DIY 48V LiFePO4 battery pack for home ESS [Energy storage system] in 2022

[Zucca]

I just started to design my off grid home system.
Since my power company does not support net metering and for other reasons, I think it will be best for me to put in place a off grid energy storage system, PV Panels and use the grid only if necessary (battery low or home asking for too much power).

I would like to use two Victron Energy Quattro 10KVA 120VAC wired up in a split phase configuration.
Yes, I am located in USA.

Now I need some battery, and I am temped (looking at the costs) to build my own 16S packs starting with 3.2V LiFePO4 cells.

Specs:

- >100AH, 200AH preferred
- M8 screw terminal connectors

Since most of the batteries are made in China, I would have no problem to buy on the www in Aliexpress or so, but I am worried not only about lemons (should not be the case), but to get bitten by the slow supply chain.

On top of that I need a good BMS system.

1 - Can you point me to a good battery shop? China or USA?
2- What would be a good BMS system for that?

I have a TTi QPX1200SP and a TTi LD400P to perform some experiments to understand if the cells are good or not. Plus some 6.5 bench DMM.

Many thanks in advance for all your help.

[Zucca]

For example

https://www.aliexpress.com/item/3256804013889748.html

 

[NiHaoMike]

Have you considered an EV battery pack? The 400V or so matches up to 240V mains (340V peak) quite well and saves a lot of cost in voltage conversion.

Then there's energy storage technologies other than batteries. Thermal storage HVAC is a good one if there's significant HVAC demand. Pound per pound, making ice gives about the same cooling as running a conventional air conditioner from lithium batteries, but making ice is several orders of magnitude cheaper.

[Zucca]

Have you considered an EV battery pack? The 400V or so matches up to 240V mains (340V peak) quite well and saves a lot of cost in voltage conversion.

What inverter you wanna use then?

Then there's energy storage technologies other than batteries.

Yes I would like after the battery are full to dump the enery in the electric water heater which is a air to water heat pump with a resistance heater if needed.
I never thought you can dump the energy in ice. Here were I live (SC, USA) it is a warm climate so not sure if it is a smart idea...

[Zucca]

How about this one?

https://www.victronenergy.com/battery-monitors/lynx-smart-bms

https://www.victronenergy.com/media/pg/Lynx_Smart_BMS/en/index-en.html

I can't understand it it monitors the single voltage cells of the 16S pack I wanna build.

[Marco]

Here were I live (SC, USA) it is a warm climate so not sure if it is a smart idea...

With whole house AC it can make sense from an abstract view, ignoring the equipment market.

The problem is that it will take DIY, even for reversible heatpumps the automation isn't really designed for hot and cold storage tanks AFAICS. Also high capacity heat exchangers for ducted air conditioning probably aren't cheap. It is mostly for really big installs where they can do some engineering to make it work, until it's standard for homes and there are commodity components it only makes sense as a hobby project.

[metrologist]

For example

https://www.aliexpress.com/item/3256804013889748.html

 

That's temp ting... 2+ months delivery

[rx8pilot]

I did a similar project about 18mos ago with Victron electronics and random Alibaba LiFePO4 cells.

My original plan was 16S but BMS options were very limited so I compromised to 8S 400Ah. In general, I found that there are a myriad of horrible rotten BMS systems out there that are super cheap. The ones that I felt were trustworthy, were super expensive. None of them (at any price) were impressive.

I ended up going with a Chinese BMS that I modified to make it passable (but far from greatness). As I went through the DIY process, I was more and more reminded of how easily the BMS can make/break the system. I really wanted a set it and forget it setup, but have been unable to get to that level of trust. The BMS can destroy your pack with only a tiny glitch in its function. Knowing that it takes months to replace and costs a substantial amount of money - I did not want that. My system is not full time, just a backup and periodic use 'portable' system. When I need it, I don't want to learn the BMS broke it.

Perhaps in the last 18mos something better has become available.

[Zucca]

You read my mind pilot!
Exactly!

