A123 20Ah Pressure Thoughts--and bits for experiments.

[999zip999]

my problem is under vacuum and water boils under my vacuum at 29 ATMs same as LFP the water a steal boil under the same vacuum even in a foil pouch. If it boils water cold, can it broil electotolight in it's pouch.

 

[liveforphysics]

Electrolyte doesn't boil when water would. Electrolyte is carbonate based and loaded with stabilizers.

 

[arkmundi]

However, I've had several cell scientists and engineers tell me that A123 cells are absolutely plating Li if the terminal voltage is held (C.V.'d) above 3.60V after charge current drops off. They "are rated" to go as high as 3.8V for up to 10 seconds at a time, at high charge rates---as long as they settle below 3.6 you're considered to be OK.

Thanks. There's one small nuisance in my current eBiking operation. I have a Smart Charger from batteryspace.com for charging. I've gotten into the habit of getting back home, socketing my two celllogs on, noting the voltage of each of the 16 cells, plugging the charger in and waiting till full charge.

Note that it charges at 58.2VDC 6A, to nominal 51.2V (3.2). Its bulk charging until it hits the last LED and then supposedly trickle charges the last 20% of the charge cycle. I've watched the celllogs and notice that its constantly charging & discharging in cycles. That 58.2VDC is 3.6375 per cell average. Actually, its upwards to 3.5V per cell, except for 1 or 2 of them, the same ones day to day, that hit 3.62V, where I've got the celllogs programmed to alarm.

I have no idea how this "smart" charger knows it just charged to full capacity. I expect its got some method of detecting internal resistance as a differential of voltage increase over time.

I could leave the smart charger to complete and it'd turn itself off at full charge. That's preferable as I could leave it be, go to bed. But my celllogs alarm me before the smart charger has a chance to end - those 1 or 2 irksome cells. So I'll unplug the charger and end it prematurely. All the cells will settle down to a comfortable 3.35 to 3.40V. I have left it alone until the smart charger decides its done. All cells are OK except the irksome ones and they'll be closer to 3.7V, which is not OK. So for now, its the celllogs alarm and me unplug. Smart enough I guess.

 

[dnmun]

the reason one cell goes high like that is because the other cells in the pack are self discharging and over time their voltage drops during each cycle which forces the high cell higher to fill in the voltage from the charger. over time the low cells get further and further away from full charge and the pack can no longer be balanced by the BMS when the voltage is so low. so then you are forced to manually rebalance the pack or use a better charger that will charge to higher voltage so the BMS can balance the pack properly.

 

[arkmundi]

the reason one cell goes high like that is because the other cells in the pack are self discharging and over time their voltage drops during each cycle which forces the high cell higher to fill in the voltage from the charger....

thanks for the suggestion, but I don't think so. these are new cells and the at-near-full-charge characteristics of the various cells in the pack emerged from the get-go. no bms. no detectable self discharging. just a few cells that want to charge up faster than the others, so reach 3.62v first.

 

[999zip999]

They charge to 3.65v but settle to 3.5v ( or lower ) "over in short time, right ark ? I don't worry of 3.41v to 3.53v on my cells as I find 3.45v as 96% and above 3.5v is un necessary. But it will it 3.8v fast, but unhooked will drop down to 3.6v as I don't think there any storage room in the cell for it.

 

[arkmundi]

They charge to 3.65v but settle to 3.5v ( or lower ) "over in short time, right ark ? I don't worry of 3.41v to 3.53v on my cells as I find 3.45v as 96% and above 3.5v is un necessary. But it will it 3.8v fast, but unhooked will drop down to 3.6v as I don't think there any storage room in the cell for it.

Exactly. Its just been a few times when I left the charging unattended and came back to find the smart charger had stopped the charge cycle and my celllogs were beeping that I found those cells had settled 3.65<SOC<3.70. That scenario is the worrisome one per wb9k. Question is what are the symptoms of cells that have begun to plate-out? Self discharge as dnum suggests? A lower than expected voltage? A higher than expected voltage? An as expected voltage that drops lower than expected during use & discharge?

