Slave ID. The Slave ID is what is called the "slave address" in the original Modbus specification, and is called the "unit identifier" in the Modbus/TCP specification. It should be in the range 0 to 255.

A zero slave identifier is used in write requests as a broadcast identifier (each slave that receives the message should store the values supplied in the request, but should not normally send a response). The serial specification limits the range of slave identifiers to 0 to 247, for no apparent reason. The Modbus/TCP implementation guide suggests that Modbus/TCP slaves that can be uniquely identified by their IP address should use slave identifier 255.

You may enter a slave identifier that is different from the slave identifier configured in General settings. In particular, you may enter 0 for the slave identifier in order to send broadcast messages.

Address. Address of the first register to be transferred, in the range 0 to 65535. You should enter message addresses into these pages - the Address mapping settings are not used.

The manuals for many Modbus devices adopt the confusing convention of adding an offset to Modbus addresses to indicate the type of variable:

The manuals for some Modbus devices simply add an offset of 1 to all addresses, so that, for example, "Holding Register 123" would have an address of 122.

ModMaster requires actual Modbus addresses (as sent in Modbus messages), so you may have to subtract the appropriate offset from the addresses in device manuals.

Count. Enter the number of values to be transferred - for example, if you want to transfer five 32-bit floating-point values you should enter 5. Note that this count may be different from the number of registers and may also be different from the count that is actually sent in the Modbus message (the settings for 32/64-bit Value Settings are used to convert the count if necessary).

Type. The type (signed integer, unsigned integer, or floating-point) and size (1, 8, 16, 32 or 64 bit) of the value(s) to be transferred. You would normally use Int 16 or Uint 16 for holding registers and input registers, and Discrete (i.e. 1-bit integer) for coils and discrete inputs - the other sizes are non-standard.

Radix. The radix to be used to display the value(s). The radix only affects the display of unsigned integer values; floating-point and signed integer values are always displayed in decimal.