A 110 kWh Powerbucket

The lead-acid batteries of my off-grid solar system are dead. I will replace them by a 18650 batteries stack housed in a big wooden box.

Similar projects worth following
My PV system works fine, but the lead-acid batteries have proven to be the weakest part of the design by far. Efficient 18650 lithium ion batteries are available at affordable prices, so why not change now ?

This new LiIon battery will replace the 8 heavy lead-acid batteries currently installed (500 Kg). Theses batteries are housed in a big battery box. See the battery box here:

Once fully populated with LiIon elements, the Powerbucket will store 110 KWh of energy, enough to cover the needs of my house during 11 days. For comparison, a Tesla Model S 85D has a 85 KWh battery. The complete battery stack will be made of 11 strings of 15 X 70 = 11 550 cells.

The first phase of this challenging project will be to design and build the first 10 KWh string made of 15 X 70 = 1050 cells. I will design the Battery Management System in DIY mode.


Here are the definitions of the words currently used when it comes to batteries...

CellOne elementary part of a battery. (Lithium Ion 18650 elements in my case)
Lihtium Ion 18650 cellA cylindrical cell having a diameter of 18 mm and a length of 65 mm. Weighs approximately 45 gramms. These cells a very common. LG, Samsung, SOny, Panasonic and others produce several billions of these cells each year. A Tesla Model S car contains 7 104 of these cells.

ModuleA group of cells wired in parallel to increase the current capacity and decrease the internal resistance.
Internal resistanceA parasitic characteristic of a battery that creates losses, heating and degradation of the battery. The internal resistance cannot be null, but should be kept as low as possible.
A group of Modules wired in series to increase the voltage and therefore decrease the joule losses at high currents. In a battery design, one always tries to increase the voltage as much as possible. 10 % of increase of the voltage will decrease the joule losses of 20 % at the same current. In a Tesla Model S, the strings are made of about 100 modules in series, giving a total of 400 Volts.
Occur each time a current flows through a cable or a battery. These losses increase with the square of the current. If you can achieve to divide the current in a system by two, the joule losses will be divided by four. In general to decrease the current, you must increase the voltage accordingly to keep the same power rating.
StackThe whole battery made of one of several strings wired in parallel.
PowerThe current flowing through the stack times the voltage of the stack (in watts)
EnergyThe power going in or out of the stack times the time (in Joules or Watt-hours).
1 JouleEnergy of 1 Watt of power produced during 1 second
1000 JouleEnergy produced by a sports rower at each stroke
1 kWhEnergy of a 1 kW (= 1000 Watts) of power produced during 1 hour. One kWh is equivalent to 3 600 000 Joules
10 kWhAverage Energy consumed by my house during one day ( = 3 600 kWh per year)
100 WhYou need approximately 100 Wh to make one liter of water boil
200 WhYou need approximately 200 Wh to drive one kilometer with an electric car
Phase AFirst step of my Powerbucket project where the total capacity will be 10 kWh
Phase BSecond step of my Powerbucket project where the total capacity will be 110 kWh
BMSBattery Management System. Takes care of the battery by measuring the individual module voltage and temperature. Furthermore, the BMS insures also proper balancing of the cells, avoiding dangerous overcharging. The BMS also maintains counters of the energy going in and out of the battery stack
Specific energyThe amount of energy stored in one Kg of battery. The specific energy of a good lead acid battery is 40 Wh/Kg. The specific energy of a good 18650 Lithium Ion element is 250 Wh/Kg, 6 times more.

Powerbucket Specifications

This table contains the main characteristics of the Powerbucket

ItemValueWhy ?
Number of cells in parallel in one module5 * 14 = 70Fits well in my wooden bucket
Number of modules in a string15Fits well in my bucket AND is OK for my inverter
Nominal Powerbucket stack voltage15 * 3.6 = 54 VoltsWell In the input voltage range of my inverter (38 to 66 V DC)
Minimum Powerbucket stack voltage
15 * 3.0 = 45 VoltsWell above the minimum input voltage of my inverter (38 V DC)
Maximum Powerbucket stack voltage
15 * 4.2 = 63 VoltsWell below the maximum input voltage of my inverter (66 V DC)
Number of cells in phase A15 * 70 = 1050 cellsNumber of modules in series times the number of cells per module
Nominal voltage of one cell3.6 VThis is the average voltage of one cell during discharge.
Capacity of one cell2900 mAhI will use PANASONIC NCR18650PF cells having 2900 mAh capacity
Energy in one cell3.6 * 2.9 = 10.44 WhNominal voltage in volts X nominal capacity in Ampere hour = nominal energy stored in Watt hour
Total energy in the Powerbucket stack in phase A
10.44 * 1050 = 10962 Wh
Number of...
Read more »

  • Second 10 KWh string put into service

    Michel Kuenemann05/02/2018 at 04:02 0 comments

    I have added the second 10 KWh string to my setup by the beginning of April 2018.

