Incorrect (and also weird) data from solar PV

[randominseattle]

I just added a solar PV system to my house, and I don't understand the data I'm seeing from the GEM.

There's a 60 amp breaker between the solar system and the main power supply for the house. I've put a micro 80 on each leg of that breaker, on the solar side, and attached them both to a single channel of the GEM. That channel is set to the micro-80 defaults (Type 210, Range 4).

So I've got two questions:

1. Correctness: During the day the GEM's reckoning of the power production is way off from what the PV system's monitor says (SunPower PVS5x). For example, the GEM might be reading 7.4kW while SunPower is showing 8.4kW. My other circuits seem to be much more accurate -- what can I do to troubleshoot this?

2. Just a question for my own understanding. At night, when the system is actually drawing from mains for the inverters, the GEM shows ~50W, but 9.7A. My Fluke 323 agrees that each leg is running at 4.8A, and it says the voltage between them is 244.4V. I understand that when computing A/C power you have to consider the phase angle between the voltage and the current. I don't have any gear capable of measuring that, but they'd have to be nearly 90 degrees out of phase for these numbers to add up. Does that make sense for solar inverters when there's no sun?

[ben]

1. How does it compare when you disregard the power use at night?

2. Inverters have very poor power factor at night that the GEM may be having issues with. The power draw as per the specs is likely something like less then 5W.

What GEM COM/ENG firmwares are you running? I'll check with Paul about the power factor issue.

EDIT: What model inverter do you have installed?

[ben]

Also is your GEM a blue or black board?

[randominseattle]

1. I compared the energy readings for the last 24 hours from the GEM, SunPower's monitoring site, and the solar production meter installed by the utility company, and they're all within 2-3% of one another. The instantaneous power readings from the GEM are off from what SunPower is reporting, but I guess they must be disagreeing in a way that the total energy production adds up the same? So far I've just been looking at the live data (I don't have a data host hooked up yet), so I'll see about getting a data host involved so I can look at some charts over time.

2. COM 4.33
ENG 1.49
Blue board

I have 35 SunPower X21-335-BLK-AC panels, each with its own 320W microinverter.

[ben]

Paul suggests flipping off the inverters at night if possible and looking at how the mains react to the inverters being off. You'll probably need a data host setup for this however as looking at one sample won't work too well (something else may go on/off right when you switch the breaker over).

Looked up the specs for the panels, they don't list the idle power. Usually it's listed as ~1W. Given each panel has its own inverter, it may very well be close to the number the GEM is given.

Under Channel Option, do you have NET enabled for the Inverter channel?

[randominseattle]

Yes, the inverter channel is set to NET. Live data shows negative watts when producing during the day, positive watts at night. The live data watt-hour counter is always going up, though.

Regarding turning off the inverters, I can do that. What effect am I looking for on the mains? The mains CTs are inside the main panel, right above the main breaker (thus measuring total consumption on all my house circuits). The inverters splice in to the mains upstream (outside the house, closer to the utility meter), so the mains reading doesn't include their consumption right now.

[ben]

Regarding turning off the inverters, I can do that. What effect am I looking for on the mains? The mains CTs are inside the main panel, right above the main breaker (thus measuring total consumption on all my house circuits). The inverters splice in to the mains upstream (outside the house, closer to the utility meter), so the mains reading doesn't include their consumption right now.

Ahh, that method won't work then.

What readings do you get from your SunPower while the inverters are idle?

[randominseattle]

SunPower very helpfully says things like "waiting for sun". I could probably extend the leads to move the CTs outside, but I probably won't bother since from the conversation so far it seems kinda likely that the inverters just have a really low power factor at night.

I'll get a data host in there so I can compare the daytime readings better.

[ben]

SunPower very helpfully says things like "waiting for sun". I could probably extend the leads to move the CTs outside, but I probably won't bother since from the conversation so far it seems kinda likely that the inverters just have a really low power factor at night.

I'll get a data host in there so I can compare the daytime readings better.

Paul mentioned there might be a way to improve things by adding capacitance to the PT. Once you get a data host setup I'll get you some more information on this so we can compare values with changes to the capacitance.