FoxESS Community - Owners & Installers Forum

Hmm.

With the consumer unit under the stairs, and the power and CT cables running up the side of the house, into the loft, then across to the opposite side of the loft to the inverter, is it expected that the import values reported by the inverter are consistently around 3.5% lower (full range 2-5%) than those reported by the meter (via a CAD)?

Is there any additional calibration that can be done on the inverter to bring it in line?

(I note that having the inverter report lower import values makes it look better than it is but I'm sure that's just coincidence...)

I think this is simply due to the 5 minute time interval between snapshots so often the inverter misses things. If you happened to boil a kettle and it falls in the 5 minutes between snapshots then the inverter data wouldn't show any grid used in that time but the meter would. Similar happens on export too.

No, I'm talking about the total transfer (in Wh) counters, not the instantaneous power (in W) numbers. These are continuously updated inside the inverter and do not "miss the kettle".

How do you know they are continuously updated? The inverter being a digital device, it seems likely it is simply taking periodic readings from the CT clamp and doing some sort of numerical integration to calculate the energy, which is one source of potential error.
The CT clamp itself also has a level of error associated with it.

My point being, the meter is designed to accurately measure the amount of energy flowing through it. The inverter is designed to keep the net energy flow to or from the grid at about zero, as long as there is sufficient PV generation or battery capacity to allow it. Absolutely accurate measurements of total energy flow are not a high priority for such a device.

Well, yes, but the integration the inverter is doing is on a much shorter timescale than the 5 minutes that it sends a snapshot to the cloud at.

(If I try to integrate the grid power numbers from the 5 minute snapshots. then I get numbers which are even more wildly out, and are out in both directions. So I know that's not what the inverter is doing to generate its transfer numbers.)

And the inaccuracy of the CT clamp is exactly what I'm getting at: it seems from googling that people believe it's supposed to be very accurate, but no one appears to mention what effect a long cable run might have, or what would constitute "long" (probably a little over 10m in my case). My point is: is a long cable run expected to cause a reduction in power readings, is my run long enough to count, and can the inverter be calibrated to account for that?

My Foxess CT clamp is usually around -2% of what my smart meter reports which is pretty good for a CT clamp.

FoxESS use the CTSA016 clamp which has a rated accuracy of 1% but it's linearity is between 1% and 5% and you will also see it's accuracy being much worse at lower loads and with longer cable runs which it sounds like you have. The CT clamp is supplied with a 10 metre cable but for best accuracy it is best to trim it as short as possible, conversely it can be extended up to 25 meters with a reduction in accuracy.

It isn't possible to calibrate these CT clamps and given your installations cable runs, if accuracy is critical I would recommend changing it for a Chint DDSU666 which is installed at the meter, it communicates over RS485 to the inverter and because it measures power factor it has greatly improved accuracy over wider ranges of load.

Ta.