Try the new test version, based on the log I do think there was a overloading push in my code:
Marstek Energy Storage | Homey
Sorry i didnāt tell that i switched it off myself:-)
Yesterday evening al of a sudden battery 3 came back alive in the app and now this morning also battery 2 did.
Only battery 1 is still not showing data.
I have installed the beta version but result stays the same.
tell me the version you are on now, and if the last doesnt come back (even with the option set to no) then send a me new diagnostic
Hi, Iām having the same time out issues as 0nl0gisch , also a venus E 3.0 (firmware v148). App version 1.3.18. Diagnostic report: aa0929ef-3a3a-44d8-ba58-df100bca4e9b
Iāve had this issue for over a week, but after updating to the latest homey software today, it started working again. Iāll monitor it to see if it stays this way.
yours seems to be a config weirdness, I see slave 1 and 2 devices? how did you set your modbus?
Yes, only the top two capacities update after restarting the app and then nothing works anymore.
What I have already tried: reinstalling the app, restarting Marstek. Update Homey to the latest version. All of this without result. No values āāare coming in anymore.
Send me a diagnostic report please
264dd478-b665-4c44-a13c-0196195b51ea
Ok, so for the V3 on UTP modbus the firmware is really⦠picky
ā¦it can not handle multiple clients and it gets overloaded really easily without having documentation on what the thresholds are. The problem I believe these devices have is that when I exceed the communication threshold it hangs the modbus, I detect a broken connection and try to reconnect, but the V3 did not release the broken connection yet so it doesnt accept the new connection.
I want to interact with short cycles, thats normally the advantage over local modbus versus local api, so energy flows can respond quicker by an external controller (Homey) like my Homey Smart EMS app does. The external serial port based controllers like the EW11 etc dont have those picky behaviors so only a group of people suffer from this.
So the balance between fast control and a weak modbus tcp controller is the fight we are having. I am building more and more safeguards and reconnect methods and defenders etc, but this balance makes even that part hard (reconnect automatically to recover from a dropped connection due to overloaded modbus while the firmware is still recovering for example) with a V3 myself.
So my pro advise is:
- if you want EMS control that quickly adapts and sends commands, use a modbus serial gateway additionally.
- if you dont need that reduce some polling intervals and disable the safeguards, its just for measurements a manual control (see device advanced controls), tips:
āPoll intervalincrease (for example. 10000ms)
āAuto-recover force-command revertsturn off
āFast-check intervalincrease if you want to keep recovery
I will use this log to improve the recovery of hung connections a bit more. expect a new version incoming.
ok, check the new test version please: It now has a device advanced setting called connection profile. It defaults to external gateway but you can now switch it to in a gentle builtin connection profile. It will reduce the load on the modbus.
If the device is still in overloaded mode it might still take a minute. So it can help to disable the app for 5 minutes after switching to let the Venus catch its breath and then enable the app. See how it behaves.
all three go available/unavailable from time to time?
And send a new diagnostic report please.
b05a19e9-4925-4487-a570-29c530c3e936
Please send me a DM with screenshots of your advanced settings of all three.
Are there other apps using the modbus ? In homey (like the slim laden etc) but also in other systems like HA/NodeRED etc.
Are you using the Marstek phone app with local connection enabled while testing (it can be a modbus client also).
What is your firmware.
While the deivces are unavailable, can you still ping them?
How are they connected? UTP to switch, all same switch etc?
I also updated the test version to force the connection profile into a more gentle mode if modbus errors persist. You will see a warning in the device if it did that.
Here the same. Until 1.3.19 (26 jun 2026) it worked fine. Also test version and changed settings as given advice doesnāt work anymore. Rollback to 1.3.18?



