Some n00b questions, pack, inverter and automation

ok, here is another n00b

after doing some investigation i decided that i want to build a powerwall.

already i have 3 suppliers where i can pickup e-bike batteries for free now al i have to do is take them part and test them.

the basics:
my first desision was that i want to go with a system where i can swap cells easy, with comon, easy to get parts. so i go with a 18650 pcb system with cell holders.

I have not fully decided yet, wich pcb i am going to use.

1. https://www.electronics-lab.com/diy-18650-powerwall-onboard-bms/

this is "my winner" so far, i like the fact that it has a bms system per cell. there are also a few thinks i don't like, mainly the lack of documentation or reviews on this board. also, the cells can only be populated on one side of the pcb. the pcb manufacturer can also take care of soldering all smd components.

2.

nice design, not much info about bms connections, or use with the jehu bms pcb. also, populaties only on one side.

3. https://jag35.com/collections/jehus...gh-power-18650-battery-module-diy-pcb-kit-10x

this can hold 2x7 cell per pcb. lot of documentaton. don't like the need of an external pcb.

still need to make the decision and order the pcb's and other components, i am quirious about your toughts.

i will store everything in a 22U 19 inch rack, mounting everything on shelfs, so that problem is solved.

the next problem is the inverter / charger. maybe this is also a good reason to explain why i want to build a wall.
in The Netherlands where i live there is a high hourly fluctuation of energy prices (spot market), see: https://www.easyenergy.com/nl/energietarieven

i want to buy power at the lowest rate from the grid, and sell back at the highest rate.
our connection to the grid is 3x35A (can upgrade to 3x80A, maybe later) 230V, so approx 23KW/h. if i want to store 6 hours of energy, i need a powerwall of at least 150KW/h. i see fluctuation between lo and high of 30 cents.

so, i need a (or multiple) inverter, that can handle this. i need to be able to set to charge /discharge or idle, preferable controlled via HomeAssistant, most ideal is fully automating charge and discharge moments (see https://community.home-assistant.io...rgy-prices-to-activate-electric-boiler/154445 for a positie example).

i see all kind of different projects, most of them has separate chargers and inverters. is there anybody that can give some advice on this? the pack will grow as i proceed, i hope that i can reach somewhere in the region of storing 300KW when finished, and be able to charge and discharge the whole pack in only 6 hours.

Thanks!
Ivan

ivan123 said:

my first desision was that i want to go with [...] a 18650 pcb system with cell holders.

Click to expand...

This is something a lot of Newbies first decide on (incl me), but I'd say that it is almost certainly a mistake. Those PCB designs certainly do have their niche, but I've yet to see one that is really suitable for use in a usefully large powerwall. Some of the issues are:

* Not really scalable. Like for 1000s ~ 10000s of 18650 cells that are required for a decent powerwall. Sure, they can be stacked to a certain extent, but then maintenance becomes a nightmare, which completely negates their primary advantage of being easy maintenance.
* Just not cost effective. Do the math on the system you're dreaming of - the PCB and components are going to cost thousands and thousands of $$$. It's likely going to cost more than the 18650 cells.

You probably decided against the "mainstream" design, because...
* spot welders cost too much; Actually super cheap, if you split the cost over the 1000s ~ 10000s of cells.
* soldering on cells seems dangerous; Certainly demands technique and care, but the only catastrophic incidents I recall seeing reports of were caused by accidentally shorting the cells.
* replacing bad cells will be impossible; Actually, several members have made videos showing that it's actually not that difficult. I myself replace perhaps 5 cells every month - usually takes no more than 30 minutes.

ivan123 said:

reach somewhere in the region of storing 300KW

Click to expand...

That's a LOT of energy. Even 150kWh is going to be massive - pretty sure that's not going to fit into a 22U 19in rack. Especially not when using PCB holders.

Note that it's "kW" when talking about power.
When talking about energy (eg. capacity of a battery) it's "kWh" (not "kW/h").

ivan123 said:

1. https://www.electronics-lab.com/diy-18650-powerwall-onboard-bms/

this is "my winner" so far, i like the fact that it has a bms system per cell. there are also a few thinks i don't like, mainly the lack of documentation or reviews on this board. also, the cells can only be populated on one side of the pcb. the pcb manufacturer can also take care of soldering all smd components.

Click to expand...

Super expensive, that's 15$ per PCB/7 cells! Wow.

For my 1120 cells it would cost 2400$, three times the cost of the cells.

Uhm.

PCB boards are fine for small projects where a single board, maybe 2, would be used for a particular project. Getting to where you'd need 100's of them, waste of money.
Not to mention every board that is added in another point of possible failure

ok, lot of good arguments to reconsider my desision.

i made some calculations on the 14 cell PCB from Jehu, costing about $15 each.

the rack i have is actualy a 30U, 55x55cm, that can take a 14s25p (am i doing it right?) per 2u, that's 5250 cells for the whole rack, when using the PCB 375pcs (€5550)

that is a huge amount of money, i did not make any calculation in the number of cells i need, but i guess it will be 2 or 3 racks full, so 15K just for a pcb....

any toughts on the inverter / charger system?

ivan123 said:

PCB from Jehu

Click to expand...

