September 2026
Changes released in September 2026.
An inverter fault now opens its outage under curtailment
An inverter that reports a fault no longer opens its outage only after a curtailment cap or low light has lifted. The event dialog shows which condition it overrode, with the maker's status and text.
Plant outages, device faults and event times now read as the devices ran. Outages no longer open from a zero combiner box reading, and event times no longer carry the timezone offset.
- A combiner box that reports zero power against its own live readings no longer opens an outage.
- A silent inverter now opens a plant alarm after 12 hours.
- Two data loggers can now be added, reading combiner boxes, strings, the grid meter, irradiation and temperature sensors, and curtailment limits.
- A plant-wide grid trip is no longer recorded as a device fault.
- Event times from the data logger are no longer off by the timezone offset.
A closed outage lists every component that was down
A closed outage previously named only the components that recovered last — 7 of 17 silent inverters — and now lists the whole incident. A device fault no longer reopens once the device has withdrawn every fault code, and no longer reads "Live" after an outage takes it over. Plants receive this with their next agent update.
The phone app's unread badge no longer keeps counting a notification already read on another device.
A plant with an alarm system now has an "Alarm system" page
A plant with an alarm system now has an "Alarm system" page, showing each contact's state, its history over 24 hours, 7 or 30 days, and the alarms raised. The intrusion panel and the communication cabinet's UPS, heating and main switch open a plant event while a contact stands and close it when it clears.
An alarm or fault reported by an inverter now stays open until the inverter itself withdraws it. It no longer disappears from the list on its own.
- An alarm raised by an inverter no longer closes just because part of its alarm list could not be read.
- A fault reported by an inverter no longer closes on its own after the on-site data collection restarts.
- A low-insulation warning no longer stays shown as active after the inverter has withdrawn it.
- A battery unit read by a second data logger no longer keeps the make and model of the previous one.
- The grid operator's ceilings on a battery storage's charging and discharging now appear on the plant's "Control Units" tab.
- Wind and battery plants now receive the site's weather data — temperature, wind, precipitation, snow and irradiation.
An inverter the data logger cannot read is no longer called faulty
A plant no longer opens an "Inverter device fault" when its data logger reports it cannot read an inverter and no readings arrive. A low-priority "Logger cannot reach inverter" entry opens instead and becomes "Inverter not communicating" after an hour of daylight, and plants receive this with their next agent update.
Push notifications reach an iPhone again when the Mirox app is also used on an Android phone.
A faulty irradiation sensor no longer triggers false inverter outages, and a plant's total feed-in energy is no longer missing after an inverter replacement or a restart of the on-site data collection.
- A daylight reading an irradiation sensor cannot physically deliver is now refused, and inverters are no longer declared in outage against it.
- A plant's total feed-in energy is published again after an inverter has been replaced, or after the on-site data collection restarted.
Station icons on the analysis page now count inverters
On a plant's analysis page, the icons on each station row now count inverters rather than items: one blue wrench for each inverter with open service work. An icon opens a list of those inverters with everything open on each, each entry linking to its inverter.
Alarms reported by inverters now open, close and are counted the way their maker intends.
- An emergency stop, a device fault or a latched arc fault no longer clears itself while the inverter is still reporting it.
- Sungrow fan messages now appear as warnings rather than faults, as Sungrow itself rates them.
- Measurements taken while the plant cannot reach Mirox are now stored and sent on once the connection returns.
- An inverter that keeps falling back into the same reported state keeps one open entry, and every recurrence is counted.
- An alarm an inverter first reported as a warning and then as a fault now shows the more serious of the two.
Component problems now show what confirmed them and how devices ran
A component problem's "Confirmed by" now shows which measurement or device report opened the case, with an explanation of what it means, and a finding still being watched reads "Under observation". A "Reported operating states" section shows how long the machine spent throttling itself, capped, switched off or reporting a fault.
A brief network glitch at a plant no longer reads as an internal problem. Battery racks, modules and cells list in position order, and the battery power graph marks AC and DC readings.
- The plant connection view now reads "Agent restarting or reconnecting" for five minutes before it calls the problem "Internal service issue".
- A device that went down during an outage on that path reads "Reconnecting" for three minutes after the connection returns.
- The details panel on that view no longer closes when a date range is picked.
- Battery racks, modules and cells now list in position order, instead of Cell 1, Cell 10, Cell 100.
- The battery power graph now marks a reading "AC · converters" or "DC · racks", and containers and racks measure DC only.
A curtailed plant now reads as producing, not as shut down
A plant held at its grid operator's or direct marketer's limit now reads as producing, not "Grid Shutdown" or "External Shutdown". One long curtailment no longer fills the event list with dozens of entries.