diy solar

diy solar

13 kWh Lishen van build

blutow

Solar Enthusiast
Joined
Dec 20, 2020
Messages
344
I'm building a battery bank for a new sprinter van I've got coming next month. Original plan was battleborns or SOK's, but I decided to go the DIY route after spending some time on this site. Thanks for the all the great info!

The plan is 400watts solar w victron mppt, Victron 12/3000 inverter, cerbo and touch 50 screen. I've got everything bought except the solar panel (picking up next week)

16 lishen 272 cells into 4 12v packs, each using an overkill BMS. Custom built battery boxes with welded steel frames and hdpe sides with built in compression.

I ordered cells at the end of December through Michael Caro group buy. I'd recommend going with Michael. Cells arrived from Basen today. I paid ~$100 extra for expedited shipping/handling and it was just under 60 days. The cells were packed 2 to a box instead of 4, that was part of the extra shipping. Double bus bars and studs w/lock nuts.

IMG_0433.jpgIMG_0434.jpg

I am pretty happy with the condition. They had the "re-wrap" issue with some bubbles that others have reported, but that doesn't bother me. A couple of them had a little gap near the top where you could see metal. Pic below is the worst one. I'm insulating between cells and against the case, so it's not really a concern for me.
IMG_0435.jpg

All the cells look identical with blue ringed vents except the 2 shown below (cells #15 and 16 on the right). These 2 had slighter higher voltage and lower IR, but I'm not going to be concerned until I test them. I wish they were all the same, but hopefully it's not an issue. One cell has a little bump I can feel on the top under the covering, but it feels like it's just some junk caught under the sticker. No issue with corrosion in any of the vents. All threads look good and I didn't notice any significant "leaners" I've seen reported recently. At this point, I'm happy with the physical appearance and I'm moving on to testing.
IMG_0437.jpg
 
The cells looked pretty consistent for voltage and IR. The 2 "non-blue" vents were a little lower on IR and just slightly higher on voltage.
1614394296970.png

Rather than starting with the top balance, I built cells 4 into a pack with the BMS to get some Ah's into them. I've been pushing ~80a into them for 2+ hours and cells are between 3.437 and 3.474.

I used red loctite on the studs for the first 4 cells. I think red loctite is the right product for decent threads, but the challenge is that it's not an open hole. It seems that regardless of home much or how little loctite you apply, it wants to push out the top of the hole since you are forcing the air out as the stud is screwed in. I wish there was a way to get more loctite in to fill in around all the threads. Hopefully it will work OK.

IMG_0438.jpgIMG_0439.jpgIMG_0440.jpg
 
Top Balance

The top balance on first 4 cells only took about an hour or so at ~7a. Doing the series charge prior was a real time saver. I made some little jumpers so I could do the top balance without breaking down the pack and keep the compression. I stopped it when the cells were at 3.64v and taking less than 2a, figured that was far enough.
IMG_0449.jpg

I used the victron cerbo and smart shunt to monitor voltage during the top balance so I didn't have to constantly watch it. I used the starter battery voltage lead from the smart shunt and set up alerts when it hit 3.6v. This victron stuff is pretty slick and easy to set up. As soon as the cells hit 3.6, things were beeping and I got an alert on my phone. I could also just watch the voltage through the VRM cloud app since the cerbo can make a real time connection through VRM. I'm not sure I really need all the connectivity and alerting capability, but it's fun to play with.

real-time view of system through VRM on my phone
IMG_0452.PNG


Cerbo alert (beeping and screen):
IMG_0453.jpg

Email alert form VRM
IMG_0454.PNG

Moving on to capacity test for the first 4 cells
 
Seems to be going well for you so far!

How are the M6 threaded terminal holes in your cells? How much thread engagement are you getting? I'm only getting 3 turns of the grub screw on my new Lishens, which really concerns me.
 
Seems to be going well for you so far!

How are the M6 threaded terminal holes in your cells? How much thread engagement are you getting? I'm only getting 3 turns of the grub screw on my new Lishens, which really concerns me.
The threads are definitely my biggest concern since they are going in a high vibration environment. They seem solid with the loctite, but it’s just not a lot of threads. Wish I could have gotten more loctite to go into the threads, but it was basically pumping out the top with the hydraulic effect. Even with just a little applied to the threads. If someone made a stud that was hollow down the center, it would be much better to allow the loctite to flow freely.
 
Capacity test for first pack

Ran into my first problem and realized I probably screwed up the initial series charge a little bit also.

