Spare power going to export, not battery

With my AC3 system set to Eco Mode and other timed charge/discharge settings disabled, I have noticed a reluctance on the part of the inverter to start up and divert spare solar power to the battery. The inverter just stays in the 'idle' state even though spare power is being exported.

Could it be that the power demand required to run the inverter is such that it avoids starting because it would cost a significant amount to run, and that would offset or possibly use up any saving? So, if this is the case, would anyone like to hazard a guess at the threshold spare power level (and possibly for a duration) that is needed in order to cause the inverter to start up and charge the battery?

666tim666 We don't have an AC-coupled, just hybrids but from everything I've read, all the inverters respond dynamically to detected export (AC3 starts charging, hybrid diverts PV to battery but doesn't automatically divert export to battery). There is a time delay from export being detected to the inverter taking appropriate action, but that should be less than ten seconds. Also, there is no minimum threshold (AFAIK) for the inverters to divert to/from battery, they just work dynamically.

If you want another toy to play with try Inverter by David Russell on the App Store (not sure if its available on Google Play). The app updates every ten seconds and for our installations shows say 28W export on one inverter and 53W import on the other. The inverters know nothing of each other but are continuously working to reduce the import/export to zero and the app shows this very well.

You don't say whether this is recent behaviour you've noticed or if has always been this way but if there is export of more than a few tens of Watts for more than twenty seconds, I'd expect the inverter to divert that to battery. It's feasible your EM115 CT isn't clipped fully. If it's a solid core CT, then that's not the issue.

Tim Thanks for your reply.

I am not sure that an AC3 is sufficiently similar to a hybrid to be able to compare this particular behaviour.

My inverter was only fitted just before Christmas and I am not sure whether it has behaved like this before. It partly depends on whether the sun is out or not, of course.

I have checked the CTs (one on the solar AC and one on the grid interface) and they seem to be correctly fitted.

Is the David Russell APP just an alternative to the GivEnergy APP? If so then that is interesting as it implies that the interface is published, or are these APPs just using the API interface anyway.

If you have two hybrid inverters, and they are unaware of one another, then why do they not work against one another with say one inverter generating power and the other storing it?

666tim666 Is the David Russell APP just an alternative to the GivEnergy APP? If so then that is interesting as it implies that the interface is published, or are these APPs just using the API interface anyway

No the app suggested polls the Modbus locally over wifi with an update period of ten seconds, similar to Givtcp. One inverter showing a small import while the other shows a small export is simply a timing issue; the inverter updates its internal registers from the EM115 every few seconds to control inverter output but reports to the app whatever the register contents are when the request is made.

666tim666 why do they not work against one another

Each inverter works independently and monitors the import/export measured by the EM115. In ECO mode, if the CT reports export, the inverter diverts more PV to battery, or reduces output if there is no PV. If the CT reports import, the inverter increases output from PV or battery to reduce the grid import. In fairness to Givenergy, if the inverters responded very quickly (within say 100mS) as some have requested, they would continuously be fighting each other as one attempted to compensate for the other. The way Givenergy have configured the response times simply means that both inverters independently see the same trend and both respond accordingly. Faster response times or steeper ramp up/down would work OK for a single inverter situation but for a multi-inverter situation, the (very slightly) slower response times of around 5-10 seconds and a flatter ramp up/down means they just "play nicely" together, even though neither knows anything of the other.