Jumps in velocity in tracker 4.85 dynamic modle for free fall.
Ami B
9 Posts
Hi, to the Tracker experts , as high school physics teacher I use Tracker a lot I am trying to study "free fall with air drag" using "Tracker (Window 4.85) " capturing video with mobile camera (30fps). I have some difficulty with the model the velocity is not smooth.
I am using dynamic model with several masses of the same shape attached are the model and graph for 5 gram and 20.3 gram: M (Kg) V (m/s) kl kt Reynolds 0.005 2.10925 0.023231 0 0 0.0203 3.794713 0.034049 0.004843 0.539689
remarks I did excel model using Euler method and got smooth curve. (this is the simplest)
also I suggest to estimate the error by cheking the differnce of the model and the real X,Y values and Velocity .
also may be using more points for the model will help ( half or third of the time)?
Re: Jumps in velocity in tracker 4.85 dynamic modle for free fall. -
Douglas Brown
450 Posts
Hi ??
My drag_viscous.trz file with your model and no video at all is at http://www.cabrillo.edu/~dbrown/tracker/test/drag_viscous.trz.
Your freeFall_5g.trz file wouldn't open for me and was not even recognized as a valid zip file, so it must have gotten corrupted.
Open my drag_viscous.trz and check video clip and model settings. They're almost identical to your 5g numbers but you can see the plots are all nice and smooth.
The difference is in the missing times that you see clearly on your position plot. I suspect your video camera has dropped occasional frames (many do this automatically when they can't keep up 30fps). This means the deltaT is not constant which is not what the dynamic model assumes.
Using a video with no dropped frames would fix the problem...
Hope this is helpful. Doug
PS I see your name as ??? ??????. Is this just my computer? Can you change it to something more widely readable?
Re: Re: Re: Re: Re: Jumps in velocity in tracker 4.85 dynamic modle for free fall. -
Ami B
9 Posts
loo,doug thanks very much . both are rigth the problem is becuase of the cammera. using an older mobile cam with constant times the problem disappeared.
I have some questsions/suggestions for future releases: 1. loo how did you change the times for the model - this could be a good solution? 2. suggest that you add pair wise least square differece to estimate the model fit. 3. as you see I have 2 constants ( linearly connected Kt=mg/v^2 -kl/v v is the terminal velocity ) i did find the best pair by minimizing the difference /
Re: Jumps in velocity in tracker 4.85 dynamic modle for free fall. -
Douglas Brown
450 Posts
Hi Ami,
Glad to hear you were able to fix the problem by using an older camera with constant times.
I like your suggestion to add an easy way to find the rms deviation between a model and marked points. I'll see if I can add that to a future release :-)