I reassembled the battery with BMS and started the capacity test. I immediately saw one of my cells showing significantly lower voltage then the others. The bus bar on the cell was hot (like 3 minutes into test). Everything else was cool. I stopped the test and pulled the bus bar. Some loctite was dried on the terminal surface. Cleaned it and reassembled, restarted test. Now a couple other bus bars were running warm and voltage wasn't consistent across cells. I stopped the test again and pulled everything off and cleaned all terminals and applied Noalox. Restarted test and everything was consistent and cool throughout. My test load was 1000 watts, so about 75-80 amps or about .28c. I used 2 old halogen work lights.

Lesson learned - you really need to watch the terminal connections. I had one cell dragging last night during the series charge and it was the one with the loctite on the terminal. Several of my bus bars were warm during charging, but none got warm during the capacity test after re-cleaning terminals and applying Noalox. The BMS never got above 97F during the test and I'll probably never pull this many amps continuous from each battery in real-world use.

Results -
278Ah is the last value I saw before the BMS cut power. It was actually 271 on the screen, but I had to add in 7Ah from the prior 2 false starts. I was running the test off my smart shunt and it would loose it's memory each time I had to kill power to clean terminals. Pleased with the results.

More encouraging is that the cells were pretty consistent all the way down, so no problem children yet.

Following were voltages at 272Ah (starting to wander a little at this point):
IMG_0461.PNG
It's just crazy how long these cells hold their voltage, and then it's like a free fall.

One more lesson learned - set the cerbo to log every minute instead of every 5 when I get near the end of the capacity test. I wasn't able to capture the final Ah number since the smart shunt died right before the next scheduled report.

Moving onto pack #2.
 
Results on 2nd group of 4 cells -
  • Charged in series at 90a for a couple hours until BMS cuttoff on single cell. This pack was less balanced compared to first, so wasn't able to put as much capacity into it. This pack included my 2 "non-blue" ringed vents that showed lower IR than all the other cells.
  • Parallel top balance at 8a for almost 5 hours (compared to under 2 hours on my first set). I did let them get all the way to 3.65 this time.
  • No hiccups on capacity test this time. I pushed it up to 100a discharge by adding another light. I turned off one of the lights near the end because amps were getting up near BMS limit of 120a as the voltage dropped. The BMS temp hit 105f and cells were a little warm, but I had no heat build up on bus bars. Cells were pretty balanced for most of the test. They were all 2.5x volts when the bms shut down.
  • Another happy result with 276Ah capacity. My BMS was showing over 290ah consumed, but I'm inclined to believe the victron shunt. Screen shot below from end of test. I probably won't have time to mess with pack #3 for a few days due to work, but I may cycle pack 2 a couple times to see how well it stays balanced.

1614565068370.png
 
I did a more "real world use" test on the 2nd pack today. I only charged the pack to 14.2v and then discharged at ~40a (and then at 10a at the very end).

At 281ah, I stopped the test when my low cell got to 2.6. Results were better than yesterday on the same pack despite starting with a lower charge and ending the test prior to BMS shutdown. Discharge rate seems to make a difference.

Also, I was really surprised how much the cell voltages recovered when I reduced my discharge from 40a to 10a at 272ah into the test. I thought I was about done at that point. You can see in the graph below that voltage jumped up and I got another 9 ah. The pack voltage at the end of the test was actually still higher than what it was showing at 272Ah.

1614648642900.png
 
Results from 3rd pack:

Series bulk charge - this group of cells were pretty well balanced out of the box. I only had my charger set to 14.1v and the cells reached that without any significant runners.

So, the parallel test took less than 2 hours @ 7a (much quicker than my 2nd pack with mixed cells).

Capacity test - 283Ah, my highest yet. Again, I didn't wait until BMS cutoff, but my lowest cell was 2.5x, so getting really close. The slightly higher result compared to the first 2 packs was probably helped by a lower discharge rate. I ran the test overnight and kept discharge to only ~30a so it would still be running when I woke up this morning.

One more pack to test and then I'll probably do a bunch more real-world pack cycling while I finish construction on my cases.
 
Just came across your thread @blutow and I am planning nearly the same setup, so I'm watching excitedly.
Good luck with your build. If you have any questions on the equipment, let me know. I'm very pleased with the victron stuff and the remote access and notification features have been fun to tinker with and very helpful while doing all the charge and discharge cycles (even the top balancing). So far, I'm very pleased with how things are going. Sourcing the cells and the long wait to see the condition/performance was definitely the most uncomfortable part. I ended up very lucky with no visual issues and the cells all seem to be performing well. I'd definitely go with one of the top recommended vendors or do the group buy with Michael if you haven't already sourced your cells. Issues do happen and I felt like Michael would have taken care of me if I had run into problems.

