Cinergi's 28 kWh / 4 kW Solar / 10 kW inverter RV build

Steve_S said:

3 Stage Charging ? If you really want a "Float" I guess but it's unusual.
2 Stage works tickety boo for LFP as all you need is CC/CV and set your end amps (pending on gear) to 0 or 1 Amp.

3 work arounds for the loctite:
1) You can use a Serrated Washer (not the split lock washers), they hold better IMO.
2) A Missus trick learned years ago for when Loctite is a problem. Once the nut is tightened on and all is set & tested as good, use "one drop" of nail polish right where the tread goes to the nut. Coloured makes it easier to clean off later so you can see residue to wipe with pure acetone.
3) Double Nuts, if you have enough thread above the nut.

Click to expand...

3 stage because I need float because I'm in an always-on RV system with DC loads. And I can't just charge to 3.4vpc average of course (no balancing). Orion doesn't have tail current capabilities that I'm aware of.

The loctite wasn't for the nuts, it's for the studs inside the cell terminals. And it was to prevent me from pushing the stud into the cell terminal.

I should note that the Victron charge profile is 100% disabled when under CANBUS (BMS) control. That's why this is presenting such a problem. The only thing I can do is tell Orion max cell voltage is 3.65 (shut down charging and contactor if necessary) and max pack voltage is 54.4 (3.4vpc average). So that basically just sets up a CV of 54.4, period. I'm waiting to hear back from Orion on this matter.

cinergi said:

Lastly, I'm abandoning the loctite on the studs. Despite only using one drop, there's excess that never bonds and some of it ends up erupting out of the cell terminal and makes a mess under bus bars and increases resistance. I'll just use an allen wrench to hold the stud while tightening it with a wrench and then the final tighten with a torque wrench.

Click to expand...

I shook up the loctite before I used it. Some loctite come up when I screwed in the studs. I wiped it off then let it sit for around 24 hours. Then I wiped off the cells terminals with acetone just to make sure they were clean. I don't see any evidence of any erupting out but I will check under the busbars when I rearrange things.

@cinergi .... Curious what you are using for a torque wrench? I did a little searching for an open end torque wrench, but most of them seemed to be designed for mini split systems.

Bob B said:

@cinergi .... Curious what you are using for a torque wrench? I did a little searching for an open end torque wrench, but most of them seemed to be designed for mini split systems.

Click to expand...

Standard torque wrench (not open) - that's why I pre-tighten with an open wrench first and then the final tighten is with the torque wrench.

Gazoo said:

I shook up the loctite before I used it. Some loctite come up when I screwed in the studs. I wiped it off then let it sit for around 24 hours. Then I wiped off the cells terminals with acetone just to make sure they were clean. I don't see any evidence of any erupting out but I will check under the busbars when I rearrange things.

Click to expand...

Yeah mine erupted after using the cells. Definitely check under there the next time you have a chance.

It's been about a week since my last update. Video:

• Swapped in all my other unused cells to create a "new" battery and get all my cells tested at least once
• Swapped Class T fuse with Breaker
• Created small + and - bus bar inside for powering relays and BMS etc; BMS etc gets shut down with breaker
• Swapped 2 1 guage wires with 1 3/0 wire coming from cells
• "new" bank under load test
• Orion BMS back-n-forth, discussing a way to achieve something similar to 3-stage charging (2-stage only possible right now)
• Swapped out first-stage relay with a better one (less power and isolated/dry contacts)

This design, however, makes even manual precharge virtually impossible. The delay between BMS power-on and contactor-on is like 1 second.
I also realized that the power draw of the relays (and technically the BMS) aren't being accounted for. I need to move the hall-effect sensor ... but it's too small to fit over the 3/0 lugs so moving it requires cutting cables and re-crimping lugs so I'll do that fix when I create my final battery.

I'm still debating how I'll set up the final battery. I'm looking into Anderson for the battery-to-main-bus-bars connection as well as Anderson connectors to break up the pack into, say, 4 pieces (which will require me to break up the sensing leads with connectors). This would make moving them into the RV doable and I think provide for a better compression setup. I still need to play with the copper bars I got (I have a cutoff saw, drill press, and something to bend it with, in theory). I'm not entirely sure of the final physical layout inside the RV, so I'm not really sure what I want to do. With that design, there's also the issue of where to mount the equipment (like the BMS, breaker, and relays) that goes with the battery.

cinergi said:

This design, however, makes even manual precharge virtually impossible. The delay between BMS power-on and contactor-on is like 1 second.

Click to expand...

1 sec is enough for precharging the caps. Unless I missed something you can use that delay to precharge them

cinergi said:

With that design, there's also the issue of where to mount the equipment (like the BMS, breaker, and relays) that goes with the battery.

Click to expand...

Choose one of the packs to mount all the accessories (probably the most negative one) or mount them all on a separate board/box/whatever. The best solution will obviously depend on the physical design in the RV, so you need to do that first.

BiduleOhm said:

1 sec is enough for precharging the caps. Unless I missed something you can use that delay to precharge them




Choose one of the packs to mount all the accessories (probably the most negative one) or mount them all on a separate board/box/whatever. The best solution will obviously depend on the physical design in the RV, so you need to do that first.

Click to expand...