    This string is made of 980 X Panasonic NCR18650PF's cells. You can see them in the foreground (the green cells).

    The first string, made of various types of cells, sits behind. The complete battery is a 140P14S stack (1960 cells in total).

    The long rods are tinned copper bus bars which insure the parallelization of the modules. As they are quite long, there is provision for addition of two more strings.

    I have also added an LCD display that indicates Battery voltage, current and temperature as well as input / output powers.

    Battery management is performed by my wireless modular BMS

    This battery pack covers now 2 days of consumption. A third string is planned for addition this summer.

  • Phase 2: Adding a second 10 kWh string

    Michel Kuenemann08/27/2017 at 09:19 0 comments

    I have purchased 1050 new PANASONIC NCR18650PF cells in order to add a second string to my PowerBucket. 

    PANASONIC NCR18650PF cells are very good and cost efficient 2900 mAh cells. They are a great choice for my application.

    The cells were bought at NKON and were delivered them within a few days. Thanks you Arjan and congratulations !

    2 Parcels delivered  in perfect shape on Thursday, 24th August 2017. Each package weighs about 25 Kg.

    Brand new premium quality cells perfectly packed and protected

    In order to build the 15 modules containing 70 cells each, I had to prepare the following stuff:

    • 450 pieces of Nickel Strips - 84 mm length each - Total: 38 m of strip
    • 60 copper bus bars - 31 cm each - Total: 20 m of bus bar

    Cutting the nickel strips with a disk cutter - The strip is wery thin and does not damage the cutter. Easy but boring work :-)

    Strips ready for use

    Bus bars before cutting - Theses are high end 10 X 3 mm tinned copper busbars

    Cutting the bars with my radial saw

    Bunch of 31 cm long bus bars ready for use

    Assembly Step 1: instert 2 BB's in the jig

    Assembly Step 2: Adding some extra tin on the bars. I use flux to improve the solderability

    Assembly Step 3: Soldering the Nickel trips on the bus bars. I am using a Weller Jumbo soldering iron with a huge tip for this operation.

    Once removed from the jig, the "fishbone" looks like this:

    (Two such fishbones are needed for one module, one for each pole)

    Assembly Step 4: Populate battery holders

    Assembly Step 5: Spot welding the positive and negative fishbones

    The fishbones must be precisely positionned on the battery frame before welding starts.

    The strips are manually pushed down on the batteries terminals to facilitate welding

    6 finished modules, ready for use

    9 Modules ready for welding

  • NEW PANASONIC NCR18650PF Module Capacity Test

    Michel Kuenemann08/20/2017 at 04:36 0 comments

    I have run a charge/discharge test on a module made of 70 brand new PANASONIC NCR18650PF cells. The cells were purchased at NKON in August 2017.

    The measurements were made with my modular Lithium Battery Manager used in "Discharge Mode"

    The parameters were the following:

    End of charge voltage4.1Volt
    Charging current10Ampere
    End of discharge voltage3.0Volt
    Discharging current - whole module10Ampere
    Discharging current - one cell143milli-Ampere

    It took approximately 16 hours to discharge the module down to 3.0 Volts  and the results are the following:

    Module capacity180Ah
    Capacity per cell2570mAh
    Restituted energy - complete module656Wh
    Restituted energy per cell9.37Wh


    The PANASONIC NCR18650PF cells are rated for 2900 mAh when charged up to 4.2 V and discharged down to 2.5 V. Taking into account that the end of charge voltage in my case was only 4.1 V and the end of discharge voltage was 3.0 V,  the measured capacity of 2570 mAh is really great. This also proves that the batteries sold by NKON are guenuine PANASONIC cells.

    The following curves show the last hours of the test

    Module equipped with the Modular BMS:

    I will perform the same test on a module which as made approximately 250 cycles in my solar system.

