# Homework Help: How to Combine Gradient Uncertainty with other Uncertainty?

Tags:
1. Mar 31, 2016

### Banker

1. The problem statement, all variables and given/known data
I did an experiment to measure the speed of sound(using two microphones and a hammer). I changed the distance between the two mics and calculated(using a fast timer) the time taken for the sound to reach from the start mic to the end mic. I made a graph(distance on x axis, time on y axis) on excel using my results and added a line of best fit. I need error bars and the uncertainty in the gradient. Also, I need to combine the uncertainty from the gradient with the random uncertainty, calibration and scale reading uncertainty(from meter stick). How can I do this?

2. Relevant equations

3. The attempt at a solution
I know the formula for random uncertainty and the Pythagoras-like formula for combining uncertainties. I just don't know how to combine all of this with the gradient uncertainty.

Last edited: Mar 31, 2016
2. Mar 31, 2016

### Staff: Mentor

In general: include uncorrelated uncertainties in the individual datapoints, make sure your fit takes those uncertainties into account (not sure if excel can do that). Correlated uncertainties need a different approach.

3. Mar 31, 2016

### Banker

@mfb Thanks for the reply I did a little experimenting and I combined the random uncertainty for each of my times with the digital reading uncertainty(using ∆w^2 = ∆x^2 + ∆y^2 + ∆z^2, x = random uncertainty, y = scale/digital reading uncertainty, z= calibration uncertainty ) and also did the same with my distances. I then plotted these in my excel graph as a custom error bar for each of my points. Is this the correct way to go? How would I go about finding the uncertainty of the gradient now, with the vertical and horizontal error bars in my graph too?

4. Mar 31, 2016

### Staff: Mentor

Excel has a function for the uncertainty of parameters of linear functions, I don't know if you can also directly get them from a trend line, and I don't know if the uncertainties are taking into account properly (change them to see if the result changes).