Sungrow Hybrid: [site] ERROR battery mode: modbus: exception '4' (server device failure), function '6' when attempting to unlock battery

[DiDu0815]

Locking house battery against discharging was once working quite well, a few months ago.
I now tried again; locking still works, but unlocking battery seems not to work anymore. I tried LAN port directly on inverter, als well as via Winet-S dongle - identical behavior. Locking works, but instead of unlocking I get

[site ] ERROR 2025/09/12 17:07:41 battery mode: modbus: exception '4' (server device failure), function '6'

I can unlock the battery again manually via iSolarCloud app, by setting Max. Discharging Power (kW) from 0.01 (which corresponds to locked) to 10.6 (which corresponds to unlocked).

Since it once worked, evcc has changed a lot, and Winet-S and hybrid inverter firmware changed as well. Right now, I am on the most recent versions for each.

Before attempting to dive deeper I wanted to ask if anyone is able to reproduce that behavior on other Sungrow hybrid inverter installations.

[harmendp]

I have a similar error, when charging the battery from the GRID. In the process several registers are set, this works quite well. However, when the charging is finished, I get below error.

[site ] ERROR 2025/09/17 14:03:57 battery mode: modbus: exception '4' (server device failure), function '6'

I need to check further how to reproduce exactly. But it seems a more general issue.

[LeonSol]

Look at my last comment and look at the code (sungrow-hybrid.yaml) - probably we will find the solution.

[DiDu0815]

Last Sungrow Hybrid template change seems to be this one from May. Since, there was - at least for me - no need to lock the battery, considering more than sufficient solar power for charging during summer. That might be an indication for this change being the potential cause.
FYI @zachelnet @premultiply

[LeonSol]

But after May Sungrow released a lot of firmware updates - only the last one (95.10) is apparently reliable. Which one do you use?

[harmendp]

There is definitely an error in the logic. Two different registers are used/set for 'hold' mode (33048 max discharge power) and 'forced charge/discharge' mode (13052 charge/discharge power), however only one of them is reset at the end. I ended up with a limit on my battery discharge, however no feed-in after the charging of the battery.

But as far as I can see, this error in the logic doesn't explain the modbus errors.

[LeonSol]

I know that a member of the core team (@naltatis) also uses sungrow. Probably he can help us here....

[DiDu0815]

@LeonSol Current firmwares; I believe the newest ones available in public.

Hybrid Inverter

Winet-S

[LeonSol]

Es gab diverse Änderungen in den Registern. Einige wurde von "unsigned" auf "signed" umgestellt usw. - ich finde nur gerade die Dokumentation nicht. Wie ich schon geschrieben habe: vielleicht kann hier ja @naltatis als zukünftig Betroffener (;-) mal rein schauen. Ich bleibe meinen alten Versionen treu und werde wegen der AC->DC->AC Verluste >= 20% und 0,22 €/kWh fix keine Batterieladung aus dem Netz durchführen.

[DiDu0815]

Laden der Batterie aus dem Netz ist auch bei mir weniger relevant. Der Hauptzweck der Batteriesperre ist bei mir, das schnelle Leersaugen der Hausbatterie bei geplantem oder Schnelladen des Fahrzeugs zu verhindern, wenn gerade nicht genug Solarüberschuss zum Laden zur Verfügung steht (i.d.R. nachts).

[LeonSol]

Ich weiß nicht, ob es zutreffend ist, aber um die Entladesperren aufzuheben, wird meines Erachtens im Code das falsche Register benutzt:

Beim Entladen wird jedoch nicht 13051 (im Zusammenspiel mit 13049 und 13050) genutzt, sondern 33047

Kann sich mal bitte jemand den Code anschauen und mit den aktuellen Modbus-Registern von Sungrow vergleichen?

[smartelmus]

Moin,

wir haben genau dasselbe Problem mit Sungrow Wechselrichtern SH10RT-V20 und SH15T. Gibt es bereits einen Ansatz? Bzw. ist es möglich, die Register im Template selbst anzupassen?

Vielen Dank im Voraus für eure Mithilfe.
Beste Grüße, Flo

[smartelmus]

Hello harmendp,
thank you for your quick response. The following values are set:

In Battery parameters:
Max. charging power: 9 kW
Max. discharging power: 9kW

In Ernergy managemant parameters:
Max. permitted charging power from AC: 9kW

Should we set these parameters to other values?
Thank you in advance for your effort.
Kind regards, Flo

[harmendp]

Thanks for the sungrow params. But there are similar ones on evcc level. Did you set those?
What battery are u using? Sungrow or other?

[smartelmus]

Hello harmendp,

ah, got it. Settings in evcc are:

In the meter settings:
maxacpower: 15 kW (cause it is a SH15T)

In the Battery Settings:
maxchargepower: 0 (to set it automatically to maxpower BMS)

I am using the Sungrow SBR160 Battery

Thank you in advance for your effort.
Kind regards, Flo

[harmendp]

No problem 😊
Could you try setting maxchargepower to 5000 and try again? Please also post DEBUG log here so we can check the error.
If it works, we can try to find a higher value by raising it. Max for SBR160 should indeed be around 9kW.

