figuire LFP life

classic Classic list List threaded Threaded
39 messages Options
12
Reply | Threaded
Open this post in threaded view
|

figuire LFP life

Peakfoto Digital Photo Still n Video
  I'm thinking for going to LFP form leadies .  my leadies for motorcycle lasted 5000+ miles thanks to easy driving and parllel bank charger thanks to the former owner!!

to replace would cost around $500 so that means, battereis cost me 10 cents mile. and 1/2 cent a mile for electricity. If i get LFP $1330 for that batts and should get a charger .

SO how do I convert from  1.5 kilowatts  per trip, with 40 amp hr LFP, 72 volt  system, to fguire out cost of batteries over life or how many trips. I'd guess ??  30%  DOD witch sould last my me 8-10 years.

I don't know how to do the math for this . I thought of getting a CA to judge my trips better. leadies are simplier but I wan the long term benies and wiegh safeings. any web site to plug into ?

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

David Nelson-5
There are a couple of things to consider in sizing your LFP pack.
Winston batteries (FKA ThunderSky) are rated to 3C current meaning 3
times the Ah rating of the battery. For example the WB-LYP40AHA
(http://www.thunder-sky.com/pdf/201072311158.pdf) is a 40Ah battery so
3C means that the current draw should be 3*40A=120A or less. You
should measure what your current draw is on acceleration to see if it
stays within spec otherwise the batteries will not last as long.

Second is the voltage of the string. My experience, using a CA, is
that using 3.2vpc (volts per cell) is a realistic nominal voltage to
use for calculations. Depending on the voltage limits of your
controller and DC-DC you could go with 22-23 cells in series. 22 cells
= 70.4V nominal, 23 = 73.6V nominal but the cells will sit at a higher
voltage than that with no load. The other thing to consider is the
ending charge voltage. I know the spec sheet says 4.00vpc but what it
doesn't tell you is that is while charging at 0.015C or 0.6A for a
40Ah pack. If your ending current is less than that your cutoff
voltage should also be lower. Furthermore, from my testing of cells on
the bench with low ending current there is very little energy above
3.45V. I recommend using a cutoff voltage of 3.5vpc or less. I'm using
3.485vpc with my pack and it has been doing just fine for the past
8000 miles. The higher your ending voltage the more stress you are
putting on the cells so why push it to the max? If you use 3.5vpc as
your ending voltage you will have to make sure your other electronics
can handle 22*3.5V=77V or 23*3.5V=80.5V. If you use a higher ending
voltage make sure you won't blow something out.

For useable energy capacity LiFePO4 cells are great because they are
nearly 100% efficient. So for a 22 cell pack of 40Ah cells the total
energy is 22*3.2vpc*40Ah=2816Wh. The most you should use is 80% of
this or 2252Wh. You should get the CA and measure your current energy
usage so you can plan your pack size for the range you need. You
didn't say what bike  you have but I found one on evalbum.com which
claimed 83Wh/mi at 35mph. Suppose yours is this efficient. This would
mean you could go (2252Wh)/(83Wh/mi)=27miles before reaching 80%DOD.

You do need to invest in a properly programmed charger. When the
batteries reach full the voltage spikes really fast! You need a
charging system which can properly shut off. At the risk of starting a
war here you have a couple of options as far as a battery management
system. One method is to balance your cells once and then stay away
from the extreme ends of the charge curve. Charging to only 3.5V and
staying well above 80%DOD will work just fine with no cell level
monitoring. (NOTE: I am only referring the the LiFePO4 type of battery
here. This is an important point!) However, if you are not techie
enough to check your cells periodically you may want to look into the
various cell level monitoring options. The main thing is get educated
about the proper care of LiFePO4 batteries. I used to charge my cells
to 4.00vpc and balance them every time. What I learned from more
research and my own testing is that this is unnecessary and it harder
on the cells to charge to such a high voltage. In your search you will
find those who say you must have a BMS and balance each and every time
and you will find those at the opposite extreme who say a BMS will
damage your pack. So far my data and experience doesn't support either
extreme. I no longer balance my cells on charge but I haven't removed
my Black Sheep Technology BMS boards, either. I will say that unless I
get different data in the future my next conversion will not have cell
level monitoring but instead use a device which compares the voltages
of each half of my pack.

At this point no one knows how many years LiFePO4 batteries will last
in an automotive application. They haven't been around long enough.
Mine are from November 2009 but there are others who have been using
them for longer. Remember, however, in any statement about the life of
a cell the way the cell was used has a significant impact on
longevity.

HTH,

On Mon, Jun 13, 2011 at 11:49 AM, Peakfoto Digital Photo Still n Video
<[hidden email]> wrote:

>  I'm thinking for going to LFP form leadies .  my leadies for motorcycle lasted 5000+ miles thanks to easy driving and parllel bank charger thanks to the former owner!!
>
> to replace would cost around $500 so that means, battereis cost me 10 cents mile. and 1/2 cent a mile for electricity. If i get LFP $1330 for that batts and should get a charger .
>
> SO how do I convert from  1.5 kilowatts  per trip, with 40 amp hr LFP, 72 volt  system, to fguire out cost of batteries over life or how many trips. I'd guess ??  30%  DOD witch sould last my me 8-10 years.
>
> I don't know how to do the math for this . I thought of getting a CA to judge my trips better. leadies are simplier but I wan the long term benies and wiegh safeings. any web site to plug into ?
>
> _______________________________________________
> | Moratorium on drag racing discussion is in effect.
> | Please take those discussions elsewhere.  Thanks.
> |
> | REPLYING: address your message to [hidden email] only.
> | Multiple-address or CCed messages may be rejected.
> | UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
> | OTHER HELP: http://evdl.org/help/
> | CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
>



--
David D. Nelson
http://evalbum.com/1328

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

martinwinlow
Hi David,

I would be interested to hear what your theory is about why TS (WBL)  
still recommends taking their LYP's to 4.0V on charge...?

Regards, Martin Winlow
Herts, UK
http://www.evalbum.com/2092
www.winlow.co.uk


On 14 Jun 2011, at 00:06, David Nelson wrote:

> There are a couple of things to consider in sizing your LFP pack.
> Winston batteries (FKA ThunderSky) ... Furthermore, from my testing  
> of cells on
> the bench with low ending current there is very little energy above
> 3.45V. I recommend using a cutoff voltage of 3.5vpc or less. I'm using
> 3.485vpc with my pack and it has been doing just fine for the past
> 8000 miles. The higher your ending voltage the more stress you are
> putting on the cells so why push it to the max?



_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Roger Stockton
Martin WINLOW wrote:

> I would be interested to hear what your theory is about why TS (WBL)
> still recommends taking their LYP's to 4.0V on charge...?

David, I would appreciate it if you could clarify just what you mean when you refer to charging with "low ending current".

A typical lithium profile is constant current to the target voltage, then hold that voltage until the current tapers to a suitably low level.  Are you observing that if you use a lower target voltage, but then hold that voltage until the current tapers to a lower level, you achieve much the same state of charge/available capacity as if you were to use a higher target voltage but terminate the charge at a higher current level?  If so, what difference have you observed by using the higher target voltage, but allowing the current to fall to the lower target level before terminating?
 
Thanks,

Roger.

> On 14 Jun 2011, at 00:06, David Nelson wrote:
>
> > There are a couple of things to consider in sizing your LFP pack.
> > Winston batteries (FKA ThunderSky) ... Furthermore, from my testing
> > of cells on the bench with low ending current there is very little
> > energy above 3.45V. I recommend using a cutoff voltage of 3.5vpc or
> > less. I'm using 3.485vpc with my pack and it has been doing just fine
> > for the past 8000 miles. The higher your ending voltage the more
> > stress you are putting on the cells so why push it to the max?

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

David Nelson-5
I'll try to articulate how I have reached my current position on
LiFePO4 charging and care. I'm going from the studying of various
sources of information I've found including NASA summary documents,
the CMU professor's videoed talk, TS documentation (and its changes
over time), various individual reports of experiences with these
cells, my experience with my own cells (TS-LFP100AHA, mfg 11-2009) in
my EV and on the bench, and combining it with my Physics background. I
do not know what differences the LYP cells have compared to the LFP
cells since I have not found much info about it. As always, my
conclusions are open to modification as more data/information is
presented. I would love to do more thorough testing but my limited
resources don't allow it at the present. I would love to get some
1-5Ah WB cells for testing as this would decrease the cost of the
batteries and make destructive testing much cheaper.

