Darmok's Power Wall build

Darmok

Member
Joined
Feb 18, 2017
Messages
96
Hi all,
Please see below details of my system and what I am doing. I am keen for any feedback and input, I appreciate the interest shown.
I have a SMA 5000TL-20 inverter and 6kW of panels on the house roof. I export 40% of what generate to the grid and get a small cheque from the retailer.
Where I am it is not in my favour to yet have the power wall connected to my house .I currently get a good feed in tariff (48 cents). The feed in scheme is a local state one and has been running for 8 years. It will run out in 2.
A requirement of that agreement i have with my electricity retailer is that i have no batteries connected to that system. Obvious reason you could milk the system;charge at night at 28c and export at day for 48c. Utility and retailers here in Western Australia are still yet to get their shit together and offer attractive peak and off peak incentives.

So instead of waiting I am building a standalone system in my workshop. which will hopefully run a 1600w pool pump, workshop socket outlets and LED lights.

My process;
When I get batteries they come as laptop batteries. I get 1-3 tubs of them.I then sort packs into manufacturer then model of the batt then do the high capacity ones first, ie above 50 Whr per laptop battery pack. Saves a hell load of time only processing batts with the higher capacity first and importantly of the sameconstruction. When you learn how to get into a pack and extract the cellsits good to stay in that mode till you've done all those pack types.
I've just completed my assembly of 15 x 100 cell packs (14 and a spare to change any out for service) with all cells in each pack better than 2300 mAh. I am going to progressively make these cells bigger when i process more cells. Anyone see any issues extending the packs longer than 25 cells long? 30 or 40?

To get to 15 x packs of 100 has taken 4months of processing.
Im very lucky that i have gotten volume batterypacksto be able to do this.Does depend what volume of packs you are able to source.

My testing methodology has changed over time. At first I was testing voltage of every cell and then binning based on voltage. I have since stopped this and bin based on manufacturer and rated capacity. I then test in 6 x Opus charger/testers. Then bin for capacity.

I have graded tested cells into bins as follows;
<1800
1800-1900
1900-2000
2000-2100
2100-2200
2200-2300
>2300 went into the packs. Best i had was 3200mAh cells.

What I've found (going from memory) is;
Panasonic Grey 3000
Samsung Light blue 2800
Samsung Pink 2600
Sony Red (light bluering 2600)
Sony Green G72550
Newer Sony Orange 2500
Panasonic purple 2500
Samsung blue 2400

Cells with rated capacity less than 2300 go into a binssorted by manufacturerand are not tested at this stage. These cells could be potentially sold to recoup costs or for other projects or expansion of system once I see how version 1.0 performs constructed with only cells greater than 2300 mAh cells.

I am yet to decide on the Inverter and other hardware. I am keen to hear thoughts from people.
Early options are ;

BMS
Batrium

Inverter
PIP 4048

Battery pack mounts yet to be decided.





Pictures to follow.




image_klciaf.jpg



image_xyztcp.jpg



image_yafjmi.jpg



image_dzkhtj.jpg



image_nflipu.jpg



image_czgeig.jpg



image_kgfjzh.jpg
 
"ie above 50000 " is wrong right or?

Batrium with PIP would work fine. Just beware of it may be recomended to have something in between that controll the pip based on the BMS. Im working on one way to do this alongside with the Batrium stuff.

ie if the Batrium BMS senses that the voltage is to high on 1 cell it will stop charging if needed so it can balance that cell.

On the other hand i think you are spot on. Though if you have plenty of 2Ah+ too you could build strings of that too. :)

I like your n+1 approach. Always keep 1 or more in spares so you can switch out if needed.
 
Correct daromer i added in one to many 0's. Fixed now. Cheers.

So Peter has gone with a separate charger to the PIP my understanding is the PIp won't charge to full 2.20v and he doesn't like the software, Is that all?
Daromer thoughts on just running the Opec to do a charge to determine capacity. Could i get cells that don't hold voltage slipping through with this methodology?
Re 2Ah+ yeah that was the plan. Although running a more longmon gets expensive. Id rather have bigger packs than multiple packs.
Daromer how many cells are you running in each pack?
 
5000mAh? There is no 18650 that hold that much capacity from laptops :)

Peter is using the PCM60x.
charge to full 2.2V?
Its kind of the same software on the PIP as the PCM60x. they just refurbished the software between their models :)

If you run the charge/test cycle you will end upp with fully charged cells. Then you just check the voltage 1-4 weeks later and you will catch the ones not holding voltage.
I have 80 Cells in each packs for the laptop packs. But i also have packs with 300Cells in each that is 390Ah LiFePo4 packs / longmon.
 
