Nissan Leaf 48v Off-Grid Solar Project


Solar Batteries, Panels, Chargers, Inverters & Mounts at www.batteryhookup.com Save 5% with coupon code "Powerwalls"

Fukuokian

New member
Joined
Oct 15, 2024
Messages
29
I'll use this place to document my build... It's gonna be a long, slow project, but I'm excited and enjoy the research, design & DIY that is going into it.

I have a small second house in the Japanese countryside; eventually I may retire there. The motivation for the project is to be completely off-grid with no bills. (other than property tax and maintenance.) I've already cut off the electrical service, and am running on 2 solar panels, 2500w inverter, victron 100/20 MPPT, 4 agm 12v batteries 2s2p for a 24v system. It's good enough for lights and a small fridge, and even a spot air conditioner during summer days. It's time for an upgrade!

Onto the project...

I've purchased 2 Nissan Leaf Gen2 packs (8 bars - 61-66% capacity) for around $220 usd each after shipping. :) Would have preferred 9 or 10 bars, but that would have doubled or tripled the price at best, if I could even find them. Also, I was only gonna do 1 pack, so I'm very happy with the new total capacity. It should give me 30kWh.

The packs have been disassembled. *** ...best advice I found was to use a multi-tool on the clamshell seam to cut through the rubber/glue holding it together. Had it open in 5 minutes. Most of you already know, but I'll put out this Warning: The pack is High Voltage (400v-- or 2 x 200v with the service plug removed) Enough voltage to kill you. Use insulated tools and gloves (rated for the voltage), eye protection, etc. I wrapped my socket wrench and extension with thick rubber hose and used several wraps of electrical tape on the sockets. I also took everything apart in steps that reduced the voltage of the connected parts as quickly as possible.

1729574537758.png


For my 48v system, I decided on 7 stacks connected together in series each stack will have 13 modules connected in parallel. The first pack of 48 modules came discharged to about 7.2v, the second stack's modules are all at 7.8v. My plan is to use about half of each to make a stack. --> Warning... they must be the same (or within 0.1v? for Lithium ion) voltage before connecting in parallel. (Double-check the voltages just before making the parallel connections, they may settle a bit after charging, or you might have mixed in a different module. I'm currently charging the first pack's modules up to 7.8v before I make the parallel connections. After that, I can let them sit for a good week or more to balance themselves.

Where am I at now... I'm still charging the modules 1 by one using a step-down converter (buck converter) and an old 9v 3a power supply. It's painfully slow, but cheap. About $7 for the converter, and the power supply was free. (Yes, my crimping is cringe! :eek:! Need to redo, or at least tape. Connection is good though!)

1729574174156.png


Still deciding on the terminal connecting busbars. WIll reuse the copper bars that came with the pack... but still need to source more. I got a lot of good advice here... https://secondlifestorage.com/index...arallel-cell-module-copper-connections.12888/ I'm considering flattening copper pipe, but I've tried this before and gave up because the surface was still too rough; I didn't like to possibility of a reduced terminal connectivity. I may do some more testing with a hydraulic jack/or modified crimping tool. It would only need to be very flat/smooth at the terminal connection points.

That's it for now... I'll update as more gets done!
Your advice, personal experience, and pointing out things you feel are, or could be unsafe, is appreciated. I'm learning as I go.
 
For my 48v system, I decided on 7 stacks connected together in series each stack will have 13 modules connected in parallel. The first pack of 48 modules came discharged to about 7.2v, the second stack's modules are all at 7.8v. My plan is to use about half of each to make a stack. --> Warning... they must be the same (or within 0.1v? for Lithium ion) voltage before connecting in parallel. (Double-check the voltages just before making the parallel connections, they may settle a bit after charging, or you might have mixed in a different module. I'm currently charging the first pack's modules up to 7.8v before I make the parallel connections. After that, I can let them sit for a good week or more to balance themselves.
7 stacks of 13 would be 14s13p. Found this and trying to decipher it - https://pushevs.com/2018/01/29/2018-nissan-leaf-battery-real-specs/ - leads me to think in terms of each physical unit has about ~60ah? So a 2s13p stack would be ~780ah? At 14s13p that would be 390ah * 52v = ~20kwh overall capacity.

For each 13p group in parallel you'll have best results if each of the 13p(s) are pretty close to the same ah capacity (and IR is good all around). Typically one does this by testing each cell and then evenly distributing them so that each 13p has pretty close the same capacity. If 780ah is theoretical then +/- 50ah (5%) would be a reasonable target for good results. If you average 70% original capacity then 70% of 780ah = 546ah so 546ah +/- 25ah would be the proportional target.

