When I mention the data type as fixdt(0,8,3) in a constant block, The total wordlength is 8bits, fractional length is 3 bits. So the remaing 5 bits allocated for integer part. So If I enter any value greater than 32, simulink throws overflow error. This is clear for simulation. However when I generate code for a particula simulink block with the datatype mentioned as fixdt(0,8,5) The corresponding signal is assigned with the datatype uint8 and no restriction on bitsize. It seems the signal with datatype fixdt(0,8,3) can hold the value 255 even. Making the datatype fixdt(0,8,3)/ uint8 has no difference in the genrated code.