Even on a fine, clear day my system will always use a little grid power. Today, for example, current grid usage (measured from midnight until time of posting) is 0.6 kw/h. Most of this, but not all of it, is consumed whilst the sun is up. This is per SEMS portal, not per the meter, so it includes any feeding back to the grid. In fact the inverter tries to zero the grid each day and so this implies that it will always be using some (the COJ meter I have doesn't recognise the back supply, but it's usually a small amount and it's not worth my while to change tariffs, meters etc). 

Following on from another recent thread, I am guessing that this is the natural consequence of having to interact with the grid all the time to 

1. know if the grid is up or down
2. to synchronise with grid voltage and phase.

Am I right or do I need to phone a friend?

Hi @Bobster

I have noticed the same both with importing and exporting,

From observations the "leakage" seems to be a bit more pronounced when there are large loads switching on and off, e.g. our one 4kW geyser, and the larger the load the more the "leakage".

While most of the PV generated goes into charging the battery the "leakage is much lower if not zero, once the system start to export and the loads come online, the "leakage" seems to increase.

Purely a guess (without running tests), is that this is a way for the inverter to help balance and protect itself, rather than absorbing the surges (and compensating for short "dips") caused by especially the larger loads switching on and off and which will require extremely fast "reaction times", it uses the grid, while connected, as a bit of a buffer.

 

 

 

 

 

20 minutes ago, WeNotGood said:

Purely a guess (without running tests), is that this is a way for the inverter to help balance and protect itself, rather than absorbing the surges (and compensating for short "dips") caused by especially the larger loads switching on and off and which will require extremely fast "reaction times", it uses the grid, while connected, as a bit of a buffer.

I'd guessed this too. There are little leakages (to use your term) all through the night and all that can possibly be is fridges turning on and off. In the wee hours the water heating and the pool pump are turned off, other stuff is ticking over and the only things that can vary in state are the fridges.

As you say, it's probably easier to deal with the spike caused by a motor starting up by drawing from the grid momentarily than by drawing more from the batteries.
 

Edited December 28, 20205 yr by Bobster

My current battery capacity is a bit low so normally drained by the time it gets to the wee hours of the morning, will monitor what happens once the capacity has been increased

Some additional hypotheses:

• If some of the loads are not part of the "backup" output of the inverter, the accuracy of the CT (current) sensor could also potentially cause "leakage" in either direction i.e. exporting or importing, would expect this to be fairly low but more obvious during low power consumption scenario's.
• If the consumption is very low, the battery may simply not supply any current at all, and just use energy from the Grid, I have got this under my "Diagnose Messages" which can be accessed in PV Master from under SETTINGS -> Diagnose Message

18 hours ago, WeNotGood said:

If some of the loads are not part of the "backup" output of the inverter, the accuracy of the CT (current) sensor could also potentially cause "leakage" in either direction i.e. exporting or importing, would expect this to be fairly low but more obvious during low power consumption scenario's.

That's not the case in my case. The non-backed up loads are the outbuildings and the pool pump. The former are not used at night and the latter is on a timer and also not used at night. 

If I look at the graph on the portal I can see the downward movements of the yellow (grid) as load suddenly increases, which, to me, suggests it's something to do with inductive loads on the backed up circuit.