I'd be careful with these boards. They have flaws. The original designer pointed them out to Jehu and he didn't upgrade the files for quite some time. @rev0 can give you more details on this particular issue.

There are better and more safe designs out there and are cheaper. You don't need a full sized PCB for holding the cells. If you want the ability to replace the cells easily, then just mount the cell holders on a piece of board and run wires for the BMS. For cheaper, Far safer, Far easier. And can be changed/upgraded/modified easily too.

Korishan said:

I'd be careful with these boards. They have flaws. The original designer pointed them out to Jehu and he didn't upgrade the files for quite some time. @rev0 can give you more details on this particular issue.

There are better and more safe designs out there and are cheaper. You don't need a full sized PCB for holding the cells. If you want the ability to replace the cells easily, then just mount the cell holders on a piece of board and run wires for the BMS. For cheaper, Far safer, Far easier. And can be changed/upgraded/modified easily too.

Click to expand...

From the €15 per pcb, about €10 is used for the holders (2x1 and 6x2 per board). I skip the holders, and i think i go for spotwelding or soldering. Thanks!

ivan123 said:

[...] actualy a 30U, 55x55cm, that can take a 14s25p (am i doing it right?) per 2u, that's 5250 cells for the whole rack

Click to expand...

There are a lot of things that needs careful planning when dealing with high power and high energy systems. Just some that pop into my mind:
* One 18650 cell typically weighs 45g. Let's say 60g when including individual wiring, fuses, solder, etc. Then just the 5250 cells for one rack is going to weigh 315kg. Can the rack handle that? The floor?
* Dealing with the heat. The cells generate very little heat when (dis-)charged at a slow rate over several hours. But can get quite toasty at high (dis-)charge rates. And it can get exponentially worse as you pile more and more cells into a small space, especially for the cells in the center or top (heat rises). The #1 killer of lithium cells is heat, so if you're hoping to quickly charge when the grid power rates are lowest, you may need to think about active cooling of the cells.

ivan123 said:

i see fluctuation between lo and high of 30 cents.

Click to expand...

You need to take into consideration that your system is not going to be 100% efficient. The Tesla powerwall claims a round trip efficiency of 92.5% when new and under ideal conditions.
If you use cheap off-she-shelf chargers, they might have an efficiency of ~90%. And inverters typically also have ~90%. In total, you'd only have a round trip efficiency of 80%.
You might also want to look into the accounting and tax rules. If you get charged VAT on ALL the electricity you bought, and have to pay income tax on all electricity you sell, the financial round-trip efficiency may fall to around 50%.

ajw22 said:

* One 18650 cell typically weighs 45g. Let's say 60g when including individual wiring, fuses, solder, etc. Then just the 5250 cells for one rack is going to weigh 315kg. Can the rack handle that? The floor?

Click to expand...

The rack is rated for 300kg. Will be placed in a outdoor shed.

ajw22 said:

* Dealing with the heat. The cells generate very little heat when (dis-)charged at a slow rate over several hours. But can get quite toasty at high (dis-)charge rates. And it can get exponentially worse as you pile more and more cells into a small space, especially for the cells in the center or top (heat rises). The #1 killer of lithium cells is heat, so if you're hoping to quickly charge when the grid power rates are lowest, you may need to think about active cooling of the cells.

Click to expand...

Space is not a issue, i want the setup to be as neat as possible, if i need some more room that's not a problem.
What is considered as slow and fast charging, in a time perspective? my plan is to charge and discharge the full pack in ~6-8 hours.

ajw22 said:

If you use cheap off-she-shelf chargers, they might have an efficiency of ~90%. And inverters typically also have ~90%. In total, you'd only have a round trip efficiency of 80%.

Click to expand...

i am looking at the Powland 5000W hybrid converter, to charge and discharge the pack.

The efficiency (yeah, i know, it's chinese) is 95% (line mode, Rated R load, Battery fully charged), with a 90% peak efficiency.


this is about a month ago, where there is a nice time window, where you get paid for taking energy
Even with those price difference, a 80-90% efficiency is good enough, in a later stage i will look at those, high end, expensive inverters.

ajw22 said:

You might also want to look into the accounting and tax rules. If you get charged VAT on ALL the electricity you bought, and have to pay income tax on all electricity you sell, the financial round-trip efficiency may fall to around 50%.

Click to expand...

