You don't need Convert to change from an integer to a byte; you can just cast it:
byte s = (byte)sum;
But ... if your sum can be up to 5000, then that will not fit in a byte - which has a fixed range of 0 to 255 inclusive (because it is defined as an 8 bit value, and that is all it can hold).
You need to look at your whole system and work out why you want a byte at all - if the value is in the range 0 to 50900, then your whole system is going to fail if you continue to try an use byte values to hold them.