diy solar

diy solar

Solved! SolarAssistant with 6x EG4-6500EX and 9x EG4-LLv2, individual connections

You might want to shut 4 of them off during the night. This is because there will be 0 PV power and the EG4 inverter draws about 100w.
6 EG4's will be a 600+ watt draw off the batteries. 2 EG4's will only draw 200 watts. Probably not an issue in the high sun months but definitely an issue in the winter with short days and lots of clouds. I have a system similar to yours except only 4 EG4 inverters but I have 12 batteries where you have 9. I thought that 2 EG4 inverters would not be enough, boy was I wrong. So I now have inverter #3 that I use only to help charge the batteries faster and provide more PV power on low sun days. Inverter #4 is strictly a spare. By the way what do you have for solar panels ? You will need 12 strings for 6 inverters. I have 6 full strings. You can see what I have in my signature.
I've thought about the idea to shut down some of the inverters, depending of day/night time and load/PV values. Because I've integrated the system via SolarAssistant and MQTT bridge to my OpenHAB smart home system, I would like to control this automatically via OpenHAB rules. Unfortunately, SA does not yet provide MQTT messages to shutdown (and maybe startup out of standby, if possible) the inverters via command. The modbus protocol of the inverters have these capabilities and SA uses them (at least the shutdown to standby) by itself in one of it's features (in the Power "Battery shutdown protection" section). I've asked Pierre from SA if he may extend the MQTT messages to be able to remote shutdown individual inverters (and maybe also startup). But I have no answer so far.

I need about 1 battery charge per night to just feed the idle consumption of 6 inverters... here in the mohave county, it's not very often cloudy... :cool: I will have 8 PV strings, all between 200-300VDC (1.8kW - 2.8kW) with a total of 18.4kWp. 2 of the inverters will be connected with both MPPT entries and the others only with one MPPT - that's the plan.
 
I've thought about the idea to shut down some of the inverters, depending of day/night time and load/PV values. Because I've integrated the system via SolarAssistant and MQTT bridge to my OpenHAB smart home system, I would like to control this automatically via OpenHAB rules. Unfortunately, SA does not yet provide MQTT messages to shutdown (and maybe startup out of standby, if possible) the inverters via command. The modbus protocol of the inverters have these capabilities and SA uses them (at least the shutdown to standby) by itself in one of it's features (in the Power "Battery shutdown protection" section). I've asked Pierre from SA if he may extend the MQTT messages to be able to remote shutdown individual inverters (and maybe also startup). But I have no answer so far.

I need about 1 battery charge per night to just feed the idle consumption of 6 inverters... here in the mohave county, it's not very often cloudy... :cool: I will have 8 PV strings, all between 200-300VDC (1.8kW - 2.8kW) with a total of 18.4kWp. 2 of the inverters will be connected with both MPPT entries and the others only with one MPPT - that's the plan.
I would have 4 inverters with 2 strings each. The other 2 inverters are spares. Believe me with 4 inverters you will have no issues regarding power availability. Wiring 4 inverters with only 1 string each does not make sense. It just consumes more power when the sun goes down.
 
All great information. I have the Sabrent usb to db9 (pl2303) sent by SS but cannot seem to get any communication from the inverters.
I am getting battery data. I have 2 EG4-6500EX-48's and neither seem to commucate. Tried Watchpower on laptop and same issue, so not a solar assistant issue. Is the com port 9600,8,n,1? I do not see any comm related settings in the inverter.
 
All great information. I have the Sabrent usb to db9 (pl2303) sent by SS but cannot seem to get any communication from the inverters.
I am getting battery data. I have 2 EG4-6500EX-48's and neither seem to commucate. Tried Watchpower on laptop and same issue, so not a solar assistant issue. Is the com port 9600,8,n,1? I do not see any comm related settings in the inverter.
There's no settings on the inverter.

Since you're not seeing any communication for either device, there are 2 different ports that are both an ethernet connection (RJ45) on the bottom of the inverter display and you need to make sure you're using the one to the far right for computer connections. The one just to the left of that port is for BMS communication. Might be something to check. I know I've plugged into the wrong port several times.
 
