1. Not finding help here? Sign up for a free 30min tutor trial with Chegg Tutors
    Dismiss Notice
Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Question about uncertainties

  1. Sep 26, 2008 #1
    (First post. Go easy on me, mods :D)
    EDIT: Seems I got the wrong forum. If a mod could be so kind to move it, I'd appreciate it :P

    Hi everyone!

    I'm working on a formal lab report for my physics class, and after propagating my uncertainties into a formula, I got an even smaller uncertainty relative to the original uncertainties (one of the indep. variables was 9% while my propagated uncertainty was 6%)

    Is this possible in ANY case?

    If it's not, I will post more information about the equations I'm using. I already spent 2 hours looking at all the numbers and using mathematica to calculate the results for me, and I just don't know what could be wrong with it (if there's even anything wrong with it).

    Thanks for your help,

    KodeK
     
    Last edited: Sep 26, 2008
  2. jcsd
  3. Sep 26, 2008 #2

    mathman

    User Avatar
    Science Advisor
    Gold Member

    Yes. In general if the random variables are independent, the uncertainties tend to smooth out.
     
  4. Sep 26, 2008 #3
    Hmm, I'd like to show you the equations I'm using just to get an okay so I know I'm doing this right. How would I copy mathematica's LaTeX output and paste it on the site? I don't want to have to copy an in-line equation :P
     
  5. Sep 26, 2008 #4
    Here's a screenshot of the problem I'm having:

    [​IMG]
     
  6. Sep 26, 2008 #5
    Yes, the errors can become less if the derivatives are small. If you evaluate a function f of independent variables x1, x2, ..., with respective errors dx1, dx2, ..., then the error in f is:

    df = sqrt[df1^2 + df2^2 + ...]

    where

    df1 = f(x1 + dx1/2, x2,...) - f(x1 - dx1/2, x2,...)


    df2 = f(x1, x2+dx2/2,...) - f(x1, x2 - dx2/2,...)

    etc.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?