18 month DIY project finally done: 39kW inverter power off grid system

I also monitor very closely the battery situation. All 12 batteries are very close regarding their SOC and voltage. Also, while charging/discharging they all take/give the same wattage of power. Here is an example while charging with 7.6kW:


I also monitor the max. highest cell delta over all batteries in my smart home system. The following chart shows the max. delta over the last 5 days (SolarAssistant only gives me the max/min/average cell delta per battery and not for each cell inside each battery, the these are all max values):

The max. cell delta will usually stay below 30mV and only rises for a short time up to 150mV just before they are fully charged in bulk mode (with about 0.15C) and while keeping them full for the rest of the day the max. delta will be between 60mV and 90mV (float?). As soon as the batteries gettting discharged again (from around 100% SOC), the max. delta will again drop below 30mV. I think this behavior is ok - but not 100% sure! As far as I understand, it's ok if the max. cell delta will be much higher just before bulk charging ends...

This is the battery temperature chart over the last 5 days:


Update
Btw. I'm using 55.2V absorption charge voltage and 53.6V float charge voltage

fmeili1 said:

Btw. I'm using 55.2V absorption charge voltage and 53.6V float charge voltage

Click to expand...

This is why they "appear" to be well balanced.
You are staying in the flat part of the charging curve.
There shouldn't be any delta variation, unless a cell is failing.

I would like to thank everyone!
Without this DIYSOLARFORUM and your help, suggestions, answers, tips, etc. the project would not have been a success.

Everything is possible with a good community behind you - share and learn from each other!

A very big thank you

Update:

• 8 days in a row with 100% sun, no clouds, 660kWh PV produced (and consumed)
• still no issues
• first time using my 3D printer and even here no LED pulsing/flickering if the printer is running (I had this previously in some tests)
• tuned the smart home rules again to control the AIO temperatures with the ducted mini split, now they usually stay below 52°C under highest previous load/charge
• today is the first day with some clouds, curious how the system will perform in not ideal sunny conditions

I have also installed a “grid outage” detection.

On the one hand, I would also like to know one of the situation for which I designed the system also, occurs - one of the goals was to have power if the grid fails - just knowing if this happens, will give me a good feeling
On the other hand, I can take the “grid outage” status into account in my smart home based solar load management rules.

It's based on a battery powered Z-WAVE door/windows sensor which has an connector to optionally connect an external switch/relay (instead of using the internal reed contact with an external magnet). In case of grid power loss, the relay changed it's state to detect to grid outage. To achieve this, I've installed a "permanent grid" receptacle on my AC-in breaker panel.

Just a short update after 17 days online:

• never connected AC-in to the grid until now
• 960kWh consumed, 1356kWh PV production results in an overall efficiency of about 71% (not really good but acceptable)
• about 430 starts of the 5 TON central heatpump (with soft start installed) - no failed starts
• max. daily load was 83kWh lowest was 39kWh
• max. peak load was 17kW lasting about 4 minutes
• max. peak power drawn from the batteries was 13kW lasting about 1 minute (0.2C with 12 batteries)
• max. peak PV power was 19kW (out of my 18.5kW STC peak PV modules) - cloud edge effect
• the remaining battery capacity in the morning before new PV productions starts is between 40% and 70%
• between 10am and 12pm the batteries are usually recharged to 100% SOC
• highest AIO temperature was 50°C, highest battery temperature was 32°C (garage ambient between 22°C and 30°C)
• active equipment cooling helps a lot and keeps the AIO's silent

So far, no faults, no dropouts, no issues, no LED pulsing/flickering, no manual interactions, slowly starting to forget that I'm running pure off-grid... this was one of the goals at the beginning of the project.

Looking forward for my next energy bill (still need to pay the base fees).

fmeili1 said:

960kWh consumed, 1356kWh PV production results in an overall efficiency of about 71%

Click to expand...

This is why these off-grid AIO inverters suck. 722 kWh per month wasted to standby loss, conversion inefficiency and battery cycling. Almost as much energy I use per month lost to losses. Ideally we would all grid-tie at 97.6% efficiency if utilities were not hostile entities.

fmeili1 said:

This is the battery temperature chart over the last 5 days:

Click to expand...

May want to extend those cooling ducts into your battery racks. Running LFP batts at elevated temps will increase calendar aging capacity loss.

AntronX said:

This is why these off-grid AIO inverters suck. 722 kWh per month wasted to standby loss, conversion inefficiency and battery cycling. Almost as much energy I use per month lost to losses. Ideally we would all grid-tie at 97.6% efficiency if utilities were not hostile entities.

Click to expand...

Yes, this is the "not so nice" truth. It looks like this is the price I have to pay by using "cheap" AIO's and wanted to be independent of the grid. It would be interesting how the overall losses would be if quality equipment had been used (e.g. Victron chargers and inverters) - maybe 80% or 85% - I don't know.

AntronX said:

May want to extend those cooling ducts into your battery racks. Running LFP batts at elevated temps will increase calendar aging capacity loss.

Click to expand...

That's a good idea. I will think about how it may be possible from the mechanical point of view. So far, some of the cold air from the duct outlets is already falling down on the battery racks which may help a bit to keep them a bit cooler, but a direct injection of cold air into the racks would be a perfect solution.
I don't know how much the difference between e.g. 32°C and maybe 20-25°C would increase the lifetime of the batteries (32°C is still in the given optimum usage range, regarding the specs). I even don't know, if the temperatures provided by SolarAssistant is the battery temperature or the BMS temperature - there is no hint about this.
The ambient in the garage is about 27°C (conditioned).

fmeili1 said:

I even don't know, if the temperatures provided by SolarAssistant is the battery temperature or the BMS temperature

Click to expand...

It appears to be the average temperature of all BMS's. (According to my calculator)
You can check yours on the battery monitoring page.
Not sure about your system. But my BMS's are always warmer than anything else.



Temperature 1 and 2 are the individual probs. I used them for cells and breaker.

timselectric said:

It appears to be the average temperature of all BMS's. (According to my calculator)
You can check yours on the battery monitoring page.
Not sure about your system. But my BMS's are always warmer than anything else.

View attachment 215381

Temperature 1 and 2 are the individual probs. I used them for cells and breaker.

Click to expand...

Unfortunately, SolarAssistant only showing the following data for my 12 batteries (even they are individually connected all via RS485). The header shows the average of all 12 battery temperatures, but each battery only shows one temperature at all. Looks like SolarAssistant does not request all possible data from the EG4-LL batteries because the "BMS Tools" from EG4 is able to show 6 different temperatures including BMS/PCB and ambient, see video.


So I have no idea which temperature SS is showing. If it would be the BMS/PCB temperature like in your case, I would not worry too much with 32°C because this would be a nice temperature for the BMS electronic and my assumption is that the battery cell temperature is lower than the BMS/PCB temperature.

fmeili1 said:

Unfortunately, SolarAssistant only showing the following data for my 12 batteries (even they are individually connected all via RS485). The header shows the average of all 12 battery temperatures, but each battery only shows one temperature at all. Looks like SolarAssistant does not request all possible data from the EG4-LL batteries because the "BMS Tools" from EG4 is able to show 6 different temperatures including BMS/PCB and ambient, see video.
View attachment 215382

So I have no idea which temperature SS is showing. If it would be the BMS/PCB temperature like in your case, I would not worry too much with 32°C because this would be a nice temperature for the BMS electronic and my assumption is that the battery cell temperature is lower than the BMS/PCB temperature.

Click to expand...

You should be able to compare between the two, to see what it's reporting.
Just to have an idea

timselectric said:

You should be able to compare between the two, to see what it's reporting.
Just to have an idea

Click to expand...

I've send SolarAssistant a feature request to read some more data from the EG4-LL batteries. I've sent them the protocol spec from EG4. In this spec there are three temperature registers/addresses listed: "Temp of PCB", "Temp Avg", "Temp Max" (also I've asked for the cycle count which is available but not shown in SolarAssistant).

Unfortunately, until this feature request may be implemented, I don't have a chance to get the different temperatures while the batteries are in use. Maybe I can try one battery with the "BMS tools" and try to compare with the latest seen value in SS - but it has to be measured short after disconnecting from SS and no load is allowed while using the BMS tools, so I don't have a real change to compare the values.

fmeili1 said:

I've send SolarAssistant a feature request to read some more data from the EG4-LL batteries. I've sent them the protocol spec from EG4. In this spec there are three temperature registers/addresses listed: "Temp of PCB", "Temp Avg", "Temp Max" (also I've asked for the cycle count which is available but not shown in SolarAssistant).

Unfortunately, until this feature request may be implemented, I don't have a chance to get the different temperatures while the batteries are in use. Maybe I can try one battery with the "BMS tools" and try to compare with the latest seen value in SS - but it has to be measured short after disconnecting from SS and no load is allowed while using the BMS tools, so I don't have a real change to compare the values.

Click to expand...

I see
Well, that's a shame indeed.
We have all been requesting enhanced functionality (for years) . But most of it appears to be ignored, unfortunately.
All of their time is dedicated to new product monitoring, instead of updating exiting functionality.

timselectric said:

I see
Well, that's a shame indeed.
We have all been requesting enhanced functionality (for years) . But most of it appears to be ignored, unfortunately.
All of their time is dedicated to new product monitoring, instead of updating exiting functionality.

Click to expand...

I have good new from SolarAssistant. I've sent them the EG4-LL (V2 and S) protocol spec and they asked me to grant admin access to my SA instance for further analyzing. So it seem they are willing to find a solution and/or extending the number of available battery data requesting from the batteries.

fmeili1 said:

I have good new from SolarAssistant. I've sent them the EG4-LL (V2 and S) protocol spec and they asked me to grant admin access to my SA instance for further analyzing. So it seem they are willing to find a solution and/or extending the number of available battery data requesting from the batteries.

Click to expand...

This is good news.

What are these? Is this a way to draw in fresh air for the cooling? I have never seen something like this applied to inverters.

Aridom82 said:

What are these? Is this a way to draw in fresh air for the cooling? I have never seen something like this applied to inverters.
View attachment 215817

Click to expand...

Very impressive.

Aridom82 said:

What are these? Is this a way to draw in fresh air for the cooling? I have never seen something like this applied to inverters.

Click to expand...

Yes, the equipment cooling keeps the units nice cool and so it keeps them quiet, also. I'm sure the lifespan of the inverters will be better also when they usually running cool. I'm very happy with the cooling/loudness results with the small 9,000BTU ducted mini split which does not cost much energy.