Solax Tripower LiFePo4, from a suspected issue to "howto charge your battery"

[wiltet]

Dear contribuitors of the homeassistant-solax-modbus intergration,

in a other thread I was writing already about a problem I have, startet from here:
#2079 (comment)
and would like to ask you developper guys, with quite a lot of experience, about following descripted occurence.

To give it a better overview I'll try to explain my issue carefully, because I couldn't reproduce it again when I started to research honestly with debug-logs etc.

systen-overview:
My power system is in use since 9.2025, build out of a Solax-X1-Hybrid-G4-7.5kW inverter, a X1-Mini-G4 connected to EPS in "Micro-Grid-Mode", a T-BAT H 17.3 battery and 6kW PV. I do have access to public grid, but use it only as backup in winter time. My system is basically autonomous and offgrid, because of grid-injection-limitations in my country. There is a second X1-Hybrid as backup and spare, because this model is EOL and I don't trust it to survive 20years+. As well this inverter surprised me with unexpected EPS interrupts and EPS-Relay-errors ending in black-outs and made me thinking this way. The spare X1-Hybrid-inverter is connected to two (spare) PV-panels in summer time to keep the capacitors alive.

issue in general:
All begann since the days became longer and we produced plenty of energy. Then I did set the max.SOC from the usual 100% to 80%, thinking that this gives a longer lifetime to the battery. Once a month starts a HA automatisation to charge SOC to 100% for ballancing and callibration.
After a while I did observe that the inverter does not really stop charging when reaching the 80% limit. He start/stop charge, visible at BDC-state, allways with 70Watts. Sometimes the charging state changes every few seconds and sometimes only a few times in a hour. Like this the SOC ends up at the end of the day with up to 84%.
This has come and gone frequently and sometimes changed with Solax FW upgrades and as well HA integration updates. Fully inverter reboots did give many times different results as well.

issue problematics:
Tree days after the last Solax FW upgrade on the 29.5.2026 the inverter started to do this again, but more intense then the days before. At the end of the day SOC was at 85% and there must have been a higher frequency then usual, because all my shelly devices reported "overload" and where in protection mode. And this I could proof that it is freq-based.
Rising the frequency when reaching the defined max.SOC is normal, when the inverter is in "MicroGrid-Mode". It is used to control the more simple on-grid inverters and stop them injecting by setting a higher frequency. The X1-Hybrid usually goes up to 53.2Hz. Sometimes I can see higher peaks in the HA sensor log, but don't know if this is very trustworth.

start of research:
After this happend I did start to search for the cause. My first presumption was the FW upgrade, because it was after that. So I activated cloud access and wrote to Solax support.
In the evening I didn't feel good about the invertes behavior and made a full reboot, cutting all power sources until it "died", did wait two minutes and started up the system again. (this is what I make after every FW upgrade as well)
The next day, everything was fine! No more overcharging after reaching max.SOC of 80%.
Hm...
Because all was fine again, I did push the update of HA integration from 2026.5.3 to the 2026.06.1. After this update the overcharging began right after the restart of HA again.
I don't know why, but this is what I observed and after I found more issues, I started to report on the linked issue thread.

Right now I can't reproduce the SOC-overcharge-issue. Everything is working fine atm.

Somehow it gives no sence to my understanding that this behavior could be caused by the HA integration.
The fact that it came present after a update, this created my doubts.
It would be a lot only to be 100% shure, that this can't be HA integration.

thanks to everybody taking the time for reading this!

[wiltet]

[wiltet]

@wiltet I just quickly screened through your 80% SOC issue. I didn't find a hint that you are using remote control from this integration. Did I overlook that?

I do not use "remote control".
I'm only setting some values by automatisations like "winter mode" set -->100%SOC, "chage from grid mode" set-->SOC60%, max. charging current 5A and similar simple automations.

[wiltet]

The answers I received from Solax support:

after detailed descriptions and screenshots similat to the once here:

yesterday:
"You can continue the test today. In my view, there are no issues with the firmware. I haven't received any other reports of similar issues.
Any questions pls feel free to let me know, thank you."

now:
"Hello. I have reviewed the data you sent, as well as the remote data. Strictly speaking, this is not a fault. Fluctuations in the load can cause the inverter's control signals to the battery to be less precise, resulting in a certain degree of error in the State of Charge (SOC) reading.
In fact, we do not recommend setting the SOC limit to 80%, as this is detrimental to the battery's lifespan."

Somehow I cant beliefe this what he is saying and pleased him for detailed explanation.

My experiences with Solax support is, that some colaborators answer sometimes fairy tales. So I began to be a bit more critical to their answers.

[wiltet]

I will record debug.logs as soon the issue is present again. atm is all working fine!

[rosenrot00]

@wiltet I just quickly screened through your 80% SOC issue. I didn't find a hint that you are using remote control from this integration. Did I overlook that?

I do not use "remote control". I'm only setting some values by automatisations like "winter mode" set -->100%SOC, "chage from grid mode" set-->SOC60%, max. charging current 5A and similar simple automations.

If you’re not using the remote control functionality, this integration is unlikely to be the cause of the issue. Likewise, if there is no automation in Home Assistant toggling the setting, the problem is probably unrelated to this integration.

The debug logs are not expected to provide any useful insights in this case. However, you could still help SolaX investigate the issue by providing them with relevant Home Assistant data and sensor history.

