My Experience with diy Lifepo4 17S bank

[frankrudis65]

Hi everyone ,as discussed this in another thread ,i will open now this new one.Here I'll state my experience with building a 17S Lifepo4 battery bank in detail.I'll explain why I've done this and the outcome of each step.Why i keep cells under compression,why I limit the ma manufacturer's recommend "safe" discharge and charge cut-off voltages and why I apply certain steps to reduce stress in general. All these steps keep in mind,that a premature degredation effects,most notably capacity loss can have a significant impact on the profitability of a home storage system

 

[frankrudis65]

Having started many years ago with a tiny back-up system,including 3 solar panels and two lead acid batteries,to overcome grid outages,I am here know with 8.5 kw of solar panels ,spread into two strings,a decent 6kw Hybrid inverter and building and testing my first 17S lifepo4 bank.What a satisfying journey
Some words to my background: I'm a technician in optics,where molding our own lenses in glas forms in our lab.I have got a small optician shop in one of the pre-alps valley in northern italy.My early start envolved studying electro-technics fkr a year before changing eventually to optics.Ad a hobby I am designing aand building audio amplifiers and tbeir speaker enclosures .....

 

[frankrudis65]

Before starting on Lifepo4 ,or even thinking ,coming from a complete different chemistry build lead carbon,I was going to read throug a lot of even scientific researches ,to get vaguely informed and somehow prepared to the coming-ups....
I was surprised ,having worked for nearly 10 years with those robust 45kg/12V lead-gel based blocks,when i first came to see those tiny lifepo4 cells and together with a friend we took one into parts.....This ultra thin layers arou the anode and cathode with those tiny M6 connectors,somewhat weldet more or less together and pressed into a very thin aluminium case with an even thinner plastic coating gave me something to think about....It was here when I decided ,once to get those celms to reduce any stress to them,to avoid premature failure

 

[frankrudis65]

Pressing or not...and how
Many people argue that compressing lifepo4 cells is an other invention of the unknown maybes and purely irrelevant .....
I would like to go to a different appe,and actually ruling out any possibilities of doupt,while going here for some researches of the "Journal of Energy Storage" which is available through ScienceDirect.Here you can find basically everything which is regarding any possible testing with any possible equipment of any possible material to have any possible data regarding those tests
Here some quick data of the build of a lifepo4 cell : the cell casing,cathode,anode and separator have thicknesses from 20 to 130 micro m.The cathode comprised a 14 micro m thick AL curfent collector coated on both sides with 43 micro m thick layer of LiFePO4,and the anode comprised a 17 micro m thick CU current collector coated on both sides with a 38 micro m thick layer of graphite.Notably much effort has been directed to toward reducing the grain size of the LiFePO4,to increase the high-rate capacity of this material....The evolution of the swelling and potential measured during discharge and relaxation periods at different C -rates.Swelling was mainly caused by electrochemical processes at both electrodes.This had an effect on the crystal structure .Most remarkably ,during discharge,the Lifepo4 battery cells swelled at medium SOC's and shrunk at low and high SOC's ,in contrast to most other chemistries......Crucially ,thermal swelling depends on the heat conduction between the cell core and cell surface and heat convection between the cell surface and air.The outcome of those test under different force of compression was,depending on how the cells are packed,enclosure thickness,chemistry aso,micro cracking on the electrodes can be reduced,alart from difficulties in stacking cells foe ex. in electric vehicles .....
With this,hoping to avoid any stability problems I decided to compress the cells,but not to have them in an enclosure for air circulation .....

