What battery do you have?
Can you also provide (along with the Battery SOC & Battery Voltage and Current graphs):
My thinking would be that there’s a cell with a high voltage that’s causing charging to stop or it’s similar to the OP where the BMS drops the voltage when the battery is full which will drop the PV charger.
Continuing from my original post. My system was set to “Keep batteries charged” around 17:00 today.
This is a bit weird to me. It is as if a control loop somewhere has suddenly become too slow.
As an experiment I will now set the ESS to “Optimized without BatteryLife” for a while, and then set it back to “Keep batteries charged” again.
Seems your batt does the same as mine, in the 18 Cell bank thread. Post 263.
Mine charge-discharge/charge-discharge with small wattages. In your case the BMS actively stops charging, and discharges the battery. In my case, the MPPTs still carry on.
Just what I assume, deduce. Could be totally wrong.
If that is the right deduction, it is good for top balancing. Your volts exceed 3.60v. So it is very good.
Idea if you want … set the max charge volts to 3.45v. See what it does then.
Ok, I found this oddity. The Victron (Multiplus and MPPT) reported voltages have drifted higher over time, from what is actually measured on the DC bus. The battery voltage is reported accurately.
I have gone to the system with my Brymen multimeter, and measured all over the DC bus - the measured voltages are the same at the MPPT, the bus bar and at the battery. Only the battery is reporting it accurately (within 0.1 V). The Victron components are all reporting a higher voltage (0.8 to 1 V) than what is actually on the DC bus.
This is from today:
This is from two months ago:
Is it possible that the system starts oscillating now, because it reacts too late / too early to the system voltage? I am not sure how this works - does the DC bus voltage affect this oscillating behaviour, or is it only using the BMS? @plonkster, do you perhaps have any thoughts on this?
I’ve heard of this before. Measured voltages increasing over time. I don’t know more than what I just wrote, that I heard about it. Don’t know if it is common place at all, and what is done about it.
I know it is possible to recalibrate the Multi, with the right software. Have no idea how to even get that.
Basically, sounds like you need support, but not the kind I can give. Ask on community, or from your reseller?
Thanks - I will ask on the Victron community site as well, and perhaps contact Current Automation too.
Question: will the measured voltage difference have an effect on the oscillation behaviour I am seeing at the 100 % mark?
I have switched off the (only) MPPT in the system to make the fault finding simpler.
The MPII appears to be measuring the DC voltage about 0.8V higher than what it is actually, and also more importantly, what the battery is measuring. I have confirmed all over the system that the BMS (battery) is measuring correctly. It also appears that the MPPT (when it is on) is just using the MPII’s voltage reading and not its own.
I observe the following, when in ESS and “Keep batteries charged”, and the battery is not full (less than 100 %):
The oscillation, where the voltage is lowered and raised multiple times, may in fact be less if it isn’t overshooting.
Not sure what you meant here - are you saying that fixing the voltage difference will make it worse or better 
I’m saying that if you don’t constantly go higher than the battery wants, the battery doesn’t have to lower the charge voltage to stop the charging, and you won’t get that sawtooth graph, or at least, it will have way less bounces in it.
I wanted to suggest switching on SVS (shared voltage sense), but then I remembered this is not possible with BYD batteries.
You can test it so long. Use that DVCC charge voltage limit feature and limit it to 0.8V less than what you see the battery ask for at the highest point. See if it squirms less 
Thanks - I will try with 0.8V less out of morbid curiosity.
I actually think it is the other way around in that in lets the battery V go too low since it (MPII) measures too high if that makes sense? So it does not overshoot, and it rather let it drop too low - then charge up, then let i drop too low etc.
I assume it is the MPII voltage that is used for DVCC yeah?
I also asked about the “voltage drift” here: Multiplus II 48/3000 - measured voltage drifting up over the last few months - Victron Community
I had a call with @warwickchapman from Sensible Solar - they are awesome by the way
He suggested disabling vsense in the MultiPlus, which does not make much sense since I did not even have a vsense cable installed. But what do you know - this fixes the inaccurate voltage reading and the repeated charging / discharging when the battery is 100 % full.
I did a bit of searching, and also found this thread: https://community.victronenergy.com/questions/196650/why-is-mppt-reporting-nearly-3v-higher-than-pylont.html
Guy Stewart made a post in that thread that was related:
This is my system diagnostics page before disabling vsense - you can see the “sense” voltage is far from the terminal voltage.
After disabling vsense:
Bus voltages after disabling vsense:
The oscillation of charge / discharge when the battery is at 100 % has now completely stopped.
It appears there is something wrong with the vsense hardware / algorithm in my MPII, and this affects the voltage reported by the MPII even if you don’t have it connected. It does not matter too much since I have a LiFEPO4 battery with BMS though.
Nice… I was unaware of that new option
So, I switched it off to, the Disable VSense.
After one charge, so preliminary, it appears my batt now does not go as “high” as before the setting was off.
Will monitor going forward.
In this whole exercise today, I thought to also upgrade the firmware, maybe my inverter is also one with the fan noise reduction … low and behold, whisper quiet now. To test it I threw 108amps at it, and the fan was still quieter, “deeper sounding”, more “controlled”, than before this firmware upgrade.
My PCs fans are now the loudest. 
So is this the setting on the GX or must I use VEConfig to double check…
Have to use VE.Config, it is on the configuration file of the inverter as per Alberts post above.
It’s worth checking your installation’s diagnostics page before considering making this change: VRM Portal - Victron Energy (change the xxxx to your installation).
Compare the VE.Bus Terminal voltage against the VE.Bus sense voltage.
Also - note that this is marked as “for diagnostic purposes” in veconfigure… change stuff at your own risk.
