Looking for used Li-ion batteries - pack or just cells

I'm looking for used batteries that are no longer good for EVs, but would be okay to use on my home solar system. I currently have 4 very large PbA batteries that give me almost 10 kWh at 48 volts (nominal). I was thinking that 16 Li-ion batteries in series would match the nominal voltage of the system. The voltages are Bulk=57.2 (3.575/cell), Float=53.2 (3.325/cell), Low-Battery Cutout= 47.6 (2.975/cell). The Equalize voltage is 57.2 (3.6/cell), but it is rarely used. Are these cell voltages okay for Li-ion batteries? Should I think about 15 in a string, to get the cell voltages up? LBCO is where the generator automatically starts so the voltage won't be below that.

The 10 kWh of PbA batteries that are installed now are new (and cost a little over $400 each (x4). Looking at the EVTV store, the cost for 10 kWh of Li-ion batteries would be MUCH HIGHER. Of course, I don't care how much the batteries weigh, since they will be holding down a box on the floor of my garage. That's why I'm looking for used batteries. Plus I could keep them out of the recycling bin. I'd like to add 10 or 20 kWh of capacity, so that we could get up to a day or two of storage.

These batteries will spend most of their life at the float voltage, only going down to LBCO if the power is out and the sun isn't shining. The inverter is only 6 kVA, the maximum discharge current is limited to much less than an EV. Something like 120 A max, but with 3 strings (1 PbA and 2 Li-ion) that's 40 Amps each, or .4C for a 100 Ah battery.

Know anyone who has old batteries? Is this a crazy idea?

You really should be using a specialized charger for the cells used in the Leaf, Volt and Tesla, or you will quickly ruin them.

Depending on your charger its possible you could use 4 LiFePO4 cells in series to replace a 12V lead acid battery.. They will be expensive..

Check back again when used packs become available on the market..

http://www.autobeyours.com sells leaf modules and packs. ~$150 shipped for a module (about half a kWh nameplate, two cells in parallel then another two parallel cells in series internally), and $6,000 for a pack (plus freight shipping cost). You'd need 8 modules, giving you around 3-3.5 kWh usable for $1200. A leaf pack has 48 modules in it, btw.

Those voltages are a bit on the low side, the cells' max voltage is 4.1V at full charge, and I think low cut out is around 3.2V. If you can adjust the set points on your charger and delete the equalization step of charging, you can use it. You'll still need a BMS to monitor any imbalance in the cells. Since you will be getting used modules from possibly different cars (unless you spring for a full pack), you don't get the luxury of "factory matched" cells and you could have to do quite a bit of balancing compared to the leaf's BMS.

I've bought a module through autobeyours, but I'm moving soon so the fun "second life" projects I'd like to start have been put on hold.

I'd honestly recommend starting with a small 4 cell "12v" LiPO4 pack to learn about lithium and BMS's. Watch the EEVBlog video on lithium charging to get an idea of how they are charged.

Keep in mind that lithium chemistries don't like sitting at 100% for long periods of time like lead does. On the other hand, cycling losses are much lower with lithium, so you can use them to do peak shaving after the sun goes down every night. They are not suitable for "float duty" like lead is, unless you want to leave them partially charged. Oh, and don't combine lead and lithium together on the same bus. That's just asking for trouble.

JeremyW said:

http://www.autobeyours.com sells leaf modules and packs. ~$150 shipped for a module (about half a kWh nameplate, two cells in parallel then another two parallel cells in series internally), and $6,000 for a pack (plus freight shipping cost). You'd need 8 modules, giving you around 3-3.5 kWh usable for $1200. A leaf pack has 48 modules in it, btw.

Click to expand...

I'll check out the autobeyours site. I had not heard of them. Are they new or used modules?

JeremyW said:

Those voltages are a bit on the low side, the cells' max voltage is 4.1V at full charge, and I think low cut out is around 3.2V. If you can adjust the set points on your charger and delete the equalization step of charging, you can use it. You'll still need a BMS to monitor any imbalance in the cells. Since you will be getting used modules from possibly different cars (unless you spring for a full pack), you don't get the luxury of "factory matched" cells and you could have to do quite a bit of balancing compared to the leaf's BMS.

Click to expand...

Maybe I'll use 15 cells in series, not 16, to get the voltages up. The inverter only "knows" PbA and NiMH/NiCd chemistries, so I'd have to choose one that was closest to Li-ion's charge profile. The inverter is rather old (11 years), and I'm sure their are better ones out there, but the wife-acceptance-factor is low so I'm going to have to go with what I have...

JeremyW said:

I'd honestly recommend starting with a small 4 cell "12v" LiPO4 pack to learn about lithium and BMS's. Watch the EEVBlog video on lithium charging to get an idea of how they are charged.

Keep in mind that lithium chemistries don't like sitting at 100% for long periods of time like lead does. On the other hand, cycling losses are much lower with lithium, so you can use them to do peak shaving after the sun goes down every night. They are not suitable for "float duty" like lead is, unless you want to leave them partially charged. Oh, and don't combine lead and lithium together on the same bus. That's just asking for trouble.

Click to expand...

I had not thought of peak shaving.... The inverter does have a mode where it sells all it can during the peak time period (something like 35 cents/kWh), discharging the battery. And then charging the battery back during the off-peak times (10 cents/kWh). That didn't make sense with PdA because they only have a small number of cycles in their life. Li-ion, on the other hand...

I was thinking of "floating" them partially charged, that's why the lower than max voltages... And I was also thinking of paralleling them, at the 48v string level, with the 10 kWh of PbA that I have. I was thinking it was all good, until your last sentence. :-(

What is the worry? I won't be charging or discharging them at more than .2C max, and the PbA string will be there to take up the load when discharging below the "constant voltage" region of the Li-ions. Can they get so out of balance treated so nicely, with just 8 modules, that they require a BMS?