On Tue, Jun 14, 2011 at 9:48 AM, Roger Stockton <[hidden email]> wrote:
> Martin WINLOW wrote:
>
>> I would be interested to hear what your theory is about why TS (WBL)
>> still recommends taking their LYP's to 4.0V on charge...?

My hypothesis on why they recommend 4.0V on charge is that it
virtually assures the cell reaches the maximum state of charge
possible. Remember that they first started by saying to charge to
4.25V then lowered it to 4.2V and now 4.0V. I think they realized that
the long term effects of high voltages were not good on their cells so
they have lowered the cutoff voltage but they still don't want to give
up that last fraction of a percent of capacity by lowering the cutoff
voltage further.

>
> David, I would appreciate it if you could clarify just what you mean when you refer to charging with "low ending current".
>

The TS docs I have read imply from the charging graphs that the cutoff
current is 0.015CA. For my 200Ah pack this would be 3A. My Zivan
charger shuts off somewhere below 0.2A. It pulls ~15W from the wall
just before shutting off and then draws ~3.5W sitting idle. Even
assuming 100% charging efficiency at this point 15W into my pack at
69.7V gives 215mA. Subtracting the idle power means the charge current
is less than 165mA. I'd call this extremely low ending current. Many
shunting BMS boards shunt somewhere around 1A so a charger has to put
out less than that if the shunting is to keep up with the charger. I'm
assuming that most chargers then are ending the charge cycle at less
than 1A which I would consider a low ending current except for very
low capacity packs like a 40-60Ah pack.

> A typical lithium profile is constant current to the target voltage, then hold that voltage until the current tapers to a suitably low level.
>

The charge profile is a combination of the target voltage and ending
current. Since only terminal voltage can be measured it has to be used
to determine end of charge but it varies quite a bit depending on the
charge current. It is important to pick the two so that cell life
isn't shortened. Taking the conservative side of the target voltage
will not hurt the cell, taking the high side may hurt the cell. If I
was in a hurry to top off a pack I might pick a higher target voltage
but stop charging at a higher ending current. With my 200Ah pack that
would be ending at something over 3A after reaching the target
voltage.

>Are you observing that if you use a lower target voltage, but then hold that voltage until the current tapers to a lower level, you achieve much the same state of charge/available capacity as if you were to use a higher target voltage but terminate the charge at a higher current level?
>

That is what I'm seeing. I did some playing around with a couple of my
100Ah cells before installing them into my pack. Here is a summary of
the charge testing I did. I do not have a way to record discharge
capacity to verify the results. Since relatively low currents were
used and the fact that these cells are essentially 100% efficient
(NASA document) I'm assuming that all the current went to charging the
cell. This assumption will actually inflate my values a little so the
real values will actually be slightly lower.

These tests were run on a TS-LFP100AHA cell manufactured on 2009 November 05
Test done by hand using a BK Precision 1761 Powersupply
Voltage readings at cell terminal using an Extech EX830
I took voltage and current readings every 30 sec to 1 min when the
values were changing rapidly and then at larger intervals when the
values changed more slowly. I then did a midpoint integration of the
voltage and current values over time to get my Ah values.
Unfortunately I didn't record ambient air temperature. I don't do well
with the cold so the air temp was very likely in the 18-20°C range.

In January 2010 I did the following tests.

3.600-4.00 Volt capacity test: 0.2213Ah
Starting with a single 100ah cell slightly discharged, charged the
cell to 3.600V and when the current dropped to below 30mA I started my
test. Maximum charge current was 3.486A which tapered off rapidly. The
test ran until the terminal voltage was 4.00V and current tapered to
23mA (0.00023C).

3.500-4.00 Volt capacity test: 0.3550Ah
The cell was discharged from the previous test until its open circuit
voltage was below 3.500V. The cell was again charged to 3.500V and
held at that voltage until the current tapered to below 30mA before
starting test. Maximum charge current was 3.485A which tapered off
rapidly after a couple of minutes. The test ran until the terminal
voltage was 4.00V and current tapered to 24mA (0.00024C).

3.397-4.00 Volt capacity test: 0.6588Ah
The cell was discharged from the previous test until its open circuit
voltage was below 3.4V. The cell was charged to 3.397V and held there
until the current tapered below 30mA before test was started. The
maximum charge current was 3.486A which started to taper off after 7
minutes. The test ran until the terminal voltage was 4.00V and current
tapered to 38mA (0.00038C).

As you can see from these tests, when the initial terminal voltage was
achieved by holding that voltage until a very low current and then
starting the test that there is very little capacity above 3.400V.
Less than 0.66% of the cell's capacity is above 3.40V. That is not
much to gain by pushing the cell to its voltage limit. For safety it
is definitely not worth it.

On July 14, 2010 I did another test on a cell to see what capacity it
had starting from 3.300ocv and going to 3.450V and then on to 3.65V.
Below I'll tell you why I decided to do this test.

I discharged a cell until it had an open circuit resting voltage of
3.300V. I kept discharging it until the voltage didn't bounce above
this point. I hooked it up and started charging it at 3.507A. This was
taking a long time to charge so at 300min I paralleled the two power
supplies and continued the test at 6.604A. This test continued with an
initial target voltage of 3.425V. I let the current taper to 1.501A
(0.015CA the TS ending current. This is coincidental, really.) By this
point the cell had received 63.18Ah. At this point I decided to go to
3.450V as the ending voltage. I let the current taper to 476mA
(0.00476CA) and recorded the energy put into the cell: 66.0155Ah. This
was 935 minutes from the start of the test.

Next I turned the current up to get a capacity value for 3.450-3.650V.
I let the current taper to 420mA. Energy added for this voltage range
was 1.0056Ah. Had I stopped at 1.5A (0.015CA) the energy would have
been 0.400Ah for this range.

>If so, what difference have you observed by using the higher target voltage, but allowing the current to fall to the lower target level before terminating?
>

I can't say what charging to the higher target voltage with a low
ending current did to my pack. When I first installed my pack I
installed only 18 cell pairs and adjusted the voltage calibration
screw in my Zivan NG1 to terminate at 4.00vpc. This is the voltage
that my Black Sheep Technology BMS boards are set to shunt. Given that
the charge profile programmed in the Zivan was for a lead acid pack
charging took a while at the end. What I did observe was that while
there were a few cells which usually began shunting before the others
they were not always the same ones. Calculating the energy difference
between the first cell to begin shunting and the last one showed a
difference in the 20mAh range, quite small for a 200Ah pack! Also, I
noticed that some cells would shunt for a little while and then stop
shunting while. This would happen just as the current would drop below
2-4A, IIRC. The next cell to start shunting was always after the
charge current dropped below 500mA (the max shunting capability of my
BMS boards) and then it wasn't always one of the first cells to start
shunting which would shunt at this point. I wish I had a video or
something of it so see if there was any change or trend to this
behavior. I ran my pack this way for a little over 6000Ah delivered
from the pack. I found that by the time I accelerated to 15mph the
pack voltage was down to less than 3.4vpc no load which is what
prompted my last capacity test: 3.45-3.65V.

At that point I had my Zivan reprogrammed for a LFP profile and
installed my last two cell pairs so I now have a 2p20s pack. At this
point I balanced all the cells at 4.00V using the BMS boards. I don't
have a charger which can charge to 80V so I did this in two passes
using my bench PS I used for the above tests. I adjusted the target
voltage on the Zivan for 69.7V (3.485vpc) and have not balanced my
cells since. At the end of charge the Zivan goes into this
on-off-on-off mode where it is trying to hold the target voltage for
an hour or so. This is when it is charging at less than 200mA. I can
hear when it shuts off and also look at my Kill-a-Watt meter to tell.
During these "off" times I take individual cell voltage readings. As
of June 1, 2011 the pack has delivered over 11,000Ah and the cells
have a voltage spread of 0.087V. The highest difference was 0.094V at
9,892Ah delivered (May 5, 2011) and the lowest difference was 0.055V
at 69Ah delivered (August 9, 2010). The high and low cells have
changed only once since I quit top balancing. I hope to see if there
is a trend but I'm not seeing one yet. The cells do trade places a bit
but otherwise not much else is happening.