The van I ordered in November is supposed to be built next week, but I heard from my dealer that they are dealing with some certification issues on the 2021 models. The vans are just sitting after manufacturing and the dealers don't know when they will be certified and shipped. It could be months, but I was kind of mentally prepared to wait a while longer.
 
The van I ordered in November is supposed to be built next week, but I heard from my dealer that they are dealing with some certification issues on the 2021 models. The vans are just sitting after manufacturing and the dealers don't know when they will be certified and shipped. It could be months, but I was kind of mentally prepared to wait a while longer.
Where are you located?

I ordered a 4x4 170 3500XD, here in Canada, which means it will be built in Düsseldorf. I'm told emissions certification issues with the 3.0l TD are holding up production and MB is not manufacturing them until that goes through. Currently I'm in line for a late June build slot and late July delivery, not nothing is guaranteed until MB accepts my order and secures my slot.
 
Pack #4 went pretty much the same as pack #3.

They seemed well balanced out of the box and took a charge in series up to 14.1v without any runners.

Top balance took less than 2 hours @ 7a.

Capacity test = 282ah.

So, all packs were 280+ except the first @278. I think that lower result was more about the slightly different test protocol on the first pack. I discharged that one at ~80a until low voltage disconnect on the BMS. On the other packs, I reduced discharge to around 30a when I started getting voltage variation and that seemed to level things out and also bumped the voltage up enough to squeeze a few more Ahs. I also killed those subsequent tests once I saw a cell drop below 2.6 rather than waiting for BMS cutoff. I'm not sure whether reducing the discharge is cheating or not, but pushing down to 2.5v at 80a seemed a little abusive to the poor little cells and there isn't any real capacity under 2.6 anyway. I'm sure it wasn't hurting anything to wait for the BMS cutoff, but it seemed pointless.

I'm going to do some additional real world cycling just to see if I encounter any weirdness, but at this point I feel like I have 4 good batteries. With all the reports of damaged and problem cells I've seen, it's a big relief to have these initial tests done with good results. I also have cautious optimism that my loctite studs turned out good and will hold up.

I've got one of those fan-based cell testers. I was originally planning to test each cell, but I'm questioning whether there is any significant value. Probably just an opportunity for me to screw something up. I bought it primarily to test any suspect problem cells and I haven't seen one. The tester can only take about 7a, so we're talking about a 40 hour test for each cell. I may monkey with a couple cells since my van isn't going to be here for a while, TBD.

As a beginner going through this process for the first time, a few observations/comments:

Make sure your bus bar/terminal connections are good. I ran into issues on my first pack. I spent a bunch of time filing, sanding, cleaning all bus bars and terminals, and applied noalox to everything. I didn't have another issue after that.​
I recognize that I'm a newbie to Lifepo4, but I question Will's standard recommendation of 14.5v for bulk/absorption charging (in the sticky at the top of the forum). I'm not saying there isn't a place for 14.5v charging, but it seems like it's on the hairy edge from the behavior I saw with my cells. 14.5v = at least 3.625v per cell. With my cells, they were not balanced enough to get to 14.5v without hitting high voltage cutoff at 3.65v on one cell (2 times for me on my second pack). With such little capacity above 3.5v, I don't get the point of charging above 14.1v or maybe 14.2v. Will says 14.1v is a good charge voltage to extend cell life, but I question whether 14.1v shouldn't be the default recommendation and make 14.5v the setting for people who want to charge a little faster at the end or get every last drop of capacity. Particularly when most folks on this forum are using unmatched cells prone to running at the extremes. Again, I'm not an expert and there is probably something more going on that I don't understand, but it's just my observation.​
Don't get complacent - The first time I was running BMS leads and bus bars, I was extremely careful and aware of the danger. After doing it so many times over the course of a week, I had a few lapses where I was maybe a a bit too comfortable. I didn't have any issues and I don't want to preach, but just a friendly reminder to be careful out there (safety glasses, tools wrapped/insulated, clean work area, etc.).​

Next step is battery boxes. I've got one pretty much built, I just need to finalize a few things and fabricate the other 3.
 
Where are you located?