You might be satisfied with manually balancing them, as opposed to a full blown BMS. The guys do that for an 85kWh pack in the dune buggy.

It's not fun, but how often would you have to do it for a small cell count and static waiting for the day you need them.

JohnO said:

I'll check out the autobeyours site. I had not heard of them. Are they new or used modules?

Click to expand...

They are used.

JohnO said:

Maybe I'll use 15 cells in series, not 16, to get the voltages up. The inverter only "knows" PbA and NiMH/NiCd chemistries, so I'd have to choose one that was closest to Li-ion's charge profile. The inverter is rather old (11 years), and I'm sure their are better ones out there, but the wife-acceptance-factor is low so I'm going to have to go with what I have...

Click to expand...

You don't really have the option of going with 15 cells, unless you split the module. The there are two cells in series per module, so you will always have an even number of cells in your pack. There is a balancing terminal that lets the BMS see the individual cells, but I wouldn't use that as one of the end terminals (leaving one cell out of the string) as it isn't meant for sustained high current draw.

For the bold above, I'm going to be a stickler. You can't get "close" with lithium. You have to be exact. You'll need something to cut off the voltage below 3.2 volts per cell. If you go below the knee of the curve into the bottom 1% or below in the cell, there's a good chance you will destroy it, even if it is just one time. If you go above the high voltage cutoff (overcharge) then you'll also destroy the cells (cobalt based chemistires catch fire when overcharged, I don't think manganese chemistries do, but you don't want to find out the hard way). There are no second chances with this stuff. The leaf itself doesn't let the cells go below 2-3%; the contactors physically open to cut off load to the pack. You will need something to that effect to protect your investment! You might be able to do manual balancing, but low/high voltage cutoff needs to be automatic, along with a charger that has the correct (NOT "close") voltage set points.

JohnO said:

I had not thought of peak shaving.... The inverter does have a mode where it sells all it can during the peak time period (something like 35 cents/kWh), discharging the battery. And then charging the battery back during the off-peak times (10 cents/kWh). That didn't make sense with PdA because they only have a small number of cycles in their life. Li-ion, on the other hand...

Click to expand...

You should be thinking of peak shaving. Since lithium is relativity expensive compared to lead, there is a desire to "get your money's worth."

JohnO said:

I was thinking of "floating" them partially charged, that's why the lower than max voltages... And I was also thinking of paralleling them, at the 48v string level, with the 10 kWh of PbA that I have. I was thinking it was all good, until your last sentence. :-(

What is the worry? I won't be charging or discharging them at more than .2C max, and the PbA string will be there to take up the load when discharging below the "constant voltage" region of the Li-ions. Can they get so out of balance treated so nicely, with just 8 modules, that they require a BMS?

Click to expand...

Well, first off, the lithium has a higher low voltage cut off than the lead pack. This means you could potentially ruin cells before you hit the low voltage cutoff of your voltage monitor. I'm also certain that the lithium pack also has a lower internal resistance. This means that under load the lithium pack will be doing more of the work. The cells in the leaf are very low resistance, so it begs the question of why is the lead even being hooked up if the more expensive batteries are supplying most of the load. Also, it goes without saying you won't be using the full capacity of the cells if you don't charge them completely.

You might as well as let your lead run it's life, in the mean time working on a small pack now to gain experience and nail down a bms. When the time comes you can purchase a spent pack cheaper than what you can get today. Then you can just go full lithium.

In summary, lithium chemistries are not something to half ass (no offense). My professional opinion as an electrical engineer is that you need proper charging and monitoring equipment to pull this off safely. If you inverter does not have adjustable set points, then you need look for a charge controller/inverter that does. There's no need to rush, because lithium prices haven't come down enough yet. The oldest leafs are only a bit over two years old. Wait until some 60k+ mile packs are in the used market. Also, LiPO4 has better "voltage compatibility" with lead stuff and probably better durability than what the leaf is using. Nissan chose it primary due to its higher energy density vs LiPO4. That's not applicable to you. You can also get much higher AH cells in LiPO4 than the 40Ah cells in the leaf.

Good luck! I hope I don't discourage you. These are just the realities of it. Better you know now then after buying and ruining a whole bunch of leaf modules!

The other option is you could just buy a battery system designed for home storage.

http://www.solarcity.com/residential/energy-storage.aspx" onclick="window.open(this.href);return false;

JeremyW said:

There is a balancing terminal that lets the BMS see the individual cells, but I wouldn't use that as one of the end terminals (leaving one cell out of the string) as it isn't meant for sustained high current draw.

Click to expand...

The LEAF monitors 96 parallel cell pairs, not every cell.

a parallel cell pair acts like a single cell to the BMS, each cell can even have a different capacity

TonyWilliams said:

JeremyW said:

There is a balancing terminal that lets the BMS see the individual cells, but I wouldn't use that as one of the end terminals (leaving one cell out of the string) as it isn't meant for sustained high current draw.

Click to expand...

The LEAF monitors 96 parallel cell pairs, not every cell.

Click to expand...

Right. There are two cells bonded together at their tabs in parallel. So they always have the same voltage. Essentially the pair is treated as one large cell. I'd call that technically monitoring every cell. I'd imagine AESC did their due diligence in matching the cells and subsequent packs. However when working with random modules, capacities will differ. You'll always be limited to your weakest module.

The inverter I plan to use is a Power One Aurora Uno. It utilizes a high frequency transformer to isolate the dc input from the mains output. I just need a house to put it on!