Shift in Reference spectrum with instrument orientation

[Ken Harrison]

I know that some instruments suffer from a lack of stability which can lead to a shift in the position of the reference spectrum relative the the CCD pixel positions...hence the concept of taking a reference before and after an observation.
My question is - has anyone actually analysed the problem and determined the variables which should be addressed??
Accepting it is one thing - finding and fixing the root cause is another.....

[Ken Harrison]

I have adopted a f7 transfer lens for the reference lamp in the latest batch of Spectra-L200.
This uses a 4mm aperture in the housing, which allows all/ any reference lamp to be used ie Neon, Relco, Flouro, Filly etc.
Testing so far has shown good repeatability - as long as the lamp illuminates the aperture the position of the wavelength/pixel doesn't change.
Onwards and upwards.