I can think of three possible factors which could contribute to cell
drift. One is the fact that I have left the BMS boards in place and
they most likely do not have _exactly_ the same drain on the cells. No
matter how small the differences may be, time is sure to multiply it.
Maybe if they do get way out of balance I'll re-balance and then
remove the boards and see what happens. The second is that I can't
perfectly seal out dirt and water from getting to the top of the
cells. They usually don't get water on them but it happens
occasionally. Finally, there is a chance that there are cell defects
so one cell may age at a different rate than another one.

>From what I've been able to find there is no charge shuttle reaction
in LiFePO4 cells so there is no mechanism for self discharge (ie.
internal to the cell). Ions merely move from one plate to the other.
This is why they are so efficient. With out a charge shuttle reaction
there isn't anything to cause one cell to drift from a neighboring
cell. The cell temperatures across the pack are likely quite uniform
given that there is only one battery box and it is mostly insulated.
At this point it looks like the most a person needs is a half pack
voltage comparison monitor (AKA Lee Hart Battery Bridge) to check the
health of a pack. This is in addition to a charger which will
terminate charge consistently at the proper cutoff voltage and
current. I read of one person, in the UK I think, who had a cell go
dead on him and he didn't know it for some unknown amount of time. It
didn't smoke or cause a fire. It basically acted like a shorted or
missing cell.

Someone reported, on the TS group IIRC, that he left a cell on his
CC/CV bench supply set to 4V and forgot it over a couple of days. When
he came back it was swollen. Some questioned whether the bench supply
was defective but I think that the cell was over charged. The CALB
spec sheet lists a "float voltage" of 3.4V. If this number is
accurate, this means that holding a cell somewhere above this voltage
could overcharge the cell causing plating of the lithium ions and/or
breakdown of the electrolyte. So far this agrees with my understanding
of how the cell works.

In summary, there is very little energy capacity in a LiFePO4 cell
above 3.4-3.5V. Charging to a higher voltage puts more stress on the
cell, brings it closer to damaging voltage/charge levels, and puts the
cells closer to a chance of a fire when no one is likely to be around
to put it out until it is too late. Even if I'm off base with my
conclusions I'm gaining potentially longer cell life at the expense of
at most 1% of my pack's capacity so I'm missing out on at most 0.7 mi
of range which can easily be gained by driving gentler and slower or
stopping and plugging in to a 120V outlet for 10 minutes.

I hope I answered your questions. Feel free to ask for clarifications
or challenge my conclusions/assumptions/methods.

> Thanks,
>
> Roger.
>
>> On 14 Jun 2011, at 00:06, David Nelson wrote:
>>
>> > There are a couple of things to consider in sizing your LFP pack.
>> > Winston batteries (FKA ThunderSky) ... Furthermore, from my testing
>> > of cells on the bench with low ending current there is very little
>> > energy above 3.45V. I recommend using a cutoff voltage of 3.5vpc or
>> > less. I'm using 3.485vpc with my pack and it has been doing just fine
>> > for the past 8000 miles. The higher your ending voltage the more
>> > stress you are putting on the cells so why push it to the max?


--
David D. Nelson
http://evalbum.com/1328

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

gottdi
In reply to this post by martinwinlow
Yttrium introduces a contaminate into the structure much like carbon is introduced into iron to make steel. Iron is soft and not so strong but with the introduction of a tiny amount of carbon the structure of iron now becomes very strong and we call it steel. It is a deliberate contamination of the parent substrate to change its properties. In the LiFePO substrate yttrium is introduced in small amounts during manufacture and it strengthens the structure in a way to allow a much slower break down and therefore increase the life of the cell and maybe even increase the max voltage allowed and decrease the internal resistance by keeping the structure open better to allow a better exchange of ions.

You get a lower internal resistance which allows a higher amp draw with little to no degradation over time.
A higher voltage limit though it really is not needed.
Much longer cycle life. From 3000 cycles to 5000 cycles. Impressive.

All that extra by adding in a tiny bit of yttrium.
http://onegreenev.blogspot.com/
No need to wait any longer. You can now buy one off the shelf. You can still build one too.
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Christopher Darilek
In reply to this post by David Nelson-5
I've worked at a LiFePO4 manufacturer and with their cells they found that 100%
charge could be achieved charging to 3.4V so long as the current tapers low
enough.  At 3.33V you would only get 60% charged.  I think the higher charge
voltages are an effort to reach 100% faster.

-Chris




________________________________
From: David Nelson <[hidden email]>
To: Electric Vehicle Discussion List <[hidden email]>
Sent: Tue, June 14, 2011 8:54:45 PM
Subject: Re: [EVDL] figuire LFP life

I'll try to articulate how I have reached my current position on
LiFePO4 charging and care. I'm going from the studying of various
sources of information I've found including NASA summary documents,
the CMU professor's videoed talk, TS documentation (and its changes
over time), various individual reports of experiences with these
cells, my experience with my own cells (TS-LFP100AHA, mfg 11-2009) in
my EV and on the bench, and combining it with my Physics background. I
do not know what differences the LYP cells have compared to the LFP
cells since I have not found much info about it. As always, my
conclusions are open to modification as more data/information is
presented. I would love to do more thorough testing but my limited
resources don't allow it at the present. I would love to get some
1-5Ah WB cells for testing as this would decrease the cost of the
batteries and make destructive testing much cheaper.

On Tue, Jun 14, 2011 at 9:48 AM, Roger Stockton <[hidden email]> wrote:
> Martin WINLOW wrote:
>
>> I would be interested to hear what your theory is about why TS (WBL)
>> still recommends taking their LYP's to 4.0V on charge...?

My hypothesis on why they recommend 4.0V on charge is that it
virtually assures the cell reaches the maximum state of charge
possible. Remember that they first started by saying to charge to
4.25V then lowered it to 4.2V and now 4.0V. I think they realized that
the long term effects of high voltages were not good on their cells so
they have lowered the cutoff voltage but they still don't want to give
up that last fraction of a percent of capacity by lowering the cutoff
voltage further.

>
> David, I would appreciate it if you could clarify just what you mean when you
>refer to charging with "low ending current".
>

The TS docs I have read imply from the charging graphs that the cutoff
current is 0.015CA. For my 200Ah pack this would be 3A. My Zivan
charger shuts off somewhere below 0.2A. It pulls ~15W from the wall
just before shutting off and then draws ~3.5W sitting idle. Even
assuming 100% charging efficiency at this point 15W into my pack at
69.7V gives 215mA. Subtracting the idle power means the charge current
is less than 165mA. I'd call this extremely low ending current. Many
shunting BMS boards shunt somewhere around 1A so a charger has to put
out less than that if the shunting is to keep up with the charger. I'm
assuming that most chargers then are ending the charge cycle at less
than 1A which I would consider a low ending current except for very
low capacity packs like a 40-60Ah pack.

> A typical lithium profile is constant current to the target voltage, then hold
>that voltage until the current tapers to a suitably low level.
>

The charge profile is a combination of the target voltage and ending
current. Since only terminal voltage can be measured it has to be used
to determine end of charge but it varies quite a bit depending on the
charge current. It is important to pick the two so that cell life
isn't shortened. Taking the conservative side of the target voltage
will not hurt the cell, taking the high side may hurt the cell. If I
was in a hurry to top off a pack I might pick a higher target voltage
but stop charging at a higher ending current. With my 200Ah pack that
would be ending at something over 3A after reaching the target
voltage.

>Are you observing that if you use a lower target voltage, but then hold that
>voltage until the current tapers to a lower level, you achieve much the same
>state of charge/available capacity as if you were to use a higher target voltage
>but terminate the charge at a higher current level?
>

That is what I'm seeing. I did some playing around with a couple of my
100Ah cells before installing them into my pack. Here is a summary of
the charge testing I did. I do not have a way to record discharge
capacity to verify the results. Since relatively low currents were
used and the fact that these cells are essentially 100% efficient
(NASA document) I'm assuming that all the current went to charging the
cell. This assumption will actually inflate my values a little so the
real values will actually be slightly lower.

