The number of bits you say indicates the length of a number or variable in PLC, and it can also be considered as the size range of the number, which determines the number of registers occupied. For example, if your physical register is 8 bits, then 16 bits occupy 2 registers, and the word length is one word; If it is 32 bits, there are 4 registers with word length of 2.

You'd better operate on numbers with the same data type and word length, or you may make mistakes.

For example, you define a BCD number of 16 bits in PLC, such as D 1, and a 32-bit unsigned integer, that is, an unsigned long integer, such as D2 (actually D2 occupies the words D2 and D3), that is, four consecutive registers. D 1 is directly added to D2 without data conversion, and some PLCs may make mistakes, because it is possible to add D 1 to the high 16 bit of D2, which is a low-level error of PLC manufacturers, but some PLCs do have such errors. You'd better convert D 1 into a 32-bit unsigned long integer, such as D5, and then add D2 to ensure this. Having said so much, I don't know if you understand. If you have time, you'd better look at the principle of microcomputer and you'll understand after reading it. It must be good for the future!