 

[megacycle]

Yes i just bought what looks like a Vicpower or similar, 16S pack, were the two end cells were killed by the BMS, in a similar fashion.
I've swapped out the dead cells and rebuilt it up as a 24S now and got it on my gokart
We're to find 24 x 3.5V/50A, small chargers, for cheap, my worries regarding BMS pains, would be over.
Might be in the LED supply department

 

[whatever]

thats interesting it was the end cells, I also had the positive end cell fail twice, I thought it was due to lack of compression,
but there might be something else going on, I wasn't using bms, I was charging 16s pack as two 8s packs, but connected as 16s whilst being used.
Since the packs are vacuum sealed the internals would have quite a large pressure on them from the atmosphere, both the cells that failed on my pack went 'soft', I'm wondering if internal release of gas is a factor of cells going from 'hard' to 'soft'.

 

[megacycle]

thats interesting it was the end cells, I also had the positive end cell fail twice, I thought it was due to lack of compression,
but there might be something else going on, I wasn't using bms, I was charging 16s pack as two 8s packs, but connected as 16s whilst being used.
Since the packs are vacuum sealed the internals would have quite a large pressure on them from the atmosphere, both the cells that failed on my pack went 'soft', I'm wondering if internal release of gas is a factor of cells going from 'hard' to 'soft'.

actually that's what the guy i bought it off said to me, it was the BMS and testing the cells i found them not recoverable from close to zero volts.
I just cut them from the pack

 

[miro13car]

Because this thread is about A123 20Ah pouch....
I cant recommend enough this PDF document from A123.
What is surprising to me is how A123 rate "nominal capacity" :
look at page 64 APPENDIX A:
19,5 Ah nominal capacity at 6 Amps discharge???? - THATS NOT EVEN 0.5C !!!
Silly me, all this time I thought A123 pouches were rated at 1C discharge!!
I was wrong.
This rating makes my Victpower cells even better, I tested them at 1C SOLID on CBA and literally all 20 cells delivered more than 19Ah , and six delivered even 19,5Ah.

 

[wb9k]

They charge to 3.65v but settle to 3.5v ( or lower ) "over in short time, right ark ? I don't worry of 3.41v to 3.53v on my cells as I find 3.45v as 96% and above 3.5v is un necessary. But it will it 3.8v fast, but unhooked will drop down to 3.6v as I don't think there any storage room in the cell for it.

Exactly. Its just been a few times when I left the charging unattended and came back to find the smart charger had stopped the charge cycle and my celllogs were beeping that I found those cells had settled 3.65<SOC<3.70. That scenario is the worrisome one per wb9k. Question is what are the symptoms of cells that have begun to plate-out? Self discharge as dnum suggests? A lower than expected voltage? A higher than expected voltage? An as expected voltage that drops lower than expected during use & discharge?

I would also be concerned if cells were settling to rest above 3.65 Volts. This won't kill your cells quickly, but it will accelerate their demise somewhat. The more time you spend with the cells in this state, the worse it gets. The two problems caused by this are reduced capacity and elevated cell impedance. Self discharge characteristics should not change...that is caused by completely different mechanisms.

Capacity loss is caused because Lithium that was previously available for "cycling" within the cells to store electric charge is now permanently plated onto the cathode. It can no longer move from one side of the cell to the other, so it is no longer available for energy storage. This is the mechanism for ACTUAL capacity loss by Li plating. APPARENT capacity losses will generally be greater than actual losses.

Elevated impedance causes a more complex constellation of symptoms, some of which may be easy to confuse with High Self Discharge (HSD). Ohm's law (E=I/R) holds the key to understanding here.
1) Elevated Peukert losses. Because more energy per unit of current through the cell is lost as heat, less of the cell's capacity is actually USABLE. Thus, apparent capacity loss can be significantly greater than actual capacity loss caused by the loss of cycleable Li alone. In low current applications, the two numbers will be close together. In high current applications, Peukert losses increase in proportion, so apparent loss of capacity breaks further and further away from actual capacity loss as current increases.
2) Greater voltage excursion under the same load. Elevated resistance across the cell means that voltage will sag more under the same load than it did when the cell was healthier. Conversely, voltage will rise higher with the same amount of applied charge current than it did when it was healthier. At the same time, rebound/settling voltages will be further away from loaded/charging voltages. In other words, the cell will rebound to a voltage further away from loaded voltage, all else being equal. Similarly, voltage will settle farther from the charge voltage with the same charge applied. This can give the illusion of elevated self-discharge, but the phenomenon is actually not the same thing. Again, the greater the charge and load currents, the greater the effect becomes.
3) Absolute max current decreases. Because the cell's series resistance is elevated, max possible current through the cell is decreased.