The caps on 2 5kva Quattros are impressively large. The 50 ohm resistor takes over 10 seconds to bring them up to about 50v (pack at 54). I'll get a new resistor like maybe 10 ohms (5 amps / 300 watts during precharge ... no problem for the contactors)

Capacity test completed. 258 Ah 2 cells fell short way before the others. I'll pull them out. Sigh .. deja-vu.

cinergi said:

The caps on 2 5kva Quattros are impressively large. The 50 ohm resistor takes over 10 seconds to bring them up to about 50v (pack at 54). I'll get a new resistor like maybe 10 ohms (5 amps / 300 watts during precharge ... no problem for the contactors)

Click to expand...

Yep, you'll need a low value resistor (I'd go as high current as what your relay can handle to shorten the time as most as you can). You also don't really need to bring them super close to the battery voltage so that will help

Forgot two things:

• I tested one of my highest cells using that tabletop tester: 278.5Ah. About time! And helps validate the tester being OK (versus just testing everything wrong at 270Ah; although that's already validated at the pack level)
• The Orion BMS beta firmware release notes says they're adding precharge relay control. Bam, problem solved.

Cinergi,

this is a very interesting build you are undertaking and the use of the Orion BMS, the relay logic for LVC and the can bus communication is great to see evolve.

I have also been watching your thread as I have four of the EVE 280 en route to me via ship from China and am somewhat apprehensive about getting a reasonably balance pack. I see you are having some challenges in this regard.

I was recommend this Lifepo4 lecture in another forum. It's very long and oriented for EV, but really gets into some of the chemistry and manufacturing details of Lifepo4 and at the end of the video they give some recommendations as to how to sort batteries of different capacities. Thought you or others may find parts of it useful. I did.

I'm spending a LOT of time with Andrew Ewert (Orion BMS) figuring out how to get effectively the same thing as a 3-stage charge so that the cells spend SOME time above a float voltage so that balancing and SOC drift correction work properly. It takes a lot of time to iterate through the changes and testing if they're effective. I'm not holding my breath, but this "float" feature does at least give me the ability to charge at a higher voltage until the pack is nearly full. Trying to charge at a float voltage of 54.4 is quite slow above ~30% SoC. Meanwhile, I'm topping off a couple of cells to swap in with the 2 low cells so I can do another capacity test on the pack, and I'll capacity test those two cells. I've upgraded the wiring on the cell tester so I'll lose less to wire resistance and run at 15-20 amps so I get faster results. This wire is incredibly flexible. I bought multiple gauges. Happy so far.

Marinepower said:

I was recommend this Lifepo4 lecture in another forum. It's very long and oriented for EV, but really gets into some of the chemistry and manufacturing details of Lifepo4 and at the end of the video they give some recommendations as to how to sort batteries of different capacities. Thought you or others may find parts of it useful. I did.

Click to expand...

Thanks. that was really useful LiFePO4 context. I watched the whole video but for those who just want the cell matching part, start at 1:04:40.

Marinepower said:

Cinergi,

this is a very interesting build you are undertaking and the use of the Orion BMS, the relay logic for LVC and the can bus communication is great to see evolve.

I have also been watching your thread as I have four of the EVE 280 en route to me via ship from China and am somewhat apprehensive about getting a reasonably balance pack. I see you are having some challenges in this regard.

I was recommend this Lifepo4 lecture in another forum. It's very long and oriented for EV, but really gets into some of the chemistry and manufacturing details of Lifepo4 and at the end of the video they give some recommendations as to how to sort batteries of different capacities. Thought you or others may find parts of it useful. I did.

Click to expand...

I just watched the video. I'm so glad I did. I don't know that it will change what I'm doing for this build but that was cool stuff!

cinergi said:

I just watched the video. I'm so glad I did. I don't know that it will change what I'm doing for this build but that was cool stuff!

Click to expand...

Well, I'm glad that you still found the video worth the time invested. I did warn all it is very long.

I found the information on topics such as why different manufacturing process yield high vs. low C rates; what chemically happens to a cell during overcharging, over heating, over discharging, etc. to really up my LiFePo4 understanding and to a certain degree inform the design of my build.

MP

OK, swapped out two cells and I got my best results so far. 268Ah according to Victron. I'm starting to think the Victron is off by maybe 8-10Ah ... the Orion counted 278. My tabletop tester is testing the strong cells around 278 as well... and the times when the pack tested to 260, the tabletop tester got 270. So now I'll begin getting capacity numbers for all my cells so I can try to match on capacity. I don't have a great way to match on IR but hopefully that won't matter with the C rates these cells will typically see.

Andrew at Orion doesn't think he can get me the charge curve I want ... I'm going to try 1-2 more things ... if that doesn't work, I'll still use the float feature so I can charge at a higher voltage until the pack is nearly full. I'll then use something to externally influence a higher-voltage balance cycle ... I guess I'll see how often I need to end up doing that based on those results (which won't be for many months).

I've been playing with the Grafana/Victron integration and also using it to help me tune the Orion to do what I want ...



In the second graph (Battery Voltage) the Red line represents an arbitrary line at 58v and the blue another arbitrary line at 54.4. The dotted yellow is the max voltage the BMS is telling the system to use and the green is the actual voltage. You can see the SoC graph above and the DC Power below to see that I started charging at 15:13, SoC hit 100% at 16:00, and how the voltage temporarily ramped up towards 56.5v and then down to around 54.4 where it will stay. I have some more tuning to do (I want the voltage to stay higher longer so balancing can be effective, and I want to visit the SoC drift correction voltage setpoints), but the "algorithm" is working out it seems! Finally!

I've tested a few more cells. Most are 270, 1 at 273, and 1 more at 280 (so 2 out of ~10 have hit 280). I'll post the final results when I'm done (early January most likely; 1 day per cell).