These tests were run on a TS-LFP100AHA cell manufactured on 2009 November 05
Test done by hand using a BK Precision 1761 Powersupply
Voltage readings at cell terminal using an Extech EX830
I took voltage and current readings every 30 sec to 1 min when the
values were changing rapidly and then at larger intervals when the
values changed more slowly. I then did a midpoint integration of the
voltage and current values over time to get my Ah values.
Unfortunately I didn't record ambient air temperature. I don't do well
with the cold so the air temp was very likely in the 18-20°C range.

In January 2010 I did the following tests.

3.600-4.00 Volt capacity test: 0.2213Ah
Starting with a single 100ah cell slightly discharged, charged the
cell to 3.600V and when the current dropped to below 30mA I started my
test. Maximum charge current was 3.486A which tapered off rapidly. The
test ran until the terminal voltage was 4.00V and current tapered to
23mA (0.00023C).

3.500-4.00 Volt capacity test: 0.3550Ah
The cell was discharged from the previous test until its open circuit
voltage was below 3.500V. The cell was again charged to 3.500V and
held at that voltage until the current tapered to below 30mA before
starting test. Maximum charge current was 3.485A which tapered off
rapidly after a couple of minutes. The test ran until the terminal
voltage was 4.00V and current tapered to 24mA (0.00024C).

3.397-4.00 Volt capacity test: 0.6588Ah
The cell was discharged from the previous test until its open circuit
voltage was below 3.4V. The cell was charged to 3.397V and held there
until the current tapered below 30mA before test was started. The
maximum charge current was 3.486A which started to taper off after 7
minutes. The test ran until the terminal voltage was 4.00V and current
tapered to 38mA (0.00038C).

As you can see from these tests, when the initial terminal voltage was
achieved by holding that voltage until a very low current and then
starting the test that there is very little capacity above 3.400V.
Less than 0.66% of the cell's capacity is above 3.40V. That is not
much to gain by pushing the cell to its voltage limit. For safety it
is definitely not worth it.

On July 14, 2010 I did another test on a cell to see what capacity it
had starting from 3.300ocv and going to 3.450V and then on to 3.65V.
Below I'll tell you why I decided to do this test.

I discharged a cell until it had an open circuit resting voltage of
3.300V. I kept discharging it until the voltage didn't bounce above
this point. I hooked it up and started charging it at 3.507A. This was
taking a long time to charge so at 300min I paralleled the two power
supplies and continued the test at 6.604A. This test continued with an
initial target voltage of 3.425V. I let the current taper to 1.501A
(0.015CA the TS ending current. This is coincidental, really.) By this
point the cell had received 63.18Ah. At this point I decided to go to
3.450V as the ending voltage. I let the current taper to 476mA
(0.00476CA) and recorded the energy put into the cell: 66.0155Ah. This
was 935 minutes from the start of the test.

Next I turned the current up to get a capacity value for 3.450-3.650V.
I let the current taper to 420mA. Energy added for this voltage range
was 1.0056Ah. Had I stopped at 1.5A (0.015CA) the energy would have
been 0.400Ah for this range.

>If so, what difference have you observed by using the higher target voltage, but
>allowing the current to fall to the lower target level before terminating?
>

I can't say what charging to the higher target voltage with a low
ending current did to my pack. When I first installed my pack I
installed only 18 cell pairs and adjusted the voltage calibration
screw in my Zivan NG1 to terminate at 4.00vpc. This is the voltage
that my Black Sheep Technology BMS boards are set to shunt. Given that
the charge profile programmed in the Zivan was for a lead acid pack
charging took a while at the end. What I did observe was that while
there were a few cells which usually began shunting before the others
they were not always the same ones. Calculating the energy difference
between the first cell to begin shunting and the last one showed a
difference in the 20mAh range, quite small for a 200Ah pack! Also, I
noticed that some cells would shunt for a little while and then stop
shunting while. This would happen just as the current would drop below
2-4A, IIRC. The next cell to start shunting was always after the
charge current dropped below 500mA (the max shunting capability of my
BMS boards) and then it wasn't always one of the first cells to start
shunting which would shunt at this point. I wish I had a video or
something of it so see if there was any change or trend to this
behavior. I ran my pack this way for a little over 6000Ah delivered
from the pack. I found that by the time I accelerated to 15mph the
pack voltage was down to less than 3.4vpc no load which is what
prompted my last capacity test: 3.45-3.65V.

At that point I had my Zivan reprogrammed for a LFP profile and
installed my last two cell pairs so I now have a 2p20s pack. At this
point I balanced all the cells at 4.00V using the BMS boards. I don't
have a charger which can charge to 80V so I did this in two passes
using my bench PS I used for the above tests. I adjusted the target
voltage on the Zivan for 69.7V (3.485vpc) and have not balanced my
cells since. At the end of charge the Zivan goes into this
on-off-on-off mode where it is trying to hold the target voltage for
an hour or so. This is when it is charging at less than 200mA. I can
hear when it shuts off and also look at my Kill-a-Watt meter to tell.
During these "off" times I take individual cell voltage readings. As
of June 1, 2011 the pack has delivered over 11,000Ah and the cells
have a voltage spread of 0.087V. The highest difference was 0.094V at
9,892Ah delivered (May 5, 2011) and the lowest difference was 0.055V
at 69Ah delivered (August 9, 2010). The high and low cells have
changed only once since I quit top balancing. I hope to see if there
is a trend but I'm not seeing one yet. The cells do trade places a bit
but otherwise not much else is happening.

I can think of three possible factors which could contribute to cell
drift. One is the fact that I have left the BMS boards in place and
they most likely do not have _exactly_ the same drain on the cells. No
matter how small the differences may be, time is sure to multiply it.
Maybe if they do get way out of balance I'll re-balance and then
remove the boards and see what happens. The second is that I can't
perfectly seal out dirt and water from getting to the top of the
cells. They usually don't get water on them but it happens
occasionally. Finally, there is a chance that there are cell defects
so one cell may age at a different rate than another one.

>From what I've been able to find there is no charge shuttle reaction
in LiFePO4 cells so there is no mechanism for self discharge (ie.
internal to the cell). Ions merely move from one plate to the other.
This is why they are so efficient. With out a charge shuttle reaction
there isn't anything to cause one cell to drift from a neighboring
cell. The cell temperatures across the pack are likely quite uniform
given that there is only one battery box and it is mostly insulated.
At this point it looks like the most a person needs is a half pack
voltage comparison monitor (AKA Lee Hart Battery Bridge) to check the
health of a pack. This is in addition to a charger which will
terminate charge consistently at the proper cutoff voltage and
current. I read of one person, in the UK I think, who had a cell go
dead on him and he didn't know it for some unknown amount of time. It
didn't smoke or cause a fire. It basically acted like a shorted or
missing cell.

Someone reported, on the TS group IIRC, that he left a cell on his
CC/CV bench supply set to 4V and forgot it over a couple of days. When
he came back it was swollen. Some questioned whether the bench supply
was defective but I think that the cell was over charged. The CALB
spec sheet lists a "float voltage" of 3.4V. If this number is
accurate, this means that holding a cell somewhere above this voltage
could overcharge the cell causing plating of the lithium ions and/or
breakdown of the electrolyte. So far this agrees with my understanding
of how the cell works.

In summary, there is very little energy capacity in a LiFePO4 cell
above 3.4-3.5V. Charging to a higher voltage puts more stress on the
cell, brings it closer to damaging voltage/charge levels, and puts the
cells closer to a chance of a fire when no one is likely to be around
to put it out until it is too late. Even if I'm off base with my
conclusions I'm gaining potentially longer cell life at the expense of
at most 1% of my pack's capacity so I'm missing out on at most 0.7 mi
of range which can easily be gained by driving gentler and slower or
stopping and plugging in to a 120V outlet for 10 minutes.

I hope I answered your questions. Feel free to ask for clarifications
or challenge my conclusions/assumptions/methods.

