GivTCP automation suddenly stopped working

I've had a regular schedule switching ECO on and off via HA automations for about 2 years with very few issues. All of a sudden it has stopped working - I notice my battery was not discharging during peak period.

I've done a load of testing and can see that even if I manually trigger the script to flip ECO on or off, the invertor doesn't respond.

Inverter logs show that until the 10th the changes to mode were triggered by "external event" which I assume is HA but since then not one has got through.

As far as I can see no firmware was upgraded and I have not upgraded GivTCP (because when I do it won't work due to requiring external MQTT where my running images use built-in). The only upgraded components were HA core and OS but I can see the automations trigger so confused.

I don't expect answers - just trying to sort my own thoughts out by writing stuff down.

I guess I'm going to have to bite the bullet and upgrade GE firmware and then get GivTCP working again both of which fill me with apprehension as when stuff works I like to leave it alone.

Ah well - it was good while it lasted!

Hmmm - just thought I'd restart GIVTCP before hitting update so did and watched the logs.

GivTCP_1 | 1739437984: New connection from 127.0.0.1:55929 on port 1883.
GivTCP_1 | 1739437984: New client connected from 127.0.0.1:55929 as GivEnergy_GivTCP_1 (p2, c1, k60).
GivTCP_1 | 1739437984: Client GivEnergy_GivTCP_1 disconnected.
GivTCP_1 | 2025-02-13 09:13:14,297 - Inv1 - write - [INFO ] - Setting Battery Mode to: Timed Export
GivTCP_1 | 2025-02-13 09:13:14,577 - Inv1 - write - [INFO ] - Setting export mode was a success
GivTCP_1 | 2025-02-13 09:13:14,867 - Inv1 - write - [INFO ] - Enabling Discharge was a success
GivTCP_1 | 1739438004: New connection from 127.0.0.1:41775 on port 1883.
GivTCP_1 | 1739438005: Client <unknown> closed its connection.
GivTCP_1 | 1739438005: New connection from 127.0.0.1:47043 on port 1883.
GivTCP_1 | 1739438005: New client connected from 127.0.0.1:47043 as GivEnergy_GivTCP_1 (p2, c1, k60).

I see constant connect/disconnects but not sure if that's "normal" or not. I changed the mode from my HA dashboard and it can be seen that GivTCP makes the change and reports as successful. However, it is not - battery still not discharging.

OK. Time to hit update and cross all my bits.

MikeBakke I’m pretty sure that GivTCP 2 doesn’t keep the connection alive so you do get these disconnects and connects. You don’t with GivTCP 3.

Doing a restart of the inverter and/or a reset to defaults is usually enough to get the inverter accepting givtcp commands. Sometimes it just stops working, I’ve had this

My experience with GivTCP3 was that i uninstalled GivTCP2 in HA and then installed GivTCP3.
I put it off for weeks worrying about how it would go.
For me it was simple and easy and the whole HA thing has been far more "stable" since then

Well... I did the firmware update of inverter and battery and both seem ok, however my battery now shows 14% rather than the 58% it was before upgrade. Can't see any discharge or charge so unclear whether I should charge it or not (before peak rate).

I assume the inverter restarted although I see no message in the logs so maybe I'll do a manual restart, same for givtcp and see if control is back.

Ref GivTCP3 - I did try switching to that but that's where the builtin MQTT was deprecated so I held off trying to figure that all out.

Oh - I'll hang fire - the charge % seems to be climbing slowly despite no charging happening - will hang fire for a bit and see what happens

MikeBakke
I am still using MQTT so maybe i did'nt do what I should have?

SteveCook I think it's fine if your HA has MQTT installed - mine doesn't and I always found it another complication but I guess I should do it as now I've upgraded the firmware GivTCP v2 is more and more likely to be out of step with the API

Well - the battery charge eventually climbed back up. I spoke to tech support and they said yeah - the firmware change can cause that. I also agreed to start the battery conditioning cycle tonight to get it out of the way.

Currently (pun kind of intended), the system is behaving and even my automation executed correctly at 16:00 and put me back in ECO. <huge shrug>. WIll monitor to see if it reverts off battery mid morning tomorrow when I hit 65% SOC again.