Context: 9 kVA single-phase installation, AC coupling of microinverters to a MultiPlus-II 48/8000, CRAE arrangement (surplus exported to the grid under an EDF OA contract).
Equipment: MultiPlus-II 48/8000 · Cerbo GX MK2 · 4× Pylontech US5000 (19.2 kWh, DVCC active) · 4× Hoymiles HMS-2000-4T (8 kVA total, AC coupling, EN 50549) · VM-1P75CT as the grid meter on the Enedis side.
Proposed architecture: the microinverters and all the loads (house plus two remote outbuildings, supplied by a single AC cable that has already been installed, with no possibility of running a second circuit) are wired to the same AC OUT, downstream of the transfer relay. In grid-connected mode, any surplus not absorbed by the battery charging system (limited by the BMS’s CCL via DVCC) is exported to the grid. In island mode, the expected control method is frequency shifting (EN 50549 P(f) curve) to curtail the Hoymiles units when the batteries are full.
My installer’s position: they recommend wiring the microinverters to AC OUT 2 (the programmable output, switched off/on according to the battery SOC threshold) rather than to the main AC OUT, considering the latter “unsafe” for battery charging — without specifying any numerical limit. However, this solution would require dedicated AC wiring, which I cannot install.
Question: is AC coupling the Hoymiles units to the main AC OUT (shared with the loads) valid and safe with DVCC active, or is there a genuine constraint (an AC OUT current limit during absorption, or DVCC behaviour with respect to a third-party AC source) that would objectively justify using AC OUT 2 in this case?
PV inverters are usually installed on AC OUT1. This is why frequency control is needed so the multiplus can throttle generation when batteries are full. The factor 1 rule is also there to protect the system. This way the PV can keep generating on grid loss. AC OUT 2 is disconnected on grid loss so you will lose any PV should that happen.
Hello,
Thank you for your reply, which confirms that coupling the PV panels to AC-OUT1 is an option. I am indeed looking to maintain photovoltaic production in the event of a grid failure.
However, my installer maintains his recommendation to use AC OUT2, putting forward the following argument: “using the same AC OUT as the house distribution board means that the batteries will not charge safely”, and refers to a risk of “letting the inverter manage things on its own, without programming a precise cut-off”. My understanding, to be confirmed or corrected: DVCC already imposes the BMS charge limits (CCL, CVL) on the MultiPlus in real time, regardless of the energy source on the AC OUT (grid or coupled PV). When the CCL drops to 0 (batteries full), the MultiPlus stops absorbing power, regardless of the energy available on the AC OUT1 side. Is there a real scenario in which the presence of PV on AC OUT1 would make battery charging “unsafe” despite an active DVCC and a communicating BMS? Or does this argument apply only to batteries without a communicating BMS / without DVCC enabled (which is not my case with Pylontech US5000s)?
Question: is the installer right to consider that there is a specific battery-charging safety risk associated with coupling PV to AC OUT1 with DVCC + an active Pylontech BMS, or is this risk already covered by DVCC?
You configure the PV inverter frequency control in the ESS assistant. It will throttle the PV inverter when it needs to, assuming the PV inverter responds to frequency control. When the BMS drops its CCL or sets it to 0, the ESS assistant will raise the frequency to reduce or stop PV production.
This is the correct way to do it.
Your installer seems to have some illogical reasons to try and force you to use AC OUT2 instead.
DVCC alone does not make a safe stable system, PV inverter control through either the ESS assistant setup or modbus/sunspec control is the other part required for it to work well.
A good read about this is the AC-coupling/Factor 1 guide from Victron. By using an MP2 48/8000 with 8kVA of AC-PV you are right on a factor of 1, but this is still fine.
Which is part of the noted gridcode EN50549-1 (schematic notes EN50549-10 which is not the gridcode itself but rather how a device is tested to comply with -1). So as long as the AC-PV is using this gridcode itself, the frequency-power-shifting will work also in island operation.
Also small info about the schematic, there is no VM-1P meter avialable from Victron. You can use VM-3P to measure a single-phase grid and another one to measure the also single-phase AC-PV. But using three phase meters is a bit of a waste. Personally i would consider using ET112s instead
