On each Channel class, the wavelength range is hard-coded at $0\lambda90$ angstroms. This ensures that every line that could fall on the detector is included in the instrument spectral cube. However, this overcompensates and results in a lot of wasted computation in projecting lines off the edge of the detector, particular at the higher orders.
Instead, the maximum wavelength range should be estimated as approximately the highest wavelength that originates from the far limb of the sun that would fall at the edge of the detector.
Similarly, while the wavelength resolution is already dependent on the order. For the zeroth order, we should just not bother with the factor of 2.
On each
Channel
class, the wavelength range is hard-coded at $0\lambda90$ angstroms. This ensures that every line that could fall on the detector is included in the instrument spectral cube. However, this overcompensates and results in a lot of wasted computation in projecting lines off the edge of the detector, particular at the higher orders.Instead, the maximum wavelength range should be estimated as approximately the highest wavelength that originates from the far limb of the sun that would fall at the edge of the detector.
Similarly, while the wavelength resolution is already dependent on the order. For the zeroth order, we should just not bother with the factor of 2.