incorrect method used to identify data fields.. ?

[jgc234]

I know the code works, but reading through it, I believe the method used to determine the field types in the QUERY_NORMAL_INFO(0x11,0x2) reply packet is incorrect.

This is my understanding:

Rather than using a hard-coded list to map each field type (%DATA_BYTES in the code), the mapping should be dynamically described by the system itself. The mapping should be obtained using the QUERY_DESCRIPTION (0x11,0x00) command. This returns a list of all the data codes (in order) that will be returned when you when you call QUERY_NORMAL_INFO later. This would be unique to each device (in theory, although it looks like most return the same order anyway).

You then use this mapping to determine what fields are returned from the QUERY_NORMAL_INFO (0x11,0x02) command.

In theory, this should eliminate the need for the difference between options_strings mappings for different DATA_BYTES mappings, and explain every field returned by QUERY_NORMAL_INFO.

That's my understanding anyway..

[solmoller]

Very, very interesting.

I haven't looked at that part of the code for years, so I am a bit rusty, though:

QUERY_DESCRIPTION was never used in the code. I can see in the documentation, that the QUERY_DESCRIPTION is properly set up with a reponse of 0x80. Unfortunately I cannot find a key to decode the reply on QUERY_DESCRIPTION

Did you look at what the inverter replied with asked with QUERY_DESCRIPTION?

BR Henrik

[jgc234]

Hi Henrik, I've written my own module in python just for fun, but I've learnt the protocol from reading you code and some PDF documents floating around the Internet. (search for eversolar_Inverter_PMU_PROTOCOL_V1-1.pdf or similar). The protocol document is difficult to understand - it's easier to learn from your code (thank you!).

Here's the results from my inverter (Eversolar TL1500)

send query_description...

2018-09-24 09:38:49 [INFO] tx packet - aa55010000101100000121
2018-09-24 09:38:49 [INFO] expecting reply of (17, 128)
2018-09-24 09:38:50 [INFO] rx packet - aa5500100100118016000d010441424344454748494a4c78797a7b7c7d7e7f0862
2018-09-24 09:38:50 [INFO] rx packet - src=0x10, dst=0x100, function=(17, 128), data=b'\x00\r\x01\x04ABCDEGHIJLxyz{|}~\x7f', len=22, checksum=b'\x08'

My inverter returns 22 bytes, which represent the single-byte codes for each position

00
0d
01
04
41
42
43
44
45
47
48
49
4a
4c
78
79
7a
7b
7c
7d
7e
7f
08
62

If you decode this as single bytes representing the data code by positional index..
eg
position 0 -> code 0x00
position 1 -> code 0x0d
position 2 -> code 0x01
etc..

So, looking up a huge table of all possible fields (from protocol PDF), and then keeping a map of index -> data code..