Just to confuse things further, there can be many factors that lead to impedance rise. Some are related to Li plating, others are not. Thus, one cannot predict a linear relationship between impedance rise and actual capacity loss. Insufficient compression of a prismatic cell is one way to raise impedance with no effect on Li availability. Certain cell defects can do the same.

Hopefully that clarifies more than it confuses.

 

[wb9k]

Because this thread is about A123 20Ah pouch....
I cant recommend enough this PDF document from A123.
What is surprising to me is how A123 rate "nominal capacity" :
look at page 64 APPENDIX A:
19,5 Ah nominal capacity at 6 Amps discharge???? - THATS NOT EVEN 0.5C !!!
Silly me, all this time I thought A123 pouches were rated at 1C discharge!!
I was wrong.
This rating makes my Victpower cells even better, I tested them at 1C SOLID on CBA and literally all 20 cells delivered more than 19Ah , and six delivered even 19,5Ah.

A little thought about all of your statements here put together should make plain the fallacy in your logic, but if you still don't get it, please start another thread on the topic. This one is about compression and compression-related issues, not how A123 grades its cells.

 

[miro13car]

I don't know what "logic" you talk about???
A123 document says it clear :
A123 20Ah nominal capacity 19.5Ah at 6 Amps discharge only.
1C is 20Amps, 0.5C is 10Amps
At just 6 A discharge curve must be really flat.
I just bring to attention what is in PDF document.

 

[wb9k]

I don't know what "logic" you talk about???
A123 document says it clear :
A123 20Ah nominal capacity 19.5Ah at 6 Amps discharge only.
1C is 20Amps, 0.5C is 10Amps
At just 6 A discharge curve must be really flat.
I just bring to attention what is in PDF document.

You disparage the method, but then go on to say that the recycled scrap you got from Victpower (which were also tested at A123 by this very same method) was WAY BETTER than this. How does that make logical sense? It reads to me like you're slamming A123's test methods, while verifying that they deliver cells capable of the instant power numbers we are used to talking about with A123 cells, even when the cells were pulled from the trash.

The primary reason for that particular test is that it is in fact a much more accurate measure of true capacity than a higher current test can yield. IOW, it is a more sensitive, higher resolution test that best divorces actual cell capacity from other nuisance variables like parasitic impedances unrelated to true capacity. This is possibly the most accurate capacity test the cell will ever see--and this would be true of any battery, of any chemistry. It would actually be even more accurate if it were slower yet, but this slows throughput to unacceptable levels for production. Maybe you should be complaining that the test is too fast. That aside, it is not the only test those cells see by a longshot, nor is it a power test. Don't confuse capacity with discharge capability. They are two different things.

 

[arkmundi]

http://endless-sphere.com/forums/viewtopic.php?f=14&t=52244&start=200#p1004360

I would also be concerned if cells were settling to rest above 3.65 Volts.. Capacity loss is caused because Lithium that was previously available for "cycling" within the cells to store electric charge is now permanently plated onto the cathode. ....
Elevated resistance across the cell means that voltage will sag more under the same load than it did when the cell was healthier. Conversely, voltage will rise higher with the same amount of applied charge current than it did when it was healthier....

Thanks, appreciate the explanation. So, my pack exhibits an early "healthy" life, with certain charge & discharge voltage characteristics, from which it departs as it ages, and that departure is unfortunately associated with a loss of capacity. Which makes sense, because in my situation, with the charger I'm using, its exhibited by cell voltage >3.65 before the charger believes the pack as a whole should be fully charged. And I'm right to just pull the plug on it when my celllogs beep.