> Thanks,
>
> Roger.
>
>> On 14 Jun 2011, at 00:06, David Nelson wrote:
>>
>> > There are a couple of things to consider in sizing your LFP pack.
>> > Winston batteries (FKA ThunderSky) ... Furthermore, from my testing
>> > of cells on the bench with low ending current there is very little
>> > energy above 3.45V. I recommend using a cutoff voltage of 3.5vpc or
>> > less. I'm using 3.485vpc with my pack and it has been doing just fine
>> > for the past 8000 miles. The higher your ending voltage the more
>> > stress you are putting on the cells so why push it to the max?


--
David D. Nelson
http://evalbum.com/1328

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.sjsu.edu/mailman/private/ev/attachments/20110615/167887c2/attachment.html 
_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

David Nelson-5
Thank you for that info, Chris. It explains what I'm seeing and why
CALB lists 3.4V as a "float voltage."

On Wed, Jun 15, 2011 at 9:33 AM, Christopher Darilek
<[hidden email]> wrote:
> I've worked at a LiFePO4 manufacturer and with their cells they found that 100%
> charge could be achieved charging to 3.4V so long as the current tapers low
> enough.  At 3.33V you would only get 60% charged.  I think the higher charge
> voltages are an effort to reach 100% faster.
>
> -Chris


--
David D. Nelson
http://evalbum.com/1328

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

David Nelson-5
As an additional point on these LiFePO4 cells not having a self
discharge mechanism take a look at
http://media2.ev-tv.me/news061011-1280.mov (the file is 3.06 GB in
size). Starting at the 40min point it shows a new shipment of
WB-LYP400AHA cells being delivered to EVTV. Jack Rickard measures the
voltage on one and it reads 3.3076V. Next he opens a box which has
been sitting unopened since he got it. Inside are 4 cells which from
their size look like TS-LFP100AHA cells. The documentation in the box
state 2008 Oct 11 as the manufacture date. The film date is 2011 Jun
10. He takes a reading on each cell and gets the following: 3.3016V,
3.3020V, 3.2995V and 3.3021V.

If you have your cells and are not ready to install them in your car,
leave them alone!

--
David D. Nelson
http://evalbum.com/1328

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Lee Hart
On 6/15/2011 10:08 PM, David Nelson wrote:
> As an additional point on these LiFePO4 cells not having a self
> discharge mechanism... Jack Rickard measures the
> voltage on one and it reads 3.3076V. Next he opens a box which has
> been sitting unopened since he got it. Inside are 4 cells which from
> their size look like TS-LFP100AHA cells. The documentation in the box
> state 2008 Oct 11 as the manufacture date. The film date is 2011 Jun
> 10. He takes a reading on each cell and gets the following: 3.3016V,
> 3.3020V, 3.2995V and 3.3021V.

This tells you that they are not dead; but it doesn't tell you their
state of charge. The voltage barely changes at all from 20% to 80% state
of charge.

This makes the self-discharge rate very hard to measure. You would have
to do something like this:

1. Charge the cell to a known state of charge (say, to 3.8v at 0.01C
    amps). Call that 100% SOC.

2. Discharge it to a known voltage to determine its amphour capacity
    (say, to 2.5v at a 0.1C rate). Call that 0% SOC.

3. The amphours you got tells you cell's capacity.

4. Charge the cell again to the same ending voltage and current.
    Now it's at 100% again.

5. Now wait... weeks... months... the longer the better.

6. Discharge the cell (without charging it first) at the same rate
    and to the same ending voltage as for step 2 above.

7. See how many amphours you got. The difference between the full
    capacity with an immediate discharge, and the capacity you got
    after letting the cell sit for X amount of time will tell you
    the self-discharge rate.

For example, if you measured it as a 100 amhour in step 3, and a 90ah
cell in step 7 after sitting for 1000 hours, then it lost 10ah / 1000
hours = 0.01 amps, so its self-discharge current is 0.01a = 10ma.

--
Lee A. Hart | Ring the bells that still can ring
814 8th Ave N | Forget the perfect offering
Sartell MN 56377 | There is a crack in everything
leeahart earthlink.net | That's how the light gets in -- Leonard Cohen

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Jukka Järvinen-2
Charging to 4V or ever higher helps to absorb the Li-Ions to places
where they normally will not go.You'll be able to 'format' the cell
afterwards. Expensive and coherent particle structures have the effect
too but it's quite mild.

TS cells have been proven (believe or not) to work better if charged
high every now and then. Keeping the cells at high voltages for longer
time will murder the electrolyte. It dissolves. How long time you keep
the cell at high voltage and/or temperature hastens the degration.

How to do it 'right':

Charge the cells and voltage balance to 3,85 V. Then take the whole
pack up to 4,0-4,25 for few minutes. After getting all cells to the
top check the measured As (Ah/3600) between 3,85 and 4,25 per cell.
This will help you to know how to balance while operating at the
'flat' which Lee reminded us.

Now .. in general.. staring the voltages is pretty much meaningless if
you do not know the cell properties. Long life of the LiFePO4 is
forgiving and many get away for many years without BMS. But
eventually.. someday... it will not be enough anymore. Cells will
float apart and pack will be unmanageable. Change on cell and do it
one by one... fine.. what ever suits you best.

Ever driven your EV 100.000 miles ? Just driving not 'hobbying' ? This
is what I'm talking about. Not arguing if anyone should sit on the
cells every time they're charged.

Personally.. I will never 'go back' to that with a daily driver.

-akkuJukka

http://www.google.com/profiles/jarviju#about



2011/6/16 Lee Hart <[hidden email]>:

> On 6/15/2011 10:08 PM, David Nelson wrote:
>> As an additional point on these LiFePO4 cells not having a self
>> discharge mechanism... Jack Rickard measures the
>> voltage on one and it reads 3.3076V. Next he opens a box which has
>> been sitting unopened since he got it. Inside are 4 cells which from
>> their size look like TS-LFP100AHA cells. The documentation in the box
>> state 2008 Oct 11 as the manufacture date. The film date is 2011 Jun
>> 10. He takes a reading on each cell and gets the following: 3.3016V,
>> 3.3020V, 3.2995V and 3.3021V.
>
> This tells you that they are not dead; but it doesn't tell you their
> state of charge. The voltage barely changes at all from 20% to 80% state
> of charge.
>
> This makes the self-discharge rate very hard to measure. You would have
> to do something like this:
>
> 1. Charge the cell to a known state of charge (say, to 3.8v at 0.01C
>    amps). Call that 100% SOC.
>
> 2. Discharge it to a known voltage to determine its amphour capacity
>    (say, to 2.5v at a 0.1C rate). Call that 0% SOC.
>
> 3. The amphours you got tells you cell's capacity.
>
> 4. Charge the cell again to the same ending voltage and current.
>    Now it's at 100% again.
>
> 5. Now wait... weeks... months... the longer the better.
>
> 6. Discharge the cell (without charging it first) at the same rate
>    and to the same ending voltage as for step 2 above.
>
> 7. See how many amphours you got. The difference between the full
>    capacity with an immediate discharge, and the capacity you got
>    after letting the cell sit for X amount of time will tell you
>    the self-discharge rate.
>
> For example, if you measured it as a 100 amhour in step 3, and a 90ah
> cell in step 7 after sitting for 1000 hours, then it lost 10ah / 1000
> hours = 0.01 amps, so its self-discharge current is 0.01a = 10ma.
>
> --
> Lee A. Hart             | Ring the bells that still can ring
> 814 8th Ave N           | Forget the perfect offering
> Sartell MN 56377        | There is a crack in everything
> leeahart earthlink.net  | That's how the light gets in -- Leonard Cohen
>
> _______________________________________________
> | Moratorium on drag racing discussion is in effect.
> | Please take those discussions elsewhere.  Thanks.
> |
> | REPLYING: address your message to [hidden email] only.
> | Multiple-address or CCed messages may be rejected.
> | UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
> | OTHER HELP: http://evdl.org/help/
> | CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
>

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Roger Stockton
In reply to this post by Lee Hart
Lee Hart wrote:

> On 6/15/2011 10:08 PM, David Nelson wrote:
> > As an additional point on these LiFePO4 cells not having a self
> > discharge mechanism... Jack Rickard measures the
> > voltage on one and it reads 3.3076V. Next he opens a box which has
> > been sitting unopened since he got it. Inside are 4 cells which from
> > their size look like TS-LFP100AHA cells. The documentation in the box
> > state 2008 Oct 11 as the manufacture date. The film date is 2011 Jun
> > 10. He takes a reading on each cell and gets the following: 3.3016V,
> > 3.3020V, 3.2995V and 3.3021V.
>
> This tells you that they are not dead; but it doesn't tell you their
> state of charge. The voltage barely changes at all from 20% to 80% state
> of charge.
>
> This makes the self-discharge rate very hard to measure. You would have
> to do something like this:

[...]

