In addition, the mean was also very close in both cases. The only difference came with the standard deviation, which was lower for the combined data set. This makes sense as the increased amount of data collected allowed the data to fall closer to the average as seen in a normal distribution. The other difference may have come in the weight of the cans, as we only measured the weight of one empty can. Figure 2: Histogram for Combined Data Set. The bin sizes for the histogram in figure 2 were determined automatically by Excel. In this histogram, there are no clear outliers.
Figure 3: Histogram for Load Cell 2’ The data from the histogram in figure 3 is normally distributed. This is because the cumulative percentage line is relatively linear.
The overall uncertainty for the data set was calculated using the following …show more content…
However, there were some errors present in the experiment. The first error would come in the measurement of the weight of the can. Although we did our best to completely empty the liquid from the can, it was difficult to dry the inside as it was not easily accessible. To rectify this, we drained the can as much as we could, and then let it air dry for some time. In addition, there may have been errors in calibrating the load cells themselves. This is because all the load cells had different calibration factors in the Arduino code, but they all had perfect outputs when measuring with the calibration weights. This was likely due to the load cell being unable to measure past 3 decimal points. Conclusions In conclusion, the listed weight on the can is accurate to the measured weights. This is because the listed volume and weight on the can falls within the range of the final measured value and its overall uncertainty. In addition, it was concluded that all the load cells are consistent with one another as the resulting values, the f statistic for the combined data set, was less than the critical f value.