  • Balancing test video

    Michel Kuenemann05/28/2017 at 08:15 0 comments

    I have made some tests to check the precision and efficiency of the balancing modules. Check the video out here

  • Balancing modules installation completed today

    Michel Kuenemann05/21/2017 at 20:43 0 comments

    I managed to make a "mass prod" of 15 balancing modules - at last. They are installed on the modules now - see the details here

  • About Battery Temperature

    Michel Kuenemann02/18/2017 at 08:38 0 comments

    Several hackers asked me about temperature management of the powerbucket, so let me give some info about this aspect.

    Currently, I have put one sensor in the middle of the string. This sensor is attached to a 35 mm² bus bar, tight to the stack. I have a very good confidence in the measurements it provides. Of course I assume that the individual temperature variations between the 1000 cells are very small and can be neglected. This assumption has to be discussed, of course.

    Thanks to my Raspberry Pi based solar data base I have produced a chart showing the temperature of the stack since its start of service in novembre 2016.

    This chart shows clearly that the battery temperature remains close to the average room temperature of my house, namely 21 °C (70 °F). One can say that these temperature conditions are very safe for the stack. One can also assume that none of the 1000 elements gets a lot warmer or colder than the measuring point.

    Why are the temperature variations so small ?

    Batteries heat up when they are crossed by a current, regardless if it is a charging or discharging current. This heat comes (mainly) from the power (Joule effect) dissipated in the internal resistance of the cell:

    P = R * I²


    P = dissipated power (Watt).
    R = Internal resistance of the battery element (Ohm)
    I = current flowing through the element (Ampere)

    Let us assume a current of 140 A flowing through the stack. This represents a power of nearly 8 kW. In my system, this can last only for a few minutes, due to inverter power limitation.

    The internal resistance of a module, composed of 70 cells wired in parallel is about 1 milliOhm.
    Power dissipated by one module = 0.001 * 140 * 140 = 19.6 Watt

    Power dissipated by one 18650 cell = 19.6 / 70 = 0.28 Watt. This is the power rating of a small through hole resistor. Such a small power cannot heat up a 18650 element significantly.

    Keep also in mind that if the current is of 70 A (half of 140), this power is divided by 4 !

    This is why my Powerbucket needs no cooling system.

  • A perfect cycle

    Michel Kuenemann12/09/2016 at 04:16 0 comments

    The power/energy curve of yesterday December, 12th 2016 is the following:

    First thing: one can see that the most part of the available energy (yellow area) has been used (green area) by the PV system. That's very positive. The flat top of the green area is due to the charging current set to 40A. I will change this parameter to 60 A today (maximum of the charger is 70 A).

    The battery state of charge (blue curve) has changed direction serveral times during the day, indicating a heavy and rapidly changing power usage.

    The interesting curves are the following ones (all of them located at the bottom of the chart, below the thick green line ):

    1. Orange: energy delivered by the charger
    2. Magenta (pink): energy absorbed by the inverter
    3. Cyan (green): energy absorbed by the battery
    4. Red: energy delivered by the battery

    At 8 AM, the state of charge (blue curve) was at 8 kWh (scale on the left)

    The 4 energy curves were at 0 kWh.

    At 8 PM, the state of charge (blue curve) was again at 8 kWh (scale on the left)

    The Orange curve and the Magenta curve are both at 11.5 kWh approximately. That means that all the power delivered by the solar panels has finaly been used by the house.

    The Cyan curve and the Red curve are also superposed at about 5.5 kWh. That means that from 8 AM to 8 PM, the battery managed to absorbe and deliver again the same amount of energy.

    Among the 11.5 kWh delivered by the sun:

    1. 11.5 - 5.5 = 6. kWh were directly used by the inverter (approximately)
    2. 5.5 - 4 = 1.5 kWh were stored and delivered again by the battery during day time (approximately)
    3. 4 kWh were deliverd by the battery after sunset

    The fact that the in/out curves are so nicely superposed at the end of the day shows the astonishing efficiency of the lithium technology. I never got this kind of curve whith the lead-acid batteries.

  • My first week with Lithium - an overview

    Michel Kuenemann12/05/2016 at 04:23 0 comments

    My Powerbucket has been working for a week now and it's time to have a first look at the curves and figures.

