DIY Electric Car Forums banner

1 - 20 of 26 Posts

·
Registered
Joined
·
70 Posts
Discussion Starter #1
although I'm not sure there are many here with serious chemistry
knowledge I will float the question anyway

is there a fundamental difference, balancing wise, between a large Ah
cell or several smaller of same combined capacity in parallel?

Dan

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #2
Dan,

I don't have much chemistry theory, but I think there's an electric circuits
explanation.

The small batteries in parallel can have slightly different values of
internal resistance. Because of this, the batteries in parallel will
discharge at different rates, so one will get to empty before the others.

Also, as they discharge at different rates, the internal cell voltages might
no longer be the same, so when you change the load, you could get
circulating currents from one battery charging another. Say one battery is
at 12.5V and the other is at 12.4V and their combined internal resistance is
15mOhm. Then You'll get 6.7A from the one battery into the other battery;
this is just a waste of energy.

Overall, fewer high Ah cells is easier to balance than more low Ah cells in
parallel.

If you use large strings of cells in parallel, it becomes much easier to
deal with these problems, and there are fewer locations where circulating
currents can occur.

This is just using my common sense and what I've read about batteries
online; I'm no expert, but I think that's how it works.

-Morgan
_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #3
My big question is how reliable it would be to string together cells into
larger packs without having an individual monitor on every single cell?
Like the person who suggested stringing C-sized Lifepos into a long tube,
you'd only be able to monitor the combined properties of the cells, not the
individual cells themselves. If an individual cell were to go bad in the
middle of the tube, it would be hard to isolate it.

-----Original Message-----
although I'm not sure there are many here with serious chemistry
knowledge I will float the question anyway

is there a fundamental difference, balancing wise, between a large Ah
cell or several smaller of same combined capacity in parallel?

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #4
[email protected] <[email protected]> wrote:
> My big question is how reliable it would be to string together cells into
> larger packs without having an individual monitor on every single cell?
> Like the person who suggested stringing C-sized Lifepos into a long tube,
> you'd only be able to monitor the combined properties of the cells, not the
> individual cells themselves. If an individual cell were to go bad in the
> middle of the tube, it would be hard to isolate it.

Without battery monitoring/equalization, once the cells start to get
unbalanced, you'll overcharge some while charging and over-discharge
others while driving; this will really shorten the life of your
battery pack.

You can just cut slits in the side of the tube and put in tabs for the
battery management system there. You still have to get the BMS,
though; I don't know of an existing solution that you or I could buy.
For anything mass-produced, it would be stupid (economically/range
wise) to not use BMS, but for the first hobbyist solutions there may
not be many available/affordable BMS options.

-Morgan

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #5
Its very simple to find a bad or low cell in a series string. Takes only
about 5 minutes to find the bad cell. I use to have ninety 2 volt cobalt
cells in series for a 180 volt pack.

Take a volt meter reading across the whole battery and lets say it reads 178
volts. Then read one half of the batteries and lets say one side reads 90
volts and the other half reads 88 volts.

You then split the 88 volt half into two more halves which you may read 45
volts in one side and 43 volts on the other. You then keep splitting the
lower voltage section until you get down to one cell.

This was the methods we use when we want to find a bad street light that
were wire all in series. Each city block had a return loop at one corner of
the block to perform this test.

I still perform this test with my 6 volt batteries about every four months
and found two of them that had a voltage different of 0.02 with the rest of
the batteries. I than finish charge these two batteries for about 1 minute
after I do a normal charge with the main battery charger.