 

[frankrudis65]

BUILDING THE STRUCTURE FOR COMPRESSION
This is,if you want a straight forward task,as one of the latest information of the so famous EVE ces give the answers.The cells are put in one or more lines,hold togeter with a non conducted material,stabilised with three threaded rods each side,6 all together pet row.If you like guessing you can tighten up with nuts,some washers and a torque wrench .....A bettet job is achieved with springs,calculated for the mechanical force,they have to put up with and a precise measure tape....
this pic was used in the testing phase,for this nothing was covered up

 

[frankrudis65]

Aging aware opration of lithium-based battery energy storing systems,including lifepo4 chemistry.....
Havi g seen the vulnerable parts of thecells while actually cutting one in half ,I decided to do everything possible to give my cells a caring home.....
Even if the cycle life exeeds those of electric vehicles ,residential storage systems have to cope with the growing home consumer applications like smart homes,cell phones,pc's aso.
As the aging mechanism eventually leeds to loss of capacity and power fade,let's have a look what is actually causing this....
Electrodes aging: this is basically a chemical reaction ,while the layers,for ex. graphite reacts with the electrolyte.This forms a certain layer,while solvent molecules may still diffuse through ,voluma change can lead to cracking and a side reaction would be such as dissolved transition metals from either cathode could form additional growth of the solid electrolyte interphase....During a rest period and tbrough discharging this reaction is partially reversible .Here the biggest enemy seems to be holding a certain charge for too long.....Notably this dendrite growth as a consequence of lithium ating is a significant safety concern,ad it can pierce the seperator,short circuit the cell and induce thermal runaway ....collector corrosion is also an aging issue .
Calendet aging: obviously here's not much we could do,apart from overpainting the white bids
Temperature accelerate cell degradation ....
State of Charge SOC: A high SOC comes with a low anode and high cathode potential .The low anode potential is known to accelerate the deposits growth ,therefore a high SOC will accelerate cell degradation.
Cycle aging: Here we can actually do something aboutwe increase capacity....
Charge and discharge rate:
To normalize for the battery capacity Cbatt,tbe charge-discharge rate is often given instead of the charge-discharge current.
A hihg Crate will accelerate particle cracking &graphite exfoliation as well as additional growth of solids on the electrodes.A high Crate will also cause the cell to generate more heat and increase its temperature ,making it challenging to distinctly seperate individual stress factors.
Average State of Charge: The average state of charge SOC around which a cell is cycled is known to influence cycle aging as well .The lowest cycle aging was found in the range of 45 to 55% SOC,at SOC=50%,and the highest in the range of 90 to 100%SOC,at SOC =95%.
The impact of the graphite anode is,that it will expand in volume when charging and discharging .The average cell terminal voltage maybe used to quantity this stress factor....

 

[frankrudis65]

Charge and Discharge Cut-Off Voltage:
The usable and nominal capacity and thereby the definition of the SOC are dependent on the set operation voltage limits of a cell: the charge and discharge cut-off voltage.A high charve cut-off voltage can lead to over-delithiation of the cathode material and thereby accelerate structural disordering on the cathode.Low discharge cut-off voltages can lead to corrosion of the anode's curdent collecter.The operating voltage window should therefore be set such that the cells deliver a high nominal capacity while retaining high cycle life.Limifing fhe manufacturer's recommended "safe" discharve and charge cut-off voo from [2.5,3.65] to [3.0,3.45] which is more o less a 20% reduction,will double your cycle life .

 

[frankrudis65]

Therefore I have introduced a charging - discharging strategy by not allowing to go under 3V and not to go over 3.43V with fits perfect with the 17 S.If you have a charger which goes over 60Vdc,you could also give it a try with 18 S.As we have got now bms's they can cope with this.
17S gives a lesser dropdown when on discharge.
With a higher voltage you have less current for the same power
If something goes wrong with one cell,it is easily rearranged for 16S.....
I found ,when a proper precharge on the individual cells are done,first to see the singular cells behaving on their first charge,secondly to bring them really up to the same SOC,no issues with cell drifting apart.Not even going down to a low SOC.My bancer is set to 1.5A and jumps in occasionally.

