jumper

Here's an image to illustrate my point. There is a difference between the bus voltage and battery voltage once the battery hits 100% and disconnects the charging circuit from the DC bus. It will only allow charge into the cells once the SOC hits 96%, which can be set in the BMS.

I'm actually running both... I have my 3kVA 24V running on LA which is where I first noticed the premature float bug, that inverter needs the float pushed up near bulk as it sometimes goes straight to float after bulk and sometimes it only charges to float and the firmware is ancient so has no equalization. I also have a 5KVA King which I'm running on a 5kW shoto with seplos BMS and on that one I use the bulk charging time set to 120mins at 56.4V and equalization every 5 days at 57V just to make sure it hits 100% as I have solar only and bad weather might mean the battery doesn't get fully charged... I want to avoid SOC drift. This is where I noticed that the BMS switches on and off when it's top balancing causing the current to go to 0 and the mppt will ramp down to around 200W while it does this. Once the BMS hits 100% it disconnects the cells from the DC bus and they settle, so the inverter voltage doesn't matter at that point anymore. It is possible to raise the voltage once the BMS is at 100%, but it is likely not actually charging the cells, merely raising the DC bus voltage on the BMS.

I charge to 56.4V (14.1V) because I have to in order to get the BMS to reach 100%. I tried lower voltages, but it would either not top balance or sit forever at 96%. I've seen a lot of problems come from not allowing the BMS to hit 100% where SOC drifts and batteries drop from 40% to 0% and shut off and need re-calibration. If my BMS was connected with inverter comms it would be charging to 57.8V (14.45V) as per the manufacturers setting, so I'm happy with 56.4V.

If you pull the cable out and the inverter doesn't complain then you are probably set to User battery mode which will run on volts. For BMS comms to work you need to change to PYL or LIB or LIC and have comms on RS485 to the battery. It is likely that it won't communicate, the axperts are a bit picky with batteries. This is also why the inverter has a different soc level to the battery, it is guessing based on your charging and cutoff voltages... the battery is correct.

That does present a bit of a problem. The only solution would be a shunt or a cheap clamp meter, but they aren't so cheap, maybe you can borrow one to do the measurements and then you can just keep the settings. With a shunt you could tie the readings into your dashboard and they will be a lot more accurate, but not sure it's worth it on such a small setup.

Nice clean UI you've made there, well done! I reckon with these black box type batteries you'll need to keep an eye on the charging current when it hits bulk voltage and see how long it takes to drop to a few amps and then set your equalization time based on that. The CV part of the graph should be about the same time always, no matter how low the battery was discharged, it's the CC phase that'll be longer and the inverter does that part just fine, it just needs a bit more time at CV which can be mimicked with bulk charging time, equalization or increasing the float voltage, depending on the inverter.

What I'm referring to is not the constant current part of the curve, I'm referring to the constant voltage part of the curve which is usually flat, the charger should stay at bulk voltage while the current drops. (source) As you can see from the above, the CV part of the charge curve is necessary for lfp cells. I've also observed on my battery that it needs to sit at bulk voltage for a while because internal cell balancing also happens during this stage and if it is terminated early the BMS doesn't report 100% SOC. On my battery some cells need to be above 3.5V for balancing to even start, so the CV voltage needs to be high enough for that. Are you getting your SOC reading from the BMS or inverter? If your BMS reports 100% then you're fine, I just know I needed to extend the CV part of the curve to hit 100%. This image is from an axpert manual which shows the algo should have the constant voltage part of the charge curve as the current falls:

Sorry, what I meant was the shoto implements the pylontech protocol over CAN, but the inverter uses RS485 for the PYL setting. Hopefully the new shoto setting will be using the pylontech CAN protocol.

Looks like your pump is trying to start too often. Perhaps try to change to frequency of the starts to only 1 per minute. P.S. I remember reading in a Franklin submersible motor manual a while back that they suggest starting the motor only once every 4 minutes.

Unfortunately the shoto won't communicate directly with your inverter as it only does CAN comms, but the inverter only does RS485. You will need to run in user mode unless you get something like solar assistant. The above settings will work fine.

Have you tried setting 16 on SbL UdC? That should force only the solar to charge the battery to test if it works. I take it the solar works to support the loads?

You shouldn't be able to break anything with incorrect addressing, it happens all the time. Imo the worse that will happen is the inverter will think you have only 1 battery instead of 2. Dip addresses are sometimes used to tell the BMS to use either CAN or RS485 comms, but mostly it is to give each battery a unique address on the comms network so they can communicate with each other and the master with the inverter.

I would use what the battery manual suggests as they will know how their addressing system works on the BMS. Your battery seems a little different as it is flipped from what is usually used.

This short spike to bulk voltage and almost instant drop to float voltage is what I was referring to above, there is no absorption stage in the charge curve. To charge properly, the battery should spend some time at the bulk voltage so it has time to do internal cell balancing in order to reach 100% SOC. To overcome this on my 5kVA inverter I use the "Bulk Charging Time" setting (32) at 120mins to get the absorption part of the curve. If there is no setting 32, then one can use the Equalization settings (33-39) set to bulk voltage to do the same thing... if neither of these settings exist in the firmware then one needs to push up the float voltage to just below bulk voltage to get a proper charge... I have to do this on my 3kVA inverter as it's really old.

Try changing your float voltage to 28V and see if you reach it that way, if you do then it's the algo going to float early. Some BMS have what's called intermittent charging when they get full and start balancing, the BMS switches off and on and this fools the inverter into thinking the battery is full because the current goes to 0.

Although not as common, they do exist. Here is one with 1500W mppt: https://solarequipment.co.za/product/kodak-3kw-24v-solar-off-grid-plus-inverter-with-ups/ I only point this out because I have a 3kw axpert (not Kodak brand though) which only takes 600W PV and I've seen Synapse branded models that take 1000W PV... I just like to be absolutely sure as going with incorrect pv specs can be very unforgiving.

What are the mppt specs of the new inverter? It could be that 3 panels in series is too high or too low voltage for the new inverter. Also, what are the specs of the panels?

You should post a pic of the sticker on the side of the inverter, these 3kW inverters come with varying specs, some only allow 600W of panels, some 1200W, some more, depending on the specific model.

I don't think it will communicate at all, most likely the inverter will give a comms error

Unfortunately this won't work because if you use CAN comms the inverter only supports this on the Weco protocol which doesn't work with the batteries. All the other protocols force RS485 on the inverter. Your best solution would probably be solar assistant as that will give control of the batteries and inverter via the interface.

It looks like you PV voltages have dropped massively. I would think this could be due to faulty panels in series strings or a damaged mppt.

Do you have communication between the battery and inverter? If you do and the incorrect battery is selected it can cause issues. I would try disconnecting the battery communication cable and manually set the inverter to charge to 56V and float at 55V to see if you can bring the battery back to life and fully charge.

Sounds like communication issues causing SOC drift, probably because the battery is not fully charging and balancing. I would suggest disconnecting the comms cable between battery and inverter and switching the inverter to a USER mode where you can enter the charge voltages and charge the battery to 53V and see if that can get you back to a proper 100% SOC.

What I find even more devious is that, as we have already established, they are counting every 20% instead of 80% which means they are counting 4x faster, not 2x faster, so that should read "4500 cycles (18000 Cycles on BMS display)".

Looks like the same BMS as the SVolt battery uses... I'm not sure what it is, but maybe that'll help in your search.