The majority of this week has been spent learning about sources of intensity noise and the methods we use to characterize it. The sources that I focused on were the "electronics noise" associated with various instruments (this week the instrument used was a photodetector, contributing fluctuations in intensity called "dark noise"), shot noise, and the noise produced by source itself (in our case, a 495 mW NPRO Laser).
The theoretical method of characterizing a noisy time series is called a Power Spectrum, which is defined as the Fourier transform of the series' autocorrelation. Intuitively, this is a frequency-domain representation of the extent to which a given frequency is represented in the time series. Experimentally, it is more convenient to take advantage of a couple of theorems relating the Power Spectrum of a time series to its Fourier transform over a finite time interval; The SRS spectrum analyzer that we used this week calculates the discrete Fourier transform of an input signal over a given time interval, squares that value, and divides by the time interval to get an estimation of the Power Spectrum. Typically, we run the signal into the spectrum analyzer for many of these time intervals, and the values computed over each of these intervals are averaged.
We measured the dark noise and the total intensity noise of a 495 mW NPRO laser using a PDA10CS photodetector, and made a theoretical prediction for shot noise based on the output power of the laser. The data we collected is contained in yesterday's entry.
Aidan has been working hard on getting the DAQ to its operating point, which will make measurements such as those described above easier to make.
The plan of action for the upcoming days is to set up a Mach-Zehnder Interferometer for the same laser and characterize the phase noise its beam acquires, hopefully with the assistance of the DAQ.