There's no settings on the inverter.

Since you're not seeing any communication for either device, there are 2 different ports that are both an ethernet connection (RJ45) on the bottom of the inverter display and you need to make sure you're using the one to the far right for computer connections. The one just to the left of that port is for BMS communication. Might be something to check. I know I've plugged into the wrong port several times.
I think the Sabrent adapters that were sent by SS are not working for me. I just got a micro USB cable and it works.
 
Any light flickering at all?
I'd like to post a short update.
  • Still no solar panels installed (still way too hot here)
  • 2-pole transfer switch (common neutral) installed
  • AC-in connected to grid
  • 24,000 BTU mini split in garage installed to reduce the ambient temperature (at least a bit down from around 100 Fahrenheit)
  • Updated the 6 EG4's to the newest firmware versions (DSP and MCU)
Last Sunday we've had a 4 hour power outage in the morning. This was the right situation for a longer running first real life test for the system.
About 6 weeks ago, I've charged to batteries from grid (via AC-in) to be prepared for such a situation to be able to use the system just as an UPS.

Everything worked without any issues!
No light flickering (most lights in the house are LED's).

My house base load is about 600W. By intention, I have not reduced the load, instead I've turned on every consumer which I thought about (60,000 BTU central heat pump, 24,000 BTU mini split in garage, opened warm water tap to let the water heater working, let the pool pump working, told my wife to switch on electrical oven and cooktop, turn on microwave, washer, dryer, etc.).

Peak usage was about 25kW and >20kW for more than one hour. The other 3 hours it was typical use (base load, pool pump and from time to time the heat pumps). The one battery rack with 5 batteries delivered about 280-300A max. and the other rack with 4 batteries about 200A-250A. SO the load was nice balanced between the batteries. Also the load difference between the inverters were only in the range of <500W.

I've checked all battery connections with a thermal camera and found only one not perfect tightened bolt, that's it.

The temperature controlled PWM fan modding in all 6 EG4's worked without issues. But, because I still don't have the concealed mini split installed to inject cold air into the inverters, the fans speed up really fast to their maximum speed (the 24,000 BTU mini split is not able to cool the RV garage ambient to very low temperatures - still around 92-95 F). But this was expected behavior.

So, overall I'm very happy with the performance and the load balancing of the system - as long as a 4 hour test will be able to proof. But this was the first test and it worked.

I've recharged the batteries again from gird to be prepared for the next power outage!
 
I want to share a short update with my findings about some additional tests (idle) with the EG4 6500EX-48 and SolarAssistant and integration with OpenHAB rules.

Environment:
  • 6x EG4-6500EX-48 in split phase (3 for each phase L1 and L2), all with newest firmware DSP: 79.71 and MCU: 61.13
  • 9x EG4-LL (V2) batteries (4-Pin-DIP-Switch version) with newest firmware Z01T13 is 2023-03-14
  • NO communication between inverter and BMS (program 5 set to “USE” in all 6 AIO’s)
  • Inverters and batteries are all connected individually via separate cables to SolarAssistant be able to get all data of all components as detailed as possible (see: diagram and post)
  • SolarAssistant’s MQTT broker is enabled and integrated in my existing OpenHAB smart home solution
  • Solar controller circuit which contains the SolarAssistant Raspberry Pi and a couple of relays and WiFi relays to control AC-in contactors, Battery contactors and PVRSS Tigo system. It also implements an emergency system shutdown by intention via emergency buttons or via smoke detectors or by initiate remotely by intention (see: post)
  • Same charging current is programmed in all 6 AIO’s (same values in programs 02 and 11)
  • All AIO’s set to “SBU” in program 01 and to “CSO” in program 16
  • All Tests are done without PV solar modules connected (still too hot outside and no spare time to install the PV panels) and without AC load!

