LiFePO4 / Li ion Parallel pack, A123 hybrid for HighPower

KWS Seuren

10 mW
Joined
May 22, 2019
Messages
22
Location
Netherlands Limburg
Hello,

I was thinking of making a parallel pack with LiFePO4 ANR26650M1B cells and Li-ion for example 35E cells.

For example: 30S LifePO4 = 3.45V max per cell = 103.5V Fully charged
25S Li-ion = 4.14V max per cell (Good for long live)

*The Max voltage of the lifepo4 is according to the charge curve of LiFePO4. (above 3.45V the voltage starts to increase fast and the current drops)

*In the example I chose the max charge voltage of LiFePO4 cells and choosed a Li-ion pack that matches this voltage. Should I do this the other way around?

The LifePO4 cells have an excellent power capability and the li ion a good capacity.

The main goal is to have the power of the LFP battery's and the capacity of the li-ion batteries. What do you guys think about the different discharge curves in parallel?

Best,

Stan Seuren
KWS Seuren
 
A picture is worth a thousand words, hopefully also in this case.

A123 M1B versus 35E.jpg

Your idea is not good, politely told. 8)
 
docware said:
A picture is worth a thousand words, hopefully also in this case.

A123 M1B versus 35E.jpg

Your idea is not good, politely told.

Have you read my example? I meant to put the 30S lifepo4 parallel with a 25S Li-ion So when the battery is fully charged the voltage is 103.5V.

HVC
Charged = 103.5V / 30 = 3.45V for the lifepo4 and 103.5V / 25 = 4.14V for the li-ion

LVC
Discharged = 75V / 30 = 2.5V for the lifepo4 and 75V / 25 = 3.0V for the li-ion

This could work right? keeping in mind that you save alot of battery life for the li-ion, keeping the soc between 20% en 90%
 
Shapes of discharging curves are too much different. Therefore you get too big voltage differences through the course of discharge (from – 2.1V to + 6.2 V in this example) :


LiFePo v 35E pack voltage.jpg

Moreover LiFePo are much bigger and heavier with lower energy density. Why anybody instead of building big power Li-ion pack would be looking for such complications ?
 
docware said:
Shapes of discharging curves are too much different. Therefore you get too big voltage differences through the course of discharge (from – 2.1V to + 6.2 V in this example) :


LiFePo v 35E pack voltage.jpg

Moreover LiFePo are much bigger and heavier with lower energy density. Why anybody instead of building big power Li-ion pack would be looking for such complications ?

But those voltage differences are equalled out because they are in parallel.... Right, I know it doesn't feel right but in theory it should work.

The main advantage of the LFP is the high power density. If we can combine the energy density of li-ion with this.

In an electric mx bike high power is needed and ofcourse capacity....

I'm going to try this out with some A123 cells and a lil pack of 18650 li-ion

Thinking of making a 6s1p lfp and a 5s2p li-ion 35e pack with a 30-50amp load.

With this setup I will be getting the same differences as with a 30s/25S pack.

Any other suggestions?
 
How are you preventing all the wasted energy via current flow from higher to lower resting voltage?

Sorry whole thing is just a silly idea for anything real.

Since Peukert coefficient is so low, absolutely no reason to parallel them

Interesting science experiment, I suppose.

Better off IRL to have two separate packs, A/B switch from one to the other.


 
There are a number of discussions about doing this that may have useful info:
https://endless-sphere.com/forums/search.php?keywords=parallel+chemistr*&terms=all&author=&sc=1&sf=firstpost&sr=topics&sk=t&sd=d&st=0&ch=300&t=0&submit=Search
Not everything in the list is relevant, but most are obvious by title.
 
Frankly I don´t see any advantage of building new pack with such configuration, mixing their disadvantages together. Moreover two BMS needed, pack built from two different size cells, uknown unpredictable lifetime.
 