Putting a 16S 15KWh battery together is not difficult, the nightmare is to protect it without to degrade its function.
Yes I burned my finger already with cheap-ass BMS in smaller project. That's why I'm posting here.

Even if it takes two months a decent LiFePO4 from china land should be okay.

The Victron energy quattro does stop charging at a settable high voltage and do not drain the battery under a low settable voltage, so it is a first layer of BMS already there.
But if I hook up the battery directly to the quattro with no BMS:
1) If the quattro fails, there is nothing to protect the battery and with such energy welll....
2) There is no protection/monitoring at cells level

Still thinking how important is to monitor the single cells 24/7, I will anyway check them every month or so to see if they are +-0.2V to each other.

[uer166]

I found that there are a myriad of horrible rotten BMS systems

Seriously, I don't understand: is it the inherent complexity of making a reliable BMS? Maybe designers under-estimate the impact of faults/glitches on a pack? Or maybe people just don't give a crap? It seems like it's impossible to determine the quality/reliability of a BMS without reverse engineering half of it, not even counting the code.

This is why I would only do a monitor-only BMS if possible with a service connector to balance the pack every couple years if needed.. Surprisingly similar to shitty "lipo" RC hobby packs with their balance chargers and cell tap connectors actually.

[NiHaoMike]

What inverter you wanna use then?

A cheap option is to get a Prius inverter and modify it. There's also a few options for ready made 400V inverters.

Also high capacity heat exchangers for ducted air conditioning probably aren't cheap.

Car radiators are pretty cheap, just use multiple to reduce the thermal resistance.

[Zucca]

My wife will never approve an ice box. 

[NiHaoMike]

My wife will never approve an ice box. 

If air conditioning is used more than occasionally, ice storage is far more economical than extra batteries.

[Marco]

My wife will never approve an ice box. 

You'd have an insulated IBC with water. The heatpump pumps propylene glycol through one set of pipes in the IBC cooling/freezing it, another set pumps propylene glycol to duct heat exchangers. Chillers with "water" loops and air handling units are common components of large installations, hydronic distribution with fan coil units is common too. Those are the most complex components and they can afford more one time engineering, so they can experiment with ice based heat storage relatively cheaply. Water loop + duct based air conditioning for consumer class equipment, not so common and the effort put into engineering not really economically viable ... but hobbies don't need economic justification.

Might have to put some closed cell foam along the inside of the IBC to take the brunt of the freeze/thaw cycle forces

[rx8pilot]

Still thinking how important is to monitor the single cells 24/7, I will anyway check them every month or so to see if they are +-0.2V to each other.

The voltage curve on LiFePO4 is so flat that the cell differential is only interesting at 85%+ SoC or so. Trying to 'middle balance' is not practical.

My pack maintains balance fairly well and I only evaluate the balance periodically. To do that effectively, I (fairly slow)charge the system to a relatively high SoC that is limited by the first cell to reach a target voltage in the range of 3.55v to 3.60v. This is where I will check the differential an allow the BMS to start a 1A balance if needed. I leave this process to run for quite a while while the pack is unloaded and it gets within some tiny amount like 20mV.

The problem is that this process is manual (for me), it takes a long time, while it is running the pack is not usable. For a home system, that would not be a good option.

I believe the more advanced systems can reliably maintain a tight balance without any user intervention or monitoring. My system runs from about 30% to 70% SoC which is a range where imbalance cannot be easily seen, but also is less critical. I intentionally oversized the the system so that I did not need to go from 0%-100% SoC and that really reduces the need for precision balancing.

For fun, I laid out my own BMS with a lot of cool and fancy features but it is unlikely I will ever take the time to actually build it. It was, however, a fun exercise learning a lot about how BMS systems work and what makes the cheap ones scare me.

[Zucca]

My wife will never approve an ice box. 

You'd have an insulated IBC with water. The heatpump pumps propylene glycol through one set of pipes in the IBC cooling/freezing it, another set pumps propylene glycol to duct heat exchangers. Chillers with "water" loops and air handling units are common components of large installations, hydronic distribution with fan coil units is common too. Those are the most complex components and they can afford more one time engineering, so they can experiment with ice based heat storage relatively cheaply. Water loop + duct based air conditioning for consumer class equipment, not so common and the effort put into engineering not really economically viable ... but hobbies don't need economic justification.