    By chance, the last week was globally very sunny, and I had enough power to fill up the stack with photon juice nearly every day.

    The following chart shows the data between Monday, November 28th and Sunday, December 4th 2016.

    The white curve near the top shows the battery voltage. It varies between 54V (3.6 V/cell) and 60.0 V (4.0 V/cell) , which is quite good and safe for the cells.

    The blue curve in the middle is the fuel gauge. It shows that the battery gains and looses 4 kWh during each cycle. This represents 4/11 = 36 % of its nominal capacity. This is also very conservative. I will probably increase the depth of discharge to at least 50 % once the new ballancing BMS is installed.

    For the moment the parameters are good and I will go on running with these settings for a while.

  • First Charge / Discharge Cycle

    Michel Kuenemann11/30/2016 at 05:36 2 comments

    Thanks to the bright and sunny day, the batteries could perform their first charge / discharge cycle yesterday:

    • A - This thick white horizontal line indicates the end of charge voltage. Currently set to 60 Volts. This corresponds to 60 / 15 = 4 volts per element. This is a very safe end of charge voltage for LiIon batteries.
    • B - This white curve is the voltage of the stack. During the night, this voltage was of about 54 V (3.6 Volts / element). At 8:50 AM the sun began to shine (brutaly) on the panels and the charge started. The voltage rose up to exactly 60.0 V which was reached at 11:00 AM.
    • C - This blue line is the State Of Charge (Fuel Gauge). Its max is set currently to 12 kWh. (Thick green horizontal line) and min to 6 kWh (Thick red horizontal line). During the cycle the SOC started at 8.5 KWh, reached approx 12 KWh and decreased to 8 KWh.
    • D - The yellow area is the solar potential - eg the maximum solar power available at any time. Its smooth envelope indicates that the sky was perfectly clear during all this day.
    • E - The green area depicts the solar energy used. The power reached 2250 Watts (see the scale on the right) during the first part of the charging cycle (constant current phase), then decreased to 1500 Watts (constant voltage phase). and then to approx 350 watts ("noise" floor of the house).
    • F - The orange curve indicates the energy delivered by the charger: it reached approximately 6 kWh at 4:00 PM.
    • G - The red curve is the energy delivered by the battery. The battery started delivering power when the sun went down at 3:30 PM. The battery had delivered 3.6 KWh when the system switched automatically back to the grid power.

    A few comments:

    The battery behavior is really great. Its voltage reflects well its state of charge.

    The system switches currently automatically back to grid when the battery has lost 4 kWh. This figure could be increased to 6 or 7 kWh to use more of the battery capacity. (Nominal capacity is 11 KWh).

  • First Power On

    Michel Kuenemann11/27/2016 at 13:49 0 comments

    The first power on occured at 9 AM on Novembre, 26th.

    The "old" lead-acid BMS was adapted to accomodate the voltages and thresholds of the lithium battery.

    I installed a temporary desk in order to perform the necessary changes in the BMS software:

    Monitoring the voltage with my DMM. The Nexus 7 tablet on the right displays the system real time status:

    The weather was very dull yesterday. The sun gave bearly 1 KWh of energy on the whole day:

    Charging of the battery started approximately at 11 AM (when the yellow area turns green). The first tests show that the modules are very well naturally balanced and behave very well under load (minimal voltage drop under heavy load). The batteries are currently at approx 50% state of charge.

    In depth electrical test will be performed later.

    See the real time curve

    The weather should be much sunnier tomorrow. This will allow a full charge of the stack.

View all 29 project logs

Enjoy this project?



Romain wrote 07/28/2019 at 16:38 point


I just stumbled on your project, and it is very impressive! I'm very interested since I started looking at building a similar solution for my sailboat!

However, I just have a question regarding the capacity of your Phases. It seems to me that your calculations are wrong.

Since there is only 70 cells in parallel in a module, it seems to me that the capacity of the module is 10.44Wh*70 = 730.8Wh. Now, even if you were to put 100 of those modules in series, this would only serve to augment the voltage, not the capacity of the powerbank, that would stay at the nominal capacity of one module (730.8Wh). Now, if you were to connect your modules in parallel, then the capacity would be 10950Wh, but with a voltage of 3.7V.

Am I mistaken here? What am I missing?

  Are you sure? yes | no

Romain wrote 11/10/2019 at 11:27 point

For those eventually reading, I've answered my own questions.

