Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
A first introduction of a to-do diy powerwaller
Hi there,

To start this introduction, my name is Bas and i live in The Netherlands(Holland).

I've first thought about using energy storage at home for PV or wind energy around 10 years ago. (just for fun)
Back then it was with lead acid battery's, with some thinking about other types like NiCd and NiMH.
Then came Tesla and one of my first thought was that a 65kWh li-ion with a 10 year life span would be great to use at home.
The thought lay dormant for a while till I ran into Jehu's channel with the samba, from there to pete's channel (HBPowerwall), Mikes DIY Tesla Powerwall and DIY Tesla Powerwall

This sparked some new idea's, this combined with the fact that i have a house, resources and understanding for electrical systems I have begun planning for a diy or homebrew powerwall.

This is a project that is going to take some time. First I will start a testing setup for a "hopefully to come" house scale system.

So here it goes.
I have a Semi-detached home, with a wooden garage and a stone "shed" The testing will be done on/in the shed using idea's from multiple sources most of the wandering this forum.

I chose to do is on the stone shed because it is free standing without any shadow hitting the roof witch is facing south what is a great position in the Netherlands as shown by the image below.

A second reason is that if something goes wrong and the battery's do burst into flames there is minimal risk of burning the house down and destroying my or my neighbors valuables.
Though I will have preventive and protective components in place.

The idea for the test setup is to be able to develop and test battery packs the connections, fuses, bms and whatever else I think of.

The first step is to start off with solar power.
So to start I will have to set some parameter/requirements. The test setup wil be based of of mike diy powerwall setup. Same chargers and inverter.
First I've looked at how much solar panels I can place. On the roof that will be 2 but should I make my garbage bin storage I could make it to fit 2 more.
Since It will be a test setup I've looked for the cheaper panels and settled for the Yingli Energy YGE YL260P-29b
This panel is a 260Wp solar panel with Umax=37.7V and an Unom=30.3V
Since these 2 panels produce 520Wp I can use a MPP-7210a. This does mean I have to put the solar panels in parallel. When in series the Umax=75.4V and Unom=60.6V.
This could work but I don't want to change capacitors like mike (though I may add some on the input should I experience the same rapid energy output problems as mike).

For the inverter I plan to use the same type inverter as mike is using. A 1000W mppt inverter-limiter with an input of 22-60VDC and an output of 230VAC.

Now it is to the more interesting part the energy storage.
Since I haven't found a lot of places to get laptop battery's over here. And because I don't want to be dissembling, testing all the battery's I have made the choice to but new ones.
I found a place where I can get the battery's for €2.15 a piece so I will be using these ones.
They are new samsung ICR18650-26J battery's with a capacity of 2600mAh with a 2c discharge and 1c charge current. This works out to 5,2A discharge and 2,6A charge current.
For a longer life span the recommend 1c discharge and 0,5c charge so 2,6A and 1,3A respectively. I gave myself a hard current cap of 1A for the design.

I want a pack with as high as an voltage as possible so the current wil be lower. since the inverter can take up to 60VDC in I've made the decision to make a 14S string.
In a 14S configuration my voltages wil be Umin=42V Unom=51.8V Umax=58.8V.
This is also the setup what pete is using with his MPP Solar inverter and I am planning (at this point) to use a hybrid V series one for the house.

At Umin the current wil be 23,8A and at Umax 17A so I've settled for a 20P pack.
This makes for 280 cells per string (I may add some more later on) the capacity of the pack would be 3,6V x 2,6Ah x 280cells = 2620Wh or 2.62kWh

Construction packs.
For the constuction of the packs I am not sure on the format, 2x10 or 4x5 seems the most likely.

For the connections I am going with pete's bus bar idea with the cable connections from mike. Although I will be making the packs as a batterypack with a single connector.
I have chosen for the EC5 connector for the reason that they seem solid and the claims are that they can handle 100A of current. Though I will put more than 50A over a connector and with the 20p pack not more than 20A. I want to use spot welding to connect all the connection of the pack. This since I really dislike soldering and it seems faster and easier.

I am also going to fuse each cell. I haven't decided yet on what wire to use. I've tested 0.22mm windingwire and on a length of 1cm (about the distance from the busbar to the cell) it handled over 10A(measured) for 10 min and 16A(calculated current) for 30s which is way to high and way to long to consider it fusing.
I've ordered 0.15 and 0.10 winding wire to redo the test.
I've tested it on my car battery with different lights and an (cheap) multimeter.

Depending on the configuration of the string, 1 string with multiple parallel pack or multiple strings.
If it is multiple pack on one string I will fuse every pack or if it will be multiple string I will fuse every string.

Battery configuration.

As said above I am considering 2 options. This is not necessary applicable on the test setup but it will if I use it on the house wide system.
1. Multiple parallel packs on a single 14s string.
This is easier to manage as there's one 14s string to manage. So if I want to check the voltage of each cell or balance them I only have to make one system to do so.
2. Multiple strings. Most likely for reason see management.
This makes it easier to do maintenance on a string and balance a string without completely disconnecting the battery. (or I should have the inline balancing in order)

I want the system to handle it's own crap.
This means that the idea is that they will keep themselves in check.
Checking and controlling their own voltage and removing charge from the high peaks of the battery.
And here comes the reason why battery configuration 2 is most likely.
I want the battery packs to stay between 3,0V and 4,1V but I want the string to be able to disconnect from the rest if a single pack goes under 2,8V or above 4,2V so if one pack for one reason or another goes way out of wack It wil just stop till I checked and fixed it.

To accomplish this I want to implement a version of DIY Tesla Powerwall's version of an arduino BMS.
The main problem is that I have never used an arduino or have programmed at all.
I do now enough of PLC system to create something like this. I have looked at something like REX for unipi or piface.
But to use it for a large powerwall it could become somewhat expensive.

For the testing I wil figure out some way to disconnect the battery based on total voltages.

This is pretty much a fast discription of my plans, if you have questions or remarks, please put them below I can only learn from them.
I've completed my testing of 3 different potential fuse wire.
0,1mm, 0,15mm and 0,22mm winding wire.

I have some data and it is different from what i've seen the most on youtube.

first the data and testing method.

I've tested the wire's on a 12V system 11,6V to be precise, using my car battery.
From here I've used my 400A jumpercables to get the power on the table for easy of use.
From the clamp of the jumpercable it goes through the fuse wire and to the load, from the load it goes though the current meter and back to the negative jumper cable.
For the load I've used car lights.
I've created 4 different sets. One H7, one H1 light, two H1 lights but different current and a collection of small lights to create smaller load.
The two H1 lights are cheap white light bulbs and the one H1 light is an standard yellow light bulb.
This makes for the different loads in Amps measured.
H7 3,5A-3,6A
H1y 4,3A-4,4A
H1w 4,2-4,3A two of these
bundle citylights 2,4A

I've measured the different wires at different lengths to address some of the lengths. These are 1cm, 2cm and 3cm.
When the wire didn't melt I kept the current going for at least 10min without my multimeter and checked the current again after. This to protect my multimeter from overheating.

Here is the data.
0,1mm@1cm 2,4A hot to touch
0,1mm@2cm 2,4A 1s
0,1mm@3cm 2,4A direct

0,15mm@1cm 2,4A cold    4,25A hot to touch  6,6A direct
0,15mm@2cm 2,4A warm  4,25A hot to touch  6,6A direct
0,15mm@3cm 2,4A 5s       4,25A direct

0,22mm@1cm 10,8A warm  13A 20s
0,22mm@2cm 8,6A 10s       6,6A warm 4,25A cold
0,22mm@3cm 8,6A 1s         6,6A 15s    4,25A warm

I got the coating off the wire with scraping it with a knife. With this the results from testing 4 pieces the same test method was the least spread out and the wired all melted in the middle that had not been touched.
Doing the same tests with the coating being burned of by a soldering iron produced a lot of different values. All lower than the scraping method.

These are my test results and they do differ a lot from other peoples tests.
This is the reason why I did the tests at different lengths.

This experiment has decided that I will be using a proper fuse between the cells and the busbars. This because I am not sure of my testing and the fact that on a 10Kwh I will spend over €2000,- on 18650 alone so another €25,- on fuses for the cells is not such a bad idea. (for pete's set up it would have cost just under AU $80,-)
At the moment I am looking at a glass fuse with leads 5A.
I also saw the self resetting fuses for batterypacks, but for some reason my gut says I don't want a self resetting fuse.

The fuses I plan to use
Resettable fuse

Test setup.
[Image: 20161201_114620_HDR_thumb.jpg]

Forum Jump:

Users browsing this thread: 1 Guest(s)