Hardware: SolarFlow 2400 AC + 3× AB3000X
Inverter firmware: v2.0.0 (as shown in the app).
AB3000X pack firmware: 1.0.13.
Setup: the battery is NEVER actively commanded — it runs permanently in Smart mode
with a fixed 20–90% SoC range, doing its own dispatch.
Grid metering is a Shelly Pro 3EM (CT clamp); the inverter follows that CT autonomously.
I only READ the device (local API via Home Assistant); it's also reachable in the app.
Battery healthy and NOT empty when this happens (e.g. 66% SoC, pack voltage 48–50 V).

Symptom

Intermittently, the inverter drops into what looks like a deep-sleep / stuck-standby
state:
ALL power outputs go to 0 W simultaneously and stay there despite strong house
consumption: AC home output, battery in/out, AND the EPS / off-grid output.
ac_status = 0, pack_state = 0.
The reported SoC / "electric level" FREEZES (stops refreshing) at the moment it
enters this state.
BUT the communication side stays fully alive the whole time: Wi-Fi RSSI, pack
voltage and temperature keep updating in real time, and the operating mode still
reads as normal/smart. So from the outside "the device is online and the mode is
correct" — yet it produces nothing.
Duration observed: from ₄₀ minutes up to 2h34 if left alone.

What seems to triggers it

It follows an abrupt power transient — a fast ramp into a strong discharge (the
device is tracking the Shelly CT in Smart mode, so a sudden load = a sudden autonomous
discharge ramp). Two clear cases:
A large load starting (heat-pump water-heating cycle, ₄.6 kW draw).
A household appliance starting: battery flipped from −680 W (charging) to +1493 W
(discharging) in a single sample (₂.2 kW swing). ₇ minutes later, all outputs cut
at the same timestamp.
It is NOT tied to one specific appliance — it's the sudden demand / fast ramp toward
high discharge that seems to trip it. Local API timeouts appear in the log around the
same time

Possible firmware correlation (v2.0.0)

I'm on inverter firmware v2.0.0. I've seen other users on the 2400 line (including the
2400 Pro) report the device getting stuck in standby specifically AFTER updating to
V2.0.0. Could v2.0.0 have introduced or worsened this behaviour?

Why this is NOT normal standby

I want to rule out the usual "legitimate standby" explanations up front:
It is NOT the battery-full case (SoC ₆₆%).
It is NOT the SoC-at-minimum / low-voltage discharge cutoff — pack voltage stays
healthy (48–50 V, ₂.0 V/cell above the empty knee) throughout the frozen period.
The device SHOULD be discharging (there is unmet home demand and the grid is being
drawn from), but it simply does nothing.

The most important part: the EPS / off-grid output also dies

When it freezes, the EPS output goes to 0 W too — so anything on the off-grid/backup
socket (in my case a fridge/freezer) loses power until the inverter is woken. This is
the real problem: a "backup" output that silently switches off with no fault, no
alert, and no self-recovery.

Recovery — I've found that a mode-cycle brings it back

The only reliable way to wake it is to re-push the operating mode:
switch the manager mode OFF, hold _30–35 seconds, then back to smart. A faster cycle
(₁₀ s) does NOT wake it. Cycling only the off-grid mode does not help. After the
mode-cycle, ac_status returns to 2, the EPS output comes back (>0 W), and the SoC/
electric-level starts refreshing again.
The fact that re-pushing a mode setpoint fixes it strongly suggests a firmware/control
state-machine issue, not a hardware protection trip (a genuine protection fault would
not clear just by re-sending a mode).

Questions to Zendure
Is this a KNOWN issue on the SolarFlow 2400 AC, and is it acknowledged?
Is it related to inverter firmware v2.0.0 specifically? Is there a fixed firmware
version, and if so which one — and how do I update the INVERTER firmware (not just
the AB3000 packs)? If v2.0.0 is the cause, is a downgrade possible?
What is the official recommended mitigation in the meantime?
Crucially: is there ANY device-side auto-recovery from this state, or a way to make
the EPS output fail-safe? Because if the only recovery is an external system
re-pushing a mode, then putting critical loads (or the very controller that would
restart it) on the EPS is unsafe.

Happy to provide timestamped logs of a specific episode on request.

Best regards,
Nicolas

Symptom

What seems to triggers it

Possible firmware correlation (v2.0.0)

Why this is NOT normal standby

The most important part: the EPS / off-grid output also dies

Recovery — I've found that a mode-cycle brings it back

Questions to Zendure