Roland


----- Original Message -----
From: "Morgan LaMoore" <[email protected]>
To: "Electric Vehicle Discussion List" <[email protected]>
Sent: Tuesday, September 11, 2007 8:25 PM
Subject: Re: [EVDL] Battery theory


> [email protected] <[email protected]> wrote:
> > My big question is how reliable it would be to string together cells
> > into
> > larger packs without having an individual monitor on every single cell?
> > Like the person who suggested stringing C-sized Lifepos into a long
> > tube,
> > you'd only be able to monitor the combined properties of the cells, not
> > the
> > individual cells themselves. If an individual cell were to go bad in
> > the
> > middle of the tube, it would be hard to isolate it.
>
> Without battery monitoring/equalization, once the cells start to get
> unbalanced, you'll overcharge some while charging and over-discharge
> others while driving; this will really shorten the life of your
> battery pack.
>
> You can just cut slits in the side of the tube and put in tabs for the
> battery management system there. You still have to get the BMS,
> though; I don't know of an existing solution that you or I could buy.
> For anything mass-produced, it would be stupid (economically/range
> wise) to not use BMS, but for the first hobbyist solutions there may
> not be many available/affordable BMS options.
>
> -Morgan
>
> _______________________________________________
> For subscription options, see
> http://lists.sjsu.edu/mailman/listinfo/ev
>

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #6
My guess/hope is that because the LiFePo4 chemistry expects a cc/cv
charge that they will play well together in parallel. During the
Constant Voltage part of the charge they will self equalize. The
critical part is not to go above a certain voltage during charge and or
below a certain voltage during operation. Since voltage is the same by
definition across cells in parallel, this should work well. I cant see
them working in a tube in series easily.

So... take pack voltage and divide by cell voltage to get number of bms
modules
ie 300V/3.2V = 94 boards.
Now to have reasonable charge times, we need a certain percentage of
capacity in balance capability that will determine the amps these boards
must shuttle or dissipate.

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #7
Jeff Shanab wrote:
> My guess/hope is that because the LiFePo4 chemistry expects a cc/cv
> charge that they will play well together in parallel.

Laptop battery packs generally parallel cells directly (because it's
cheap and easy). The drawback is that if a cell shorts, the others in
parallel also dump into it, leading to fires or explosions.

> During the Constant Voltage part of the charge they will self equalize.

I'm not so sure. The voltage-vs-SOC charge curve is so flat that the
self-balancing effect may be negligible. Even for lead-acids it takes
days for cells in parallel to self-balance to the same state of charge.
With lithiums it could take weeks or months.

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

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #8
[email protected] wrote:
> My big question is how reliable it would be to string together cells
> into larger packs without having an individual monitor on every
> single cell?

With lithiums, you need to have individual monitors on every single
cell. If you don't, you have a high risk of catastrophic failure (fires,
explosions, etc.)

> is there a fundamental difference, balancing wise, between a large
> Ah cell or several smaller of same combined capacity in parallel?

Yes. When it's all one big cell, it tends to all be at the same
temperature, and all parts of the cell were obviously made at exactly
the same time and of the same material composition.

With many small cells in parallel, they can be at different
temperatures, and from different production runs or batches of
materials. Some battery chemistries (like lithium) do a poor job of
staying at the same state of charge if simply connected in parallel. If
you connect (say) an 80% charged and a 20% charged cells in parallel, it
takes an impractically long time for them to move to the same state of
charge (50%-50%).

The big problem in balancing lithium cells is that you must prevent each
and every cell from every being overcharged or undercharged. This is
hard to do if you directly connect them in parallel.

There are also problems if one cell shorts or opens. If you can't detect
such failures and keep using the pack anyway, they can lead to much
bigger failures.
--
Ring the bells that still can ring
Forget the perfect offering
There is a crack in everything
That's how the light gets in -- Leonard Cohen
--
Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.net

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #9
I won't know until I actually test, but I was hoping that it wasn't
going to be that bad. The lower internal resistance of A123 type cells
will need less of a voltage differential that corresponding lead acids
to self equalize. Barring the catastrophic failure, handled differently,
the under and over voltage protection should still work as in parallel
the voltage is equal by definition. The lowest average capacity will
determine the effective capacity of the paralleled cells. equalizing
will increase during a discharge if the state of charge is imbalanced
And temperature is very important. It becomes important to have well
matched cells that are evenly cooled.(or heated)

Building a fuse into the interconnects will allow a catastrophic single
cell failure to not start a fire, taking that cell out of the pack.

