L1 error (sum of the absolutes of the differences) for a chosen species and process when changing the number of steps n = {1, 2, 4, 8, 16} for each solver for the current NuPPN implementation. Each solution is compared to the solution yielded by the BD solver for n = 32 steps with a tolerance of 10⁻³. Left panel: final abundance for ¹⁶O in a 12 M_⊙ He-burning trajectory. Right panel: final abundance for ¹³⁸Ba for a ¹³C pocket trajectory. The BD and PED converge to the same solution. The slower convergence of the BE solver is due to its lower convergence order.