So as my inverters soc readings are such poorly defined I'm contemplating building my own shunt as a test for possibly building more for my bigger system can anyone point me towards some reference material as I'm unfamiliar with how the get accurate readings. When I say building my own I mean sticking together a few modules and programming an esp with c++ no PCB design involved I'm just unsure how to turn voltage, current, power and energy readings into an accurate soc
Building my own shunt
- Thread starter ThaiTaffy
- Start date
Because I doubt any off shelf shunt I buy can configure to my smart home without alot of messing about. I can buy a pzem module and esp32 for less than $10 put it together in 10mins and program it in 2mins but I'm just unsure if I'm supposed to use voltage as my parameter for soc or there's a better way to do so.
I'll take a look not 100% on if the pzem DC meter can interpret direction of current flow but you've atleast given me something I can look at tyvm
Edit: yes the pzem DC does record current flow direction
wpns
Solar Joules are catch and release
As meetyg said, voltage doesn’t work, count amp-hours in and out of your battery. Precision and resolution are key.
mikefitz
Solar Wizard
- Joined
- May 28, 2020
- Messages
- 3,180
Seems Interesting, use with an optional shunt up to 300 A. The specification suggests hardware compute of power. It's up to you to provide the softwere to drive it. With the low cost shunt, accuracy with temperature maybe an issue , as will resolution at low currents and possible pickup in sence wires to shunt . What algorithm you develop to measure SOC still needs a battery reference point, battery full or battery empty. The softwere needs to be 'told' this by monitoring battery current and voltage, or feedback from chargers or battery BMS.
It's this issue that causes most problems in determining battery 'full' or 'empty' as a reference point. Over time small errors will cause a SOC drift, so some mechanism needs to 'reset' SOC every so often.
That seems to be the pzem017 which sadly doesn't report current direction I've looked at the other models but I'm currently looking at the junctek kg-f (Hope it's not junktech) it seems to be able to show directional flow and I should be able to hack the rs485 output. It's just a test idea for proof of concept as it stands and will be used on a FLD so no bms.
42OhmsPA
What's in a title?
If it makes you feel better, it's also sold as Won-Von.
I've been running the 600A since 2022, very limited issues, like Bluetooth disconnecting a few times per year.
It's within a tenth of a volt compared to my clamp meter.
DIYrich
Solar Wizard
Resistor in parallel. Measure voltage at the resistor, and current flowing through the resistor. That will be a fraction of what flows through the other part of the parallel circuit.
From what I can see most of the work has already been done as far as registers and modbus and the big thing is it's available here which pretty much all the recommended ones aren't.
I'm not that interested in the data to look at Bluetooth isn't an issue its more being able to use it for automation IE being able to burn load depending on projected PV production the following day.
Similar threads
