Wavelength Shift vs Temperature in LHiRes III
Posted: Thu Feb 26, 2015 7:19 am
Just some quick thoughts regarding the wavelength drift in the LHiRes III wrt temperature changes;
The micrometer that is used to adjust the grating turret has a spindle that appears to be some sort of stainless steel. Looking up the coefficient of thermal expansion for stainless steel I found various values (depending on the type of stainless) from around 10 to 17 *10e-6 meter/(meter*T) where T is degrees K or C. So I assumed an average value of 13.5 e-06. It appears the spindle is around 50 mm long. So, for each degree Celsius it should expand (or contract) by:
dL = 0.05m * 13.5e-6 m/(m*C) = 6.75e-07 meters/C = 6.75e-04 mm per degree Celsius.
I see that adjusting the micrometer by 10mm changes the central wavelength (using the 2400 line/mm grating) by about 3500 Angstroms - so about 350 A/mm.
So, 350 A/mm * 6.75e-04mm = 0.23 Angstroms. So I'd expect to see a wavelength drift of around a quarter Angstrom for each degree Celsius change in temperature. Now, I don't have any way to measure the actual temperature of the micrometer but looking at the wavelength drift I've seen as a function of air temperature I'm seeing about half of the value calculated, above. The spectrograph is attached to a substantial mass of aluminum, steel, and concrete that surely cools and warms much more slowly than the air, so that's not too surprising.
Can you tell it's been cold, cloudy, and snowing here of late?
Mike
The micrometer that is used to adjust the grating turret has a spindle that appears to be some sort of stainless steel. Looking up the coefficient of thermal expansion for stainless steel I found various values (depending on the type of stainless) from around 10 to 17 *10e-6 meter/(meter*T) where T is degrees K or C. So I assumed an average value of 13.5 e-06. It appears the spindle is around 50 mm long. So, for each degree Celsius it should expand (or contract) by:
dL = 0.05m * 13.5e-6 m/(m*C) = 6.75e-07 meters/C = 6.75e-04 mm per degree Celsius.
I see that adjusting the micrometer by 10mm changes the central wavelength (using the 2400 line/mm grating) by about 3500 Angstroms - so about 350 A/mm.
So, 350 A/mm * 6.75e-04mm = 0.23 Angstroms. So I'd expect to see a wavelength drift of around a quarter Angstrom for each degree Celsius change in temperature. Now, I don't have any way to measure the actual temperature of the micrometer but looking at the wavelength drift I've seen as a function of air temperature I'm seeing about half of the value calculated, above. The spectrograph is attached to a substantial mass of aluminum, steel, and concrete that surely cools and warms much more slowly than the air, so that's not too surprising.
Can you tell it's been cold, cloudy, and snowing here of late?
Mike
