diy solar

diy solar

Solar Assistant - data logger

Does anybody use the Victron Smart Shunt with Solar Assistant? I'm curious if it only uses the SOC info, or if Solar Assistant also refers to it for Charge/Discharge current?
How would Solar Assistant connect to both the inverter and a smart shunt?

I'm using the "emulated BMS" option although I have my doubts about its accuracy on SOC value.

For example, yesterday (a crappy rainy day all day for solar PV) the SolarAssistant battery power chart reported a total discharge of 1415Wh, and a total charge of 205Wh, so that would suggest at loss of at least 1210Wh from the battery for the day.

My battery nominal capacity is 18200Wh, so that suggest battery SOC should be reporting as 1-(1210/18200) = 93%. Yet the SOC reading never went below 97% and by evening and this morning it was back to reading 100% SOC despite no overnight charging.
 
How would Solar Assistant connect to both the inverter and a smart shunt?

I'm using the "emulated BMS" option although I have my doubts about its accuracy on SOC value.

For example, yesterday (a crappy rainy day all day for solar PV) the SolarAssistant battery power chart reported a total discharge of 1415Wh, and a total charge of 205Wh, so that would suggest at loss of at least 1210Wh from the battery for the day.

My battery nominal capacity is 18200Wh, so that suggest battery SOC should be reporting as 1-(1210/18200) = 93%. Yet the SOC reading never went below 97% and by evening and this morning it was back to reading 100% SOC despite no overnight charging.
Uses a separate usb cable to a different port on the Pi running SA.
 
I think I am seeing quite a large anomaly with Solar Assistant's Emulated BMS state of charge (SOC) values. I hadn't really noticed it before as we are just experiencing our first periods of extended bad weather for solar PV output, and so have been relying on the battery more.

Here is a chart I did up from yesterday's data:

Screen Shot 2021-10-12 at 10.44.20 am.png

The blue line is battery power in watts, which is either positive or negative depending on whether the battery is charging or discharging. Yesterday because of the poor weather it shows the battery was mostly discharging. A few brief periods when it was charging.

The red line uses the battery power value for each interval to calculate an energy value in Wh, again either positive or negative, and then accumulate those interval energy values to show the trend of energy in or out of the battery. The chart shows that over the course of the day the battery discharged a little over 1200Wh more than it charged.

My battery has a nominal 18000Wh capacity and was fully charged at the start of the day, so being 1200Wh down for the day you would expect the SOC value to have dropped by 1200/18000 = 6.7% and hence show the end of day SOC to be ~93%.

But look at the SOC (green) line as calculated by Solar Assistant's Emulated BMS. It only drops to 97% at the lowest, and at end of day returns back to 100%. Which just can't be give the battery discharged significantly more energy than it charged.

I've emailed Pierre at SA to see what he makes of it but I'd have thought the coulomb counting of an emulated BMS would track the rise and fall similar to the accumulated energy line, just expressed relative to the battery's nominal capacity and perhaps also some allowance for charging inefficiency.
 
I've emailed Pierre at SA to see what he makes of it but I'd have thought the coulomb counting of an emulated BMS would track the rise and fall similar to the accumulated energy line, just expressed relative to the battery's nominal capacity and perhaps also some allowance for charging inefficiency.

For the record, Pierre got back to me with this response yesterday:

When the charge/discharge rate goes below 0.01C the software starts to use the "Voltage at 30% SoC" and "Voltage at 80% SoC" settings to adjust the SoC slowly over time. For your system 0.01C is 18.4kW x 0.01 = 0.184kW which is just over 3 amps. Most likely the charge/discharge rate was low and then this made SoC inaccurate. I plan to remove this mechanism from the emulated BMS as it seems to only make it less accurate.

So looking forward to an improvement in the accuracy of the SOC measurement. It won't be perfect (e.g. Peukert effects on the discharge side and inefficiencies on the charging side) but it will certainly make it a be more reliable value.
 
Does anybody use the Victron Smart Shunt with Solar Assistant? I'm curious if it only uses the SOC info, or if Solar Assistant also refers to it for Charge/Discharge current? Right now, Solar Assistant does not add in the idle consumption of the inverters, giving you sort of a false Load value. I also have a 48v->12v converter that has varying load values and is not reported through solar assistant. In my case, I could have 200-300 watt load that the Smart Shunt reads, but solar assistant is reporting 0w.

I ordered the VE Direct cable, so I'm curious.
Got my cable last night. Gonna hook it all up today and find out more info.
 
Got my cable last night. Gonna hook it all up today and find out more info.
Keep us posted, per Pierre they are working on Gyll battery integration, can't wait, this is MotionEyeOS on Home Assistant viewable from anywhere, the Battery monitor is spot on, Joe.
1634235017060.png
 
It does everything I thought and more. It does use the smart shunt for reporting charge/discharge amperage, and follows it with pretty high refresh rates it seems. Shows more information than I was hoping for too! Now I never have to open the victron app, its all in one place!!