[wiltet]

Allright, thanks for your advice. That's the direction what I was guessing as well. Perhabs I can show you a log in futute to proof.

About helping Solax:
I hope they care (X1-Hybrid-G4 is EOL), because it does not look like when you read the last answer from support. Well, I will find out.

Are you or anybody else experienced about increasing battery life of LiFe4Po's?
Can this be true what support says that limiting to SOC80% will reduce lifespan?

[pmannk]

Are you or anybody else experienced about increasing battery life of LiFe4Po's? Can this be true what support says that limiting to SOC80% will reduce lifespan?

Battery chemistry is a complex topic. To be clear I'm not an expert in the field, but when I was researching NMC vs LiFePO4 before buying an EV a couple of years ago there were a few considerations with LiFePO4 that studies all seemed to agree on:

• The charging curve is very flat, which means SOC accuracy is much harder to manage. Most manufacturers who recommend deep discharging ( < 10% ) followed by high charging ( > 90% ) do so as it helps the BMS SOC calculations remain calibrated (voltage alone can't always predict SOC with LiFePO4, which is where coulomb counting and calibration matters)
• In addition to the above, 100% charging is necessary to balance cells. This could be why Solax advised against limiting charging to 80%
• LiFePO4 will tolerate 100% charging much better than NMC, however it does still accelerate degradation**

Practically speaking what I do is:

• Charge to 100% periodically so that batteries can balance
• Avoid holding the batteries at 100% for extended periods of time
• Every 6 months or so I discharge down to ~10% and charge back to 100%

I'm in a warm temperate climate so cold-weather charging is never a considerations for me, but if you're in an area that experiences below-freezing temperatures then charging when cold will be a far greater problem than any of the above.

**Regarding degradation, there seems to be less of a consensus on how much this matters with LiFePO4. A lot of what I read suggested that it's immaterial given the expected lifespan of an EV or residential storage system.

[wiltet]

This is a very helpful description, thanks for sharing your knowledge.
(I'm still waiting for a answer from Solax support about the point of not charging to 80%.)

Meanwhile I think I understood what support did say with:
"I have reviewed the data you sent, as well as the remote data. Strictly speaking, this is not a fault. Fluctuations in the load can cause the inverter's control signals to the battery to be less precise, resulting in a certain degree of error in the State of Charge (SOC) reading."

And your explanation about the problematics of calculating SOC-state reaffirms the statement from Solax.
So the problem about "Solax X1-Hybrid-G4 not fully respecting defined max.SOC of 80%" can be valued as "normal" and is based on the behavior of LiFePo4.

So there is only the question left about how to real increase battery life. You @pmannk gaved already a lot of input.
I live as well in a warmer zone in south, so deep temperatures are not a point nedded to be respected.
Solax suport actually knows that I charge once a month to 100% SOC. How often do you do it?

I'm curious how big the benefit of lifespan could be when treating the batts in the right way.

[pmannk]

Solax suport actually knows that I charge once a month to 100% SOC. How often do you do it?

For the EV it's every couple of weeks due to our driving and charging patterns.
For the Solax it varies - in winter I don't generate enough solar so I have to top it up from the grid. Roughly every 4-6 weeks.
In the warmer months it's more frequent, up to once a week if conditions are good. This is based on my dump load (hot water cylinder) preventing the batteries from charging right up, rather than lowering the SOC/manipulating the Solax itself.

I'm not super rigid about it. A couple of weeks variation here and there doesn't bother me.

I'm curious how big the benefit of lifespan could be when treating the batts in the right way.

Therein lies the complexity of the topic and the realities of the real world vs laboratory tests.

[wiltet]

Today I got an answer from Solax support about how to treat the LiFePo4 best. It is a short statemant only, but I think it's on the point.

"We recommend not setting a maximum SOC limit for the battery; instead, allow the battery to work with the normal charge and discharge cycles—ranging from 10% to 100% and back to 10%. This ensures uniform usage of the battery cells, thereby maximizing the battery's lifespan."

Whith this information I try to find the best way to achieve this in real-life-scenario.
In my case it is easy to get close to it in the "PV-summer-time". That's about 7-9 months of the year.
"PV-Winter-time" will be much different and I think I will have the focus on keeping the batterys allways as full as possible.

p.s @wills106
Now that it is clearly no issue at all, specially not about HA and this integration, would it be possible to move this thread to "Discussions"?
thanks

[wiltet]

further conversation with support:

me:
"With that information I will try to find a way to get the closest as possible to that in real life scenario.
What about would be the value of the benefit in % when comparing 10-100% use cycle and daily 80-100% (or 60-80%)? Only to get a idea about necessity and profit."

Solax:
"We do not have this test datas for your question. But I think it is about 20%~30%."

Okay, I know, it is impossible to answer my question, because there are so many unknown factors. But this answer gives me the direction to know that I can easely win some battery-lifetime for my usecase with an autark power system and a (for the summer-time) oversized battery.
Perhabs it is as well a suggestion howto handle car batterys to not have them fully charged all the time, if they are not hardly used.

The main point I have learned now is that LiPo and LiFePo4 are much more different than I thought.

[wiltet]

Meanwhile the charge statistics looks much different.
It is possible to deal with it easily with the help of some automatisations.
There is still the "unwanted" charging during "max.charge current = 0 amp". But like Solax support explained, this is based on the battery type in combination with load fluctuations. Still a bit strange that this happens sometimes stronger then other times, with no recognizable logic behinde the behavior.