How to prevent implicit promotion from single to double when using CAN Pack block?

I am using the CAN pack block for my embedded system, and am trying to figure out how to prevent an implicit conversion from a single to a double in the code generation. When trying to evaluate the conditional statement (outValue > 2147483647.0), the packed variable gets implicitely promoted to a double.

I had a similar issue with the CAN unpack block assuming all variables would be doubles, and I was able to overcome this by using the signal specificaion block as shown here: https://www.mathworks.com/help/vnt/ug/control-can-unpack-block-output-type.html However, using the signal specifications for CAN packing does not achieve similar results

Answer 1

[Fangjun Jiang]

This code is likely coming from an implicit "single" to "int32" data type conversion operation, as the data value range of a "single" is larger than that of an "int32".

Also, "packedvalue" is "int16". Its range is not sufficient for the range of "int32" even with the factor of 0.1.

Basically, you need to arrange and figure out the proper data value range and data type, to avoid this type of data type conversion that needs to check the boundary values. Adding a saturation block outside of the CAN Pack block might be a solution.

intmax('int32')

ans = int32
2147483647