This function returns the minimum scale needed to represent the value exactly. This is equivalent to the number of fractional decimal digits in the value excluding trailing zeroes.