I made a confusion between capacity (expressed in Amps-hour, Ah) and energy (in Watt-hour, Wh).

When you put the cells in series, the capacity doesn't change (thanks captain obvious) BUT the energy stored does. You calculate the energy by multiplying together capacity and voltage, so if the voltage goes up, so does the energy.

So, yeah. capacity, energy, two different things.

  Are you sure? yes | no

Duane Grzyb wrote 01/24/2019 at 03:25 point

Thumbs up on your dedication and commitment. 

You are working very hard to hand build what will someday be available to the masses as prepackaged units.  You are a pioneer.

I have worked on some industrial lead acid ups systems and the challenges are great.  

  Are you sure? yes | no

ActualDragon wrote 05/10/2018 at 01:03 point

what. a. savage.

  Are you sure? yes | no


[this comment has been deleted]

Michel Kuenemann wrote 05/03/2018 at 01:27 point

Hi. Balancing is performed by the electronic cards that you can see on top of battery modules. Please see my project #A Lithium Ion Battery Management System  for a complete description of the BMS. 

  Are you sure? yes | no

Michel Kuenemann wrote 05/03/2018 at 06:47 point

Each group of cells wired in parallel has its own balancer.  I have 14 groups in series, so there are 14 balancers.

What do you exactly mean by P-L Layout ?

  Are you sure? yes | no

Michel Kuenemann wrote 05/04/2018 at 03:14 point


I made some verifications about the parallelling of many cells. I guess you should check this video out :

You will learn that Tesla Model S batteries are made of 16 modules wired in series. Each module is a 76P6S pack. Finally this makes a large 76P96S battery containing 7104 cells.

Assuming that Tesla are competent regarding batteries, one can say that there is no issue wiring many cells in parallel. By the way, If you think about it, there is no other way to build a large pack.

Regarding the safety in case of a cell failure or even a hard short circuit, I made short circuit tests on "sacrifical" modules, and no fire occured at all.

You can also check this out:

These guys build there own battery packs, and many of them are much larger than mine. I hope they are all smart guys who want to keep their house safe !

My advice: get informed about the technology and run real tests by yourself.


  Are you sure? yes | no

Bruno wrote 02/17/2017 at 17:30 point

Hi Michael,

I just now stared following your project.
Congratulations on the progress and results so far. 

I would like to know a couple more things about your project, if you dont mind.

How do you perform the thermal managment of the batteries ? One of the major concerns regarding lithium batteries is allways the safe temperature range.

I'm thinking since your load will not request huge amounts of power you never have te problem of the overheating . Am i right or do you have any kind of cooling system supporting the battery pack ?

Best Regards,


  Are you sure? yes | no

Michel Kuenemann wrote 02/18/2017 at 09:40 point

Hi Bruno,

Thanks a lot for the follow and the encouragements.

I have added a project log entry dedicated to the thermal aspects of the Powerbuck.

As you mention, the power load is shared among the 1000 cells, so each cell is submitted to a low stress in my application. Imagine ants, each of them producing a small amount of the huge work produced by the anthill.

So no special thermal management has to be performed in my application.



  Are you sure? yes | no

Loic Cuenot wrote 02/16/2017 at 09:23 point

Hi Michel.

Can you tell me a rough idea of the total price of this battery pack (cells + battery holder + wires + fuses) ? (I've found 18650 cells for 2.5€ each over there : )

I've read that these battery were losing 20% of their capacity after only 200 cycles, and that its charge must be kept between a certain range. what range do you use to keep your battery at its best, and did you see any difference since the installation?

  Are you sure? yes | no

Michel Kuenemann wrote 02/16/2017 at 10:37 point

Hi Loïc,

You can get a better price (close to 2 €) if you purchase 1000 cells.

Battery holders: about 50 € - Copper bars: 100 € - Fuses: neglectible

About life: if you fully charge/discharge the cells (4.2 down to 3.0 V), you will have this life expectancy. If you stop charging at 4.1 V or less and stop discharging higher than 3.0V, the life will be a lot much longer.


  Are you sure? yes | no

Loic Cuenot wrote 02/16/2017 at 12:57 point

Thanks for your feedback! 

I have another question regarding your battery pack that you want to be 110kWh.