I ordered a 4x4 170 3500XD, here in Canada, which means it will be built in Düsseldorf. I'm told emissions certification issues with the 3.0l TD are holding up production and MB is not manufacturing them until that goes through. Currently I'm in line for a late June build slot and late July delivery, not nothing is guaranteed until MB accepts my order and secures my slot.
I'm in Texas. Yeah, non of the diesel v6's will be built for a while as I understand it. Hopefully they get the allocation and they get you a build date. Mine is a 2wd gas 170. A buddy and I both ordered similar vans in late november. His got built last week, mine is scheduled for the 10th. I'm just hoping I get the options I ordered, they are notorious for adjusting the build based on what they have at the plant on a given day. It's good that yours is being built in Germany, I hear they continue to have quality issues at the US plant. The Mercedes ordering process and their ongoing struggles with production, quality, and certification is like watching a clown show. This has been going on with the sprinters for years. If there was another model van that would work for what I'm trying to do, I would be running quickly in another direction. I've got low expectations, so hopefully I'll be pleasantly surprised.
 
I picked up my solar panel this week and figured I do a little testing on it this nice sunny morning. It's a solaria xt400 panel. I'm going to start with just the one panel on the rear roof section of the van and see how it goes. I'll have room for a couple more skinny panels around my fan and AC up front if I decide to add more later. I really wanted to squeeze 2 of the 400 watt panels on the roof, but they would be hanging over the sides more than I can live with.

The panel was pushing over 300 watts at 9am this morning. I'm new to solar and I wasn't sure what to expect compared to the 400w rating, but that seemed good to me. I assume it will get closer to the rated watts during summer and when the sun is direct overhead.

IMG_0505.jpg


IMG_0504.jpg
 
I had a head start on the battery boxes. I designed and built a prototype based on a mock up "dummy" pack last month while waiting for cells to arrive. I decided to hold off building the other 3 boxes until I confirmed everything is going to work. I didn't encounter any surprises when I put the actual cells in, so all good.

I used a combo of 3/4 angle for the frame and hdpe for the walls. I welded in some captive nuts and used screws and lock nuts for end compression and as a hold down across the top. My wife was giving me crap about the blue color since basically everything in the system is blue.

Between cells and on the ends, I put some 1/16 neoprene to provide a little cushion, grip, and insulation.

IMG_0514.jpgIMG_0509.jpg

Pics of how top and end compression work:
IMG_0513.jpgIMG_0512.jpg

Still working on permanent BMS mounting, but this is the idea
IMG_0511.jpgIMG_0510.jpg

I've got hdpe panels that will go on the top and end to close everything up. I'll also drill some holes above/below BMS for ventilation and for cable exits. I also plan to install a safety toggle switch to switch the BMS on and off. Still a few details to wrap up, but I think the design is going to work well. I'm pretty sure you could roll this box down a hill and the cells aren't going to budge. I hope I never have to test that.
 
One of the best battery boxes I’ve seen. Nice job! Do you plan to use the stock bus bars?

Are you ultimately paralleling 4 12v batteries?
 
One of the best battery boxes I’ve seen. Nice job! Do you plan to use the stock bus bars?

Are you ultimately paralleling 4 12v batteries?
Thanks, really appreciate the kind words. The boxes were probably my biggest concern when debating DIY for a mobile install.

Yes, I'm planning to use the stock bus bars. I cycled one of my packs multiple times and I didn't get any significant movement, so I'm not that concerned about stress on the terminals. I'm hoping the neoprene between the cells gives just a little buffer for expansion/contraction. I was also reviewing cinergi's build thread and it sounds like he's studied it a bit and is going with solid bus bars. My biggest concern with the stock bus bars at this point is that they are all a little cupped. If you run them on a flat file, you get marks along the edges on one side and marks down the middle on the other. I'm not just talking about the stamping rough edges, but the entire bus bar is just slightly cupped. I debated filing them totally flat, but you'd be removing most of the plating, so I just corrected them a little bit. I decided to do "middle bulge down" and they seem to be working well in my testing.

Yes, 4 batteries parallel in 12v system. I considered 24v, but my big amp wire runs are really short and I didn't want the added complexity.
 
Nice read! Recently order 16 272 cells from Michael and I will be installing into my 2015 Roadtrek190.
 
Nice read! Recently order 16 272 cells from Michael and I will be installing into my 2015 Roadtrek190.
Thanks. Good choice going with Michael, it made the waiting a little less stressful for me.

So are you a fan of TX the state or the Longhorns? I bleed orange and have 2 aggie kids, so we're one of those divided families.
 
Back
Top