And, if you did so, the literature indicates that you would very likely find that the self-discharge rate is in the 2-8%/month range, which is what is considered characteristic of lithium ion chemistry.

Also, when the sort of test Lee describes is performed (it is the standard method of measuring self-discharge), what will also be found is that when the battery is discharged (without charging) after the storage period, it will yield somewhat less capacity than it had provided when tested prior to storage (this difference is the amount lost to self-discharge), and when recharged and then discharged again, it will still not yield the same capacity as it had prior to storage!

Some (small) amount of capacity is permanently lost after being left stored.  (Some, also small, amount of capacity is lost each time you cycle the battery, so you can't really win; whether you use the cells or not, their capacity will slowly decrease over time ;^)

While you can't "win", you can at least try to tilt things in your favour.  The way to do that, if you must store some cells, is to store them at a partial state of charge.  They will still self-discharge (though perhaps at a slightly slower rate), however, they will recover to virtually their original capacity when you recharge them after storage.

Cheers,

Roger.


_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Roger Stockton
In reply to this post by gottdi
Gottdi wrote:

> Yttrium introduces a contaminate into the structure much like carbon is
> introduced into iron to make steel.

Your analogy probably is effective at communicating the general idea, though I don't believe your suggestions of just how/what the yttrium doping does are completely accurate ;^>

As I recall, doping the cathode with very small amounts of such a material (yttrium is just one possibility) does change the structure such that there are some additional irregularities that provide additional sites for lithium ions to intercalate into (= improved capacity).  The doping can also result in better dimensional stability of the cathode material (i.e. less volume change as ions move in/out of the structure over a charge/discharge cycle).  Better dimensional stability = longer cycle life as the cathode material experiences less mechanical stress during cycling.

I don't believe that doping the cathode material in this way significantly effects max voltage; typically the voltage is limited by the electrolyte (which decomposes at higher voltages).

If the doping changes the cathode structure such that lithium ions can move in and out of the structure more easily, then this will effectively appear as lower internal resistance.

Cheers,

Roger.


_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Mike Nickerson
In reply to this post by Roger Stockton
I had my conversion on blocks for some rewiring this past winter.  It wasn't
driven from about November to early March.  In about February, I got curious
about the battery state and got out my small charger to "top up" individual
banks of 5 cells.  I put about 20 Ah into the cells.  These are LFP100
cells, so they had lost about 20% in about 4 months.

Mike

> -----Original Message-----
> From: [hidden email] [mailto:[hidden email]] On
> Behalf Of Roger Stockton
> Sent: Wednesday, June 15, 2011 11:12 PM
> To: 'Electric Vehicle Discussion List'
> Subject: Re: [EVDL] figuire LFP life
>
> Lee Hart wrote:
>
> > On 6/15/2011 10:08 PM, David Nelson wrote:
> > > As an additional point on these LiFePO4 cells not having a self
> > > discharge mechanism... Jack Rickard measures the voltage on one and
> > > it reads 3.3076V. Next he opens a box which has been sitting
> > > unopened since he got it. Inside are 4 cells which from their size
> > > look like TS-LFP100AHA cells. The documentation in the box state
> > > 2008 Oct 11 as the manufacture date. The film date is 2011 Jun 10.
> > > He takes a reading on each cell and gets the following: 3.3016V,
> > > 3.3020V, 3.2995V and 3.3021V.
> >
> > This tells you that they are not dead; but it doesn't tell you their
> > state of charge. The voltage barely changes at all from 20% to 80%
> > state of charge.
> >
> > This makes the self-discharge rate very hard to measure. You would
> > have to do something like this:
>
> [...]
>
> And, if you did so, the literature indicates that you would very likely
find that
> the self-discharge rate is in the 2-8%/month range, which is what is
> considered characteristic of lithium ion chemistry.
>
> Also, when the sort of test Lee describes is performed (it is the standard
> method of measuring self-discharge), what will also be found is that when
> the battery is discharged (without charging) after the storage period, it
will
> yield somewhat less capacity than it had provided when tested prior to
> storage (this difference is the amount lost to self-discharge), and when
> recharged and then discharged again, it will still not yield the same
capacity as
> it had prior to storage!
>
> Some (small) amount of capacity is permanently lost after being left
stored.
> (Some, also small, amount of capacity is lost each time you cycle the
battery,
> so you can't really win; whether you use the cells or not, their capacity
will
> slowly decrease over time ;^)
>
> While you can't "win", you can at least try to tilt things in your favour.
The
> way to do that, if you must store some cells, is to store them at a
partial state
> of charge.  They will still self-discharge (though perhaps at a slightly
slower
> rate), however, they will recover to virtually their original capacity
when you

> recharge them after storage.
>
> Cheers,
>
> Roger.
>
>
> _______________________________________________
> | Moratorium on drag racing discussion is in effect.
> | Please take those discussions elsewhere.  Thanks.
> |
> | REPLYING: address your message to [hidden email] only.
> | Multiple-address or CCed messages may be rejected.
> | UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
> | OTHER HELP: http://evdl.org/help/
> | CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Roger Stockton
In reply to this post by David Nelson-5
David Nelson wrote:

> I hope I answered your questions.

Yes; thank you.

> Feel free to ask for clarifications
> or challenge my conclusions/assumptions/methods.

Well, OK ;^>

> The TS docs I have read imply from the charging graphs that the cutoff
> current is 0.015CA. For my 200Ah pack this would be 3A. My Zivan
> charger shuts off somewhere below 0.2A. It pulls ~15W from the wall
> just before shutting off and then draws ~3.5W sitting idle. Even
> assuming 100% charging efficiency at this point 15W into my pack at
> 69.7V gives 215mA. Subtracting the idle power means the charge current
> is less than 165mA. I'd call this extremely low ending current. Many
> shunting BMS boards shunt somewhere around 1A so a charger has to put
> out less than that if the shunting is to keep up with the charger. I'm
> assuming that most chargers then are ending the charge cycle at less
> than 1A which I would consider a low ending current except for very
> low capacity packs like a 40-60Ah pack.

[...]

> The charge profile is a combination of the target voltage and ending
> current.

I find it somewhat ironic that you acknowledge this, yet above admit that while the manufacturer recommends terminating charge at 0.015CA (3A) and (I presume) 4.0V/cell, you are charging to a completely different ending current and voltage ;^>

> Since only terminal voltage can be measured it has to be used
> to determine end of charge but it varies quite a bit depending on the
> charge current.

This I don't follow.  If the charger is holding constant voltage, then the voltage should not be varying very much at all with variations in current.

For instance, if I use a bench supply and set it to 10A max and 3.6V max, then it will charge at 10A max until the voltage reaches 3.6V, and then will hold 3.6V precisely while the current tapers off.  If I connect the supply to my cell with long, thin wires, then when the supply first "sees" 3.6V at its end of the wires (at 10A current), the voltage at the cell terminals will indeed be somewhat *less* than 3.6V due to the voltage drop in the wiring.  As the power supply holds 3.6V at its end of the wires, the current drops and so therefore does the voltage drop in the wiring.  The result is that as the current drops, the voltage at the cell end of the wires *rises* until it is virtually at the same 3.6V as the power supply end of the wires (and the current is small).

> >Are you observing that if you use a lower target voltage, but then hold
> >that voltage until the current tapers to a lower level, you achieve much
> >the same state of charge/available capacity as if you were to use a higher
> >target voltage but terminate the charge at a higher current level?
> >
>
> That is what I'm seeing. I did some playing around with a couple of my
> 100Ah cells before installing them into my pack. Here is a summary of
> the charge testing I did.

[...]

> As you can see from these tests, when the initial terminal voltage was
> achieved by holding that voltage until a very low current and then
> starting the test that there is very little capacity above 3.400V.

