lcj

Sharing this here in its own thread.
This is my effort to replicate the Sunsynk power flow screen as a custom Home Assistant card as well as some Dashboard ideas I've been using. Improvements and suggestions welcome.
Features
Option to switch between two card styles: lite or full. Animated power flow based on positive/negative/zero sensor values with configurable dynamic speed. (Supports inverted battery, AUX and grid power). Dynamic battery image based on SOC (empty->low->medium->high). Grid connected status. Inverter status (standby, normal, self-test, alarm, fault). Configurable battery size and shutdown SOC to calculate and display remaining battery runtime based on current battery usage and system time slot setting i.e. SOC, Grid Charge. Can be toggled off. Daily Totals that can be toggled on or off. Hide all solar data if not installed or specify number of mppts in use. Set custom MPPT labels. “Use Timer” setting and “Energy Pattern” setting (Priority Load or Priority Battery) shown as dynamic icons with ability to hide if not required. If setup as switches can be toggled by clicking on the card Panel mode for bigger card AUX and Non-essential can be hidden from the full card or assigned configurable labels Customisable - Change colours and images Most entities can be clicked to show more-info dialog Optional data points include self sufficiency and ratio percentages, battery temperature, AC and DC temperature Display up to two non-essential, essential and AUX loads Display energy cost per kWh and solar sell status  You can find the latest version and details here
Documentation
Refer to https://slipx06.github.io/sunsynk-power-flow-card/index.html
Screenshots 
*Compact Version*


*Lite Version*


*Full Version*


 
I've also shared my Home Assistant Dashboard that works well for me. You can find the latest version and details here https://github.com/slipx06/Sunsynk-Home-Assistant-Dash

 

So the good thing about the Solar Assistant integration is it obfuscates the mqtt - so all I get are a set of sensors that i can toggle or values i can change / set

 
Then based on this its a simple automation to change a value. In this automation i charge the geyser from the grid for the early morning showers based on a set of conditions and then i set the house to run off inverter again once this is done 
alias: Geyser morning charge description: "" trigger: - platform: time at: "04:10:00" - platform: time_pattern enabled: true minutes: "5" - platform: template value_template: >- {{ timedelta(minutes=(state_attr("sensor.load_shedding_area_tshwane_3_garsfonteinext3", "starts_in"))) == timedelta(minutes<65) }} enabled: true condition: - condition: time after: "04:15:00" before: "07:15:00" weekday: - sun - mon - tue - wed - thu - fri - sat enabled: true - condition: and conditions: - condition: state entity_id: binary_sensor.grid_available state: "on" - condition: or conditions: - type: is_battery_level condition: device device_id: 117ead9d1883967 entity_id: sensor.battery_state_of_charge domain: sensor above: 30 - condition: or conditions: - condition: template value_template: >- {{ timedelta(minutes=(state_attr("sensor.load_shedding_area_tshwane_3_garsfonteinext3", "starts_in"))) == timedelta(minutes<65) }} - condition: time after: "04:15:00" before: "07:00:00" weekday: - sun - mon - tue - wed - thu - fri - sat enabled: false action: - device_id: 117ead9d1883967 domain: number entity_id: number.max_grid_charge_current type: set_value value: 2 - type: turn_off device_id: 117ead9d1883967 entity_id: switch.use_timer domain: switch - delay: hours: 0 minutes: 5 seconds: 0 milliseconds: 0 - type: turn_on device_id: f58645638cce28e4 entity_id: switch.geyser_switch_1 domain: switch - delay: hours: 2 minutes: 15 seconds: 0 milliseconds: 0 - type: turn_off device_id: f58645638cce28e4 entity_id: switch.geyser_switch_1 domain: switch - delay: hours: 0 minutes: 5 seconds: 0 milliseconds: 0 - type: turn_on device_id: 117ead9d1883967 entity_id: switch.use_timer domain: switch - delay: hours: 0 minutes: 0 seconds: 15 milliseconds: 0 - device_id: 117ead9d1883967 domain: number entity_id: number.max_grid_charge_current type: set_value value: 120 - device_id: 117ead9d1883967 domain: number entity_id: number.max_charge_current type: set_value value: 120 mode: single  

@tertiuscpt, happy to report that I got it working by getting the right RJ12 connector for the RS232 port.

Unfortunately not. I instead got a RS232 to usb cable and connected the integrated BMS to my solar-assistant raspberry pi instead, letting it read the battery and inverter separately.

It looks like I fouled myself. The unit continuous with the bad behavior.
I have now installed an additional array of 3 panels connected to the second inverter. Now that the sun has come back to more efficient elevation on a good sunny day the batteries got fully charged in the afternoon. When reaching floating voltage the inverters are switching the panels on and off in a well synchronized manner. But after a number of such cycles the inverter with the 9 PV panels remains with the panels switched off whereas the other one continuous working correctly. I thought that this is an inherent fault of that unit. So I swapped the panel arrays connection between the inverters. To my great deception the fault had also gone to the other unit. I conclude that this fault is an inherent one in the firmware (74 40) of the inverters. I wonder what is the criterium for that behavior with the large array and not with the small array. The large array (3s3p) outputs around 110V at full power, the small (3s, different type of panels) only around 90V. The floating voltage (27) is set to 53V, bulk charging (26) to 54V. I also noted that at the stage of reaching floating voltage the reading at the inverters is significantly higher than the reading on the batteries BMS whereas under normal conditions the readings are within measuring inaccuracy.
I have now cranked up the floating voltage setting to 53.5V and bulk to 54.5V. At first glance it looks like it behaves better. I will observe it closely.
Has anybody a clue?