Screenshot_20211014-151524_Samsung Internet.jpg
 
I'm running mostly victron stuff, but I will be using both the smart shunt and the Overkill BMS, curious which will give me better info, Solar Assistant or Venus OS. Gonna start with Venus VRM tho, cause it's "free"*.

*free with cost of expensive victron hardware
 
FYI - I got excited about earlier posts on this this thread about Growatt, so I bought a license and set up a pi3.

Caution 'support' for Growatt is status only. There is no ability to change settings. And the Power management tab says not supported.

Screen Shot 2021-10-30 at 5.05.41 PM.jpg
 
FYI - I got excited about earlier posts on this this thread about Growatt, so I bought a license and set up a pi3.

Caution 'support' for Growatt is status only. There is no ability to change settings. And the Power management tab says not supported.

View attachment 70823
The LV6548 has power management, but some of it is hit and miss with bugs, so be weary when changing settings, I suggest you compare with the actual control panel.

My favorite is the ability to change it to Solar/Battery/Utility or Solar/Utility/Battery and Utility Only. You cannot do this easily from the actual control panel.
 
My favorite is the ability to change it to Solar/Battery/Utility or Solar/Utility/Battery and Utility Only. You cannot do this easily from the actual control panel.
On my clone PIP inverter, via Solar Assistant I can change it between Utility First, Solar First, or Solar/Battery/Utility, the same settings as the inverter panel provides.

Screen Shot 2021-10-31 at 2.09.55 pm.png

Also in the Power tab settings there is this nifty Override feature I have used a couple of times. This allows me to choose to change to Utility first mode for a preset duration, after which it will automatically revert back.

Screen Shot 2021-10-31 at 2.19.45 pm.png

I used this option recently when we had a grid outage one afternoon. Once I flipped the main circuit board transfer switch over to backup power supplied via this inverter, I then changed this setting using my phone from SBU mode to Utility First mode.

I do that because usually these storm induced grid outages tend to result in multiple outages overnight, so I let the inverter automatically change power source between using the battery and using the grid, just like a UPS. In between outages when the grid is available it tops up the batteries keeping them ready to to deal with any longer outages. But as I expect solar PV to be back supplying some power the next day I set it to override for 12 hours so it automatically reverts to normal operation mode (for me).
 
Can anyone explain this? Is this a possible bug in SolarAssistant? Or is anyone else seeing something similar? This is a shot where I ran out of battery power overnight and switched over to the grid, then the inverters wait to get to 51v to switch back to battery power. In the mean time, all power to loads comes from the grid (so I assume). But in this shot, it looks like I have an 1100w load, and 1190 coming from the grid (typical because of efficiency losses I assume), and 1382w coming in from Solar, but only 423w going into the battery??

Its almost as if the solar is supplying the 1100w load and the rest is going to battery, but then how come its drawing 1190w from the grid? Is there 1000w being burned up somewhere and doing nothing?? This stuff drives me bananas.

Screenshot_20211028-085457_Samsung Internet.jpg
 
On my Growatt SPF-6000T-DVM setting 14 determines charge priority, I select OSO for Solar charging only, if i select SNU - Solar and Utility will supply load and charge Battery, joe. Edit I see what your talking about now, I don't think Solar Assistant reports Grid correctly anyhow.

1636045677498.png
 
Last edited:
On my Growatt SPF-6000T-DVM setting 14 determines charge priority, I select OSO for Solar charging only, if i select SNU - Solar and Utility will supply load and charge Battery, joe. Edit I see what your talking about now, I don't think Solar Assistant reports Grid correctly anyhow.

View attachment 71296
Yes I have mine set to Only Solar as well. I guess I will have to catch it in the act and try and see if there is actual amperage being drawn from the grid. My bet is the "Grid Input" is innacurate or not actually measure and is just a calculated equation based on other inputs.
 
Yes I have mine set to Only Solar as well. I guess I will have to catch it in the act and try and see if there is actual amperage being drawn from the grid. My bet is the "Grid Input" is innacurate or not actually measure and is just a calculated equation based on other inputs.
I would start by ensuring you are on the latest release. The changelog for the latest version lists fixes for both LV/LVX inverters and Solar first mode, both seem like what you are describing: https://solar-assistant.io/help/updates/changelog
 
I would start by ensuring you are on the latest release. The changelog for the latest version lists fixes for both LV/LVX inverters and Solar first mode, both seem like what you are describing: https://solar-assistant.io/help/updates/changelog
You are correct - @Lt.Dan - under the configuration tab scroll down and see if there is a "update available" click it and wait, it will tell you to update and reboot, once you click that wait up to 5 min for it to come back online. Posted about it 2 days ago here, Joe.
 
Back
Top