Perhaps, however, I'm not sure that these tests prove quite what you think they do. ;^>

Remember, the manufacturer's recommendation is to terminate the charge at 0.015CA (=3A), not 1/100th of this, which is what you used in your tests.  Also, remember that the recommended float voltage for these cells is 3.4V, which is at or below the voltage to which you charged *prior* to raising the cell voltage to 4V.

I think that what you have successfully demonstrated is that the cell can be essentially fully charged using a end of charge voltage at or above the nominal (and float) voltage, provided the voltage is maintained until the current is *very* low.

This is certainly useful information, but may not really answer the original question, which I think is more along the lines of "how much reduction in capacity might result from terminating the charge at (say) 3.5V and 0.015CA rather than 4.0V and 0.015CA".  And, indeed, there is a distinct possibility that this regimen could in fact *over* charge the cells compared to what the manufacturer specifies.  That is, charging to 4.0V and 3A could actually result in the cells being *less* fully charged than charging to 3.5V and 0.03A.  It is entirely possible that your "conservative" charge approach might lead to a greater amount of electrolyte decomposition as a result of routinely delivering more energy into the cell than the manufacturer's recommended profile.

Another possible shortcoming of your test regimen is that you are only shallowly discharging the cell prior to recharging, and so the test may not reveal any potential benefits of the higher termination voltage that might result when recharging after a deeper discharge.

I think I would come at this test from the other direction.  I would discharge the cell a fixed amount, perhaps 100Ah or 50% of rated capacity.  Then recharge using the manufacturer's termination criteria.  Note how many Ah are returned.  Then, repeat the discharge and then recharge using my preferred lower termination voltage.  When the Ah returned reaches the same value as in the first test, note the charge current and terminate the charge.  This termination current and voltage pair result in the same charge return as the manufacturer's recommended profile, while allowing me greater peace of mind by keeping the maximum voltage lower.

You raise a valid point regarding the limitations of readily available shunt regulators with respect to terminating the charge at the sort of final current recommended by the manufacturer for larger cells such as this.  That would lead me to a further variation on this test.  Perform a benchmark cycle using the manufacturer's termination criteria, as describe previously.  Then repeat the test using a more conservative termination voltage and note the current at the point when the same Ah have been returned as in the first test.  If this termination current is still above what my shunt regulators can tolerate, then repeat the test using successively more conservative termination voltages until the termination current is observed to be within the reg capabilities.  This termination criteria pair will likely result in a longer charge duration, however, you will have the peace of mind knowing that not only is it unlikely that any of your cells will reach the reg setpoint, but if!
  they do, the reg will be able to comfortably keep the cell in check.

Finally, bear in mind that it is Wh, not Ah that matters when it comes to usable capacity.  It is possible that differences in the charge termination voltage have only a small effect on Ah, but if they also affect the voltage on discharge then the effect on usable Wh may become significant. (So, it may be useful to record voltage throughout each discharge cycle in any testing so that differences in usable Wh can be considered as well.)

> From what I've been able to find there is no charge shuttle reaction
> in LiFePO4 cells so there is no mechanism for self discharge (ie.
> internal to the cell).

I've seen nothing in the literature to suggest that LFP is immune to the 2-8%/month self-discharge that is characteristic of all other lithium ion variants.  The self-discharge rate typically varies with temperature, so not only is it likely to vary from cell-to-cell due to internal variations, but it will also vary due to variations in temperature between cells.

> Someone reported, on the TS group IIRC, that he left a cell on his
> CC/CV bench supply set to 4V and forgot it over a couple of days. When
> he came back it was swollen. Some questioned whether the bench supply
> was defective but I think that the cell was over charged. The CALB
> spec sheet lists a "float voltage" of 3.4V. If this number is
> accurate, this means that holding a cell somewhere above this voltage
> could overcharge the cell causing plating of the lithium ions and/or
> breakdown of the electrolyte. So far this agrees with my understanding
> of how the cell works.

IIRC, it takes something like 160-200% overcharge (or more) before lithium plating would occur.  In your tests, holding the cell at 4.0V, the current tapered to about 0.03A.  160% overcharge for a 100Ah cell is 160Ah, and at 0.03A the cell would have to be left on charge for 5333hrs (222 days and change) before this amount of overcharge could be delivered.  I'm assuming that "a couple days" is much less time than this? ;^>

I don't know if the electrolyte decomposition simply occurs at exponentially increasing rates as the voltage is increased, or if there is some threshold voltage below which decomposition is negligible and above which it is extremely rapid.  I expect that the rate of decomposition is fairly modest below 4.25V/cell.

> Even if I'm off base with my
> conclusions I'm gaining potentially longer cell life at the expense of
> at most 1% of my pack's capacity

Even at 11000Ah discharged, this is equivalent to only 69 80%DOD cycles for your 200Ah cells.  You have barely scratched the surface of your cells' rated life, and so it may be far to soon to make any assumptions about just what the life will be, or why.

As I explained above, you may actually be doing your cells more harm than good with your charge regimen if your termination current is too much less than appropriate for your charge voltage.  You have used so little of your cell's life that you really can't yet tell if they are on track to even approach the rated life, much less exceed it.

If you truly want to be conservative in your charge termination, then use the manufacturer's recommended termination current, but your preferred lower charge voltage.  This will give up a somewhat greater amount of your possible capacity, but probably still not enough to be of practical importance to you.  Continuing to use the same termination current despite lowering the charge voltage ensures that you are returning less energy to the battery than the manufacturer's recommendation.  You are still taking the gamble that never charging to the manufacturer's recommended termination criteria won't result in battery degradation...

Cheers,

Roger.


_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

gottdi
In reply to this post by gottdi
gottdi wrote
..........and maybe even increase the max voltage allowed................
To be accurate I did say maybe it allows for higher voltages but I would agree about the electrolyte. It could be that they actually did some changes to the electrolyte too. That may have to come from Winston themselves to know for sure.

As for self discharge. Well I have a cell that is in a very very very low SOC and it has been at that state of charge since I have received the cells I am using. After being in my possession since last year there is no change what so ever. On the other hand my lead acid deep cycle cell that was at a full SOC is now almost dead. The cell is clean and full of electrolyte and used maybe three times before I pulled them from the vehicle when we dismantled the vehicle. So in a shorter time frame the full (soc) lead battery just up and died while the very low (soc) battery has remained the same. I think that is a very accurate statement. There is no self discharge. If there happens to be a self discharge it is so slow I will be nothing but a bag of bones before that happens.

Pete :)
http://onegreenev.blogspot.com/
No need to wait any longer. You can now buy one off the shelf. You can still build one too.
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Lee Hart
On 6/16/2011 8:36 AM, gottdi wrote:
> As for self discharge. Well I have a cell that is in a very very very low
> SOC and it has been at that state of charge since I have received the cells
> I am using. After being in my possession since last year there is no change
> what so ever. On the other hand my lead acid deep cycle cell that was at a
> full SOC is now almost dead... So in a shorter time frame the full (soc) lead
> battery just up and died while the very low (soc) battery has remained the
> same. I think that is a very accurate statement. There is no self discharge.
> If there happens to be a self discharge it is so slow I will be nothing but
> a bag of bones before that happens.

There is no self-discharge from a dead cell, because there is no energy
left to leak away.

A good lead-acid AGM can easily take years to self-discharge. I have
bought Hawkers that have sat in a sealed box for 3 years, and still had
half a charge when I took them out. It isn't good for the battery; but
they survived it with only moderate loss of capacity.

Good lithiums can be even better. I have a few 10-year-old Tadiran
lithium primary cells that are unused, and still read over 3.6v. I have
some A123 cells about 3 years old that are still over 3.25v.