I've done exactly that with 23s LiFePo4 packs and old 20s packs, and it's worked just as expected. The 20s pack helped keep the higher voltages stiffer and then the A123's took over the heavy lifting. That was without BMS's other than my human BMS.

I would choose a ratio and top of charge voltage that better fills the LifFePo4 and charges the lithium ion to 4.1v/cell or less.

I've gone entirely to automotive grade cells now (...still no BMS) and can't see changing until something much more energy dense becomes available. High power and packs that remain at +/- 0.01V/cell after over a year of use prevent me from ever going back, except for car starter batteries, where expensive A123's really shine.
 
John in CR said:
I've done exactly that with 23s LiFePo4 packs and old 20s packs, and it's worked just as expected. The 20s pack helped keep the higher voltages stiffer and then the A123's took over the heavy lifting. That was without BMS's other than my human BMS.

I would choose a ratio and top of charge voltage that better fills the LifFePo4 and charges the lithium ion to 4.1v/cell or less.

I've gone entirely to automotive grade cells now (...still no BMS) and can't see changing until something much more energy dense becomes available. High power and packs that remain at +/- 0.01V/cell after over a year of use prevent me from ever going back, except for car starter batteries, where expensive A123's really shine.

The question is : would you put together intentionally paralel set of new LiFePo4 cells and set of new Li-ion cells for new battery pack building ?
 
John in CR said:
I've done exactly that with 23s LiFePo4 packs and old 20s packs, and it's worked just as expected. The 20s pack helped keep the higher voltages stiffer and then the A123's took over the heavy lifting. That was without BMS's other than my human BMS.

I would choose a ratio and top of charge voltage that better fills the LifFePo4 and charges the lithium ion to 4.1v/cell or less.

I've gone entirely to automotive grade cells now (...still no BMS) and can't see changing until something much more energy dense becomes available. High power and packs that remain at +/- 0.01V/cell after over a year of use prevent me from ever going back, except for car starter batteries, where expensive A123's really shine.

So with your 23S LFP / 20S Li-ion you got the advantages of both right?

LiFePO4:
High discharge power
Minimal voltage sag at the 3.3V per cell region

Li-ion:
High energy density

What do you mean by automotive grade cells? like Nissan leaf, Chevy, Tesla and etc.?
 
KWS Seuren said:
John in CR said:
I've done exactly that with 23s LiFePo4 packs and old 20s packs, and it's worked just as expected. The 20s pack helped keep the higher voltages stiffer and then the A123's took over the heavy lifting. That was without BMS's other than my human BMS.

I would choose a ratio and top of charge voltage that better fills the LifFePo4 and charges the lithium ion to 4.1v/cell or less.

I've gone entirely to automotive grade cells now (...still no BMS) and can't see changing until something much more energy dense becomes available. High power and packs that remain at +/- 0.01V/cell after over a year of use prevent me from ever going back, except for car starter batteries, where expensive A123's really shine.

So with your 23S LFP / 20S Li-ion you got the advantages of both right?

LiFePO4:
High discharge power
Minimal voltage sag at the 3.3V per cell region

Li-ion:
High energy density

What do you mean by automotive grade cells? like Nissan leaf, Chevy, Tesla and etc.?

I forgot about one of the packs I combined, which was new RC lipo + A123's, whose combination picked up another advantage, improved safety. It was obvious the few times the A123's approached full discharge, but there was still plenty of juice left in the lipos to avoid deep discharge of the lipos.

Nissan Leaf batteries aren't high enough power for my taste (too saggy). I really like the Panasonic cells Ford and some others use, and wish Chevy would have kept making the Volt. I've always preferred used cells to new, because perfect balance means no weak cells and all terminations are proven. I have no problems giving up some quite temporary brand new capacity for packs that stay balanced even with multiple years of daily use, especially when so much of the assembly work is already done and the per wh price is less than half (less than 1/4 of insane prices of A123 cells. The biggest issue is fitting them due to dimensions that are generally unfriendly for ebike use.
 