Its battery pack spec not cell. 50Whr for a laptop battery pack which has 18650 cells in it. Avoid lesser Whr packs 40-48.

I heard Peter mention that the PIP 4048 doesn't charge to full voltage, might be 4.22v I unsure maybe he coould provide?
 
Max voltage my PIP4048 does is 58,4V
On 14s that is = 4.17V. With some losses that could be 4.1/Cell.
On the other hand you may not want to go that high to save some cycles. So thats the reason :)
 
Interesting re the pip voltage, thanks for that.

Daromer how do you calculate KWhr for cell packs and strings?
 
Note - you can't balance cells at max voltage with pip, i tried :) video to come soonish !
 
I'm currently working my way through testing a bunch of Sanyo red light blue ring cells. I think they're 2600 cap from memory.
It always amazes me how many of these go what I call "runaway". They just keep taking full charge current and get from mild warm to too hot to touch. I pull them as soon as they show any signs of heat. What are others experiences?
 
Yes, some will go into furnace mode. Odd thing is, some of them, once they have cooled off over a few days and settled, you can drain them a little and put them back in the charger and they act normal from then on.

However, if you use those cell, I'd recommend you mark them and put them in an easy to get to place in case of failure. Those cells would probably be one of the main reasons for fusing each cell.

I actually had one where it was making an odd smell cuz it was so hot. That was the only reason I knew it was hot for I was in the living room when I smelled it. After that, I got a thermal gun so I could monitor them a little closer.
 
I'm putting them in the box marked "hot" for now and not using them. Then may revisit them later.
 
Update
I am now looking at 120 or 140 cell packs. Layout 4 x 30 or 4 x 35 per pack. Reason being a BMS like Batrium is a lot less expensive for 1 x string of 14x140 than 2 x strings of 14x80.
They will be quite long which will cause volt drop if I go with the standard size battery buss bar wiring of 2.5mm or 4mm. So Instead I will go with 6mm cable on each pack.

Another idea I'm pondering is making packs extendible. So as you get more cells you can extend the packs. So start with 14 x 80 cell packs then extend those 80 cell packs to 100,120,140....
This obviously won't work with Peters style mounts the way they are designed with fixed ends.


Design ideas under consideration for pack mounts

Provision for fan mounts on top to draw air up. I'm thinking PC cooling fans 120mm. Could be thermostat controlled. Or BMS controlled.
Dividing panels between the packs? This could be in cheap, non conductive fire retardant cement panelling.
Mount provision for a removable ( Perspex ?)front cover to Chanel air flow directly over the packs?
space on mount system for din rail to mount isolators, protection devices, volt meters or BMS etc.



Re the smell from batteries when getting hot, I've found it can come from the leftover glue if any is on the cell.
 
You can always do 14s80p and then just paralell them. Due to the nature of those types of cells where they follow the voltage pretty good based on the SOC that do work even though its not optimal. You need decent size cables between the packs if doing it.

2.5mm2*6 that I use is more than single 6mm2 ;)

Neither less it sounds like a good plan.
 
Daromer can you explain your bus bar cable sizing I'm unclear on what 2.5mm2*6 is
 
Yeap. On my busbar i have 2 sides each with 3 strands of 2.5mm2 twisted solid core copper wire.

 
Can you confirm your testing process?
 
Select batteries that have cells that started life with 2400mAh or greater capacity.
Test Voltage and bin into the following
0-1v
1-2v
2-3v
3-4v
4v+
Cells that are greater than 3v are discharged to 3v. Then the cells are cycled in the Opus @ 500mAh.

The cells that are less than 3v are charged to 4.2v then stored for a minimum of 2 weeks. Then the voltage checked and binned as above and then the Opus procedure undertaken.

My power wall packs are now 140 cells in size. Minimum accepted cell is 2300mAh. They range from 2300 to 3200. I have 15 of them and am currently finalising the balancing of the packs.

I've been practicing building 4s and 6s packs for camping and backyard projects like a system on the chicken coop. With lights and auto poultry door open and closed by an electric car antenna. Charged with a solar panel and controlled by a 12v time clock.

I love Banggood. If you guys haven't heard of it it's a similar thing to ebay. Grab the app or go to the website.
Heatshrink
Anderson plugs
MC4 solar connectors
Crimpers
MPT7210A solar charge controllers
Heaps of arduino gear
Lots of other stuff
 
Yes, most of us use Banggood. I actually compare prices between eBay and BG. One of the great things about BG is that things come from 'warehouses'. So, if you order a bunch of products from the same warehouse, they pretty much will show up together in the same package. :) And there are a few US warehouses where you can get some stuff cheaper than eBay.
 
Back
Top