However, it's not clear you're planning on testing each cell - you're comments about taking a long time suggest you may not be planning to test. In this case, you could wind up with a13p stack at say 700ah and another at 800ah and this may lead to sagging. What I do when I have an imbalance like this..... (e.g. good to have fall back plans) .... is to add capacity. I have 126 packs of ~100p in my powerwall. About 5% of them they showed various levels of 'sag' - e.g. would fall behind the others - after putting them online. In these cases I added cells (5-15% capacity boost) after the fact and this fixed the issue and smoothed things out. *BMS with good visibility is very useful to diagnose, fix, verify this.

In you're case, adding a 14th unit on some stacks might be a useful fallback plan... so suggest a physical build that let's you do this - e.g. don't put a shelf so tight over the stack you could add a 14th unit.

However, if you can, it's best to test each cell and evenly distribute capacities for lowest risk. If you develop the dishcharge/charger infrastructure for this and make it part of you're routine, it will serve you well as you build the current battery and expand in the future.
 
Last edited:
7 stacks of 13 would be 14s13p. Found this and trying to decipher it - https://pushevs.com/2018/01/29/2018-nissan-leaf-battery-real-specs/ - leads me to think in terms of each physical unit has about ~60ah? So a 2s13p stack would be ~780ah? At 14s13p that would be 390ah * 52v = ~20kwh overall capacity.

For each 13p group in parallel you'll have best results if each of the 13p(s) are pretty close to the same ah capacity (and IR is good all around). Typically one does this by testing each cell and then evenly distributing them so that each 13p has pretty close the same capacity. If 780ah is theoretical then +/- 50ah (5%) would be a reasonable target for good results. If you average 70% original capacity then 70% of 780ah = 546ah so 546ah +/- 25ah would be the proportional target.

However, it's not clear you're planning on testing each cell - you're comments about taking a long time suggest you may not be planning to test. In this case, you could wind up with a13p stack at say 700ah and another at 800ah and this may lead to sagging. What I do when I have an imbalance like this..... (e.g. good to have fall back plans) .... is to add capacity. I have 126 packs of ~100p in my powerwall. About 5% of them they showed various levels of 'sag' - e.g. would fall behind the others - after putting them online. In these cases I added cells (5-15% capacity boost) after the fact and this fixed the issue and smoothed things out. *BMS with good visibility is very useful to diagnose, fix, verify this.

In you're case, adding a 14th unit on some stacks might be a useful fallback plan... so suggest a physical build that let's you do this - e.g. don't put a shelf so tight over the stack you could add a 14th unit.

However, if you can, it's best to test each cell and evenly distribute capacities for lowest risk. If you develop the dishcharge/charger infrastructure for this and make it part of you're routine, it will serve you well as you build the current battery and expand in the future.
Thanks for all this info!

I found mostly similar data (the modules were changed a bit between the different battery generations) Off the top of my head, I think Nissan reported 65ah for the Gen2 modules; but I read somewhere that an EV shop found it to be more like 59ah in the real world. I think part of the difference is a safety margin that Nissan built into the maximum charge they allow.

In my overall capacity, I figured 21.5kWh(usable from 24kW rated) x 2 then .65% of that). But... if forgot to account for the 5 modules not being used. (ie there are 96 total modules from the 2 packs, but 7 stacks of 13 means I'm only using 91.) 30kWh was only a ballpark figure; I'm happy with anything over 20kWh. Accounting for the 5 missing modules, I should then get about 24KWh for my finished battery. I'll use my PZEM-015 with 200a shunt to do the final capacity testing. :)

Yeah, I'm not planning on testing each cell or module. I know that is optimal, and probably best practice; but, the two packs had reportedly similar SOH when they came out of the cars. I plan to equally distribute the modules between the stacks, to keep the stack capacities similar, in theory. I also plan on monitoring the BMS data (JIKONG BMS JK-PB2A16S15P) to see if a stack is always balancing up. If that's the case, I'll break apart the stack and test those ones individually. Finally, I don't plan on pushing the top charge voltage to the max. (I'll probably keep it to 4.1v per cell. Maximum rated charge is 4.2v) Similarly on the low end, I'll cut off if the cells drop to 3.65v. I lose some capacity this way, but I think it allows for a safety buffer. What do you think? Am I making a mistake here?