Observations:
  • Different power consumption if grid AC-in is connected or not connected
    • Each inverter consumes about 93W idle power independend of working mode
    • If grid is not connected the battery consumption is about 560W for 6 AIO’s
    • if grid is connected, the battery consumption is about 270W and the AC-in consumption is about 290W for all 6 AIO’s.
      • Finding 1: If grid is connected, each AIO consumes about 50W from grid, even grid is not yet used for battery charging. I guess this is to prevent grid back feed – even if the inverter is not able to feed back energy – which is strange!
  • Finding 2: If AIO’s trigger program 12 (point back to utility), all six inverters are switching to bypass mode (you can hear the relays inside the AIO click at the same time) and all 9 batteries will be evenly charged from grid. If program 13 (point back to battery) is triggered, all six inverters are switching back from bypass mode to battery mode. SolarAssistant shows “Solar/Battery mode” in battery mode and “Grid mode (SBU)” if the inverters are in grid charging (bypass) mode triggered via program 12.
  • Finding 3: If all AIO’s are switched off (round buttons popped out – not pressed), they will automatically be activated when AC-in is connected and start charging the batteries. Even SolarAssistant is able to communicate with the AIO’s in this mode and SolarAssistant shows this mode as “Standby mode” (it also shows AIO’s individual temperatures, grid voltage and frequency, bus voltage, battery voltage, etc. in standby mode).
  • Finding 4: This standby mode charging (see Finding 4) does not stop if the voltage setting in program 13 has been reached, instead it charges until the batteries are full!
  • Finding 5: As soon as the AIO switched on again (round buttons pressed), while AC-in is still connected as described in Finding 5, the charging stops if voltage in setting 13 is or has been already reached!
  • Finding 6: SolarAssistant send the following MQTT messages when AIO modes are changed
    • solar_assistant/inverter_<inverter#>/device_mode/state Solar/Battery
    • solar_assistant/inverter_<inverter#>/device_mode/state Grid
    • solar_assistant/inverter_<inverter#>/device_mode/state Standby
  • Finding 7: I have done a pretty good job with equal length battery wires and same torque for all battery connections (I’ve double checked all connections with a thermal camera also to verify that I don’t have bad connections). All batteries draw the same current in discharging and charging mode (I’ve already checked this before under a heavy load test, also)!
  • Finding 8: Battery discharging diagrams at 0.01C with all 6 inverters at idle without load over about 3.5 days:
    • All 6 AIO charts are nearly exact overlaying each other, so I only show for inverter #1
    • The battery voltage has two nearly horizonal sections (100%-70% SOC and 63%-35% SOC)
    • The first inverter of the three L1 phase AIO and the first inverter of the three L2 phase AIO will always be about 5-6 degree Celsius hotter than the other 4 inverters – independent if AIO’s are in battery or grid mode or standby, if charging or discharging!

BatteryStateOfCharge0.01CDischarge.pngBatteryVoltage0.01CDischarge.png

BUSVoltage0.01CDischarge.pngBatteryPower0.01CDischarge.png
InverterTemperatureSeparate0.01CDischarge.png
Ambient temperatures-data 0.01C discharge.png


Grid charging situation in detail:
  • The batteries should only be charged from grid if they are nearly empty (about 7%) and grid charging should stop if they are about 25% charged again.
    • The inverters program 12 (point back to utility) is set to 50V (average of 3.125V per cell)
    • The inverters program 13 (point back to battery) is set to 52V (average of 3.25V per cell)
    • To reduce grid energy consumption, the grid should only be connected to the AIO’s if required for (emergency) charging the batteries (e.g. in the early moring) and disconnect the grid after the batteries are charged.
    • This is achieved by two AC-in contactors for L1 and L2 which are closed if the first CellVoltageLowest of all 144 cells of all 9 batteries will drop below 3.000V – this is to prevent a BMS battery shutdown, in case one of the cells will have a very low voltage compared to the average battery voltage (The SolarAssistant MQTT message for these values are “solar_assistant/battery_<battery#>/cell_voltage_-_lowest/state <cellVoltage>” ). The AIO programs 12 and 13 using the average battery voltage which is seen by each of the 6 inverters. Because there is no AIO to BMS communication, the AIO’s just use one voltage value for all batteries as an average. The values are selected in a way, that the AC-in contactors are definitively closed, before the inverter activates the “point back to utility” mode!
    • The AC-in contactors are remotely controlled by OpenHAB rules which gets all inverter and battery data via MQTT from SolarAssistant.
      • Rule 1: The AC-in contactor will be activated if the first battery cell drops below 3.000V while the inverter is in “Solar/Battery” mode.
      • Rule 2: The AC-in contactor will be de-activated if the master inverter will change it’s mode from “Grid” to “Solar/Battery” to save grid energy consumption.
    • The contactors are activated via MQTT integrated “Shelly Plus 1” dry contact relays which are running on Tasmota firmware. The Shelly activates the two AC-in contactors which are 110Amp 3-pole version.

