and we have measured both x and y and determined their mean values and standard deviations, then we can find the standard deviation in z using the formula below. Note that the derivatives with respect to x and y are evaluated using the mean values of x and y.
Here is a problem for you to try. In
lab #7 you will measure
the charge over mass ratio (e/m) of the electron. You will do so by
shooting
electrons from an electron gun through a uniform magnetic field, which
is
perpendicular to their velocity. The electrons will move in a circle.
You will
measure the magnetic field B, the accelerating voltage of the electron
gun V
(this will determine the velocity of the electrons) and the radius of
the electron
orbit r. You will find e/m from the formula:
a)
( 20 points) Write an expression for the
standard deviation in the quantity e/m in terms of the measured
quantities and
their standard deviations.
Now, divide both sides by f to get a relation between the fractional errors.
b) ( 20 points) If you can measure each V, B and r with an accuracy of 1%, what would be the relative (%) error in your measurement of e/m ?
Period [s] | Length[cm] | |
2.026 | 102.8 | |
2.021 | 102.7 | |
2.024 | 102.7 | |
2.020 | 102.6 | |
2.025 | 102.8 | |
2.024 | 102.6 | |
2.022 | 102.9 | |
2.022 | 102.8 | |
2.024 | 102.5 | |
2.019 | 102.8 | |
2.023 | 102.7 | |
2.024 | 103.0 | |
average |
2.0228 |
102.74 |
STDEV |
0.0021 |
0.14 |
I determined the error using the error propagation formula. After
taking the derivatives, I found that:
Then, I plugged in the numbers above.
This tells me that the probability that my result is consistent with
the accepted value is much smaller than 5 %. Remember that +/- 2
standard deviations corresponds to a confidence interval of 95.4 %,
i.e. - the probability to get outside this interval just by
chance is less than 5% . I also notice that the
accepted
value is quoted with many significant digits ( this means that the
error is just in the last digit), while my results is not so precise.
Since I observe a large deviation from the accepted value, I must
have made a systematic error that I have not taken into
account yet. This can be due to a miscalibrated instrument or a
procedure that gives offset to one or both measured quantities.