MODBUS TCP/IP communication reading register blocks

[TomaszDEBICA]

Hello,

Congratulations on the good work.

I have a question.
How can I configure a communication channel in MODBUS TCP/IP?
I have a device that supports 50 registers in one query.
In the application, I need to read 130 registers from the device in 3 queries.
How can I configure the communication channel to separate the queries.

Thank you for your answer.

[unocelli]

Hi, I think that you can enable fragmented or define 3 devices

[TomaszDEBICA]

Hi,

After enabling the FRAGMENTED option, data above 50 registers is still not read. And a request to a device above 50 registers blocks communication.

I also tried to add a second communication channel, but the device from which I read variables has only one socket.

[unocelli]

I'm not sure if it work but you can try to set the TOKEN_LIMIT to 40 in FUXA\server\runtime\devices\modbus\index.js
line 11: const TOKEN_LIMIT = 40; // 100

[henjoe]

I'm not sure if it work but you can try to set the TOKEN_LIMIT to 40 in FUXA\server\runtime\devices\modbus\index.js
line 11: const TOKEN_LIMIT = 40; // 100

Does this work?

[TomaszDEBICA]

Ok I will try.

Thanks for your help

[henjoe]

Hi I wonder if setting the token limit to 40 works on you?

[GreenDragonSlayer]

Hello,

I have a similar problem with modbus TCP.
I am using a Codesys based PLC, V3.
I want to read
240 coils which I placed with byte addressing from %MB0.0 to %MB30.x and then I have
133 words from MW100 (fuxa addr 400101) to MW233 and finally some double words from MW234 and onward.

I am autogenerating the configuration and it imports the device and tags fine and the addresses seem to match.
I can read the coils, but no other types (words and double words/floats etc.) at the same time.
Meaning that I can not read coils, words and dwords. I tried reorganizing variables in different ways, mixing as autogenerated and in segments for each type.
I changed the token length to 300 which allowed me to read beyond 100 coils. But I can only read one data type, the coils. The words and double words will not read. I have tried many different combinations and also each datatype alone.
I enabled the fragmented option, but it does not change anything, I still cannot read words or double words.
I tried adding 3 devices, each reading the same ip address but different memory type (coils, MW, MD), but this approach does not work, I believe because they are reading the same IP address.

I compared my output to the output from an export file I generated to see the format and everything looks good and is imported correct.

Could it be my addressing that is off? I would think that I would just be reading weird values if it was off, but I get this error from the console:
"DATETIME [ERR] 'PLC' _readMemory error! [object Object]"

My import file output for double words:
T@,d_4e589200-7a620024,t_78447ac9-39bda5a5,RHV_PRI_MV1_ACTT_IN,,Uint32,400000,102,,,,,false,500,
T@,d_4e589200-7a620024,t_481d6642-508443cb,RHV_PRI_TFREM_HYST_LIMIT_IN,,Float32,400000,104,,,,,false,500,
T@,d_4e589200-7a620024,t_15838032-7baeb588,RHV_PRI_TFORW_RAWMAX_IN,,Float32,400000,106,,,,,false,500,
T@,d_4e589200-7a620024,t_856fe3bd-1de85f62,RHV_PRI_TFORW_RAWMIN_IN,,Float32,400000,108,,,,,false,500,
T@,d_4e589200-7a620024,t_da96f298-5b0a3d9e,RHV_PRI_TFORW_HYST_IN,,Float32,400000,110,,,,,false,500,

I tried deleting the last variables in the scada read list to check if I was reading outside the memory area, but no difference.

[GreenDragonSlayer]

Well... this is weird.
Now I seem to have communitation, but not for Uint32 types.
I think it helped to restart the system,

[GreenDragonSlayer]

I wuld like to know if there ar eany special considerations when using bool, word and double word size variables?
What is the best way to structure the memory mapping?