HW #3
- Due Oct 26, 2022 by 11:59pm
- Points 10
- Submitting a file upload
- Available after Oct 18, 2022 at 12am
(L) indicates light HW option (everyone should do), whereas (H) is for those pursuing the heavy HW option.
Problem 1 (L)
Pick a plot you would like to improve, and make it better.
-
- Show what you started with (if you had an original), or describe the problem you were trying to plot.
- Describe the steps you went through to improve your plot
- Show final plot, and describe why it better answers your needs
Problem 2 (L)
Assume you are looking for faint gamma-rays in a gamma-ray spectrometer. The instrument works by looking at 1000k independent energy channels; and recording a hit in each channel. However, as always, there are both instrumental and environmental backgrounds. Assume you have a background rate per energy channel of 5.3 hits/channel in each 1 second measurement period.
-
- Assume you are looking for gamma-rays at a known energy (one channel), and your signal is steady in time. Pick an emission rate (0.1-10 gamma's per second), and calculate how long you would need to average to obtain a 5 sigma detection.
- Assume you are looking for rare bursts of gamma-rays at a known energy. For homework convenience, assume nature has provided gamma-ray bursts that are exactly as long as the integration time you found in part A. If you are searching through 2 months of data, how bright must a burst of gamma-rays be for you to have a 5 sigma detection? Compare to answer in part A.
- Now repeat problem B, but this time you don't know which of the 1000k energy channels the mono-energetic burst of gamma-rays might appear in.
- Describe (no need to actually do the calculation), how you would search for a burst of unknown duration, and how to determine the trials factor.
Problem 3 (H)
Take some of your data, and come up with a question or 'worry' that would require searching over multiple locations/times/parameters.
-
- Come up with a parametrized model of your background. The more accurate the better, but if this will take too much time feel free to come up with a partial or toy model for this HW, just explain the corners you are cutting.
- Perform the search, including the trials factor. Describe what you find.