• 3.1 Optics

• 3.2 Mirror Support and Handling

• 3.4 Telescope Structure

• 3.5 Enclosure and Equipment Handling

List of Figures:

Figure 1	Figure 2	Figure 3	Figure 4
Figure 5	Figure 6	Figure 7	Figure 8
Figure 9	Figure 10

Abstract

1. Introduction

1.1 Project Status

1.2 Background

Optical problems In pondering the telescope and instrumentation for the last year, we have identified the following ``problems'' with the baseline concept. (Note that ``problems'' means ``lack of perfection'' more than it means ``things that won't work''. The previous design was a perfectly workable telescope, but an unexpected year to think has brought some new insight.)

1. ``Nobody''* wants to use the Bent Cassegrain focal stations. For each type of instrument that might mount at Bent Cass, there is some reason why direct Cassegrain will give better performance. These reasons include: induced telescope polarization, lower throughput from a third reflection, a smaller field restricted by tertiary size, more scattered light from the additional surface, higher infrared thermal background, and a more restricted instrument mount. At some level this is a psychological issue since some other telescope projects have chosen only to work at a Nasmyth focus with three reflections. The major argument in favor of the Bent Cass is to provide a quickly accessible backup or optional instrument to maximize the efficiency and flexibility of the telescope under changing conditions. (* Nobody does not include sub-millimeter observers.)

2. A tertiary large enough to feed a substantial (any) F/5 wide field to Bent Cass introduces unacceptable central obstruction at F/15 and F/33. This leads naturally to a two tertiary concept with the resulting hardware and optical complexity or cost.

3. In addition to the three baseline secondary pairs at F/5.2, F/15 and F/33. It is clear that many projects (i.e. long slit spectrograph) would benefit from an additional pair of F/15 secondaries equipped for an extended (10 arcmin) optical field. This introduces additional cost and exchange complexity, but could save substantial instrument money. The F/15 secondary would nominally have only a small unvignetted field and an IR selected coating (Ag or Au).

4. What we perceive as the two most heavily used focal stations: wide field F/5 and IR optimized F/15 have the problem that the Wide Field Corrector in the Cass hole prevents the observer from switching between these two configurations during the night. Rapid changeovers can only occur between direct cass and bent Cass or combined focus.

5. Rapid advances in active and adaptive optics technology means that we must continually re-evaluate our strategy in this area. This could significantly impact our selection of preferred focal stations. In fact the question arose as to whether it was even worth building a non-adaptive F/15 secondary (even the baseline secondaries are already active). Some uses of the combined focus would like to have an additional real pupil to place an adaptive element. As another example, the flip-top secondary exchange between F/15 and F/33 would make it more difficult to place the feed mirror for a laser guide star behind the secondary.

1. The handling of mirrors and equipment around the telescope was to be done with trolleys on the telescope and a large elevator on (outside) the enclosure. Initial engineering suggested that this was a rather expensive solution (corollary: all handling of 8m mirrors is expensive.).

2. The classic Columbus telescope had an impressive superstructure to hold the secondary mirrors and beam combiners in place. An ideal telescope would not want to have all this metal out there in the wind around the incoming beam.

3. Since the change of primary diameter from 8.0 to 8.4 meters, we had been unable to get the elevation structure resonant frequency back above 9 Hz.

4. Stiffness of the secondary spider attachment points continues to be a difficult issue (true for all designs).

1.3 Agenda and Attendees

The following people attended at least some part of the meeting:

2. Executive Summary

The meanings of these cryptic decisions are discussed later in this memo.

3. Technical Summary

3.1.1 Should first light have an adaptive F/15 secondary?

• By 1998, nearly every instrument will benefit from the addition of a ~ 100 element adaptive mirror at F/15. Such an adaptive mirror would allow the telescope to be diffraction limited most of the time at 2 microns.

• If adaptive secondaries are in common use, the newly fabricated rigid secondary will quickly become obsolete.

• The secondary is a very efficient place to locate the adaptive mirror, since no additional optical surfaces are required and all instruments can take advantage of it.

• A permanent adaptive secondary would drastically alter the specifications on many of the instruments.

• It would be technically attractive for many scientists and engineers to develop this secondary.

Con:

• Developing the adaptive secondary will add additional costs to the early phases of the project.

• Relying on a new adaptive secondary for telescope testing and debugging would place the telescope schedule at risk.

• We are already late in starting the adaptive program, if we want a secondary by first light.

• The precise direction of the adaptive mirror design is not clear at the present time.

• A much simpler tip-tilt system will already achieve the diffraction limit at 10 microns.

Result:

The Project Office has decided not to pursue an adaptive secondary for first light. At least one rigid (non-adaptive) mirror will be fabricated. But we strongly endorse the development of an adaptive secondary for implementation as early as possible in the life of the telescope. (no change from previous plan)

Adaptive Optics Strategy

The following is a summary of the Columbus Project adaptive optics philosophy extracted from an email message a few weeks before the engineering meeting.

• Columbus will take full advantage of adaptive optics in both Cassegrain (or Gregorian) and Interferometric mode. All aspects of the telescope design have attempted to take this into account --- starting with producing the best possible image w/o adaptive correction and with very clean infrared focal planes.

• Adaptive secondaries are being planned to minimize the number of optical surfaces in the optical train, and to make the corrected focal plane available to as many instruments as possible. (That doesn't mean we are ignoring the possibility of making reimaged pupils.)

• Active optics is completely integrated into the baseline design. Active in this context includes primary figure control, coarse wavefront analysis, active alignment and focus of the secondary mirror, and rapid (tip-tilt) guiding. (These are all budgeted separately from the $2M for adaptive secondaries. So the total budget for all the active and adaptive systems is more like $4M out of the total $60M (1989$)).

