Can I tell pylontech.py to use e.g. ttyUSB11 instead of ttyUSB0? Everything is working fine except my USB devices (I have multiple) are differently enumerated...

I thought it might be useful to open a pull request for discussion rather than spamming other people's issues :)

I've added parsing which I think works for all of for US2000/US3000/UP2500 as well as a few simple tests to make sure the code behaves as expected to help future refactoring.

I don't have a US2000/US3000 (or any multi battery configuration) to test with, so it someone could verify it works on real devices it would be helpful!

Again, many thanks for this code! It's great that I managed to talk to my battery with relatively little effort thanks to this :)

Hi, @Frankkkkk this is what I've got from the very beginning:

root@raspberrypi:~# pip3 install python-pylontech Looking in indexes: https://pypi.org/simple, https://www.piwheels.org/simple Collecting python-pylontech Using cached https://files.pythonhosted.org/packages/19/29/caedde3663ae093c2046ed70925b4f2cc673817f7255d7b2ae9e4dc92e63/python_pylontech-0.0.6-py3-none-any.whl Requirement already satisfied: pyserial in /usr/lib/python3/dist-packages (from python-pylontech) (3.4) Requirement already satisfied: construct in /usr/local/lib/python3.7/dist-packages (from python-pylontech) (2.10.67) Installing collected packages: python-pylontech Successfully installed python-pylontech-0.0.6 root@raspberrypi:~# python3 Python 3.7.3 (default, Jan 22 2021, 20:04:44) [GCC 8.3.0] on linux Type "help", "copyright", "credits" or "license" for more information.

import pylontech p = pylontech.Pylontech() Traceback (most recent call last): File "", line 1, in AttributeError: module 'pylontech' has no attribute 'Pylontech' p = import pylontech p = pylontech.Pylontech() Traceback (most recent call last): File "", line 1, in AttributeError: module 'pylontech' has no attribute 'Pylontech'

I thought the temperature data reported looked a bit suspicious, so I took a closer look at the protocol specification and found temperatures specified like this (typos included :)

"Kelvin temperature: K Temperature blow 0 the value is negative e.g.: 25.5°C = 25.510+2731=2986 -12.4°C = -12.410+2731=2607"

Something like this seems to give more reasonable temperatures for my battery:

 class ToCelsius(construct.Adapter):
     def _decode(self, obj, context, path) -> float:
-        return obj / 100
+        return (obj - 2731) / 10.0  # in Kelvin*10

I have implemented this and updated tests. This builds on top of PR #16, so if you merge this you can ignore the other one.

wonderfull,

Exactly what I'am looking for to translate JK-BMS rs485 protocole (or other BMS) to pylontech rs485 protocole , and then have my voltronic/mppsolar inverter talk to my DIY battery.

Thank you.

Traceback (most recent call last):
  File "pylontech/pylontech.py", line 250, in <module>
    print(p.get_values_single(2))
  File "pylontech/pylontech.py", line 236, in get_values_single
    f = self.read_frame()
  File "pylontech/pylontech.py", line 186, in read_frame
    f = self._decode_hw_frame(raw_frame=raw_frame)
  File "pylontech/pylontech.py", line 167, in _decode_hw_frame
    assert got_frame_checksum == int(frame_chksum, 16)
AssertionError

Hooked up via instructions in readme. Dumping raw frames shows some data, but doesn't seem to decode:

b'\xb0\xf7\xf7\xf7\xf7w\xf7\xf7\xf7\xbf\xbf\xbf\xbf\xbf\xf7\xff\xbb\xbb\xef\x00~20024600C06E11020F0CA30CA50CA30CA30CA20CA40CA50CA30CA30CA50CA60CA50CA50CA50CA4050B0F0AF10AF10AF10AF10000BD9D157C02C3500283E517\r\x00'

it's a vanilla US2000, made in 2019

As title says: has anyone successfully communicated with US2000C? When talking to US2000C I only get

File "/usr/lib/python3.10/site-packages/pylontech/pylontech.py", line 270, in get_values f = self.read_frame() File "/usr/lib/python3.10/site-packages/pylontech/pylontech.py", line 200, in read_frame f = self._decode_hw_frame(raw_frame=raw_frame) File "/usr/lib/python3.10/site-packages/pylontech/pylontech.py", line 181, in _decode_hw_frame assert got_frame_checksum == int(frame_chksum, 16) ValueError: invalid literal for int() with base 16: b'' Tested with two different setups (RPI, batteries) - same result...

