Physics307L F09:People/Joseph/Lab2

From OpenWetWare

Jump to: navigation, search

Summary

The point of this lab was to determine the e/m ratio of the electron. e being the charge of the electron, and m being the mass. The procedure and data collection can be found in my notebook.

The best guess we have at the actually e / m ratio for electrons would be: 3.13 x 10^11 C/kg. This value was the least off from the accepted value (78%) even though I could not say what we did differently when collecting this data from the other. Our overall range of data was anywhere from 78% to 130% off. Now this is a large discrepancy, it is surprisingly consistent. This implies some systematic error, which I discuss in the data analysis portion of the lab report.

If I had the time to redo the lab I'd do a couple of things differently:

  • SJK 00:20, 11 October 2007 (CDT)
    00:20, 11 October 2007 (CDT)More data is a good thing.  However there is that  that makes you compete with the law of diminishing returns.  Further, that only reduces the effects of your random error, whereas systematic error clearly dominates.  I think you should spend your efforts on taking more data, but with the purpose of isolating and then minimizing the sources of systematic error.
    00:20, 11 October 2007 (CDT)
    More data is a good thing. However there is that \frac{1}{\sqrt{N}} that makes you compete with the law of diminishing returns. Further, that only reduces the effects of your random error, whereas systematic error clearly dominates. I think you should spend your efforts on taking more data, but with the purpose of isolating and then minimizing the sources of systematic error.
    I'd take a ton more data. We had a total of 16 useful data points. I'd like to have over a hundred. As discussed in class, the central limit theorem illustrates how this would help us get some consistent, usable data.
  • SJK 00:25, 11 October 2007 (CDT)
    00:25, 11 October 2007 (CDT)Good ideas, and worth a try.  I agree, I'd also like to know more about the electron gun...maybe we could study it with the lights on, take pictures or something?
    00:25, 11 October 2007 (CDT)
    Good ideas, and worth a try. I agree, I'd also like to know more about the electron gun...maybe we could study it with the lights on, take pictures or something?
    I'd like to isolate this experiment a bit more. Move it to a location away from any interfering light sources, or possible magnetic influences. Set up the power sources a good distance away and isolate the apparatus as much as possible.
  • I'd like to know what the electron gun is made up of. This would help us determine a work function that we could account for and reduce our error by truly understanding what phenomena we are manipulating, and what phenomena is just something to deal with.


Overall, this is a great lab. Its difficult and inaccurate, but teaches you a lot about questioning your data and procedure. Its better for learning the actual process of taking good data, than for getting the true answer. This is the kind of lab where you make your mistakes so you don't make them so easily in the future.SJK 00:29, 11 October 2007 (CDT)
00:29, 11 October 2007 (CDT)Those are great lessons to learn!  Overall a great job with this lab: you were creative in the lab and writeup with ideas for how to do it better, and I think the data you did take were done carefully.  Take a peek at Bradley's notebook and summary for my comments which would apply to you as well.  Also, this important note on his summary page:Even though you know your number is way off the "accepted value", you need to report your uncertainty due to random error! Or, you could even report some estimate of your systematic error, if you had a way of doing that.  Otherwise, you really have no way of claiming that your answer is "wrong," because if your error bars were +100% / -50%, then the accepted value would be in that range.  Well you see what I mean.
00:29, 11 October 2007 (CDT)
Those are great lessons to learn! Overall a great job with this lab: you were creative in the lab and writeup with ideas for how to do it better, and I think the data you did take were done carefully. Take a peek at Bradley's notebook and summary for my comments which would apply to you as well. Also, this important note on his summary page:

Even though you know your number is way off the "accepted value", you need to report your uncertainty due to random error! Or, you could even report some estimate of your systematic error, if you had a way of doing that. Otherwise, you really have no way of claiming that your answer is "wrong," because if your error bars were +100% / -50%, then the accepted value would be in that range. Well you see what I mean.
Personal tools