[smartelmus]

Dear harmendp,

perfect, it´s working. I set the maxchargepower gradually from 5000, 7000 to 9000 W. Works like a charme. Thanks man.
In order to reproduce the the problem i will set it to 0 again. What Debug Mode should i use to give you the above mentioned Information?

Kind regards, Flo

[harmendp]

No need, I know what is wrong. BDC register is 0 on latest firmwares I believe.
I opened a ticket @ Sungrow, we´ll see what they say.

[smartelmus]

Thank you very much for your time and help on this. If you need assistance with the gsp ticket portal. We have an Installer Account and can handle tickets as well.

Have a nice week.
Kind regards, Flo

@DiDu0815 Can you mark the thread as solved. Thank you

[DiDu0815]

Let me change maxchargepower from 0 to 4200 now and test it using fast charging tomorrow. Then I will mark the thread as solved (provided it worked).

[aeroralf]

Hallo , ich habe das selbe Problem.
Eine Entladung der Batterie ist nicht mehr möglich .
Ich habe einen Sungrow SH20T in Verbindung mit einer SBH 250.
Auf die Einstellungen von Charging und Discharging Power habe ich als normaler Benutzer ohne Installateur Account keinen Zugriff.
Kann das Problem mit einer neuen evcc Version behoben werden ?
Danke !

[TJ22]

I've asked my Installer to tell me the sungrow configuration - in my case (SH25T, SBH200) 14080 maxIn as well as out. Setting these values in evcc solved the issue.
Thank you for providing the workaround!

[harmendp]

Important to set the max charge power to a correct value! This is then copied by evcc to max discharge power. This will work without a doubt!

Before the change we were just setting 10kW, but wasn't working for everybody either. PR will make it a param in the config.

[Funkymaddox]

my two cents:
1st: Now i have to estimate my battery max charge/discharge rate and will probably cut it off ... calculated, it should be between 10.5 and 14.0 kWh, cause the Sungrow SBH sheets give you a range of V and you need to multiply it with A, not a fix value ... have seen 11,6 kWh in reality. So what do i put in as a good value? Why not leave i.E. 20 kWh so that bat. can max perform ?
2nd: If #25529 goes live in some days, do i have to revert the manual dialed-in values? Or are these necessary anyways?

[harmendp]

Before this change, 10kW was set by default. So if you want to be safe, just set that value and it should be ok.

You can also see what the max value is that is accepted for max dischargepower, and set that for both.

[Funkymaddox]

so ... yesterday 11pm i tried 10kWh in iSolarCloud, which did not started the discharge of the battery. (PV System was in "Zwangsmodus"). In EVCC should be 5kWh set via the standard template.

BUT, this morning i noticed, that my battery was discharging. In iSolarCloud i can see, that the Battery startet discharging at the very second when our Tesla stopped charging (Tesla charged from 3-5am via a rep. charging plan). So it seemed, that this planned charge plan blocked the battery till 3am like during the active charging process from 3-5am. Strangely our rep .charging plan for our Volkswagen (charged from 6-7am) did not block the Battery in advance, only during active charging.
(PV System is now in "Netzbetrieb")

i am totally confused. Not sure what the actual problem is. Load plan issue, charge or discharge rate in evcc or isolarcloud?

[Funkymaddox]

(SH20T / SBH15) Nachdem es ein paar Tage problemslos mit den eingestellten 15kWh max. Charge/Discharge in EVCC, als auch iSolarCloud, lief, gab es heute früh morgens ein komisches Verhalten ohne Fehler im Log.

Unser Tesla läd mit einem Ladeplan morgens bis 5 Uhr auf 70% (wobei er neuerdings zu 90% eine komische 20min Pause mittem im Ladevorgang einlegt). Am Ende des Ladevorgangs stoppte zwar der Tesla, aber es wurde weitere 10min mit 11kWh in den Speicher bis 6% geladen (welcher vorher bei 0% war).
Das größere Problem, auch in den Screenshots sichtbar, dass der Speicher nach den 10min Ladung keinen Strom mehr abgibt, also wieder Discharge-Probleme hat.

[harmendp]

Hello goodmorning

A few versions ago, minsoc & maxsoc were introduced. Did you set those in the config? I think 5% is the default for minsoc. Might be related as it would explain this behavior.

[Funkymaddox]

Good morning :)

not in EVCC. In iSolarCloud the battery has a minSoC of 5% in the Settings, which shows in EVCC and in iSolarCloud overview correct 0% when at internal 5% SoC.
If i add 5% in EVCC this would mean a total minSoC of 10%.
And it worked for several days without any issues: Every night the battery drains to EVCC 0% (internaly 5%) SoC.
I'll see today, if all energy is trapped in the battery, or if it starts discharging after an small PV Charge.

[Funkymaddox]

hmn, at 11:19 local time all of a sudden the battery startet discharging again without any changes in the system. No idea if a super little charge from the sun unlocked the "barrierer".

I downloaded the trace and log files for anyone better skilled at this than me. If this keeps a problem for Sungrow SHXXT or SBHXX its a start to look at (11:19 is the switch)

evcc-20251210-112857-debug.log
evcc-20251210-112939-trace.log

[harmendp]

Try setting minsoc to 0 in evcc, if not 5% is used afaik. Not sure if that is the cause, but should be easy to test. I can´t really test myself today, as my battery is still half full.