I hope that i have to pay lots, and lots of taxes. When the system is build, it's a passive income.
When i calculate at a price difference of 20cents, 3x60A 220V grid connection, over a time frame of 6h charge / 6h discharge a day (240kw total), with a efficiency loss of 10%, i still have €40 in earnigs per day. (14K per year).

now, lets be real. i first have to build it all. 32600 cells to go...

ivan123 said:

now, lets be real. i first have to build it all. 32600 cells to go...

Click to expand...

Yes sir. After 3 1/2 years, I have deployed ~12,000 cells to build a 108kwh 18650 battery bank. Working on the next 2,500 cells now.... maybe a year from now I'll have them online

OffGridInTheCity said:

Yes sir. After 3 1/2 years, I have deployed ~12,000 cells to build a 108kwh 18650 battery bank. Working on the next 2,500 cells now.... maybe a year from now I'll have them online

Click to expand...

You are really lucky, that you can get those cells so easily. Its so hard to find any non-overpriced cells here in germany, even if they are literally trash.

Oberfail said:

You are really lucky, that you can get those cells so easily. Its so hard to find any non-overpriced cells here in germany, even if they are literally trash.

Click to expand...

These company's have to pay when they bring to recycling. It's a win-win situation.

I just mailed e-bike battery refurbishement companies, only 1 of them knew the value of the cells, the others replied that i can pickup for free.

ivan123 said:

These company's have to pay when they bring to recycling. It's a win-win situation.

I just mailed e-bike battery refurbishement companies, only 1 of them knew the value of the cells, the others replied that i can pickup for free.

Click to expand...

The majority here in germany refuses to give them out, because they legally aren't allowed to do it.

ivan123 said:

What is considered as slow and fast charging

Click to expand...

6h for charging, and 6h for discharging is not too bad. Heat from the cells should not be too big an issue as long as you leave a bit of space for airflow. Perhaps add a thermostat + fan just in case.

ivan123 said:

I hope that i have to pay lots, and lots of taxes. When the system is build, it's a passive income.

Click to expand...

Worst case, if you do it as a private household, you may have to pay 21%VAT on electricity to charge the battery, and 40% income tax on the electricity you sell... very little left after that. If it's setup as a business, you can of course write off all expenses and just pay tax on the net profits. Better make proper, long term cost analysis calculations. It's probably going to be a lot less profitable than you think. Also don't forget that the cells may only last 3~6 years, if you're starting off with ones that are already old to begin with. The electrical equipment also has a finite lifespan, and may need replacing after 10~20 years.

ivan123 said:

i see fluctuation between lo and high of 30 cents

Click to expand...

Be careful not to look at just a few cherry picked days. I see quite a number of days where the difference is just 10~15 cents.
On top of that, you need to take the average price of the top 6 hours, and average price of bottom 6 hours.
I only skimmed the data of the last 90 days, but I reckon the average profit margin will be closer to 15~20 cents / kWh.

Last week i picked up the first batch of E-Bike batteries. Had a few other appointments, but they all canceled last minute or stopped responding to the email.

I have took most of the ebike packs apart, and put them into 3-d printed packs, waiting to be tested. (thanks @ajw22 for the design, i think i will use it )

Today i had a few cells that look liked they released presure, and looks like they have holes at the welding points.
Even one of them (bottom right in the picture) released some clear liquid.

I did not had this problem with other cell packs, this is the only one of this type of pack with this type of cell.
Before i have to trow them all away, how do i avoid this? do i need another technique to get rid of the strips?
It's not safe to use the cells with holes in them i guess?

Thanks!

ivan123 said:

thanks @ajw22 for the design

Click to expand...

ivan123 said:

how do i avoid this?

Click to expand...

Happens with thick nickel tabs and strong welds. Surest way to avoid making holes is to leave the welds alone, and just cut off any excess nickel.

ivan123 said:

not safe to use the cells with holes

Click to expand...

Definitely not. The liquid electrolyte is pretty nasty stuff. I think corrosive, flammable, and possibly carcinogenic? Get rid of the fumes, seal the hole best you can, and get it to the recyclers ASAP.

ivan123 said:

I did not had this problem with other cell packs, this is the only one of this type of pack with this type of cell.
Before i have to trow them all away, how do i avoid this? do i need another technique to get rid of the strips?
It's not safe to use the cells with holes in them i guess?

Click to expand...

That is correct it is not safe to use them. Best to put them in some sand and recycle.
Those are most likely NCR18650B, or maybe NCR18650BE. Panasonic cells and they have a soft bottom and do tear very easy,.
The thickness of the nickel strip doesn't help either. Do as @ajw22 says and leave the strip.
If you are into experimenting and loosing a couple of these cells there is a cutting/rolling technique that has been successful for me but it takes a very sensitive touch and a bit of finesse.

Also @ajw22 is correct, the lithium salt being used in Li-ion batteries, as the electrolyte, which is the somewhat sweet smelling liquid, is lithium hexafluorophosphate, which is toxic, additionally posing a fire and safety hazard.
Best to avoid it.

Wolf