A practical tip about missed strokes, noise and rower placement

[JaapvanEkris]

I had quite an issue on my rower with missed strokes. Looking closely at my recordings I understood why: instead of being a nice smooth wave-pattern, the recorded data contained a lot of "spikes" on top of the waves. This throws off measurements and that is what drove me to develop the noise-filter and stroke-detector, which I designed to keep finding that wave regardless of that noise. I always assumed that the spikes originated because the Raspberry Pi was too busy to handle them at the right time and couldn't be fixed.

But in an unrelated matter, my rower moved about too much on our slippery concrete Attic floor, scratching it in the process. So I bought a soft mat to prevent that (simple puzzle-like type, 1 cm foam stuff). And now almost all noise has gone in my recordings: no more noise, no more missed strokes! Seems that the rower had too much vibrations to reliably detect the magnets in time, instead of the Raspberry being too late. As these mats are relatively cheap and also reduce mechanical noise which is transferred to the structure of my house, it might be a consideration for the rest.