 

[frankrudis65]

Floating:
Here I've got most people saying you are wrong.....
Let me explain now this again in detail.
As we know a high state of charge comes with a potentially accelerated degradation of the anode.During rest periods and through discharging ,the reactionof this lithium plating
is partially reversible ,called lithium stripping .There are other aging factores due to high average SOC like decomposition of the binder or collector corrosion.....
Therefor,and this is entirely my opinion ,and I'm certainly not forcing anyone to do so,
I am going down with the float as much as possible,if not turned off completely .Another reason is,that after a decent charge with a decent capacity bank,which should be calculated to serve your nightly needs( My lowest SOC at the morning was about 70%with only 1 out of two banks connected) the percentage of discharge without any load is so little,that in my opinion holding the SOC to a high level is not even needed.Keeping in mind that a pure voltage charge/holding the voltagewith a very low resistance itself is very difficult.....without any current flow ...

 

[frankrudis65]

Balancing:
First to understand ... balancing is nit a cure to repair cells from drifting apart.If you observe your cells while charging,you'll see,because of cell differences you'll have different internal resistances
-towards end of charge,some cells might have different voltage ,at no or very little current
- towards end of charging ,other cells might have higher voltage with no or very little current .
Therefore: End of charge =all cells have top- voltage AND there's no or very little current flow.
Active balancer ,which start balancing without regarding a starting voltage(to balance) are balancing too early and even at a far to low SoC and therefore you'll have a mismatch .Balancing is hardly necessary ,once your pack is "idle" .Imbalances should disappear ,even while charging with higher currents ,towards the end.Wrongly balancing at still high currents is happening even with low power balancers.This is not a problem ,though : High charging currents at "nearly full" is only a matter of short time.In this phase might be a bit of differences ,caused by wrongly balancing .This is usually max. 15 to 30 minutes .Keep the batteries charging ,the current goes down,parts of the imbalance disappears and the rest thebalancer will do.Some limit the absorption to a certain time,pushed then into float: Bad,because you might take again time of the balancer to do its job! Charging and balancing is clearly done at a certain point ,maybe at 3.55V,or at -I would say - 3.42-3.40V ,going further down at some point won't work anymore,this is very clear
charging and discharging: the most is achieved between 3.0V and 3.45V ,considering your batteries are not pushed too high to their limit and you use decent sized cables .
BMS balancing start 3.40/3.45V,because above there's no gain anymore .Only top balancing
Voltage over protection at 3.6V
reason : the balancer has been given a lot of time and keeps the BMS not getting nervous .
Calibrating: I have seen the victron smart shunt to be the most accurate in voltage and current readings.So I have calibrated the BMS and my inverter in base of the shunt.This gives me know similar readings and similar SOC's.
Those applyings and settings,for myself ,cuts out most of the unnecessary risks,by pushing things to the limit

 

[frankrudis65]

The BMS is a fantastic tool to point out ,once you have a chronically imbalanced cell ....that's what I do:
take those cells out of the bank and charge them fully ,slowly,so the balancer has got time to catch up....Most probably the reason in the first place ,the BMS cut off the cell was that the charging current was higher,then the balancer could cope with,removing access energy.
Discharging until almost empty.Same thing like above.Discharging finishing,waiting until the voltage rises slight and checking if and which diffetences are regarding to the rest of the bank.Measuring every single cell ,if drifting away .Now you can determine the defect cell.Those cells with the reduced capacity needs to be removed ,because it's going to continue loosing capacity,as there will be always a balancing problem That's why you always should have some spare cells

 

[frankrudis65]

Balancing cables:
I've had the problem once with a tiny cable shoe from the victron smart shunt,trying to squeeze it on a m10 connector,braking literally into bits,so I decided not to put pressure again on those cable shoes but soldering the cables onto the copper part of my flexible cell bridges.I have got a steady iron ,I presoldered to have a tiny solder point and the finally soldered those cables on....This seems to work fine and keeps another worrying out of the way