Re: [Phys-l] error vs uncertainty

[Bernard Cleyet <bernardcleyet@redshift.com>]

On 2011, Oct 04, , at 09:34, David Marx wrote:

> For example, in the e/m experiment we used to do, we had the electron beam curve around to strike a 
> series of 5 pins at different distances from the e-beam source.  Many students guess that the reason 
> they had a large percent difference between their value for e/m and the accepted value is due to their 
> inability to accurately determine where the pin was relative to the center of the e-beam.  I suggested 
> they go back to lab and make measurements with the beam on each side of the pins to see how much 
> their results varied due to this factor.  They were quite surprised to learn that this was not the source of 
> the error.
>
> David


This was, for a time, an advanced lab. experiment.  Peter Scott's write up is rather detailed in a discussion of error sources.  If I remember correctly, it included methods to reduce the Boltzmann distribution and other methods to reduce the spot diameter.  My principle worry was trusting the mfg.'s measurement of the pins.  I was going to place a large film sheet behind the tube and X-ray it using our GE XRD-5  We'd necessarily evacuate that part of the building, as the source (collimated for back Laué) would be on the other side of the room.  I never calculated if the ratio ~ 0.2m / 10m would be sufficient to render the beam parallel.

bc