Hi all.
Have noticed that when grid feed-in is activated the putput is up and down constantly, doesn't settle at a certain level. See video here: https://youtu.be/954Oy9VKfuE
The setup: 
Grid and sun are stable so they can be taken out the equasion.
Any input welcome.
Hi @szdan,
This behaviour has been corrected with a firmware update from Pytes. I have just got in touch with them asked for advice on the next steps. Please wait for a follow up.
Hello,for this question， you can contact us. Our after-sales email is: ess_support@pytesgroup.com
Hi @Guy Stewart (Victron Community Manager) and @audrey.
The issue has been resolved for the first two customers with upgraded PYTES firmware. Makes a huge difference feed-in wise!
Great to have you guys here helping out!
@Audrey @szdan @Guy Stewart (Victron Community Manager) is this the latest firmware SPBMS16SRP2205V1.5.23.C16.bin this happens only with a MPPT connected?
i'm pre-charging now for our customer 6 packs - i'm getting only 54,4V CVL - after 2h it's was raising to 54.8V
now after some hours the CVL went up to 55.9V but the Multiplus-II never changed to Bulk Charging now... looks like a new charge logic. 
Yes, that is the firmware version I also got from Pytes technical support. To be honest haven't looked at these. Was focused on monitoring grid feed-in which was very unstable, since the upgrade it is back to normal now. Values are consistent and stable.
Let me get the same stats from one of the customers and post below. Maybe it helps:
I am not familiar with exactly what is going on there.
But generally the battery will give a lower Charge Voltage Limit (CVL) when it has a reason that it wants to restrict the speed of the charge to a particular cell. Once the battery balance is improved, the CVL will increase.
This can cause a situation like you describe where the unit is already in Absorption (or float) mode because it reached the 'target' for bulk stage to end, and then doesn't restart the charge cycle automatically.
In this case you will need to manually do this, fastest way is power cycling the charge source (your MultiPlus-II) by turning it off and then on again. This should begin a new charging cycle starting with Bulk again.
This would also correct itself automatically eventually, but that might take a day instead of immediately.
After a discharge test from 100 - 5% i started a charge test to 30% -
The System goes immediately into Absorption.
Only 1 MP-II 48/5000 with 16A input limit on 6 x 48100R.
What a bad charging algorithm.
I think you can do that dynamic CVL Voltage on + 90% SoC but not before!
Balancing on 3.300 Volt/Cell with 0,03V diff. makes also no sense - at 3,45V it is enough.
I'm precharging and balancing all battery's for customers, never had any high voltage alarms - took around 48h!
Good useful info, thanks I've forwarded it on to my contact at Pytes and asked them to please investigate.
I agree with your conclusions from looking at data.
Had some issues in the beginning. 23% 53.07V makes no sense.
I have restricted charging to 52.4 as 53.x can trigger high voltage in my system.
Replugging the coms cables may help - looks like an imbalance somewhere
To support balancing restricting of charge current for one or two cycles may help as well giving more time for internal balancing.
What is the reported min max cell drift?
