I’m running a home battery and a Victron system with a shunt. I believe the wiring and calibration are correct. The shunt works very well and accurately. However, in my opinion, there’s a third weak point in the accurate SoC measurement. This is the standby power consumption when using integrated BMS and/or balancers. This power is constantly running and therefore accumulates over time. During periods without regular automatic calibration, this leads to the SoC drift. It would be beneficial if Victron could find a way to account for this standby power consumption in the basic configurations. Perhaps this could be a suggestion for a future software update.
@Sarowe1990 The only way would be to add a user input for a permanent unmeasured draw, which will be more or less accurate. What you could do is create a Virtual battery in Node-RED with the same parameters as your main battery, create a flow that calculates the SOC from the SmartShunt reading and a manual input for the BMS self consumption. With the upcoming Virtual switches (currently in V3.70 beta) you could add a slider switch which would allow you to update the phantom current from the GX switch pane.
I’m out of my depth when it comes to Node-RED programming. Personally, I can live with correcting the SoC every few days for now. However, after reading posts here and elsewhere, I think I’m not alone in experiencing this problem. The vast majority of home energy storage systems have more or less the same issue. Therefore, I believe it’s a fundamental problem and perhaps Victron should address it. Since I’ve been following the development of the DESS with great interest and active interest for the past two years, I now assume that some very angry comments about malfunctions can also be attributed to this phenomenon. A frequently voiced criticism is that the stored and calculated energy in the morning isn’t sufficient to cover peak demand. This, of course, can’t work if the SoC isn’t accurate and the system reaches its minimum voltage before reaching the calculated SoC value. As mentioned above, I’m not a programmer, but from my understanding, the basic shunt configuration should include an additional input option for precisely this standby consumption, alongside the Peukert exponent and efficiency. Please don’t misunderstand, I’m not someone who makes excessive or unjustified criticism. Consider this more as constructive feedback.