Oh, I did check all the cells in the first pack with a multimeter, and they were almost all identical. Basically, 3.6v each. However, two of the modules were lower... (3.55, 3.53) and (3.52, 3.54) so I kept those modules in the 5 I'm not using. I still have to check all the 2nd pack's cells.

In you're case, adding a 14th unit on some stacks might be a useful fallback plan... so suggest a physical build that let's you do this - e.g. don't put a shelf so tight over the stack you could add a 14th unit.
Great advice! I've left room in my stack design to allow for adding 7 more modules. (ie. I can add 1 more leaf battery pack + 1 module)
 
Oh, if anyone is using the JIKONG BMS JK-PB2A16S15P, make sure you upgrade the firmware. Older versions had various safety problems that were fixed/patched. (You can google for more info on them.)
 
Finally, I don't plan on pushing the top charge voltage to the max. (I'll probably keep it to 4.1v per cell. Maximum rated charge is 4.2v) Similarly on the low end, I'll cut off if the cells drop to 3.65v. I lose some capacity this way, but I think it allows for a safety buffer. What do you think? Am I making a mistake here?
If you're talking about day to day operations... then this is on the right track. However the discharge curve knee (cut-off) is closer to 3.5v/cell than 3.65v - depends on the load. This is illustrated on a chart from that link above.
1729692700285.png


I have 18650 cells but same voltage range / general chemistry as you're Leaf cells. My goal is to maximize long life. Early on I found this infamous chart from a page at "Battery University" - https://batteryuniversity.com/article/bu-808-how-to-prolong-lithium-based-batteries - to help explain longer life. Basically it makes the case that using cells 'gently' (avoid extremes) will greatly increase the number of cycles you can expect.

Here's the key chart lower on that web page that I built my system toward.....
1729692282652.png


I operate between 4.0v/cell max (14 * 4.0v = 56v hi) and 3.54v/cell min (3.54 * 14 = 49.5v) and <0.06C (low stress). In my system, a typical Solar day doesn't use the whole 3.5v - 4.0v range but rather in summer, things stay near the upper end and in winter the lower end. My lifetime average is 36.1% DOD and I just went over 2,000 cycles (7yrs) on 40% of my powerwall this year with no obvious sign of degradation.

My hope is to get 7,000 or more cycles as the chart suggests but there are many variables and lack of long term info. If I live long enough, I hope to report 21yrs and 6,000+ cycles - or at least until DIY home fusion is possible. :)
 
Last edited:
If you're talking about day to day operations... then this is on the right track. However the discharge curve knee (cut-off) is closer to 3.5v/cell than 3.65v - depends on the load. This is illustrated on a chart from that link above.
View attachment 32468

I have 18650 cells but same voltage range / general chemistry as you're Leaf cells. My goal is to maximize long life. Early on I found this infamous chart from a page at "Battery University" - https://batteryuniversity.com/article/bu-808-how-to-prolong-lithium-based-batteries - to help explain longer life. Basically it makes the case that using cells 'gently' (avoid extremes) will greatly increase the number of cycles you can expect.

Here's the key chart lower on that web page that I built my system toward.....
View attachment 32467

I operate between 4.0v/cell max (14 * 4.0v = 56v hi) and 3.54v/cell min (3.54 * 14 = 49.5v) and <0.06C (low stress). In my system, a typical Solar day doesn't use the whole 3.5v - 4.0v range but rather in summer, things stay near the upper end and in winter the lower end. My lifetime average is 36.1% DOD and I just went over 2,000 cycles (7yrs) on 40% of my powerwall this year with no obvious sign of degradation.

My hope is to get 7,000 or more cycles as the chart suggests but there are many variables and lack of long term info. If I live long enough, I hope to report 21yrs and 6,000+ cycles - or at least until DIY home fusion is possible. :)
Love this info... thanks. 🤗 Without knowing my indivual cells capacity, would you think my use of a higher low voltage cutoff is warranted? I know the stack's cells 'self-balance' because they are in parallel, but I don't know how quickly it happens. For example, under a heavy load that's draining the battery quickly is there a danger of 1 cell getting pulled into the life shortening range. Is it safer for me to keep my low voltage cutoff a bit higher? (I might not be understanding what is really happening, so it's ok to smack me with the truth! 🤪)
🤪
 
because they are in parallel, but I don't know how quickly it happens.
Virtually instantaneous.
Under heavy loads it is possible to have cells closer to the bus bars deliver more Amps than those further away. But their voltages will be the same.