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #10
Jeff Shanab wrote:
> I won't know until I actually test, but I was hoping that it wasn't
> going to be that bad. The lower internal resistance of A123 type cells
> will need less of a voltage differential that corresponding lead acids
> to self equalize.

Try an experiment. Get two cells. Charge one, and discharge the other.
Connect them both in parallel for (say) 24 hours. Now separate them, and
load test each one to see how many amphours it contains.

If you're right, each cell would contain 50% capacity. My hypothesis is
that they will be more like 10%-90%.

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

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #11
>----Alkuper=E4inen viesti----
>L=E4hett=E4j=E4: [email protected]
>P=E4iv=E4m=E4=E4r=E4: 14.09.2007 17:29
>Vastaanottaja: "Electric Vehicle Discussion List"<[email protected]>
>Aihe: Re: [EVDL] Battery theory
>
>Jeff Shanab wrote:
>> I won't know until I actually test, but I was hoping that it wasn't
>> going to be that bad. The lower internal resistance of A123 type =

cells
>> will need less of a voltage differential that corresponding lead =

acids
>> to self equalize.
>
>Try an experiment. Get two cells. Charge one, and discharge the =

other. =

>Connect them both in parallel for (say) 24 hours. Now separate them, =

and =

>load test each one to see how many amphours it contains.
>
>If you're right, each cell would contain 50% capacity. My hypothesis =

is =

>that they will be more like 10%-90%.

I think that this really would be the outcome. Li-Ions have so flat =

voltage curve that they would self-equalize very slowly.
- But does it really matter? Try experiment no. 2 ( let's say that the =

cell capacity is 100 Ah each):

First:
>Get two cells. Charge one, and discharge the other. =

>Connect them both in parallel for (say) 24 hours. =


Now, one cell has 10 Ah and the other 90 Ah.

Next, _don't_ separate them. Just connect a load resistor, start to =

discharge and see what happens. =

The outcome should probably be somewhere between the two extremes:
a) you get out 20 Ah and one dead cell
b) you get out 100 Ah and two healthy cells

I don't have even a guess where the result will be, but this would be =

a very interesting experiment!

Seppo

>
>-- =

>Ring the bells that still can ring
>Forget the perfect offering
>There is a crack in everything
>That's how the light gets in -- Leonard Cohen
>--
>Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.
net
>
>_______________________________________________
>For subscription options, see
>http://lists.sjsu.edu/mailman/listinfo/ev
>



_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #15
Ian Hooper wrote:
> Well, speculate no longer! I set up the experiment (sort of) on my
> battery tester, just using a fully discharged cell as the load from a
> fully charged cell, and got the following discharge curves out of the
> charged cell:
>
> http://www.zeva.com.au/Equalisation.jpg

Thanks! This is the way to proceed, folks. We can speculate all we want,
but actual data is the only way to move forward with confidence.

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

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #16
Steven Ciciora wrote:
> I don't believe this is as big of a problem as some
> people might think it is. It shouldn't matter how
> long it takes, if the multiple cells are _always_
> connected to each other.

It still matters, because the cells are never identical.

Let's suppose their amphour capacities are identical, but their internal
resistances are different. Start with both fully charged. Discharge them
in 1 hour. If one has twice the internal resistance, it only supplies
half the amphours; so it ends up with one cell at 33% and the other at
66% SOC.

> If you started charging your buddy paired cells, one starting off fully
> charged, the other starting off fully discharged, as the more-charged
> cell reaches full, it's internal impedance goes up, and it will draw
> less current than the discharged cell.

But what if the internal impedance does *not* go up as it reaches full?
For nicad and nimh cells, their impedance goes *down* when they reach
full! For two cells in parallel, the fully charged one has the *lower*
voltage -- it hogs more current, so it overcharges more, gets hotter,
gases more, etc.

> Yes, the paralleled cells will always be at different
> temperatures (nothing is ever exactly the same), will
> have different capacities, and hence different states
> of charge. But their impedances will always be
> different, too, so they will draw and put out
> different amounts of current.

For lead-acids, the hotter the cell, the lower its internal resistance
and the higher its amphour capacity. This means the hotter cell hogs the
load and/or charging, causing it to get still hotter, etc.

