Frequency-based fatigue methods have gained traction due to increased computational efficiency compared to time-based fatigue approaches. Multiple methods exist, and studies have shown that the S-N curve slope and the analyzed stress response power spectral density affect the performance of the methods. The present study focuses on weld fatigue, where only the slopes of 3 and 5 are relevant for specific applications. Power spectral densities are obtained experimentally by measuring with strain gauges located close to welds on an agricultural mower. Using these S-N curve slopes and spectra, the damage values from well-established uniaxial frequency-based fatigue methods are compared against the time-based approach of rainflow counting and Palmgren-Miner damage accumulation. Generally, the results show that the frequency-based methods perform better for low values of the S-N curve slope and low spectral widths. Following from the results of this study, the best performing methods are: Tovo-Benasciutti, Park, Dirlik, and \(\alpha _ {0.75}\) .