This is why it's good practice to have Neg and Pos either on opposite ends of the pack or opposite corners to force current to flow through all cells equally
 
Virtually instantaneous.
Under heavy loads it is possible to have cells closer to the bus bars deliver more Amps than those further away. But their voltages will be the same.

This is why it's good practice to have Neg and Pos either on opposite ends of the pack or opposite corners to force current to flow through all cells equally
Awesome! Thanks!! So I won't need to sacrifice so much useful capacity. I'll just keep things in the same range OffGridInTheCity uses, to maximize battery life!
 
Picked up some metal cabinets for the batteries. (2 x $22=$44 usd) 😙 Other ideas I had: find an old freezer chest (great option for it being well sealed and having insulation), computer server cabinet (hard to find, and usually not cheap), garden shed (made in all sizes and shapes)

1730093415685.png


I'll have to reinforce the bottoms of each to support the 150kg & 200kg weights of the batteries... probably just 2 wood planks running lengthwise, and cut to fit into the lips of the cabinet base.

1730093776207.png

These go outside, in a semi enclosed area next to the house. If there's room, I'll keep them side by side (176cm total length); otherwise, I'll build a reinforced shelf system to stack them. I prefer the side by side option, as I live in Japan where earthquakes happen all the time. ⚠️ That keeps the center of gravity as low as possible. Either way, I still have to build a small concrete block support to which the cabinets can be bolted. I do want the cabinets raised a little off the ground 1 or 2 blocks high, in case there's ever flooding.

Other important consideration:

- I have to make sure they're well sealed! Lots of critters around that like dry places to set up their homes. Wasps are especially plentiful in my area. (Recommend googling 'fried mouse in electronics' or 'fried snake in electronics' 😵‍💫)

- need to consider cooling/ventilation and heating. My first cabinet had thermostat controlled computer fans at the top to draw out hot air. Vents are located at the bottom to draw air from outside. In the picture below, there's a shelf missing with holes cut on the far right side. The idea was to have the air from outside be pulled to the right and across the batteries, then through the shelf and up across the electronics and out through the top left. I didn't get to the heating part of it, as I was only using it during warmer seasons. However, I planned to look into using seedling tray heating mats hooked up to another thermostat controller. (reference: stc-1000)



1730096312156.png
 
I'm considering flattening copper pipe, but I've tried this before and gave up because the surface was still too rough; I didn't like to possibility of a reduced terminal connectivity. I may do some more testing with a hydraulic jack/or modified crimping tool. It would only need to be very flat/smooth at the terminal connection points.

I am in the midst of upgrading my 7 year old 18650 with the nissan leaf packs myself. I already built one pack and in the midst of the second pack. I use copper pipes and I'm not worried about the amp draw. If I was pulling 200A out of a single cell then I'll be worried but across so many parallel packs I'll be surprised if it pulls even 10A. So even with the lesser quality of the conductivity I have not noticed any issues. I scan them with my FLIR camera and they're nice and cool.

The technique to get a good bar is to sandwich it between two metal plates to get nice flat bars. My first attempt was using a vice and clamping them a little at a time but the easiest is to use a press and just press it all at once.

Image6031225956873388375.jpgImage6470506548130014106.jpgImage8128572621144052004.jpgImage3534232084253404757.jpg
 
I am in the midst of upgrading my 7 year old 18650 with the nissan leaf packs myself. I already built one pack and in the midst of the second pack. I use copper pipes and I'm not worried about the amp draw. If I was pulling 200A out of a single cell then I'll be worried but across so many parallel packs I'll be surprised if it pulls even 10A. So even with the lesser quality of the conductivity I have not noticed any issues. I scan them with my FLIR camera and they're nice and cool.

The technique to get a good bar is to sandwich it between two metal plates to get nice flat bars. My first attempt was using a vice and clamping them a little at a time but the easiest is to use a press and just press it all at once.

View attachment 32529View attachment 32530View attachment 32531View attachment 32528
Looks good, nice technique! Someday I'll have a press too! (y)
 
The technique to get a good bar is to sandwich it between two metal plates to get nice flat bars
Oh, I love that technique!

I'd love a press tool! but not having a garage I suppose my wife would say "Or me or the press tool". And I might become a single man again :ROFLMAO:
 
  • Haha
Reactions: kje
Back
Top