I've seen that your house's av erage consumption is 10kWh per day. I'd understand that you want to have 40, or 50kWh battery pack, in order to cover the peak consumptions, but why choosing 110kWh? it seems much more than your house's need. Is it in order to get a charge when you have some days without any sun?

If this is the case, ins't the cost prohibitive (around 12000€ for this extra 60kWh) regarding the number of time it would be used? Wouldn't it be more interesting to buy a power generator (around 3 to 4000€ for a good one, that could be ran automatically) for the very few times it would happen during the year?

Or is there another reason for that 110kWh?


EDIT : Oops, sorry, I've just seen that Kent asked the same question. But if this is just for cloudy days, what do you think about the power generator solution then?

  Are you sure? yes | no

Michel Kuenemann wrote 02/16/2017 at 22:18 point

110 kWh is the max capacity of my "Bucket", filled up with Li Ion cellls.  This has to be compared with the 10 kWh I had with the 500 Kg lead acid batteries. 11 times more energy with the same weight and volume. 

110 kWh is a good thing at the start and at the end of winter. In autumn it can extend the period of autonomy for maybe 2 weeks. In the same way it allows you to become autonomous again a bit earlier by the end of the winter. But it will certainly NOT increase dramatically the autonomy ratio. To become 100 % autonomous, I would need a 1 MWh (1000 kWh) storage to use the surplus of energy received during summer. But this is really not affordable and it would weigh 5 tons, even with lightweight Li Ion cells ! 

In the future, to store a large amount of energy for a long time (several months), the solution will probably be to produce hydrogen from water during summer, to capture it on nickel foam, and then to burn it again in a fuel cell during the winter. But this is very complicated and also unaffordable for an individual... Currently - Maybe in 10 years, it will be as easy as to buy a smartphone...

Check out Macphy's technology:

The efficiency of the hydrogen production X the fuel cell will not be very good (less than 50 %, I guess), but there is so much available free power during summer, that even if you loose 60 % of it, what remains will be enough to cover your needs during the winter !

Nature is so: either abundance in summer or shortage in winter !

New Lithium Ion batteries cost about 200 € / kWh currently, so at this price, I will never fill-up the bucket. If I find a way to get them at 50 € / kWh or so, it would change everything. 

I have bought new cells for the first 10 kWh string to validate the concept and be sure that in case of a failure, it was not due to the fact I used "second hand cells". I have heard of batteries from wrecked Tesla model S cars. This could be a way to get very good cells for cheap. I will start digging into this.

Power generator: you mean a genset with a diesel or gas engine ? I guess I will not do that. Burning fossile fuel to get electricity is not a good solution in my case. I think that would be much worse for the environment and much more expensive than nuclear power from the grid.

I hope I have answered your questions.


  Are you sure? yes | no

brian wrote 11/29/2016 at 15:30 point

I have been so focused on lead acid batteries that it never occurred to me to string together thousands of rechargeable 'torch' batteries !! 

Are you at any point going to release the board layouts and source code for the electronics ?.....

I was also interested in the construction of the On-Line Source Selector, are you going to make a hackaday project of that too?

  Are you sure? yes | no

Michel Kuenemann wrote 12/02/2016 at 15:45 point

Hi Brian,

I hope you are doing well and I thank you so much for following several of my modest projects... 

Regarding the documentation, source code and all the related stuff, I will do my best to push them in the different HAD projects entries before the end of this year. Everything is (more or less) up to date, it will cost a little effort to gather the files and to click the "upload" button.

I am rather proud of the On-line Source selector and, as you mention, it is probably worth a separate project. I must admit that It works really well and it never let my house in the dark ! The documentation of this project is ready and up to date, so creating an HAD project entry from scratch is not a tremendous effort.

About batteries: I did not make a good experience with lead-acid batteries: the first set died sulfated due to deep discharges in the beginning and the second set has been overcharged during months (bad charger settings partially due to lack of documentation from the manufacturer). 

I must admit that I was a bit frightened when I turned the main power switch on the Lithium batteries. After almost a week, I must say that the result is beyond my dreams. The efficiency of these little things is overhelming. Nearly every single Wh pushed into them can be retrieved.  So, no regrets. :-)

Best regards,

  Are you sure? yes | no

brian wrote 12/08/2016 at 12:52 point

Hi Michel

I'm very well thanks.

Good to hear that everything is working really well and that you'll be uploading the files, and making a new HAD project for the On-line Source selector, I'll look forward to reading them.

