RikusW

Hi
This site seems to be run by a Growatt employee, so for those capable of board level repairs here is the service manual:
https://www.amosplanet.org/spf-3500-5000es-09fault/
https://www.amosplanet.org/spf-3500-5000es-spf-5000tl-hvm-wpv-troubleshooting/

Or indeed a few thousand amps. Inrush current is hard on the contactor or fuse contact that makes the initial connection. If you're making the final connection by hand, there is the danger of flash injury, burns from copper snot, damage to contacts or fuse casings, and the cost of new underpants 🩲🥴
I use high pulse power rated aluminium clad resistors, and a circuit that automatically shorts it out after a suitable time, so I never get lazy. Around 3.3 Ω and 50 W continuous / 1250 W pulse is good for 50 V systems. That limits the in-rush to some 15-18 A, depending on the battery voltage, and will pre-charge two 5 kVA inverters in less than a second.
I note that some packaged LFP batteries (e.g. Pylontech) may automatically limit the current to a manageable value for a short period of time (e.g. 1C for the battery), thereby automatically performing a reasonably graceful pre-charge. For a single US2000 Pylontech, that's about 50 A. I have no idea how hard that is on the internal MOSFETs, which have to dissipate half of the energy stored in the capacitors. If you have a lot of Pylontechs in parallel, as you eventually are likely to, the current limit will still be very harsh on contactors, fuse holders, etc.

Except if the coil is already energized, the contacts closed, and grid fails, you could still be powering the relay coil from the inverter through the contacts. There is a specialized anti-islanding device that is a bit more complicated for this sort of application:
https://www.victronenergy.com/accessories/ziehl-voltage-frequency-sensitive-relay-ufr1001e
Aside from that, inverters that comply to IEC 62109 (or perhaps VDE-AR-N-4105, not sure) and NRS097 should have ample anti-islanding redundancy that you don't need something like this.

The Growatt manual recommend USE 2 for lithium voltage based setup, equalization needs to be disabled.. The lithium protocol for Pylontech is L52 for the CAN connection, you also need a 120 Ohm terminator on the unused CAN connector on the battery.

Not for wind turbines, I think. My understanding is that the separate wind turbine controller would connect directly to the battery, and it would manage charging the battery via wind. The Axpert inverters would manage charging via their PV panels.

In December 2020, I unfortunately fell victim to a lightning surge from the grid. Various items in my house was impacted. Thankfully my Axpert inverters(x2) survived with no issues, although the Raspberry Pi that runs ICC was effected, Pi booted but USB and network ports dead. My Pylontech bank(5 x US2000 Plus) was impacted as well. The communication ports(console, RS485, CAN etc) was impacted on all 5 Pylontech modules. Fortunately they were still charging and discharging normally but i could not get any info from them. I sent them back to the supplier to get a full report. The report stated that the CMU boards in all 5 modules had to be replaced and Pylontech warranty does not cover surges, so I had to claim from insurance. Got all approvals and all boards were replaced. I commissioned the entire bank after the repairs was done. After a few days I was noticing that 2 modules cell voltages was going over 3.65v during charge and these units were throwing High Voltage and Over voltage errors in the logs. When this happens, it stopped accepting any current and waits for those effected cells voltages to drop below 3.5v before accepting a charge again. This happened over a few days and the SOH counts were increasing for these 2 modules. I decided to pull them out of the bank and to my surprise they were swollen! I was actually was shocked(no pun) when I saw this. I returned those 2 modules back to the supplier and there new report concluded that the cells are damaged from the surge and irrepairable. Their recommendation was to replace those 2 modules, which i did with a US3000C and a US2000. 
I collected the damaged modules and in the spirit of science, being inquisitive and sharing knowledge I decided to salvage some parts before disposing the cells. I took the following pics:
Bottom cover off, we can see 3 x 16v battery packs(notice them being swollen), connected in series. Power Management Unit(PMU) board on the bottom left and Communication Management Unit(CMU) on the bottom right. 

 
PMU Top: 

 
PMU Bottom, notice all the MOSFETS here.... this side of the board is attached to a heatsink situated on the front right top side of an installed module: 



 
CMU Top:


 
CMU Bottom:


 
LED Strip:

 
Internal power cable size 8AWG - 10mm2:


 
Battery Packs:





 
If you counted, there are 10 pouches per pack. A total of 30 pouches in a single US2000 Plus module, equates to 1.6v per pouch. BatteryView software shows 15 cells, so i am guessing that 2 pouches make up a cell of 3.2v.
Enjoy the pics. 

In User 2 mode voltage based setup it stops charging once the bulk voltage is reached, no constant voltage charge, then restarts charging multiple times throughout the day and once solar output is low enough not to reach the bulk voltage remain charging until the end of the day.

I have found the bms modbus protocol for the narada bms and am busy writing some code to extract the data and use it on my stats portal.
As soon as I have it working I will share it with the community.
I am so over all these guys keeping the code "secret"
Data and full control must be free and easy to extract especially given the high prices we pay for the equipment