New DIY solar project, I "dunnow" what to believe!

[eddie3000]

Hello, and good day! I am a new solar DIYer. The title may intriguing.

I have just got my solar panels up and running. I have three ginlong solis mini 4g inverters, each with their own string of panels. I have them all configured to not inject any energy to the grid. I do not want to do this as the I still have not found any electricity company in my area that will give me a fair deal. My intention is to use as much of my solar energy my self.

In order to make my life easier, I am making a system that will automatically turn stuff on and off depending on the solar production. I still haven't finished my project though, I am using an arduino with some ADCs (ads1115) connected to some SCT013s to measure currents at different points of my electrical system. In my test bench, the SCT013 produced surprisingly precise measurements. But, after putting these devices in my home electrical cabient to measure the current between my house and the grid, and the total power consumption of my house, I noticed large discrepancies between my readings, the inverter's readings and the Ginlong Solis app. They all differ by quite a bit actually!

At this moment, my 3 inverters say they are producing 20+20+360W= 400W. The EPM is limiting the power generation on two of the inverters because my demands are not too large.

The ginlong solis app says they are producing 330W, my house is using 380W, and I'm only getting 38W from the grid (these figures don't sum up).

And my DIY device say my home is guzzling 600W, and 380 from the grid, leaving my production at about 220W.

It is possible that at such low amperage the sct013s aren't as precise and can produce way off figures. But I would like to know if anybody else has any experiences with a setup like mine to see if I'm missing something.

The ginlong solis app seems like crap. It very often says my home consumption is way higher (two to three times) than what it really is. And in spite of my DIY project proving to be precise on the workbench table, it's figure's do not match the inverter's on many ocassions.

Thank you for your patience and reading my post.

 

[brum]

Get a clamp meter and measure the inverter and the grid AC current.

Check the surrounding of your current measuring device. I suppose you have some EMF sources around it and maybe getting interferences. A good hardware design would have considered these.

 

[eddie3000]

Get a clamp meter and measure the inverter and the grid AC current.

Check the surrounding of your current measuring device. I suppose you ha

Yes, I will compare all the readings with my clamp meter and see what that says. A reading that surprises me is that the one I get from the grid side. Even though the inverters are limiting their production due to lack of power consumption, and the inverter's meters saying there are nearly zero amps being injected to the grid, my sct013 measures a considerable ammount of current (I can't tell in which direction it's going though). When I turn the breakers off, the sct013 measures exactly zero amps, only for a few seconds until the inverters detect the absence of grid power and shut off.

As I said, I will double check those readings with a clamp meter when I have some time.
Thanks for your reply.

 

[eddie3000]

Good morning.

Yesterday I got my clamp meter and the readings were very close to the sct013. At that time, about 5 o'clock pm, I had 5.6 amps coming out of the inverters, 5.9 amps going into the house, and about 2 amps between my house and the grid. According to my inverter's meters, no power was going to the grid.

Unless I'm missing something very basic, these figures don't sum up. If my inverters are producing 5.6 amps, my house is drawing 5.9, there should be about 0.3 from the grid, not 2 amps.

Scrolling through the inverter's advanced menus, I have seen a few adjustment-calibration options that I have not touched. All I have done is select the country standard. The manual isn't clear either as what these options do apart from saying that they should only be used by technicians! If anybody could shed some light on these obscure features, I would be very grateful. My inverters are solis mini 4g 3KW.

Thanks for such a great forum! I am eager to finish my project and get it all working nicely!

 

[eddie3000]

At the time of measurment, most of the load was caused by a 1hp electric motor. Could the inductive load cause errors in the readings?

 

[brum]

Yes, you may have issues with your Arduino code. You need to measure the current flow multiple times per half sine wave and do the proper math.

 

[eddie3000]

I don't think it's the arduino code. I am sampling way above the Nyquist rate for 50Hz, and for a period of 333ms, and calculating the rms from that. Increasing that time window would provide a less noisier measurement, but I settled for that value because it seemed like a good balance between speed, performance and precision. And as I've said in my prevvious post, my clamp meter readings are very similar to the arduino ones. If I have time, this evening I will repeat the measurement with a mostly resistive load (my oven).

 

[brum]

Are you measuring only the AC current or you are measuring both the current and the voltage?

 

[eddie3000]

Good morning. I am only measuring current. Yesterday I checked all the sct013s again, comparing them with my clamp, and they all seem to agree with each other. I have changed some sensor placements and get more readings today when the sun is out. I am going to put more load on the system to actually drain all the solar power possible. Then I will see what the readings are.

 

[brum]

With your current approach, you can measure only can only measure the power to resistive loads. There are three different types of loads - resistive, inductive, and capacitive.

To calculate the actual power you need to sample both the voltage and the current at a high frequency. Check the image below - there is an offset between peak current and voltage. This is what inductive load looks like. Power at each sample would be voltage * current.

If I may advise - use dedicated DIN rail energy meters with pulse output. Each pulse is usually 1 or 0.5 watts. The pulse output can be connected directly to the Arduino.

 

[eddie3000]

Thank you very much for your answer. Yes, you are correct. After lots of messing and a long easter holiday, I have decided to do exactly what you suggest. Each one of my inverters has a meter with both an rs485 output AND a pulse output. That should be very easy you connect to an arduino via optocouplers.
Thank you again. This forum is great and very helpful.