I found my problem. The scaled_float parameter isn't the number of figures after the decimal point, it's just the number to multiply the input value by - so if you want 6 digits of precision you need to use 10^6 or 1000000 in this field.
1 Like