Thursday, December 23, 2010

Toasting Time is Still not Linear

Well, I've finally completed my toaster experiment.  I went through three loaves of bread and four or five days.  My family is happy to have me back.

First, a bit about the toaster.  The settings run from zero to five on the toaster.  Between each number, the dial is divided into four equal parts.  To get the toaster setting for the data, I took the setting on the dial as a fraction and multiplied by four.  That is why the data runs from 1 to 19.5 in the spreadsheet.  I used 19.5 instead of 20 because the dial wouldn't turn that far.
The International Standard Toaster

I made sure to go through each of the settings on the toaster.  The order was scrambled by a Python program.  I did write down one of the numbers incorrectly, so I did four trials on setting four, and two trials on setting fourteen.

I used a stop watch to time the toaster.  I rounded the toasting time to the nearest second.  After each trial, I let the toaster cool to 27 degrees Celsius.  That is to ensure a consistent starting temperature.

For each trial, I weighed the bread before and after toasting.  This is to measure the percent loss of mass during toasting.  My hypothesis is that most of the mass loss is due to the water in the bread evaporating in the toasting process.

The raw data can be downloaded from Google Docs at this link.  The graphs of the data are below.
Time (sec) vs. Toaster Setting
Percent Decrease in Mass vs. Toaster Setting
Percent Decrease in Mass vs. Time (sec)
I haven't applied any regressions to the data yet.  I am planning on doing so in the next couple of days.  I will let you form your own opinions first.  The only conclusion I will draw at the moment is that the toasting time is not a linear function of the toaster setting.