2018-09-24 09:38:50 [INFO]  map [00] -> code=0x00, var=temp, multiplier=0.1, units=°C, descr=Internal inverter temperature
2018-09-24 09:38:50 [INFO]  map [01] -> code=0x0d, var=e_today, multiplier=0.01, units=KW.Hr, descr=The accumulated kWh of day
2018-09-24 09:38:50 [INFO]  map [02] -> code=0x01, var=v_pv1, multiplier=0.1, units=V, descr=PV1 voltage
2018-09-24 09:38:50 [INFO]  map [03] -> code=0x04, var=i_pv1, multiplier=0.1, units=A, descr=PV1 current
2018-09-24 09:38:50 [INFO]  map [04] -> code=0x41, var=i_pv, multiplier=0.1, units=A, descr=PV current
2018-09-24 09:38:50 [INFO]  map [05] -> code=0x42, var=v_ac, multiplier=0.1, units=V, descr=Grid voltage
2018-09-24 09:38:50 [INFO]  map [06] -> code=0x43, var=f_ac, multiplier=0.01, units=Hz, descr=Grid frequency
2018-09-24 09:38:50 [INFO]  map [07] -> code=0x44, var=p_ac, multiplier=1, units=W, descr=Power to grid
2018-09-24 09:38:50 [INFO]  map [08] -> code=0x45, var=z_ac, multiplier=0.001, units=Ω, descr=Grid Impedance
2018-09-24 09:38:50 [INFO]  map [09] -> code=0x47, var=e_total_hr, multiplier=0.1, units=KW.Hr, descr=Total Energy to grid
2018-09-24 09:38:50 [INFO]  map [10] -> code=0x48, var=e_total_l, multiplier=0.1, units=KW.Hr, descr=Total Energy to grid
2018-09-24 09:38:50 [INFO]  map [11] -> code=0x49, var=h_total_h, multiplier=1, units=Hr, descr=Total operation hours
2018-09-24 09:38:50 [INFO]  map [12] -> code=0x4a, var=h_total_l, multiplier=1, units=Hr, descr=Total operation hours
2018-09-24 09:38:50 [INFO]  map [13] -> code=0x4c, var=mode, multiplier=1, units=, descr=Operation Mode
2018-09-24 09:38:50 [INFO]  map [14] -> code=0x78, var=gv_fault_value, multiplier=0.1, units=V, descr=Grid Voltage Fault Value
2018-09-24 09:38:50 [INFO]  map [15] -> code=0x79, var=gf_fault_value, multiplier=0.01, units=Hz, descr=Grid Frequency Fault Value
2018-09-24 09:38:50 [INFO]  map [16] -> code=0x7a, var=gz_fault_value, multiplier=0.001, units=Ω, descr=Grid Impedance Fault Value
2018-09-24 09:38:50 [INFO]  map [17] -> code=0x7b, var=tmp_fault_fault, multiplier=0.1, units=°C, descr=Temperature Fault Value
2018-09-24 09:38:50 [INFO]  map [18] -> code=0x7c, var=pv1_fault_value, multiplier=0.1, units=V, descr=PV1 voltage fault value
2018-09-24 09:38:50 [INFO]  map [19] -> code=0x7d, var=gfci_fault_value, multiplier=0.1, units=A, descr=GFCI current fault value
2018-09-24 09:38:50 [INFO]  map [20] -> code=0x7e, var=error_msg_h, multiplier=None, units=, descr=Failure description for status
2018-09-24 09:38:50 [INFO]  map [21] -> code=0x7f, var=error_msg_l, multiplier=None, units=, descr=Failure description for status
2018-09-24 09:38:50 [INFO] --

Then.. using this mapping to decode the two-byte values from normal_info (using the position as an index to loo up the previous mapping).. normal_info also has 22 fields for mine - which matches the response from query_description.

2018-09-24 09:38:50 [INFO] tx packet - src=0x1, dst=0x10, function=(17, 2), data=[], len=0, checksum=291
2018-09-24 09:38:50 [INFO] tx packet - aa55010000101102000123
2018-09-24 09:38:50 [INFO] expecting reply of (17, 130)
2018-09-24 09:38:51 [INFO] rx packet - aa550010010011822c007c00480b4600110013097d138d01cdffff0002452e00007ae3000100000000ffff000000000000000000000acb
2018-09-24 09:38:51 [INFO] rx packet - src=0x10, dst=0x100, function=(17, 130), data=b'\x00|\x00H\x0bF\x00\x11\x00\x13\t}\x13\x8d\x01\xcd\xff\xff\x00\x02E.\x00\x00z\xe3\x00\x01\x00\x00\x00\x00\xff\xff\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00', len=44, checksum=b'\n'
2018-09-24 09:38:51 [INFO]   [00]->[0x00] Temperature = 12.4 °C (Internal inverter temperature)
2018-09-24 09:38:51 [INFO]   [01]->[0x0d] E-today = 0.72 KW.Hr (The accumulated kWh of day)
2018-09-24 09:38:51 [INFO]   [02]->[0x01] Vpv1 = 288.6 V (PV1 voltage)
2018-09-24 09:38:51 [INFO]   [03]->[0x04] Ipv1 = 1.7000000000000002 A (PV1 current)
2018-09-24 09:38:51 [INFO]   [04]->[0x41] Ipv1 = 1.9000000000000001 A (PV current)
2018-09-24 09:38:51 [INFO]   [05]->[0x42] Vac = 242.9 V (Grid voltage)
2018-09-24 09:38:51 [INFO]   [06]->[0x43] Fac = 50.050000000000004 Hz (Grid frequency)
2018-09-24 09:38:51 [INFO]   [07]->[0x44] Pac = 461 W (Power to grid)
2018-09-24 09:38:51 [INFO]   [08]->[0x45] Zac = 65.535 Ω (Grid Impedance)
2018-09-24 09:38:51 [INFO]   [09]->[0x47] E-Total_H = 0.2 KW.Hr (Total Energy to grid)
2018-09-24 09:38:51 [INFO]   [10]->[0x48] E-Total_L = 1771.0 KW.Hr (Total Energy to grid)
2018-09-24 09:38:51 [INFO]   [11]->[0x49] H-Total_H = 0 Hr (Total operation hours)
2018-09-24 09:38:51 [INFO]   [12]->[0x4a] H-Total_L = 31459 Hr (Total operation hours)
2018-09-24 09:38:51 [INFO]   [13]->[0x4c] Mode = 1  (Operation Mode)
2018-09-24 09:38:51 [INFO]   [14]->[0x78] GVFaultValue = 0.0 V (Grid Voltage Fault Value)
2018-09-24 09:38:51 [INFO]   [15]->[0x79] GFFaultValue = 0.0 Hz (Grid Frequency Fault Value)
2018-09-24 09:38:51 [INFO]   [16]->[0x7a] GZFaultValue = 65.535 Ω (Grid Impedance Fault Value)
2018-09-24 09:38:51 [INFO]   [17]->[0x7b] TmpFaultValue = 0.0 °C (Temperature Fault Value)
2018-09-24 09:38:51 [INFO]   [18]->[0x7c] PV1FaultValue = 0.0 V (PV1 voltage fault value)
2018-09-24 09:38:51 [INFO]   [19]->[0x7d] GFCIFaultValue = 0.0 A (GFCI current fault value)
2018-09-24 09:38:51 [INFO]   [20]->[0x7e] ErrorMesssageH = 0.0  (Failure description for status)
2018-09-24 09:38:51 [INFO]   [21]->[0x7f] ErrorMesssageH = 0.0  (Failure description for status)
2018-09-24 09:38:51 [INFO] kafka send json - {"v_ac": 242.9, "e_today": 0.72, "gz_fault_value": 65.535, "h_total_h": 0, "error_msg_l": 0.0, "timestamp": "2018-09-24T09:38:51.444815", "mode": 1, "i_pv": 1.9000000000000001, "gfci_fault_value": 0.0, "tmp_fault_fault": 0.0, "temp": 12.4, "gv_fault_value": 0.0, "pv1_fault_value": 0.0, "error_msg_h": 0.0, "v_pv1": 288.6, "gf_fault_value": 0.0, "device": "8881500A10B07509", "z_ac": 65.535, "h_total_l": 31459, "f_ac": 50.050000000000004, "e_total_hr": 0.2, "p_ac": 461, "i_pv1": 1.7000000000000002, "e_total_l": 1771.0}

