Very nice, Ed.I would suggest taking readings at one-second intervals and use data smoothing (moving average sample size of 60 samples) and that would give you a smoother curve with fewer large jumps in once every 5 seconds per sample as happend especially during the eclipse in your graph. I save my Spectrum Lab data in CSV form and import that into Excel (or other spreadsheet) to plot the data using moving averages and various other data analysis.
There's a total solar eclipse running SW to SE through the US on 8 April 2024 in my mid-afternoon when I'll have just over 75% totality. The CHU time signal in Ottawa on 3.33 MHz will have to pass through totality in southern Ontario to reach me. So hopefully, I will collect some good data.
You didn't say if your receiver was GPS-frequency-locked. My RSPDuo accepts a 24 MHz GPS-locked reference signal, which is generated by a Bodnar GPDSO to lock the carrier and prevent any +/- ppm errors inherent in your receiver that can affect your data. HamSCI requires this in order to analyze data received from hundreds of radio hobbyists sending in their data.
It would also be interesting to compare any differences in received signals using verticals, small loops, dipoles, etc. to see which design(s) would be more sensitive to detecting minute variations. I use a small receiving loop (W6LVP). The advantage of the small loop is the ability to rotate it for maximum signal and minimum noise, which is what we really want. Having a small antenna rotator makes this easy but the "armstong" method and a fixed time signal station means that you can "set it and forget it".
73,
Robert