All we're saying is that things like type of cell, internal resistance,
temperature, and all the other variations between them complicate
things. You can't blithely assume they will "play nice" together. You
have to *confirm* it through testing, and then have systems in place to
catch problems and deal with them before they become disasters.

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

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #17
yes, how refreshing, someone actually doing something. thanks!

Lee Hart wrote:
> Ian Hooper wrote:
>
>>Well, speculate no longer! I set up the experiment (sort of) on my
>>battery tester, just using a fully discharged cell as the load from a
>>fully charged cell, and got the following discharge curves out of the
>>charged cell:
>>
>>http://www.zeva.com.au/Equalisation.jpg
>
>
> Thanks! This is the way to proceed, folks. We can speculate all we want,
> but actual data is the only way to move forward with confidence.
>


_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #18
Self equalize was a bad term. I am talking about equalization when in
parallel and exposed to a charging voltage.
There must be a differential or current won't flow.

li-ion like CCCV charge.
hypothetically we start with these two cells at 20% and 90% and apply a
regulated charge voltage to them while connected in parallel.

wouldn't the 90%full tend to climb rapidly and take less current while
the 20% cell would tend to draw more charging current?

And I disagree with the 90/10% estimate. I have a pdf here of a study
that illustrates this is not the case.

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #19
--- Jeff Shanab <[email protected]> wrote:
<snip>

>
> li-ion like CCCV charge.
> hypothetically we start with these two cells at 20%
> and 90% and apply a
> regulated charge voltage to them while connected in
> parallel.
>
> wouldn't the 90%full tend to climb rapidly and take
> less current while
> the 20% cell would tend to draw more charging
> current?
>

That was the point I was trying to make. While I have
not made direct measurements (can't really put an
emeter on each of the two cells, I think the
resistance of each shunt would be significant), I
believe the above to be true. I guess I could
parallel the full and discharged cells, wait, and then
charge the buddy pair for various amounts of time,
then disconnect and discharge, measuring how full they
were, but I barely have enough time to read the EVDL
list, much less do all the experiments I'd like to.

I think what it boils down to, what can you do instead
of paralleling several smaller cells? Yes, you can
parallel each string of cells, but then you would need
a BMS module for _every single_ cell. For what
currently interests me (2.3 amp hour A123 cells), that
would be too many, keeping in mind that the odds of a
BMS failure rise exponentially with the number of
components added. For example, the Killacycle has 110
cells in series, and the latest pack will have 11 in
parallel. Right now, we need 110 BMS modules. If he
managed each string individually, we would need 1,210
modules. We can't afford the cost, size, or weight.
So that's not an option.

I am not claiming to be an expert, so if someone
disagrees with me, please let me know, so we all can
learn. I hope I don't sound like I'm saying "any
other way is too hard, so my way is best", 'cause
that's not a logical argument.

- Steven Ciciora

<snip>


____________________________________________________________________________________
Don't let your dream ride pass you by. Make it a reality with Yahoo! Autos.
http://autos.yahoo.com/index.html



_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 

·
Registered
Joined
·
70 Posts
Discussion Starter #20
Thanks Ian!
I have two questions.
Hey, how did you test that, what hardware/software?
Do you have the information for the cell next to it?

Here is the test I would like to perform, but don't have the stuff.

Although in reality I believe it would be hard to get here, connecting
them together and letting them stabilize over night should put us at the
worst possible SOC difference between 2 cells in parallel.

Cell A discharged
Cell B charged
measure amps and volts of each after connection and then a side by
side comparison of individual charge currents and estimated state of
charge as a charge voltage is applied of 3.65 until current drops below
blah,blah,blah. Whatever ever the recommended charge is.

I also expect an asymptotic graph on the first cycle.

Now discharge and charge daily, I think you will see them converge not
diverge. That is the real question. How do they behave over numerous cycles.

_______________________________________________
For subscription options, see
http://lists.sjsu.edu/mailman/listinfo/ev
 
1 - 20 of 26 Posts
Top