That's my take on it anyway.. The Impedance field seems broken on mine - it returns 0xFFFF (hence the 65.535 Ohms.) Also, I haven't re-assembled the multi-field high/low values yet.

Also, I've trascribed all the definitions in, just incase they pop up.. eg:

data definitions

DataItem(0x00, 'temp',      'Temperature',     0.1,   '°C',      'Internal inverter temperature')
DataItem(0x01, 'v_pv1',     'Vpv1',            0.1,   'V',       'PV1 voltage')
DataItem(0x02, 'v_pv2',     'Vpv2',            0.1,   'V',       'PV2 voltage')
DataItem(0x04, 'i_pv1',     'Ipv1',            0.1,   'A',       'PV1 current')
DataItem(0x05, 'i_pv2',     'Ipv2',            0.1,   'A',       'PV2 current')
DataItem(0x07, 'e_total_h', 'E-Total_H',       0.1,   'KW.Hr',   'Total Energy to grid')
DataItem(0x08, 'e_total_l', 'E-Total_L',       0.1,   'KW.Hr',   'Total Energy to grid')
DataItem(0x09, 'h_total_h', 'H-Total_H',       1,     'Hr',      'Total operation hours')
DataItem(0x0a, 'h_total_l', 'H-Total_L',       1,     'Hr',      'Total operation hours')
DataItem(0x0b, 'p_ac',      'Pac',             1,     'W',       'Total power to grid')
DataItem(0x0c, 'mode',      'Mode',            1,     '',        'Operation Mode')
DataItem(0x0d, 'e_today',   'E-today',         0.01,  'KW.Hr',   'The accumulated kWh of day')
DataItem(0x20, 'sur_temp',  'surTemp',         0.1,   '°C',      'Ambient Temperature')
DataItem(0x21, 'bd_temp',   'bdTemp',          0.1,   '°C',      'Panel Temperature')

.. etc.. There's about 87 of them.

[moalmo2020]

hi jgc234
could you please post the Python-code ?

[jgc234]

I dumped my rough code into https://github.com/jgc234/eversolar .. I'll clean up the code a bit more in the next few days, but it should give you a rough idea how I decoded the fields.