Today is...
Sunday, December 16, 2018
Welcome to the Modbus Community, about
the world's leading automation protocol.
How to get Modbus slave legal adresses (Modbus TCP)?
How to get which Modbus server (slave) registers can be read/written on a device providing it?

Hi all,

I searched the web and the forums here but didn't find the answer to my question, may have missed it, sorry if so.

Context is OpenModbus on TCP/IP.

I'm programming a 16 analog channels acquisition device. This device provides a Modbus server and a list of pre-programmed functions to interact with it (a sort of API).

For example:

A function 'modbus_call_MSXE301x__AnalogInputGetAutoRefreshValues' is encapsulating the call to 'modbus_send_FC3_query' function corresponding to the read multiple registers Modbus function, with the pre-programmed registers dedicated to store input channels values. In my case starting register is 100.

When I use this function, it works just fine and I get the values of each of my analog input channels.

What I want to do for the needs of my programming is, in addition of using the preprogrammed functions, read 2 Modbus registers (set by another device on the modbus slave) and set 3 Modbus registers (which will be read by the other device). All I want to store in these registers is unsigned integer values (16 bits will fit).

How can I know which registers are allowable/readable other than the ones referenced by the device documentation for pre-programmed functions?

Thanks forward for your help, I hope I was clear enough for this to be understandable.

Have a nice day :)

The short answer is that you can't. Each device has a memory map onboard that is setup by the manufacturer. They generally provide this map with the device documentation. Any register that is not documented as available for read/write functions will be reserved for internal use. The only way you can figure out if any of those off grid registers are useful is to try and read/write to each of them and hope that you don't destroy the devices functions in the process. This kind of trial and error is hazy at best and never guaranteed. I advise against it.