But, many lithiums I have tested have not been nearly that good. Their
rates are more like Roger mentioned, a few percent a month. Worse, it
varies considerably between cells, and with temperature, etc. This is
one of the reasons I think a BMS is needed.
--
Lee A. Hart | Ring the bells that still can ring
814 8th Ave N | Forget the perfect offering
Sartell MN 56377 | There is a crack in everything
leeahart earthlink.net | That's how the light gets in -- Leonard Cohen

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

Steve Clunn
In reply to this post by Peakfoto Digital Photo Still n Video
Message: 8
Date: Tue, 14 Jun 2011 18:54:45 -0700
From: David Nelson

>I'll try to articulate how I have reached my current position on
LiFePO4 charging and care.<

Your doing a good job of keeping data , I've wrote about the abused
100ah fox cells I got from a customers car and wonder if what I'm
seeing with this pack half of them are in Audrey's green bean and half
in my pick up.,    is what might happen to a pack after many cycles .
They don't stay balanced very long and run down on their own pretty
quick and do it to different degrees. I had separated  20 of the worst
one's out and had them sitting in front of the shop trying to breath
some life back into them. . Of the 20,    9 seemed to have some life,
( less than 40 amps, if that can be called "life".   I hooked two
Anderson's plugs on them to make it easy to charge and discharge them
when I would charge one of my EV's  ( I did this with a funny looking
3 set plug Anderson series plug that would put this test pack in
series with the charger so to discharge the test pack when charging
some EV.  I have a e meter and timer, ( from an old close dryer max 80
min)  on all this so  to turn things off if I forget . This last month
has been a whirl wind of activity  as we where getting ready to hit
the road . There are many people who have projects at my shop and even
though I've had little work all year , now everybody wants their
project done before we go ( explaining reason for careless behavior) .

 As I'm passing these cells in and out of the shop I check them with
my meter .most I had given up on but I had 9 in 3 groups of 3 to make
a 12V battery ( they where all compressed together  ) and one group
had fallen below 2v. the others where about 3.2. . So in a hurry I
decided to put a 12V car battery  charger  on them and pep them up for
a few hours before leaving  . It was one of those shuemocker smart
chargers which are hard to keep going on a lead acid battery if it
isn't good . So when I saw 9v on the charger I though "It won't stay
on for long "and went back to work, till I was so tired that I forgot
and went home .

Next day back to work in the other shop and didn't notice the pile of
melted and burnt batteries in front of the green shed till noon.. So I
went to the house and told Audrey I has a little surprise  in the shed
she should see and videoed this candid moment .  I'll get her to post
the video.  I wasn't going to write about it but hopefully others can
learn from my mistake. What made me put these batteries out side  a
few months ago was hearing about the stories of the fire in someone
else's garage.

He is a little joke I told Audrey  to  make her feel better and make
her laugh....

" There once was an EV conversionist, who did a wonderful job on a
conversion, but had miswired the motor,
when the customer got in the car and put it in 1st gear, and hit the
GO peddle, he went flying in reverse into the back of the garage
causing heavy damage to the newly converted car.  The conversionist,
realizing his mistake wandered off into the woods and spoke to GOD.
and GOD look down at this man crying out.. WHY??? WHY:??

 And god looked down taking pity on him and said...
" my son, you have tried hard and done so with a good heart because of
this I am going to grant you one wish".
The  poor feller, said  IF ONLY we could have a battery that had
1000's of AMP hours and somehow would just charge itself on its own,
just by sitting, we would have Electric Cars everywhere.!

God took a deep breathe and said... I can't do that for many reasons,
..Not that it is not possible for ME to do this...but....first  It
violates the laws of Physics,and second,  I have many people working
on Battery Technology right now, there are great things to come in the
future if you will just be patient.

So God said, pick another wish.
So the man thought and replied..... " Well, how about just letting me
understand my mates moods.  She can be so happy one minute and so sad
the the next.  How one silly statement can up send her to the point of
tears......and then a few kind words can bring tears of joy.

God replies....... What Voltage did you want that Battery???


Steve Clunn


 I

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

gottdi
In reply to this post by Lee Hart
Cell voltage at 2.2 volts for 8 months with no change has no self discharge when the lead acid battery next to it has a significant voltage loss. The 2.2 voltage of the cell is at the knee of the discharge rate so  if there is a self discharge rate even at 1% we should see a significant voltage drop across the pack of 38 cells if that is the size pack you have.

So a cell at 2.2 volts that sits for 8 months and has a 3% self discharge rate would be sitting at 1.672 volts.

And if that same cell were only getting 1% self discharge then after 8 months of sitting it would equate to 2.024 volts. That is significant. Multiply that by 38.

2.2 volts each times 38 is 83.6 volts.
After 8 months that would be 76.912 volts.

That would be a 6.68 volt drop in 8 months time doing nothing. At 1% self discharge.

So you are saying that a cell will only self discharge to a certain point then just stop self discharging? I thought you could self discharge all the way to absolute zero. If there is voltage then there is something to discharge.
 
I am also going to assume you mean that as the self discharge voltage drops the self discharge rate also decreases to a point that you really never quite get to zero.

Mmmmmmmm.

I am not seeing that at all.

Pete :)

Remember that these are not Lead Acid. Don't assume they act like them. They don't.
http://onegreenev.blogspot.com/
No need to wait any longer. You can now buy one off the shelf. You can still build one too.
Reply | Threaded
Open this post in threaded view
|

Re: figuire LFP life

David Nelson-5
In reply to this post by Jukka Järvinen-2
2011/6/15 Jukka Järvinen <[hidden email]>:
> Charging to 4V or ever higher helps to absorb the Li-Ions to places
> where they normally will not go.You'll be able to 'format' the cell
> afterwards. Expensive and coherent particle structures have the effect
> too but it's quite mild.

Isn't this done at the factory with the first capacity test? I
understand that they capacity test the cell and then take it to
~50%SOC before shipment.

> TS cells have been proven (believe or not) to work better if charged
> high every now and then. Keeping the cells at high voltages for longer
> time will murder the electrolyte. It dissolves. How long time you keep
> the cell at high voltage and/or temperature hastens the degration.

How often is "every now and then?" I would think this would depend on
some Ah delivered by the cell figure. Also, by "work better" do you
mean they output power better, have higher capacity, or have a longer
cycle and/or calendar life?


> How to do it 'right':
>
> Charge the cells and voltage balance to 3,85 V. Then take the whole
> pack up to 4,0-4,25 for few minutes. After getting all cells to the
> top check the measured As (Ah/3600) between 3,85 and 4,25 per cell.
> This will help you to know how to balance while operating at the
> 'flat' which Lee reminded us.

Is there data on only charging to ~3.5vpc to compare this to? I
haven't seen any. Maybe it isn't available yet. How do you know that
the method you outlined is the "right" method? I wasn't able to find
the data on your site last time I looked but it has been a few months
so maybe it is there now.

> Now .. in general.. staring the voltages is pretty much meaningless if
> you do not know the cell properties. Long life of the LiFePO4 is
> forgiving and many get away for many years without BMS. But
> eventually.. someday... it will not be enough anymore. Cells will
> float apart and pack will be unmanageable. Change on cell and do it
> one by one... fine.. what ever suits you best.

The question I have is this. When the cells become "unmanageable" is
this also the end of life of the pack where it would be replaced any
way? Maybe a half-pack voltage comparison monitor is all that would be
needed.

> Ever driven your EV 100.000 miles ? Just driving not 'hobbying' ? This
> is what I'm talking about. Not arguing if anyone should sit on the
> cells every time they're charged.

If I could have done the conversion I wanted then I be at 40k mile by
now but no, I'm only at ~8000 miles in the last 18 months. My pack is
very young in the cycle life department. It is at ~90 cycles
equivalent looking at delivered Ah.

> Personally.. I will never 'go back' to that with a daily driver.

Are you saying you will never go back to manually monitoring cells on
a daily driver? That you will have some sort of automated monitoring
and balancing system?

>
> -akkuJukka
>
> http://www.google.com/profiles/jarviju#about

Thank you for your input. I hope I'm not sounding like I'm attacking.
I really want to know but it is difficult to sort through the "believe
me because I said so" stuff from the statements backed by data.
Furthermore, there are so many variables in how LiFePO4 batteries are
charged and used it is difficult to sort out what factors caused the
results.

--
David D. Nelson
http://evalbum.com/1328

_______________________________________________
| Moratorium on drag racing discussion is in effect.
| Please take those discussions elsewhere.  Thanks.
|
| REPLYING: address your message to [hidden email] only.
| Multiple-address or CCed messages may be rejected.
| UNSUBSCRIBE: http://www.evdl.org/help/index.html#usub
| OTHER HELP: http://evdl.org/help/
| CONFIGURE: http://lists.sjsu.edu/mailman/listinfo/ev
12