Problems:
  • If the AIO’s are running in mode “Solar/Battery” and as soon as the AC-in will get connected to the grid (to prepare grid charging), it sometimes happens, that one of the AIO’s (mostly the master AIO, but not always) will show an Error F60 (manual tells it’s either “power feedback protection” OR “battery not allowed to charge/discharge”, depending where you’re looking).

    I don’t know why this happens! I will do some more testing to dig deeper into this problem.
 
Last edited:
Just chiming in as a datapoint; Ive just installed the EG4 6500 (Replacing a Growatt SPF LVM-ES). The GROWATT Indeed flicker My LED lights, so far; this EG4 has NOt flickered any lights......... FWIW

~RandomPlanet
I just joined this board and thought I would add my own feedback after reading this thread for anyone considering the 6500ex inverters. I have yet to tackle the SA integration, but that is next and I appreciate all of this information, very helpful.

I have 4 x 6500ex, 24 x EG4 batteries and 14,400 watts of PV. I am running a fully off-grid system with my largest draws being 4 x 24,000 btu mini-splits that can draw up to 4,000 watts per mini split. I too was worried about light flickering and load inconsistencies. But, I can tell you that I am very impressed with the 6500ex output. I have not yet experienced any light flickering, even on my 1000w workshop LEDs. I even added dimming on several of my lights, which tends to be the achilles heal of many of these systems and the dimming capabilities seem to be equal to the grid system in my main house.

I have only been running this system for a few months, but so far, so good. My only complaint is the noise level, they are loud, so I am interested in any tuning results made here.

Lastly, I have a Kohler 20Kw generator waiting for my 500 gallon propane tank to get filled. I am very interested to see how the generator performance works for poor weather days.
 
I just joined this board and thought I would add my own feedback after reading this thread for anyone considering the 6500ex inverters. I have yet to tackle the SA integration, but that is next and I appreciate all of this information, very helpful.

I have 4 x 6500ex, 24 x EG4 batteries and 14,400 watts of PV. I am running a fully off-grid system with my largest draws being 4 x 24,000 btu mini-splits that can draw up to 4,000 watts per mini split. I too was worried about light flickering and load inconsistencies. But, I can tell you that I am very impressed with the 6500ex output. I have not yet experienced any light flickering, even on my 1000w workshop LEDs. I even added dimming on several of my lights, which tends to be the achilles heal of many of these systems and the dimming capabilities seem to be equal to the grid system in my main house.

I have only been running this system for a few months, but so far, so good. My only complaint is the noise level, they are loud, so I am interested in any tuning results made here.

Lastly, I have a Kohler 20Kw generator waiting for my 500 gallon propane tank to get filled. I am very interested to see how the generator performance works for poor weather days.
Thats a big system :) ... I have modified my single 6500ex to deal with the noise and can report major improvements..... Have you seen this thread ?


I added some temperature fan control sensors AND reversed the airflow direction of the fans .... Everything is working fantastic.
Keep us posted, cheers

~RandomPlanet
 
It has been several months since my last post in this thread, but I thought it would be worthwhile to provide an update.

While I was initially impressed with my setup of 4x 6500ex, I now have to backtrack that statement. In the past 4 months, I did get my generator functioning, but the setup through the 6500's was too inconsistent. I understand this is a known issue, but the generator I am using was well within the guidelines suggested by SS.

