Todo list

[wills106]

• Dynamic times for Charger Times
• Revisit disabled by default sensors as a number of them return funny values.
• Workout why number & select entities don't show in the main Integration panel.
  Document Gen4 Modes of Operation #40 Moved to Wiki
  Help document Gen4 connection #42 Moved to Wiki
  - Allow for setting Unit address up instead of hard coded to 1
• Move away from using Tick boxes for Setup, read the Serial number from the Inverter and configure based off that

[wills106]

Release 0.4.9 should start to address some of the hidden sensors returning random values.

[infradom]

Thanks for the feedback, strange that you get answers from a non-connected inverter.
You absolutely need to enter a different name for the second inverter (I saw a similar mixup when I used the same name and needed to remover both to resolve it)
I could only test by declaring my inverter twice (with a trick).
When I used 2 times the 1 address, I saw the same data twice on both instances, which is expected.
When I used modbus address 2 on one of the integrations, this second one's attributes remained unknown, which is expected

I will reverify the code, but I am still busy with another development, and in the coming days, I may not have much time, so I cant promise a solution quickly.

[stringkillersa]

@infradom No worries, I have two waveshare in my setup, but for the future it would be nice to just need one.
But it is not a big issue. More for coming users...

[wills106]

@stringkillersa are you able to try out "0.5.0b7 Pre-release" does the new sensor return anything:

• group_read_test

Just wondering if the Inverter knows if an additional Inverter is connected or not?

[stringkillersa]

@wills106 Just tested. Both inverters return 0:0W:0W:0W:0A

[wills106]

Ok, so it must only return values when in a true Master / Slave setup.

Was worth a try though.

[wills106]

Looking at moving away from the various Tick Boxes used during initial Setup.
Might still need one for additional EPS configuration?

But was thinking of something like the following:

if seriesnumber[0:4] == 'L50E':
    model = gen2x1
    max_charge = 100

elif seriesnumber[0:4] == 'U50E':
    model = gen2x1
    max_charge = 50

elif seriesnumber[0:4] == 'H34T':
    model = gen4x3
    max_charge = 25

elif seriesnumber[0:4] == 'HUE5':
    model = gen3x1
    max_charge = 25

else:
    model = gen3x1
    max_charge = 20

This takes the first 4 digits of the serial number (seriesnumber[0:4]) and compares it against the known list of inverters from #26 It sets the model to Gen2, Gen4 etc and also the maximum charge / discharge allowed for the battery.

Instead of having three lots of the following:

["Battery Charge Max Current",
    "battery_charge_max_current",
    0x24,
    "f",
    {
    "min": 0,
    "max": 20, #or 25, 50, 100 etc
    "step": 0.1,
     "unit": ELECTRIC_CURRENT_AMPERE,
    }
],

We can replace it with using a variable:

["Battery Charge Max Current",
    "battery_charge_max_current",
    0x24,
    "f",
    {
    "min": 0,
    "max": maxcharge,
    "step": 0.1,
     "unit": ELECTRIC_CURRENT_AMPERE,
    }
],

This will reduce the code repetition, also when new / alternative Inverter models get discovered we only have to update the main IF statement.

I need to workout where to put this main IF statement so it works across __init__.py and const.py etc
When I first placed it in const.py it didn't seem to to anything.
I then placed it after the "seriesnumber" register in __init__.py and it correctly detected my Inverter as a H1E5 but I was unable to pass back "model = gen3x1" & "max_charge = 20" back to const.py

Needs further investigation...

[wills106]

That's fine, no rush.

[infradom]

I have introduced 2 new attributes in const.py: (in my fork )

• blacklist: for sensor, number, select, button entities
• max_exceptions: for number entities

The other files have been adapted accordingly
Please have a look if you like the approach, if so, I can issue a PR

[wills106]

I have noticed a couple of small corrections I need to do between sensors. But I understand what's going on now and I can correct them following the PR.

[infradom]

I will fist convert the blacklist and max_exception dictionarys to lists and then send the PR

[wills106]

Quick question, whats the difference between allowedtypes = X1 | X3 | GEN2 | GEN3 and allowedtypes = GEN2 | GEN3
I was expecting them to be the same?

[infradom]

Yes, it is the same, same result, different approach
I prefer the shortest form now.

[infradom]

OK, you are right, if no other box is ticked, the old code will do nothing

[infradom]

sorry for the confusion, you are right,

[wills106]

It's ok :)

[infradom]

you could add the block of code again (except the last line)
As long as the tick boxes are there, it can function as a fallback in case we forgot some SN codes

[wills106]

I did add the following to the 0.5.0b6 release notes:

If the Integration fails to load with the following error in your log “unrecognized inverter type - serial number : {your_serial_number_here}”
Please see [#26](https://github.com/wills106/homsassistant-solax-modbus/discussions/26) providing the details asked for.

As I wasn't sure at the time if the Tick Boxes worked or not.
Need to ideally start removing them at some point though.