for years if kept in the correct voltage range? Has anyone had a cell just wear out?

Is 2000 plus cycles really going to happen?

Thanks

Mike
 
That s a good one Mike B,

2000 recharges, if recharge every day (rare) would be 5.5 years, so in normal use i think i would never see the end of my a123 ...at least i will buy another kind of pack or technolgy before.

but maybe someone did, i would love to hear about it too.
 
I could be a poster boy for abusing these things:
- lots of amps used up big hills on the way home on my commute (600 foot elevation gain over 7 miles)
- getting the cells too hot when soldering them
- discharged too low and overcharged as well when one of the packs was imbalanced for about a month
- two packs caught on fire and I lost about 12 cells, still kept the ones that were directly adjacent to these burnt ones

Originally I charged using the hyperion 5i and 6i chargers with balancers on an 3* 6s5p setup. Now I'm doing a 20s5p all single cell charging like a1234life's setup here:
http://www.endless-sphere.com/forums/viewtopic.php?f=6&t=3171&p=45746&hilit=+a1234life#p45746

I've got about 400 cycles on my cells (20s5p). I've lost on average about 100mah out of the 2300mah. My daily commute leaves me with 2ah reserve so I can see getting at least another 5 years out of the current setup. If the capacity falls off too much, I'll just make the pack smaller and charge at work and at home.
 
My experince, and abuse patterns ( :D ), pretty much tracks with what Swede has seen. In my case, I've got some cells that originally were in packs that went in large RC electric helicopters that routinely hit multiple 100A+ peaks and were drained down to where the helicopter wouldn't stay in the air anymore. These packs were also recharged at 20A and even 30A rates, in order to get the helicopters back in the air as soon as possible, and we didn't worry about balancing at all. I had one spectacular mid-air explosion, when a blade link let go, which caused one of the blades to drop down, slicing right through the helicopter. My 10-cell a123 pack was the biggest piece left, and it flew 200 feet and ended up half-buried under a cactus plant. I cleaned it off and the only visible damage was some torn shrink wrap and a pretty good dent in two of the cells. I recharged the pack (at 30A...), put it back in one of my other helicopters and it was flying again within 20 minutes. I'm still using these cells in one of the packs that is on the bike in my avatar.

Anyway, I have not seen any real loss of capacity in any of my a123 cells, even for most of the early "abused" ones, except for the ones that got overdischarged on my bikes. For the latter ones, I've found that cells that had resting voltages that were still above about 2.5V would recover fully, with no apparent loss in capacity. Cells that got even a little below 2V, however, all seemed to immediately lose about 10% of their capacity, and have reduced "C" ratings as well, because they will have larger voltage sags, under load. An easy way to tell if an a123 cell has this reduced capacity is by checking the resting voltage about a half-hour after charging. Healthy cells will be able to hold the surface charge voltage up over 3.60V almost indefinitely, but "stressed" cells will bleed this off down to the 3.38-3.45V range within minutes.

-- Gary
 
Slayer, Swade, Gary,

You guys would agree the cells are virtually abuse proof, if kept in the correct voltage range.
Loss of capacity is minimal and within specs for the M1 cells http://a123systems.textdriven.com/product/pdf/1/ANR26650M1_Datasheet_AUGUST_2008.pdf
Good point, that we are likely to have even better cells before the M1s hit 2000 cycles.

I am still curious tho, about what will happen two or three years from now. What about starting a fund, each of chip in a few dollars, and have someone (Gary, Fechter) take several original, well used M1s, and put them on continuous cycle?

These cells are just plain amazing. At two thousand usable cycles, the delivered cost per kwh is about fifty cents. Coulombic efficiency (EDIT 8/23/08) 94%. Partial charges, partial discharges, all OK. No need to make sure every cell is reformed at the end of the charge, every time. Power densities through the roof. No more 'venting with flame'.

And to top it off, things you can do to A123s and they still work:

unplanned ejection from aircraft followed by ground impact/burial with minor dents--OK
spot welding wedding band--OK
unplanned containement of battery fire--OK
plus 200 deg C on terminals--OK
dead shorted for less than (?)(how long it takes to get off the bike and rip the wire out of the pack :shock: )--OK
 
I had no idea that my company was testing these cells.
I will now have to dig though the phonebook to see who is doing this work because I need to get in on this!

I am thinking that they need some NI equipment to help run those tests and I am just the one to develop it. . . Maybe a datalogging system written in LabView (that I will maintain the files on =)

I will see if I can get any gossip (that is suitable for the public domain of course)

-methods


EDIT: Emanuel P. Roth as found via Google ->

Third article.
http://news.cnet.com/greentech/?keyword=%22lithium-ion+batteries%22

Too bad he is at the ABQ site! I do travel there often, maybe he will allow me to visit. The God of hell fire eh?
 
Methods,

The second article is even more interesting:
A123 Systems has signed on electricity utilities to use its lithium-ion batteries for short-term energy storage, according to a company executive.

Thanks for the link!

Mike
 
You must log in or register to reply here.

Similar threads

E
Learning about BMS
Replies
4
Views
197
James340
J
O
Repaired my Grin pack, I think...
Replies
8
Views
273
999zip999
999zip999
M
A123 ANR26650M1B Spot Welding
Replies
24
Views
1,181
gromike
G
V
Have new battery packs that I am stripping the proprietary BMS and adding one that is capable of more amps. What do I need to look out for?
Replies
2
Views
95
vengox
V
Skyler
LFP reverse polarity.
2 3 4 5
Replies
105
Views
4,938
DaLanMan
DaLanMan
Back
Top