# NTC error propagation: what I missing?

Reading through a Vishay datasheet of a NTC, I found the guide to devise the temperature error:

I don't understand why the errors of X and Y, for the non-approximated formula, were multiplied instead of summed. I am sure it is a silly thing that I forgot.

Both effects are accounted independently. The effect of the B-Tolerance is relative to the actual value already corrected by the X tolerance. As an example: Assuming X is 8% and the effect of the B-Tolerance is 10%. The X value would account for +8% and the Y value would have be calculated on the already increased value by the X term. Resulting in +18.8% not the 18% as you expected.At the end of the day, given the range of X and Y, you can approximmate by X+Y, as the equation suggests, withou t big impact in the result. In Math language: $(1+x)(1+y)-1=(x+y)+xy=x+y$ if $xy\rightarrow 0$