Might have to put some closed cell foam along the inside of the IBC to take the brunt of the freeze/thaw cycle forces

Too much work. In the future I will go geothermal closed loop... so I prefer to have the energy in the batteries. Also maybe an EV car will arrive.... thanks for the suggestions, surely an impressive idea/project!

[Zucca]

30% to 70% SoC which is a range where imbalance cannot be easily seen, but also is less critical. I intentionally oversized the the system so that I did not need to go from 0%-100% SoC and that really reduces the need for precision balancing.

Oversize is a must and for just 18.45Wh/$ not too bad. 30-70% SOC is a sweet point, agreed.
Temped to go for 3P 16S (49Kwh) or 4P 16S (65.5KWh) with 320Ah 3.2V.

Also if something goes wrong I can disconnect the offending battery, and service it while the other are operating.

I am almost sold with protecting the 16S with just a 320A fuse. I will put my money on good quality charger MPPT solar and discharger/inverter so I should be safe.

but if I buy lemons from china all my castle project crumbles in dust....

Also.... worth to consider...

[Marco]

Too much work. In the future I will go geothermal closed loop... so I prefer to have the energy in the batteries.

Might still want a boiler as energy storage regardless, depending on system costs at the time. A heatpump or resistive heater which can take shower water straight to legionella killing temperatures can be hard to find or expensive. The boiler seems to be the more common solution.

PS. personally I'd like to see systems which just use a heat exchanger near the shower to heat cold water to shower temperature using warm water. Then you would only need hot water at the kitchen sink, for which a small close in boiler suffices. That system only exists in my mind though.

[Faringdon]

You could buy a Powervault system....i think you can buy for <£6k.....they contain big batteries which are well checked over from China/Taiwan.
Or just ask Powervault if you can buy a battery off them...because they have been through the QA steps and have found the best qulaity battery people in China........i think they also give you a warrantee for the battery.

To check if you have a rogue cell when you make a pack......put it through full charge/discharge at max power 10 times on the run...monitor each cell voltage while discharging/chargiug....if any cell looks deviant then pull it out and replce it.....dont be too bothered about the cell voltage when they are not charging/discharging, as this is always quite variant.

I am not sure if a powervault system can go off grid or not.....maybe it can if you do some mods to it?...i am pretty certain this can be done with Powervault...ie, go offgrid....btu you probably need some elec knowledge, which you have  anyway

[Zucca]

straight to legionella killing temperatures can be hard to find or expensive.

I do not have hot water re-circulation at home, I heard the the legionella are almost not present in system with short pipes (home not building) and  where the water in the pipes stay cold.
I could be wrong I am an EE.

[Zucca]

You could buy a Powervault system...

this should be better
https://batteryhookup.com/pages/contact-us

[Zucca]

interesting

https://www.batteryspace.com/prod-specs/6411-Canbus-v4.pdf

https://www.victronenergy.com/live/battery_compatibility:can-bus_bms-cable#introduction

https://community.victronenergy.com/questions/48924/can-bms-protocol-specification.html

[NiHaoMike]

I do not have hot water re-circulation at home, I heard the the legionella are almost not present in system with short pipes (home not building) and  where the water in the pipes stay cold.
I could be wrong I am an EE.

I assume you're on well water or the city water doesn't contain chlorine? In the US, city water contains chlorine (usually in the form of chloramine nowadays, more effective than chlorine gas) specifically to prevent that problem.

[Zucca]

Yes I am connected to the public water grid and I can confirm there is chloramine/chlorine, so jay I shoud be safe thanks NiHaoMike for pointing that out...

Anyway here my main concern, I do not know if 888$ for 16x1024Wh LiFePO4 on Aliexpress is too good to be true....

[Zucca]

For example

https://www.aliexpress.com/item/3256804013889748.html

 

After some research, it must be a scam. Too good to be true.