John in CR said:
I forgot about one of the packs I combined, which was new RC lipo + A123's, whose combination picked up another advantage, improved safety. It was obvious the few times the A123's approached full discharge, but there was still plenty of juice left in the lipos to avoid deep discharge of the lipos.
What sort of voltages you talking about?

Even 3.4Vpc at rest has **very** little working capacity left with LiPo, while that is truly Full 100% SoC with LFP.

I would avoid going anywhere near 3Vpc with either chemistry, even under high-C loads.


 
John in CR said:
I forgot about one of the packs I combined, which was new RC lipo + A123's, whose combination picked up another advantage, improved safety. It was obvious the few times the A123's approached full discharge, but there was still plenty of juice left in the lipos to avoid deep discharge of the lipos.

Nissan Leaf batteries aren't high enough power for my taste (too saggy). I really like the Panasonic cells Ford and some others use, and wish Chevy would have kept making the Volt. I've always preferred used cells to new, because perfect balance means no weak cells and all terminations are proven. I have no problems giving up some quite temporary brand new capacity for packs that stay balanced even with multiple years of daily use, especially when so much of the assembly work is already done and the per wh price is less than half (less than 1/4 of insane prices of A123 cells. The biggest issue is fitting them due to dimensions that are generally unfriendly for ebike use.

You mean that the LiPo battery saved the A123 cells from deep discharge? Or the other way around?

Thats a good way of thinking but like you said already: when you use the automotive batteries the layout is predetermned and in most cases not ideal for e-bikes and e-motorcycles. The cost isn't the biggest problem with my project but the power and size is.


I'm starting with a project to convert a MX bike to electric with a Motoenergy Me1616 watercooled motor combined with a sevcon gen 4 size 6.
For the battery I'm looking at a mixed chemistry battery like I said in the beginning.

Because the max power that wil be drawn wil be around 50KW. and 15KW continuous. So just li-ion 18650 cells wont cut it, or atleast have a reasonable lifetime. Because the pack size will be limited to around 400-500 cells due to the limited space available.

24S20P would pull 25A per cell at max. power and 7.5A continuous. :(


One solution to this is an entire A123 pack but the drawbacks.... Less than half of the capacity and 2 times as expensive....

Other solution would be the mixed chemistry with LFP and Li-ion.


LiPo is not the way I wanna go maybe take a brief look at it but just not a fan of the puffing and cycle life....
 
You simply need to build more room into your design for bigger packs.

No way it will be a practical build otherwise with that thirsty a motor.
 
john61ct said:
You simply need to build more room into your design for bigger packs.

No way it will be a practical build otherwise with that thirsty a motor.

But the LiFePO4 parallel packs may be a solution because if you add a 4p pack of A123 you've got 200Amps continuous extra with a relative small pack. If I add that to the 20P li-ion pack that is capable of around 200A continuous, i've got 400Amps
 
I literally had this dilemma a couple of weeks ago and settled for the Samsung 20S cells which have a CDR of 30A with 250 cycles, now because the majority of the time you won't be at peak CDR all the time and you can under charge them to 4.1v to gain a few more cycles its not as bad as you may think.

You could always go for the Samsung 30T which are rated at 35A at 3AH and have similar ratings for 250 cycles at 35A CDR

Samsung 20S - https://mahbattery.com/content/samsung-20s-datasheet.pdf
Samsung 30T - https://mahbattery.com/content/samsung-30t-datasheet.pdf
A123 26650 - 55A~ (From specs / Nuxlands figures)

At 100V (24S) 500A the weight / power / volume figures look like;

Samsung 20S - 24S18P (540A - 432 Cells - 48g - 20.8KG - 36AH - £2.51~ - 16.54cm3 volume) - £1,084.32 - 7145.28cm3 total volume
Samsung 30T - 24S16P (560A - 384 Cells - 69g - 26.5KG - 48AH - £2.83~ - 24.25cm3 volume) - £1,086.72 - 9312cm3 total volume
A123 26650 - 28S10P (550A - 280 Cells - 76g - 21.3KG - 26AH - £3.69~ - 35.85cm3 volume) - £1,033.22 - 10,038cm3 total volume

These look like the most viable cells for your application, for my ME1616 I decided to go with the Samsung 20S because of the weight / power / volume, you have more flexibility to downrate the cells to get more cycles.
 
