Modbus data types

The data type is the third argument of MODBUSR, MODBUSW and MODBUSWNE and of UNPACKVALUE and PACKVALUE. It decides how many 16-bit registers a value takes and in which order its bytes are stored. Data type names are not case sensitive.

Base data types#

Each base data type takes a fixed number of 16-bit registers and gives the script a value of a fixed type.

Many devices store a decimal value as a whole number with a fixed scale — divide it after reading, for example MODBUSR(H, 20, Int16) / 10. BCD-coded values are converted with FROMBCD.

Byte order variants#

Every type exists in a big-endian and a little-endian variant; the 32- and 64-bit types also in byte-swap variants.

1
2
3
4
5
6
7
8

BigEndianInt16      LittleEndianInt16
BigEndianUint16     LittleEndianUint16
BigEndianInt32      LittleEndianInt32       BigEndianInt32ByteSwap      LittleEndianInt32ByteSwap
BigEndianUint32     LittleEndianUint32      BigEndianUint32ByteSwap     LittleEndianUint32ByteSwap
BigEndianFloat      LittleEndianFloat       BigEndianFloatByteSwap      LittleEndianFloatByteSwap
BigEndianInt64      LittleEndianInt64       BigEndianInt64ByteSwap      LittleEndianInt64ByteSwap
BigEndianUint64     LittleEndianUint64      BigEndianUint64ByteSwap     LittleEndianUint64ByteSwap
BigEndianDouble     LittleEndianDouble      BigEndianDoubleByteSwap     LittleEndianDoubleByteSwap

Remarks

• Reads and writes apply all flags the same way, so writing a value and reading it back with the same type gives the same value. On cores older than 2026.1, MODBUSW and MODBUSWNE ignored ByteSwap for Int32 and UInt32.
• When a value reads as nonsense, try the variants in the manual Modbus operations of the app, which switch the data type over the same registers.
• In a byte collection the flags mean byte order instead — see Byte order in collections.

Example

MODBUSR(H, 0x0100, LittleEndianFloat);
MODBUSR(A, 30, BigEndianUint32ByteSwap);

64-bit data types#

Int64, UInt64 and Double take four registers; Int64 and UInt64 are read as 64-bit whole numbers, exact over the whole range, because a double holds whole numbers exactly only up to 2^53.

Remarks — A 64-bit value keeps all 64 bits in the bit functions, so bits 32 to 63 of an alarm word are reachable.

Example — energy counter in Wh, shown in kWh

VAR energy := MODBUSR(H, 100, UInt64);
energy / 1000;

Errors — Modbus data type ‘X’ not supported. Please update your core unit to be able to use this feature. — when the script is saved for an older core.

Available — Core 2026.1, build 40595 or newer.

String#

String reads text stored two characters per register, the high byte first; the number of registers is the fourth argument.

Remarks

• MODBUSR reads 0 to 123 registers; the text ends at the first zero byte. Each byte is one character.
• String can only be read. MODBUSW and PACKVALUE do not write it.

Example

MODBUSR(H, 14, String, 4);
# Result: text of up to 8 characters, for example a serial number

Writing values#

MODBUSW, MODBUSWNE and PACKVALUE convert the value to the data type before writing.

• Integer types round to the nearest whole number; a value exactly halfway goes to the even number, so 2.5 is written as 2 and 3.5 as 4.
• A value outside the range of Int16, Uint16 or Float stops the script: Overflow when writing X as UInt16.
• NaN can be written only to Float and Double. To an integer type it stops the script: RegisterValueToWrites: Can’t write double.NaN to modbus register (MODBUSW, MODBUSWNE) or PACKVALUE: Can’t write double.NaN to modbus register.
• Bool writes 1 for any value other than 0. TRUE and FALSE can also be written to number types as 1 and 0.

Example — set point with one decimal place

MODBUSW(H, 20, Int16, ROUND(Te * 10));

See also — Modbus functions, Modbus scripts and prefetch, Binary data.