Half of true frequency measured by Spring experiment - Printable Version +- phyphox Forums (https://phyphox.org/forums) +-- Forum: App-Feedback (https://phyphox.org/forums/forumdisplay.php?fid=6) +--- Forum: Bugs and Problems (https://phyphox.org/forums/forumdisplay.php?fid=9) +--- Thread: Half of true frequency measured by Spring experiment (/showthread.php?tid=1748) Half of true frequency measured by Spring experiment - CJRiley - 05-25-2023 I am measuring a simple cantilever system with a 10Hz natural frequency. The Spring experiment raw data clearly shows a 0.1-second period and the autocorrelation plot also indicates a 0.1-second period. But, the results, autocorrelation, and resonance tabs display a 0.2-second period and 5Hz frequency. Is there a bug or something I am missing in using the tool? The Acceleration Spectrum experiment correctly identifies the 10-Hz frequency for the same system and I have confirmed with other utilities. RE: Half of true frequency measured by Spring experiment - Sebastian Staacks - 05-26-2023 You mean that you see a 0.2s period in the graph of the autocorrelation (maximum at 0.2s), but the number below shows 0.1s? Could you share the data or a screenshot of the autocorrelation? RE: Half of true frequency measured by Spring experiment - CJRiley - 05-26-2023 I see a 0.1s period but the number shown is 0.2s. I am attaching screenshots. RE: Half of true frequency measured by Spring experiment - Sebastian Staacks - 05-26-2023 Ah, I see. I am quite sure that the reason is that the acquisition rate is too low in this configuration. Phyphox looks for the first zero crossing of the autocorrelation as t0 and uses it to estimate the location of the first maximum expected in the range 3*t0 to 5*t0. Then it calculates the highest multiple of that maximum that is within the autocorrelation and uses that maximum to determine the periodicity. The problem now is that the second value already crosses zero and I assume that the actual maximum is numerically outside the 3*t0 to 5*t0 range (hard to tell from the screenshot). However, the acquisition rate is artificially limited in this experiment, because older phones (especially several years ago when we released this experiment with phyphox) struggled with calculating the autocorrelation at higher rates while "typical" spring oscillators usually do not reach 10Hz. I have attached a version that does not limit the rate - please try it by either opening the link on your phone or scanning the QR code from the "+" menu in phyphox. RE: Half of true frequency measured by Spring experiment - CJRiley - 05-29-2023 This works! I appreciate you looking into the details so quickly. I have been conducting vibration experiments with much stiffer systems than the loose spring that you demonstrate in the video associated with the spring experiment. This is a very handy way to quickly find a resonance frequency for many systems!