A question on the battery welding - Did you have any concerns regarding the welding of the batteries together? in case the electrical surge damaged the batteries in anyway?



  Are you sure? yes | no

Michel Kuenemann wrote 12/08/2016 at 13:13 point

Hi Brian,

Regarding battery welding:

In fact the welding current crosses only one terminal of the battery, so no surge can occur at the battery level. It is totally safe for it.

Imagine a bird on a wire - no danger for him as long as he does not get in contact with the other wire.

Further on, the welding open circuit voltage is very low (less than 2 volts), so there no danger for the guy who welds neither :-)

Nevertheless I had some issues welding the fuse wires on the negative side of some batteries. Sometimes the battery got perforated due to the very localized heat density at the soldering point. Panasonic NCR18650PF were very sensitive to this. Next time I will try soldering the fuse on the positive side. it seems to be made of thicker metal.

I think it was a very good idea to build this spot welder. Thanks to this tool, assembling the packs becomes a fast and reliable process, to my point of view.


  Are you sure? yes | no

Michel Kuenemann wrote 02/16/2017 at 22:38 point

Hi Brian,

The project entry for the #An On-Line AC power source selector exists now. Check it out, please,



  Are you sure? yes | no

Kent wrote 08/28/2016 at 23:42 point

I'm really curious just _why_ you want to make such a massice power bank ? Why do you need to run the house for 11 days at normal consumption on battery ? A wide area power outage of that magnitude in central europe  is unheard of .

  Are you sure? yes | no

Michel Kuenemann wrote 12/08/2016 at 14:46 point

Hi Kent,

I hope you apologize my very late replay.

Regarding the huge size of this energy storage, my main motivation is not related to the fear of a long lasting mains power outage, but rather to a "sun outage".

Keep in mind that my goal is to use as much as possible the solar energy rather than the power from the grid. 

It happens often that in autumn and in winter we have several very sunny days followed by a few dull days (few meaning 5 or so). In this perspective, a big storage could make the bridge between the sunny days, and so avoiding the use of grid energy as much as possible.

The main obstacle to this approach is financial. Currently, the cost of the batteries I am using is about 200 €/KWh. A 110 KWh storage would cost about 20 K€. That's a huge amount of money. At this cost level, economical amortization is really impossible. Maybe the prices will drop and it will become possible to have big storages at home.

I hope that I have answered your questions.


  Are you sure? yes | no

K wrote 07/27/2016 at 21:16 point

Oh wow! I'll be following this project, very cool. My first thought though was a nightmare case of just a single cell (perhaps with a manufacturing defect, what's the chances) developing a short and reaching the point of no return only to have the whole 11,550 of them burn up like a bomb!

Reminds me of the technique Tesla and others have been using in their ev cars and home power walls to isolate a single misbehaving cell from the rest, using essentially a piece of wire rated to melt and disconnect when the cell experiences too much power flow. Check this: (at about the 5 minute mark)

Still though - 11,550 cells to wire individually!! I hope your method of closely monitoring the temperature and other vitals using wireless tech will be as safe..

  Are you sure? yes | no

Michel Kuenemann wrote 07/28/2016 at 06:20 point

Hello Kristan,

Thanks for the compliments. 

The safety of the stack is, of course, the most important point to consider. First, one must take in account that the cell manufacturers make a lot of efforts to make the cell as safe as possible, by design. The cells comply with strict safety standards. There are billions of cells in service on earth, and accidents are very rare.

So, I keep thinking that the catastrophic "chain reaction" you describe is very unlikely.

Nevertheless, better be safe than sorry, and I have planned to add a "Tesla Style Fuse" on every single cell since the very beginning of this project. I have watched the videos you mention and they have inspired me a lot.

I have purchased a roll of 0.5 mm tinned copper wire for this purpose:

Monitoring: the temperature of each module will be monitored with a 0.1 °C resolution. Any deviation with respect to the ambient temperature will be detected. There will be a warning threshold and an alarm threshold. When the alarm threshold is reached, the BMS will disconnect the charger and the load.

BTW thanks a lot for the skull.

  Are you sure? yes | no

dustinthurston wrote 08/09/2016 at 20:14 point

I don't understand the function of the fuse. How would one cell be shorted? The only thing I can think of is that the failed cell may short the other cells in the circuit. Is that the failure mechanism?

