search

ljean / modbus-tk

Create Modbus app easily with Python
Other
566 stars 212 forks source link

"unpack requires a buffer of 4 bytes" when trying to read more than 2 registers over FC4 over TCP #117

Closed ElspecAuto closed 2 years ago

ElspecAuto commented 5 years ago

Hi, I have a PLC that "talks" Modbus TCP. I'm accessing it with 2 Modbus simulator and able to read blocks of data with 4,6, and even 12 registers on the same request without any problem. I have a simple Modbus client written in Modbus_tk (wrote it 2 years ago), that used to be able to do the same. Today, after upgrading my Python from 3.4 to 3.67 and my Modbus_tk to the latest version, i can read 1 float parameter (=2 registers) easily, but when asking for 4 registers - i get this message: "unpack requires a buffer of 4 bytes". (just to clarify: those next registers are also valid and if i'm asking them without the 2 before them - it'll also work)

my code: start_address = 383 num_of_registers = 2 Modbus = ModbusClient.TcpMaster(host= IP,port=Port,timeout_in_sec=Timeout) Modbus.open() Modbus.execute(slave=159,function_code=4 , starting_address=start_address , quantity_of_x=num_of_registers ,data_format='>HH')

any ideas?

ljean commented 5 years ago

Hello, You code looks good. It's difficult to debug like this but maybe, you should enable logging (see https://github.com/ljean/modbus-tk/blob/master/examples/tcpmaster_example.py) and check what is returned by the PLC. I hope it helps Best luc

meetAndEgg commented 2 years ago

I also encountered the same problem. Have you solved this problem?

ElspecAuto commented 2 years ago

Hi, Actually i did very recently.. I found out that in some PLCs you can simply send it as None and it works, and if there is a need to format it - you should have format per register like so: data_format = '<'+''.join(['HH']*num_of_registers)

Hope it helps..

meetAndEgg commented 2 years ago

thanks! I just solved this problem! At first, I realized that this was not an issue with Modbus-TK, as I also had errors copying the output directly to unpack. And then, I found the same thing you found: we should have each data correspond to one character in the data_format!