I have 7 2000C with firmware 1.7 flashed onto them. They all functioning as-expected and talk quite happily to BatteryView and the Victron Inverter.

When I try get details for pack 1 (the master) the python crashes on the line

    assert got_frame_checksum == int(frame_chksum, 16)

If I run pack 2-7, works just fine. How can I debug this?

Hi!

I've hacked a bit further and I think the get_management_info message seem to work for me now, so we can see when the battery requests charging and when it allows charging / discharging - which could be used for inverter control.

Please let me know if you think this looks right or if there's any comments or concerns :)

Hi,

first, thanks for implementing this crappy protocol.

Initially I thought my RS485 connection was not working as I was not able to establish any communication. But as soon as I did send the example "string" from the protocol documentation the battery started talking. So I thought that something was wrong with the code or the checksum (or python version) and started reversing and debugging. It does also NOT help that the checksum explanation in the PDF is wrong (wrong sum for the sample characters given). Anyway, the code is/was ok.

I'm not able to query anything except "get_values_single" (I'm getting ignored by the battery, no response). And the first battery is 2, for whatever reason. So no get_protocol_version, no get_module_serial_number. (?)

Is this normal? I do not have a gateway. Just 4 linked batteries us2000C of the same type.

Also I found a bug in the code:

According to the PDF (Page 15 "M+N+6") there is a field "UserDefinedData" in the data returned by get_values_single (42H), that is missing in the code/struct. This leads to data garbage after "RemainingCapacity".

Sooo... around Line 99:

     99     get_values_single_fmt = construct.Struct(
    100         "NumberOfModule" / construct.Byte,
    101         "NumberOfCells" / construct.Int8ub,
    102         "CellVoltages" / construct.Array(construct.this.NumberOfCells, ToVolt(construct.Int16sb)),
    103         "NumberOfTemperatures" / construct.Int8ub,
    104         "AverageBMSTemperature" / ToCelsius(construct.Int16sb),
    105         "GroupedCellsTemperatures" / construct.Array(construct.this.NumberOfTemperatures - 1, ToCelsius(construct.Int16sb)),
    106         "Current" / ToAmp(construct.Int16sb),
    107         "Voltage" / ToVolt(construct.Int16ub),
    108         "Power" / construct.Computed(construct.this.Current * construct.this.Voltage),
    109         "RemainingCapacity" / DivideBy1000(construct.Int16ub),
    110         "UserDefinedItems" / construct.Byte,
    111         "TotalCapacity" / DivideBy1000(construct.Int16ub),
    112         "CycleNumber" / construct.Int16ub,
    113     )

See line 110 of my code snipped. Also this is still not correct according to the protocol (PDF), as the following numbers are an array of the size of "UserDefinedItems". Did not investigate further into this. I do not know if this breaks something I'm not able to call. For me this fixes "TotalCapacity" and "CycleNumber", but I do not have batteries > 65ah.

Protocol version is still 2.0 for me (b'~200246...).

Greetings Frank

Hi! I’ve tried all combinations of A/B And DIP on/off (changing baud rate in code) and using L0/L1 But I don’t seem to get any answers from the battery. Currently the battery is not connected anywhere, but SOC does flash so I suppose it should reply in RS485?

Could this be just because the model is UP2500 vs US2000 and therefore it has some different commands?

or do I have some other issue?

what do you think, Frank?

thanks, Dennis