See the extensive discussion in LBT Technical Memo UA-94-01, entitled "Error Budget and Wavefront Specifications for Primary and Secondary Mirrors" by J. M. Hill, for a more comprehensive discussion of the telescope optical specification.
For active optics we envision a slow (not more frequently that once per minute) correction of the static and quasi-static errors of the telescope optics by the adaptive secondaries. The spatial sampling of the active wavefront sensors provides the limit for how precisely the wavefront can be corrected. At all spatial scales smaller than the wavefront sampling (70 cm for 12x12 slopes) the residual wavefront error is dominated by either the atmospheric wavefront, or by residual figure errors in the primary and secondary mirrors. Atmospheric wavefront errors (with the possible exception of tip-tilt) are not corrected by active optics. According to the formulation of UA-94-01, we would want to figure the secondary to correspond to a structure function of 70 cm r0 (on the physical secondary) with a suitable allowance for scattering losses at small scales. Scaling (x9) this wavefront onto the primary mirror shows us that the contribution to the telescope error budget is completely dominated by what roughness we allow for scattering from small spatial scales on the secondary. The specification for the secondaries defined below represents a higher performance active optics telescope than was envisioned for the seeing-limited telescope of UA-94-01.
For adaptive optics, the atmospheric wavefront is sampled at millisecond intervals and corrected rapidly using the 672-actuator adaptive secondaries. The level of correction and how closely we approach the diffraction limit is limited by the number of photons, the spatial sampling of the wavefront and by the spatial correction from the adaptive secondaries. The quality of the wavefront on spatial scales smaller than twice the secondary actuator spacing (2x31 mm on the secondary scaled to 56 cm on the primary) is set by the residuals in the adaptive secondary figure and by residuals from the uncorrected atmosphere. The goal is that these residual figure errors in the secondaries are never the dominant degradation in the adaptively corrected image. Piero Salinari, Armando Riccardi and Ciro Del Vecchio have calculated that the total residual error after the best possible correction of an R0=45 cm (at 500 nm) atmosphere amounts to 27 nm rms wavefront. We have allocated 1/3 of that error to the figure of the adaptive secondary shells. This residual correction error is all on scales near the secondary actuator spacing (31 mm on the secondary, or scaled to 28 cm on the primary).
In practice, the adaptive goals outlined above proved slightly too ambitious. The revised specification stated below is relaxed by approximately a factor of two from the original goal.
Because the atmospheric structure function contributes relatively little to the specification of the secondary figure, we have opted to return to a conventional specification of the rms wavefront error. An implicit structure function is implied by the dual specifications for active and adaptive optics.
The deformable secondary shells for the LBT672 F/15 adaptive secondaries should be polished to achieve both of the following two specifications:
Both the front and back surfaces of the shell should be ground and polished for strength (grit sizes selected to remove sub-surface damage from the previous grit).
The edges and bevels of the shell shall be finished to a #320 surface or better.
Failing that, the scratch-dig specification for both the front and back surfaces of the shell is 60-10 (after MIL-O-13830A). Our goal is to avoid any defects in the shell that are deeper than 100 microns in order to preserve mechanical strength.
On the optical (front) surface, the total length of scratches which are less than 6 microns wide shall be no more than 30 mm within the clear aperture.
The secondary asphere coefficient (conic constant) is -0.7328 (corresponding to an eccentricity of 0.8560) to make a Gregorian optical train with the focal plane 3.050 meters behind the vertex of the parabolic primary. (The secondary conjugate distances are 13713.7 mm and 1063.7 mm.)
The tolerance from UA-94-01 allows the secondary asphere to vary over the range of -0.7323 to -0.7333. This tolerance can also be relaxed within the bounds of active correction specified above.
The optical axis shall be aligned to the mechanical axis of the shell within 1 mm.
----