(Updated info further down in the post)
I create a very simple Python script to interrogate the Pace BMS as used in some popular lithium battery brands. I've also packed this as a Home Assistant add-on, but can just as well be used as either a standalone script / within a docker container.
It reports the voltage of cells, temperatures, the pack's current, voltage, SOC, SOH, and a few other metrics.
Use at own risk. 
Either a serial link to the RS232 port or via IP (using some port server over TCP) will work.
Repository: https://github.com/Tertiush/bmspace
The script serves by needs atm, so any further developments can be done by the community.

UPDATE:
I've rewritten most of this script using the official PACE RS232 Protocol definition. The script now supports multiple packs!
I've only implemented informational messages / retrieving data, no commands.
Use at own risk!
Many new fields are now retrieved such as warnings, balancing data, status indications, etc.

Just wanted to say thank you for pulling this together. I was about to pull the trigger on an RS485-USB and Pi4 to setup HA following this guide GitHub - kellerza/sunsynk: Sunsynk Inverter Python library and Home Assistant OS Addon but looks like this will give me the basics for now, will contribute more as I learn.

Regarding (One), the serial strings of modern inverters can carry up to 500V, some even over that. It's far more efficient than low voltage, because thinner cable can be used. A DC arc is not an AC arc, very different, it is sustained and as a result a whole lot hotter, which is a source of fire. This over and above electrocution. And regarding low voltage / high amperage strings, if we believe Mike Holt, 35mA is all that it takes for the heart to reach fibrillation.
Regarding (2), I don't think that protection from an actual direct strike is called for, there's not much that one can do cheaply if that is the case. When people talk about lightning strike and roof / panels, they are mostly referring to the induced voltage that a nearby strike could cause.    
   

You need to first understand the Schlemburger method of earthing,as this is the way that you test earth resistance.With this method,a minimum of 2 earth spikes must be used.Secondly,single point earthing is a must on most residential TNC electrical connections.Thirdly,you must bond all the panels and then ground to the earth spike with 16mmsq earthing cable.Drpending on your lightning zone,you might be required to have an air termination rod and/or Type 1 lightning protection.

One cannot be sure that 1 spike is enough, depends on many things, the only way it to teste it with an earth tester, a good earth should be less than 6 ohms. 

Both. It's a 6.2 kW inverter, just with two load relays, that's all. You can set up the relays to come on at various times of the day, I assume.

1. Do I look at the specs at STC or NOCT?: Both, It depends on what you want to measure. This is a good guide: https://unboundsolar.com/blog/string-sizing-guide
2. Is the MPPT going to stop working above 425V? Keith Gough answered this question here. I have never been over that, so I cannot tell you from experience. It is clear that anything over 500V (Keith quotes 480V) will result in expensive repairs (blowing the DC-->DC converter), and general consensus is to keep things below 450V. Leave some headroom for voltage rising when there is no load (use VoC, not VmP), and cloud edge effect. What we do know is that the MPPT does clip the amps, Sunsynks with latest firmware are limit it to 13A per MPPT. Voltage is a little more difficult to control, so make very sure that you get your calculations right. On the 5.5K Sunsynk you basically have 2.75kW per MPPT to play with, that's quite generous.
3. What do they mean by "No. of Strings Per MPPT Tracker 1+1"? Can I put 2 strings per MPPT and use a combiner box or only 1 string per MPPT, as there is only 2 PV inputs on the inverter?
😯 Is that a genuine Sunsynk 5K or a knock-off? As per Sunsynk installer manual, there are 2 inputs for each MPPT:


The guide should answer all of your other questions, take a look, and draw up a spreadsheet with your panel specs. Bear in mind that STC and NOCT are calculated using pristine conditions in a lab. You already alluded to the fact that the panels are 1 year old, that's using good judgement, look at the degradation graph on the specs and use it, also use your own judgement when it comes to azimuth and pitch / tilt. Bear in mind that STC and NOCT are calculated using pristine conditions in a lab.
Wishing you the all the best with the install. Post some pics!
 

This is not documented by the manufacturer.
 I was surprised to find, when reading an older firmware, that it appears that the rated power of the inverter is what determines the limit before an overload fault comes up. In other words, you can't bypass any more than you can generate from the inverter via the battery alone. It's not clear if this applies to later models, but I suspect that they are running up pretty hard against the limits of what currents you can out through printed circuit tracks. Unless they change the design to use off-PCB relays, I don't see this changing.
So for a 6.2 kW rated model, it seems that you can only get 6.2 kW continuous through it, regardless of where the power comes from, and only about 10 seconds of 12.4 kW (100% overload), and perhaps 30 seconds of some intermediate overload (perhaps 30% overload, I can't recall the exact figure).