Calculating error when constant is divided by variable

[akzo]

Hello everyone,

I have a very simple question but for some reason, I couldn't find the exact answer to it on the internet.
Let's say we have a formula y = k/x where k is constant and x and y are variables. Assume that I am given the absolute uncertainty in x how should I calculate the uncertainty in y?

Thanks in advance.

 

[Vanadium 50]

The relative uncertainty on y is the same as the relative uncertainty on x.

 

[Stephen Tashi]

Assume that I am given the absolute uncertainty in x how should I calculate the uncertainty in y?

Are you using the term "uncertainty" to mean the standard deviation of a random variable? Or are you using "uncertainty" to mean a definite bound without any implication of a probability being involved? (For example, compare the different sets of "rules" given in notes: http://web.mit.edu/fluids-modules/www/exper_techniques/2.Propagation_of_Uncertaint.pdf rules 1&2 vs rules 3&4 )

Is the phrase "absolute certainty" used because you wish to distinguish it from "relative uncertainty" ?

 

[Khashishi]

If you know calculus, just calculate dy/dx and replace dx by your uncertainty in x. This is a general trick you can use to propagate an error through a function. It won't help you to combine two errors though. You need some knowledge of the error distribution for that.