KWS Seuren said:
john61ct said:
You simply need to build more room into your design for bigger packs.

No way it will be a practical build otherwise with that thirsty a motor.

But the LiFePO4 parallel packs may be a solution because if you add a 4p pack of A123 you've got 200Amps continuous extra with a relative small pack. If I add that to the 20P li-ion pack that is capable of around 200A continuous, i've got 400Amps

LFP are half the energy density of the other LI type. So if your battery space is limiting factor, for every cu ft occupied with LFP you drastically reduce Ah capacity of the whole.

Plus the other LI have much greater power density.

Plus their SoC range is spread over different voltage ranges,

**if** there was any advantage to mixing chemistries, which IMO does not exist

you might as well switch from one bank type to the other rather than live paralleling them

the energy transfer from the higher voltage packs to the lower voltage ones is very very wasteful, reducing overall range.

Just use the other LI chemistry, putting in the right sized efficient motor+controller for the vehicles' spec'd requirements, and enough Ah capacity.

Which I believe will require a larger volume compartment



 
Which is also true for all LFP. Besides A123 (now Lithium Werks / Valence / Super B), excellent cell makers include Winston/Thundersky/Voltronix, CALB, GBS and Sinopoly
 
The ANR26650m1b cells especially with their 4C fast charging capability with almost no difference in cycle life, then when used under 1C charging.

The 15min charging time almost makes me wanna make a second pack consisting of only A123 cells for super high power and ultra fast charging between races. charging a at 8KW :shock: :mrgreen:

I think i'm gonna make a li-ion battery pack first, in a sligtly bigger compartment.

Later this week i'm probably going to pickup the frame to start my mx project.
 
I have A123 20ah pouches 24s. 1480 cycles I can charge in one hour. I have drain them to flat I think 2.8v when I got around to monitor. I have charge to 3.8v but I charge to 3.50v . In one hour or less. 6 years old but cells were made 2012 or so. They can at 17.5ah at my discharge rate. And still deliver 17ah. Yes big and heavy. Maybe 4 more years of use. I was looking at 18650 cells but voltage drop is a concern. I like the size and from fitting. But low cycle life as I ride 6 days a week. Let us know what you decide.
.
 
KWS Seuren said:
The ANR26650m1b cells especially with their 4C fast charging capability with almost no difference in cycle life, then when used under 1C charging.

The 15min charging time almost makes me wanna make a second pack consisting of only A123 cells for super high power and ultra fast charging between races.
Of course they would last **much** longer if charged at a lower rate whenever possible.

And best to pre-warm before charging if ambient is below 25°C, maybe just standardize on 30° when using such crazy-high rates.

 
john61ct said:
KWS Seuren said:
The ANR26650m1b cells especially with their 4C fast charging capability with almost no difference in cycle life, then when used under 1C charging.

The 15min charging time almost makes me wanna make a second pack consisting of only A123 cells for super high power and ultra fast charging between races.
Of course they would last **much** longer if charged at a lower rate whenever possible.

And best to pre-warm before charging if ambient is below 25°C, maybe just standardize on 30° when using such crazy-high rates.

According to this document the cycle life isn't much longer with a lower charge rate, They even say the batteries operate better at high currents.
http://www.mdpi.com/2032-6653/6/3/653/pdf

As for the temperature maybe watercooling is an option, i'm looking for a cheap cooling method between cylinderical cells. Maybe some thin flat tube? I think it's defenitly needed with a 18650 li-ion pack where 10A> is pulled per cell
 
Back
Top