I'm not yet experiencing any noticeable sag on discharge. Cells all seem to settle in a 3.5v<SOC<3.3v range and within ~0.05v of one another. Will continue on. Regards.

I don't know what "logic" you talk about???
A123 document says it clear :
A123 20Ah nominal capacity 19.5Ah at 6 Amps discharge only.
1C is 20Amps, 0.5C is 10Amps

miro, he means you've come to the wrong set of conclusions. Begging your pardon, but you have. A123 cells in my experience deliver on their promise of very high C-rates or as wb9k says "instantaneous power." I am quite confident that I can get 100 amps continuous from my packs, which is at least 5C, not that my controller will pull that. When I get around to concluding what controller I want to match to the MXUS 3000w, I will be pulling that, from a 72V pack. Certain.

 

[wb9k]

http://endless-sphere.com/forums/viewtopic.php?f=14&t=52244&start=200#p1004360

I would also be concerned if cells were settling to rest above 3.65 Volts.. Capacity loss is caused because Lithium that was previously available for "cycling" within the cells to store electric charge is now permanently plated onto the cathode. ....
Elevated resistance across the cell means that voltage will sag more under the same load than it did when the cell was healthier. Conversely, voltage will rise higher with the same amount of applied charge current than it did when it was healthier....

Thanks, appreciate the explanation. So, my pack exhibits an early "healthy" life, with certain charge & discharge voltage characteristics, from which it departs as it ages, and that departure is unfortunately associated with a loss of capacity. Which makes sense, because in my situation, with the charger I'm using, its exhibited by cell voltage >3.65 before the charger believes the pack as a whole should be fully charged. And I'm right to just pull the plug on it when my celllogs beep.

I'm not yet experiencing any noticeable sag on discharge. Cells all seem to settle in a 3.5v<SOC<3.3v range and within ~0.05v of one another. Will continue on. Regards.

I should have said before that the impedance rise from Li plating happens because not only has Li been taken out of circulation, the cathode real estate onto which it has plated is also lost. These areas of the cathode become effectively dead. This means the physical ions have to travel a longer, more torturous path (around the dead areas) to move from one side of the cell to the other, hence larger loss is associated with the electrochemical process. The lost energy is expressed as heat. That said, the issue happens in tiny increments in the situation you describe. All of this might not add up to enough for you to notice for a very long time if you're not driving the pack very hard to begin with in its day to day life.

Proper cutoff for charging is determined not only by the voltage, but also the current at which charging stops. When I installed a 15A (1/4C for the system) charger on my motorcycle, I was a little worried because shutoff of the charger occurs at 58.4 Volts. There is no ramp down of current or CV voltage after this happens, you just go from 15 Amps to nothing after reaching 58.4 Volts. That's only 3.65V/cell in a perfectly balanced pack, but potentially more if balance gets a little out of whack. My previous charging arrangement could only source about 7 or 8 Amps of current, but was limited to a max voltage of 57.4 V, or 3.588 V/cell group. Even if out of balance a bit, you were always well within the safe zone, as the starter battery boards I have balancing the cells on the bike would keep that from getting away very far at all. I could let the pack soak til current fell to near nil and I knew I had a perfectly balanced pack.After a typical soak, the pack would settle to around 57.2 in the first 30 minutes or so.

I needn't have worried about overcharging with the higher cutoff voltage. With the auto shutoff of the 15 Amp charger, pack voltage would typically land around 56.4 in the same period of rest time. I was finishing with a more fully charged pack with the lower cutoff voltage than I was with the higher voltage at double the max current, but with no rampdown of charge current. If I want to get the same level of balancing with the 15A charger I was getting before with the little outboard setup, I have to run one or two more automatic topoffs by restarting the charger after the pack had rested a while. Then, I can achieve a short-term pack rest voltage of 57.4. It's nice to be able to run low current at right around the voltage for 100% SOC as a finish to a charge cycle, but you can find ways around not having it if you understand how your charger works and how the cells respond--and are lucky or smart enough to have a combination that works.