  Are you sure? yes | no

Michel Kuenemann wrote 08/10/2016 at 14:16 point


Q: How would one cell be shorted ?

A: during the manufacturing process of the cells, particles of metal can enter the battery and create a short when they migrate. The risk is low, especially with famous brands, but is not null.

See some interesting background information here:

Q: Is that the failure mechanism?

A: Whenever a cell gets shorted internally, the current flowing through it is mainly limited by its own internal resistance, because the 69 other cells of the module (my modules comprise 70 cells in //)  have a very low equivalent resistance (less than 1 milli Ohm) and push hard into the faulty cell. Without a fuse, the current and therefore the power dissipation inside the faulty cell can become huge (more than 100 A, hundreds of watts) and may lead to a fire. With a fuse the current may reach maybe 20 or 30 amps, but only a during a few seconds,  mitigating a lot the fire hazard.

Without a fuse, the duration of the high current phase is impossible to predict. With a fuse (even a very basic one), one has a much better "control" of its duration.

I hope my explanation was clear.

Edit - 2016/08/13 - A few thoughts about an external short circuit.


An external short cannot be excluded and I imagine four scenarios for it:

First scenario: a piece of light gauge wire (cross section of 1 mm² or less) gets connected across the terminals of a module. Assuming a resistance of 20 milli ohms for the wire and a voltage of 3.7 Volts for the module, the current would reach 3.7 / 0.02 = 185 Amps. This current would make the wire burn and melt within a few seconds. The individual cell current would be 185/70 = 2.6 Amp, well below the "Tesla Style Fuses" rating. No damage would occur to the module.

Second scenario:  the same wire shorts the whole stack of 15 modules in series, the new current is: 15 * 3.7 / 0.02 = 2775 Amps. The real current would probably lower than this, but this calculation gives the order of magnitude of the surge current.

This current would make the wire burn instantly with an explosion noise, creating a very hot plasma and a dangerous flash of ultraviolet light. No damage to the modules making up the stack neither.

Third scenario: I drop accidentaly a piece of heavy gauge copper bar (section of 100 mm²)  accross the terminals of a module. The resistance of the copper bar is about 1 milli ohm. the current reaches 3.7 / 1E-3 = 3,700 Amps. (Theoretical calculation). This current would make the bar get very hot but it would not melt. The individual cell current would reach about 50 amps (good cells are able to deliver this current), well above the "Tesla Style Fuses" rating. The weakest fuse would open first after a few seconds. The 69 remaining cells would share the current, with a corresponing increase of the current in each fuse. The "domino effect" would make all the fuses melt one after the other within a few seconds. After the last fuse opens, the short circuit condition no more exists.

Fourth scenario: the same kind of copper bar shorts the whole stack, the current would be limited by the current capability of each cell to about the same current as the short of a single module. The behavior of the cells and the fuses would be the same.

After such a catastrophic event, the cells of the module would certainly be severly damaged, but I guess that the risk of cells "venting with a flame" is low.

I did never test the Third scenario neither the Fourth in "real life" and I cannot be sure that everything occurs like I describe it above.

I have purchased several spare cells for my stack and I will consider testing the Third scenario with a small module of 5 cells or so. I will keep you updated, with a video of the experiment.

Regards, Michel

  Are you sure? yes | no

dustinthurston wrote 08/11/2016 at 13:15 point

Thanks Michel. I wasn't clear on whether the protection was for an internal or external short. Internal is the only thing that made sense, but I wasn't aware of the cause.

  Are you sure? yes | no

Thomas Smith wrote 06/29/2016 at 18:32 point

I will catch up this weekend...  I just got really good pricing on LifePo 16340 cells, and am probably going to update for that...   'til then, happy hacking...

  Are you sure? yes | no

Michel Kuenemann wrote 06/29/2016 at 20:23 point

OK - looking forward hearing from you.

BTW I have ordered 16 sets of 5 * 14 battery holders here:

  Are you sure? yes | no

Michel Kuenemann wrote 06/23/2016 at 21:16 point

Hello Thomas,

Thanks a lot for your proposal. Of course I am interested. Send me a private  message so we can exchange our email addresses.



  Are you sure? yes | no

Thomas Smith wrote 06/23/2016 at 18:32 point

I have a 12V power module design for using 18650 cells.   Designed to be cheap and easy to slap together/replace.   Let me know if you are interested.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates