Another DIY Powerwall in Germany (15 kWh)

Chicken

New member
Joined
Jul 3, 2017
Messages
25
Hi there!

name's Alex :)
Just finished my powerwall project, which I have started in the end of february.

All cells were harvested from used/bad e-bike batteries. Most of the time, about 2-3 cells in each pack were bad (high inner resistance or very bad capacity), the rest usually was just fine

image_dukdiu.jpg


image_doptrl.jpg


I tested the cells using 5 Opus chargers in parallel.

image_kwjqto.jpg



In total, I've tested about 2300 cells, some were bad, some mechanically damaged. Some got hot during charge (some sanyo, but also some panasonic). Of course, I threw out all these bad cells. Still, I've ended up with 1960 good cells, the average cell capacity is 2100 mAh. My cell configuration is 14s140p.

image_rcrrhf.jpg


The wiring of the cell holders was straight forward. I used 1.6 mm copper wire for the bars and 0.2 mm tinned copper wire for fuses (double fuses + and -)

image_rrtrcs.jpg


image_kvjbto.jpg


The Powerwall is hooked onto a PIP 5048MS (PF1). Charge end voltage is 4.09 V / cell and discharge end voltage is 3.3 V / cell. Power is coming from a 4.5 kWp solar array.

image_mdelvs.jpg


I'm not using a BMS. Each cell block has a little voltage readout and a passive balance board that kicks in at 4.1 V (2A current).

image_txpahb.jpg


I have found a software, that automatically calculates the pack layout for the 1960 cells. This worked so well, that the voltage drift between the cells is minimal. Deepest discharge so far was 3.55 V and the difference between each pack was less than 0.02 V. I didn't expect that, but of course I'm happy :D
Also when recharging, the balancer modules did not yet get active. All packs equally charge up and perfectly end up at 4.09 V without the need for balancing (yet).

I'll see how the pack behaves the next weeks / months.

best regards,
Alex
 
Where did you get that passive balance board? Or did you build them yourself? In that case please share more details.
 
Hi Alex,

erstmal herzlich willkommen hier :) !

Great work, nice shelf. Since you've more cells than I have in my powerwall, you'll push me down one position in the cell counter... :dodgy: :D .

Where exactly do you live?

I'm also interested in the passive balancers you're using. And I would like to check out the software you've found. I wrote it by myself, can be configured for packs size and number and serial rows, also with min. capacity, min. voltage and max. voltage drop between two meassures.

And maybe you wanna check out my project...

Have sun!
Oliver
 
Hallo Alex,
auch von mir herzlich Willkommen :)

Super work you have done there. I also use batteries from e-bikes, but unfortunately I have not as many as you.

Sometimes you have also bad luck with the E-bike batteries and does not get 18650 cells. They are Li-poly or very different formats.

for example cells from Koga Akku :
image_nctxzm.jpg

image_xxbjsw.jpg


The best experiences I have made with the batteries of Bosch and Gazelle. There were always 18650 in it.

Viel Glck noch mit deinem Projekt , vielleicht kann man sich ja mal helfen bei den Akkus.

Schne grsse aus NRW

Iron
 
Hi!

Thank you for the nice welcome :)
I'm located in the south of Germany near Munich.
The passive balancers are from http://shop.lipopower.de/ they basically can be set at any voltage and as I don't want mine to fully charge up to 4.2 V, I have them set to 4.1 V.

Was trying to attach the software here, but it didn't accept the file format :huh: . It's a simple java app and cell data can be imported from .txt

image_ouzfmv.jpg


Check the screenshot how this looks like. I have a text file with cell number and cell capacity. Using the java app I have made the excel sheet which tells mit which cell goes into which cell block :)

best regards,
Alex
 
Wow, it stinks when you can get a cordless drill cheaper than some wires:

Cordless screwdriver 3,6 Volt Bruder Mannesmann M17730 Special price : 17,85
Special charging cable for 1-6 Lipo BEC connector system 40,00

The cables look to be just XT connectors on the ends. Supply and demand, I suppose.

Welcome to the group, and nice layouts and construction :)
 
Hallo Alex,
auch von mir herzlich willkommen im Forum!

Very nice work you have done so far.
I'm very curios to see your progress without the balancer. Let's see the behaviour of the packs.

Alles Gute
BigBen
 
Did you check the internal resistance at all and include this in how the the packs were built?
 
Hi!

No, I did not check the internal resistance. I was trying this first, but I think the Opus charger is not very accurate at this. If you measure 4 times, you get 4 different readings depending on how good the contact is on the sliders.
As my load per cell is very low, way below 0.5 A / cell even at full load and below 0.1 A under normal conditions, I guess the impact of internal resistance is minor.

best regards,
Alex
 
Chicken said:
Hi!

Thank you for the nice welcome :)
I'm located in the south of Germany near Munich.
The passive balancers are from http://shop.lipopower.de/ they basically can be set at any voltage and as I don't want mine to fully charge up to 4.2 V, I have them set to 4.1 V.

Was trying to attach the software here, but it didn't accept the file format :huh: . It's a simple java app and cell data can be imported from .txt

image_ouzfmv.jpg


Check the screenshot how this looks like. I have a text file with cell number and cell capacity. Using the java app I have made the excel sheet which tells mit which cell goes into which cell block :)

best regards,
Alex

Hi Alex,

very interesting, I'll check out the specs of the balancers later. I'm using these balancers and have a max. voltage difference of about 30mV in my 200p7s powerwall.

The software also looks interesting, too bad I already wrote my own :cool: . It's based on Access and I made a table to enter capacity, voltage after a full charge and voltage after 2-3 weeks. This way, I can sort out cells with a high voltage drop due to self-discharge without meassuring the internal resistance. Then I can enter the number of cells per pack, the number of packs I want and the desired capacity per pack including some parameters like minimal capacity per cell, minimum cell voltage and maximum voltage drop between the two meassures.
Already made 70 packs of 20 cells (420Ah@24V overall) this way and it's working great. Here're two screenshots:

image_piwvgd.jpg

image_aqehzn.jpg


How long is your powerwall in production? And how do you feed your battery power back into your house?

Have sun!
Oliver
 
Any particular reason you chose the PF1 version?
 
yes, 5000 W continious and 10000 W surge power. The PF 0.8 versions only have 4000/8000. The price difference was minor. So it was an easy choice :)
 
For me the PIP is not an option because of the no load consumption of 50Watt, which means if during the night I have something around 100-200watt, the efficiency is horrible and for powerwall of for example 5kwh it will consume a bit chuck of energy....
I'm considering a small efficent victron, maybe 1000watt for constant small loads(fridge, router, water pump, lights etc), AND a cheap 3000watt (maybe SHI-3000w) for loads which can be turn on explicitly, like the laundry, dish washer, oven etc.
 
Hi!

got some more cells (~ 500), mostly samsung and panasonic.

image_cxnwya.jpg

thinking about building the same layout again and upgrade to a total of 3920 cells (~ 30 kWh)
So I'm starting to measure cells again :)

best regards,
Alex
 
Chicken said:
Hi!

got some more cells (~ 500), mostly samsung and panasonic.

image_cxnwya.jpg

thinking about building the same layout again and upgrade to a total of 3920 cells (~ 30 kWh)
So I'm starting to measure cells again :)

best regards,
Alex

Some of those packs look like they would be a pain in the butt to disassemble.
 
Yes, indeed. Especially those stuck together with glue :(
But if the glue is too sticky, I usually just cut the heat shrink wrapping and afterwards just rewrap them. This is faster than removing all the glue.
 
Back
Top