 

[dnmun]

i don't like the auto shutoff chargers because they don't allow the balancing current to flow long enuff to balance the battery.

if you set the charger voltage to just a little above the 57.6V needed for 3.60/cell and let the BMS balance it then the battery can continue running on the balancing current and eventually they all reach the same state of charge. there is little chance of them overcharging above the 3.60V level then. of course i charge higher so they are always over 3.61 or so but i usually time it out so i can unplug from the charger and go.

anyway that's what i like about balancing networks on the BMS. it can keep them all close so you don't have some that are high and some low.

 

[miro13car]

I didn't come to any conclusion , simply brought to attention.
I am using A123 pouches on my 2 bikes, I know what are capable of.
was surprised A123 would not rate capacity at 1C discharge,
but let us have this back and for maybe more members will join discussion.

 

[wb9k]

I didn't come to any conclusion , simply brought to attention.
I am using A123 pouches on my 2 bikes, I know what are capable of.
was surprised A123 would not rate capacity at 1C discharge,
but let us have this back and for maybe more members will join discussion.

I'll admit I was surprised when I learned this too, but it turns out it doesn't really matter. This test is not aimed at finding power capability. That is addressed with other tests.

If A123 was putting cells out there that didn't perform as advertised, the tone of your first post here would be more justifiable. After almost 4 years of working here, I've gotten used to people crawling out of the woodwork bearing torches and pitchforks aimed at A123, usually for political reasons. I usually respond with similar intensity since the complaints so rarely amount to what they are purported to.

 

[whatever]

One thing that is surprising is that the ahr capacity does not seem to change much even if the current draw is quite high,
there are plenty of graphs on this forum and others that show high amp draw tests and still getting high ahr capacity.
For example:
http://endless-sphere.com/forums/viewtopic.php?f=14&t=26072&start=165
The graph on this page shows a 100amp discharge test, and still getting up near 19ahr at 2v cut off.
Though useable ahr before dropoff point is about 17ahr.

 

[pohjeremy]

One-time digression..

... needed for 3.60/cell....

U know, that company claims 3.357V to 100% SOC and charging current =<3.650V. Others say LFPs accommodates past 4V bursts w/o issues. Guess I'll never really know. Maybe it's a different chemistry but I've got a feeling all LFPs are nearly identical..

 

[wb9k]

i don't like the auto shutoff chargers because they don't allow the balancing current to flow long enuff to balance the battery.

if you set the charger voltage to just a little above the 57.6V needed for 3.60/cell and let the BMS balance it then the battery can continue running on the balancing current and eventually they all reach the same state of charge. there is little chance of them overcharging above the 3.60V level then. of course i charge higher so they are always over 3.61 or so but i usually time it out so i can unplug from the charger and go.

anyway that's what i like about balancing networks on the BMS. it can keep them all close so you don't have some that are high and some low.

Agreed on all counts. In a pack with no balancers, I like to aim a tad under 3.6 per cell and let the slop keep everyone under 3.6 between period checks for imbalance/manual balancing. If you select cells carefully, watch pack voltage, and cycle the battery fully a time or two to build confidence that you can predict its behavior before you stop checking balance every single cycle, this can work nicely. I manage my lawn mower and scooter packs this way.

 

[dnmun]

i would never ever try to manually balance a battery pack. that is what the BMS does.

i am not too concerned about the excess voltage because the balancing current is so tiny compared to the size of the pack.

i am dealing with maximum of 350mA of balancing current on an 87Ah 24S lifepo4 pack and the pack is at full voltage for maybe an hour max.

these are all ping packs that someone else threw away and i bot from them so this is all lagniappe.

 

[granolaboy]

I've been going through this document:

https://www.buya123products.com/uploads/vipcase/b24d4f5b63934c59d43e93b3bb4db60a.pdf

Forgive me if this was already covered in this thread, but I can't find it. The doc mentions:

Compliant Pads Between Cells
To Maintain proper cell support and account for expansion during the charge and discharge cycles, A123 uses a compliant pad between cells.

I'm wondering what makes for a "compliant pad" and what a DIY'er could use...