Pylontech x Sofar - battery in standby mode far too early

[Caesium]

I do find that the cell balancing is one thing that could be done better by the BMS.

The currents involved are so small that it could take a long time. I find that the cells are still balancing when the BMS tells the inverter to stop charging. As soon as the inverter gets that signal the voltage drops and so the cell differential drops and so does the balancing.

Because I can hold the voltage high using a setting I have on node red I can continue the balancing until I see it has stopped by watching the individual battery pack temperatures in batteryview. This is normally less than 10 minutes. If you leave it longer the packs just individually do into idle. I can't say that this is an improvement in the balancing over what the BMS actually does though. If the battery cells are way out, which mine haven't been other than on one brand new battery, the cell balancing will start earlier.

With the battery that did have cells way out I had to use my way of balancing at a constant voltage (52.5v) to stop a cell over voltage error.

Other than cost, pretty much this entire post is a good summary of why I'm going completely DIY.

I'm not completely convinced that the Pylontechs are as hands-off as most would hope - they need just as much tinkering with as a DIY stack in order to perform well sometimes. If the built-in BMS isn't doing its job properly out of the box and you have to hold its hand like this, I might as well build my own, save some money, learn a few things about how it all works, and be in a much better position to replace parts of it cheaply if anything fails.

It looks nigh-on impossible to replace a single cell in the Pylontechs for example. I opened one up recently and the cells are 3 packs of unidentifiable black blocks. I didn't take the tape off an unidentifiable black box but I rather doubt they were 18650s or anything off the shelf in there. And if the BMS failed in a Pylontech well.. its pretty much going to the recycling.

[dangermouse]

I'm thinking if it's another parameter (e.g. actual voltage) rather than SOC which prevents for further discharge. I've switched batteries around and am charging them to 100% for the last couple days. I've seen maybe 1 day when it allowed for 19% SOC but most of the time it stops at exactly 29%.
When I updated the software (3.36) and increased DOD to 90% (from 80%) it allowed for SOC around 15-16%, but that was only once, at the day of change. Annoyingly I can't get my diy cable to work to see what happens inside and I don't remember how it worked when temperatures where higher (now it's 2-5 in garage).

Getting back to the main subject...

I'm still seeing a similar effect to you. Yesterday I individually charged both packs to 70%, then connected them together to balance, then reconnected to inverter.

Just like your system, the batteries went to standby at 29%. I then reset them and they continued to discharge to 20%, which is the limit I've set in the Sofar.

The good result is that at 20% they just stayed in standby, there was no spurious charging like I had seen previously. Then overnight the inverter charged them back to full.

However this morning I inspected the batteries individually again and found that one is at 100% but the other is at 95%. Connected the 2nd battery to the inverter but it refuses to charge beyond 95. The inverter starts to supply charge for a couple of seconds and then stops. Then a short time later it charges for another few seconds, and stops. This cycle repeats endlessly.

No events in the inverter log.

[Tinbum]

The actual difference between 95 and 100% is minimal. The batteries stick at 89% for ages then rise rapidly from 90% to 100%. I'll post some graphs when I'm next on the computer.
If it takes a long time to get to 100% from 95 then it is probably trying to balance. Can you see any rise in reported temperature? I wouldn't realy expect the inverter to stop the charging if the battery isn't at 100%. Is it the battery going into idle?

[Oldgreybeard]

The actual difference between 95 and 100% is minimal. The batteries stick at 89% for ages then rise rapidly from 90% to 100%. I'll post some graphs when I'm next on the computer.
If it takes a long time to get to 100% from 95 then it is probably trying to balance. Can you see any rise in reported temperature? I wouldn't realy expect the inverter to stop the charging if the battery isn't at 100%. Is it the battery going into idle?

Exactly what I've seen. 89% seems like a bus stop, they sit there like that for ages whilst taking charge, then I turn my back on looking at the numbers and next thing I know they've jumped right up towards 100%. Never been sure what's going on, TBH, just assumed that there is some sort of glitch either in the Pylontech BMS data or in the Sofar. Doesn't seem to cause any issues, other than being a bit of an oddity.

I've also noticed that there's a few percent close to 100% where the batteries won't charge at all. If they've charged to 100%, then we've pulled some out so they've dropped to, say, 96%, vrey often they won't accept any more charge when the load goes away, even though there's PV available and going to the grid. Looks like there's some sort of buffer or hysteresis that may be intended to stop short cycling close to 100%, perhaps?

I've also solved the "only discharge to 20% SoC then shut off" mystery. Finger trouble by me, I'm afraid. at some point I must have gone into the battery menu and set the discharge depth to 80% rather than the maximum of 85% for the firmware version I have. No idea how, I can only assume I pressed the down button in error before exiting that menu and stepped it down by 5%. Looks like the cut off the other evening was just proper behaviour and not one of the battery packs signalling to shut down, so all is now well with the world (and the sun has come out and is charging the batteries!). Thanks for all the useful tips that are emerging from this thread, though.

[Tinbum]

[Oldgreybeard]

Brilliant illustration! Exactly what I've seen a few times. Is that data directly from the Pylontech BMS?

[Tinbum]

Not directly,as it's via MQQT, but it is the reading from the BMS. That's over 5 days.

[Oldgreybeard]

Not directly,as it's via MQQT, but it is the reading from the BMS. That's over 5 days.

Thanks again, that confirms that what the Sofar is repeating (over RS485/MQTT) is most probably just a relay of the Pylontech BMS data, which is useful to know.

[Tinbum]

Not directly,as it's via MQQT, but it is the reading from the BMS. That's over 5 days.

Thanks again, that confirms that what the Sofar is repeating (over RS485/MQTT) is most probably just a relay of the Pylontech BMS data, which is useful to know.

Yes, the sofar doesn't calculate it when connected to the battery BMS.

[dangermouse]

The actual difference between 95 and 100% is minimal. The batteries stick at 89% for ages then rise rapidly from 90% to 100%. I'll post some graphs when I'm next on the computer.
If it takes a long time to get to 100% from 95 then it is probably trying to balance. Can you see any rise in reported temperature? I wouldn't realy expect the inverter to stop the charging if the battery isn't at 100%. Is it the battery going into idle?

Thanks, useful info.

The battery that won't charge past 95 is indeed going to idle mode once it gets to 95%, so I don't think it's trying to balance its cells?

Battery temperature has been pretty constant at 13C, they are indoors so never get too cold.

I suspect it probably doesn't matter if the 2 batteries appear unbalanced at full charge, due to the tiny difference from 95 to 100. After a few more cycles I'll see how they compare at 50% or thereabouts.