That said, I changed the configuration to route the generator through 3 Chargeverters, which has worked well.

However, I have now encountered several other issues with the 4x 6500ex. I have added loads to the system (EV charging and others), and the inverters are too unreliable. My problem is that they cannot run reliably within their stated parameters. If I run them at 66% capacity (i.e., 17Kw output) and have PV input at 9Kw, they overheat every time (1 or 2 of them), throwing an E02 error and shutting the system down. This is very concerning since I still have several additional electrical requirements to come in the future (electric on-demand water heater, RV hook-up, well pump, etc).

I am at a crossroads and would love any input. I want to upgrade the system, but should I go to a 3x EG4 18KPV Inverter, or pay $10K more and go with Victron (3x 15K Quattro and 3x 450/100-Tr)? I have used Victron systems before, without issue, but on a smaller scale.

Any input would be appreciated.
 
Personally between the eg4 and victron I'd go victron since I'm not exactly an eg4/ss fan.

That said 10k is a lot of money and the 18kpv has very good reviews here from what I have seen. So ignoring past eg4 things and just judging it from what Ive read the 18kpv would probably be the way to go.
 
Are you planning to export to the grid?
If not, why spend the extra money on hybrid AIO's?
 
No exporting to the grid, completely off grid. The cost of running a line is prohibitive.

@Crowz - I agree, tough to ignore the positive reviews and the fact that SS has a very attractive upgrade program ($3k for the 18KPV w/ trade-in). I am trying to justify Victron, but can't do it so far.
 
No exporting to the grid, completely off grid. The cost of running a line is prohibitive.
Then why spend the extra money for hybrid AIO's, like the 18kpv?
Instead of off grid AIO's like the 6kpv.
Especially when cost seems important to you.
 
Then why spend the extra money for hybrid AIO's, like the 18kpv?
Instead of off grid AIO's like the 6kpv.
Especially when cost seems important to you.
I would need 6x the 6000 units to get the same output rating as the 3x 18kpv. They run $1550 x4 and $750 x2 (trade-in program) or $7700. The 3x 18kpv would run $11,250 (2x trade-in @ $3000 +$5250). The $3550 difference seems to come with a lot of advantages.

I am only configured for 4 units now, and adding the additional 2 would require a lot of rework/rewiring (downsizing is easier), which would also cost money. Also, I don't have the wall space for 6 units. With the 18kpv, I have the space to upgrade to 4x units if I needed, which is a nice option. I am going to ask SS to see if they will allow me to trade in all of my 6500EXs even though the trade-in program seems to allow for 2, and if they do, I may just get all 4 for $12k.

In addition, the 6x 6000 units run at 50w idle (300w) compared to 70w for the 18kpv (210w). I also like the idea of fewer units; I feel like there is less to break. Also, after running the 6500EXs hard, I am becoming biased against lower-tuned inverters. I feel like they tend to break whenever I am running them hot (50%+); I am hoping the higher-tuned inverters are more capable. A big assumption on my part, but that is where I am at right now.
 
I am only configured for 4 units now, and adding the additional 2 would require a lot of rework/rewiring (downsizing is easier), which would also cost money. Also, I don't have the wall space for 6 units. With the 18kpv, I have the space to upgrade to 4x units if I needed, which is a nice option. I am going to ask SS to see if they will allow me to trade in all of my 6500EXs even though the trade-in program seems to allow for 2, and if they do, I may just get all 4 for $12k.

In addition, the 6x 6000 units run at 50w idle (300w) compared to 70w for the 18kpv (210w). I also like the idea of fewer units; I feel like there is less to break. Also, after running the 6500EXs hard, I am becoming biased against lower-tuned inverters. I feel like they tend to break whenever I am running them hot (50%+); I am hoping the higher-tuned inverters are more capable. A big assumption on my part, but that is where I am at right now.
These are all valid points.
And you have done the research to make an informed decision.
You have satisfied my curiosity, thanks.
 
Back
Top