Calculating error in pendulum motion

10 replies · 3K views · last reply
asadpasat
Messages
41
Reaction score
1

Homework Statement


I have to calculate the propagated error on g of pendulum. I pretty much measured the T of pendulum and now calculating g while increasing the number of cycles.

Homework Equations


I used the equation of propagated error and i included picture of it and my calculations.

The Attempt at a Solution


I plugged all the values into the equation but what bothers me is the size of the error. For 20 cycles I get g=9.776, but my error is 0.02. This seems like a huge error, and would mean that my g would equal to 9.78. Anyone sees any mistake?
 
Attachments
  • Handwritten error propagation calculations for a pendulum experiment across cycles 2, 5, 10 and 20, showing L = 0.5135 m, uncertainties in T, and resulting uncertainties in g boxed at the end of each.
    error1.JPG
    58.7 KB · Views: 517
To avoid a crick in the neck
 
Attachments
  • Handwritten notes on lined paper calculating uncertainty in g for four pendulum cycles, using delta g = sqrt((4pi^2/T^2)^2 (0.0006)^2 + (-8pi^2 (0.5135)/T^3)^2 (T error)^2), with results 0.2, 0.1, 0.03 and 0.02 m/s^2.
    Pend error prop.png
    58.5 KB · Views: 489
TSny said:
To avoid a crick in the neck
Thanks! Makes it more convenient.
 
Here is picture of the equation and detail calculations.
* I don't know how to rotate the picture, sorry.

Edit: Image rotated by moderator:

Handwritten derivation of uncertainty in g for a pendulum: g = 4π²Lp/T², Lp = 0.5135 ± 0.0006 m, T = 1.21 ± 0.01 s, with partial derivatives and the error propagation formula for δg, ending with δg = 0.00746.
 
Attachments
  • Handwritten derivation of uncertainty in g = 4π²Lp/T² using partial derivatives, with L = 0.5135 m ± 0.0006 m and T = 1.21 s ± 0.01 s, ending with δg = 0.00746.
    IMG_2593.JPG
    43.7 KB · Views: 503
asadpasat said:
* I don't know how to rotate the picture, sorry.
Most platforms have ways to rotate an image. What type of computer are you working from?
 
asadpasat said:
I plugged all the values into the equation but what bothers me is the size of the error. For 20 cycles I get g=9.776, but my error is 0.02. This seems like a huge error, and would mean that my g would equal to 9.78. Anyone sees any mistake?
I don't see any mistakes (but I didn't check all your calculations). Why do you think your error is "huge"?
 
Here, I figured it out.
 
Attachments
  • Handwritten derivation of uncertainty in g for a pendulum: g = 4π²Lp/T², with Lp = 0.5135 m ± 0.0006 m and T = 1.21 s ± 0.01 s, partial derivatives, and propagated error δg = 0.00746.
    IMG_2593.JPG
    28.1 KB · Views: 452
TSny said:
I don't see any mistakes. Why do you think your error is "huge"?
Well I think it's huge because it would mean that I can trust my g only to 3 sig figs. So I pretty much throw away one sig fig.
 
3 sig figs seems OK to me. For 20 cycles you have a fractional error in T of about .001/1.44 ≅ .0007. But note that T is squared in the formula for g. So, the contribution of the error in T to the relative error in g should be roughly 2(.0007) ≅ .001. The contribution of the error in L to the relative error in g is roughly .0006/.51 ≅ .001. So, I don't think it's surprising that when you use your more accurate formula to determine the error in g, you find that you are only getting 3 sig. figs. for g.
 
asadpasat said:
Here, I figured it out.
Handwritten error propagation calculation for g from a pendulum: square root of (4π²/(1.21)²)²(0.0006)² + (−π²(0.5135)/(1.21)³)²(0.01)², with intermediate values 2.6173... × 10⁻⁴ and 5.3063... × 10⁻³, and the result δg = 0.00746.

You made an error in grinding the numbers here.