The official Columbus budget has a line item of $2M for a pair of adaptive secondaries. This was an arbitrary decision made in 1989 intended to be a strategy decision rather than a technical design. Because the technology is changing so rapidly it is still difficult to say exactly what form these secondaries will take. We want the initial adaptive system to work over a broad wavelength range --- say 1 to 10 microns. The capabilities will clearly evolve over the lifetime of the telescope.

3.1.2 Should the F/15 Secondary be Cassegrain or Gregorian?

• A Gregorian adaptive secondary would be conjugate to a layer in the atmosphere 100 meters above the telescope rather than the Cassegrain conjugate 100 meters below the telescope (below prime focus). This would allow a factor of perhaps 3 increase in the effective isoplanatic angle for seeing just above the telescope; thereby increasing the adaptively corrected field-of-view. (``conjugate'' in this context means the reverse image of the secondary formed by the primary.)

• A concave Gregorian secondary which is ellipsoidal is much easier to test during figuring since it does not require a Hindle sphere or a null corrector. This might mean a 50% reduction in cost.

• The Gregorian secondary allows access to prime focus for artificial stars or polarizing elements or occulting disks.

• Because of the F/1.14 primary, the Gregorian makes only a minimal increase in the length of the telescope. The secondary is 1.8 m higher, but the spider fits on the old telescope structure.

• The primary is easier to bend into an aplanatic ellipsoid by using a tension band around the edge of the mirror (compared to an anti-tension band to bend the RC hyperboloid).

• If we adopt the wide field F/4 focus, then the Gregorian secondary swings over the smaller wide field secondary. This simplifies the swingarm design.

• The Gregorian secondary may permit innovative instrument designs such as the wide field refractive collimator being pursued by Shectman.

Con:

• The Gregorian secondary increases the height of the dome by 1 to 1.5 meters.

• N. Woolf reports that the amount of ``seeing'' 150 meters above the ground is very small, so there isn't much isoplanatic patch gain.

• The secondary diameter increases from 73 cm to 84 cm. The increased moment of inertia reduces the performance of tip-tilt and chopping.

• Flip-top geometry becomes more difficult for F/15 to F/33 exchange.

• Lack of a Hindle sphere doesn't matter if we already built one to make MMT/Magellan secondaries.

• The tradeoff of increasing the telescope length to add a Gregorian secondary should be compared to the tradeoff of increasing the primary focal length with a Cassegrain secondary.

  Result:

  We adopted a Gregorian strawman for the baseline. But, the implications are still being studied.

View Figure 1 here

3.1.3 Should we use reimaged beam combination?

• A reimaging beam combiner would allow us to eliminate the F/33 secondary and the flip-top exchange by using the F/15 secondary to feed the combined focus.

• A reimaging beam combiner allows the creation of a real pupil image where additional adaptive elements could be placed.

• A common F/15 secondary reduces the number of adaptive secondaries that would be needed.

Con:

• More than the minimum 4 mirrors in the combined focus would increase the emissivity in the thermal IR unless cooled optics were used after the tertiary.

• Can the reimaged beam combiners really deliver enough unvignetted field at the required image quality?

Result:

A reimaged beam combiner is being designed. See the future memo on beam combiners for all the gory details and specifications.

View Figure 2 here

3.1.4 Move wide field focus above the primary?

Pro:

• Moving the focus higher will remove the 3-element refractive corrector from the central hole. Thus, enabling more rapid changeover from wide field optical to F/15 infrared work.

• Moving the focus higher will make the 1.9 meter secondary much smaller and much less expensive (1.2 -- 1.3 m).

• F/4 is a better match to CCD pixels for imaging.

• F/4 is a better match to fiber focal ratios for spectroscopy.

• By moving the corrector out of the central hole, we can decrease the hole size slightly (to match the MMT/Magellan hole diameter at 0.89 m). This reduces the central obscuration for the F/15 focus.

• If optical design permits, we can expand the field without worrying about the diameter of the central hole.

• For a fixed field angle, the size of the corrector elements is reduced.

• The platescale matches the corrector on the MMT Conversion F/5 focus.

Con:

• The short ``trapped'' focus may actually increase the central obscuration in the converging cone rather than at the secondary baffle. The central obstruction of the ``old'' F/5 corrector was approximately 2.6 meters (from the secondary baffle).

• The observer access to the wide field instrument would be greatly inhibited by being above the primary.

• A very compact and specialized instrument rotator is required.

• Maybe the same excellent image quality cannot be achieved.

• Lack of an accessible naked Cassegrain focal station might reduce flexibility or make it difficult to design high-throughput ultraviolet instruments.

• A smaller instrument envelope is available (comparable to that on the MMT Conversion).

View Figure 3 here

Result:

We will explore the optical and mechanical design possibilities and give it a try. The following acceptance criteria have been laid out for the trapped F/4 focus: (These are not the full corrector specifications.)

• Field diameter remains at least 50 arcminutes or increases to near 1 degree.
• Polychromatic image quality is at least 80% encircled energy in 0.5 arcsec over 90% of the field.
• The central obscuration of the telescope does not increase beyond 3.2 m.
• Instrument diameter can be at least 2.0 meters.
• Instrument height can be at least 2.0 meters (above tertiary?).
• The inner 30 arcminutes of the field has broadband images with 80% encircled energy in 0.25 arcsec diameter.
• Not more than one asphere is permitted.

View Figure 4 here

3.1.5 Other Optical Discussions

Domenico Bonaccini reviewed the work on wide field correctors in the normal Cassegrain position. Good solutions exist for both all-spherical and single aspheric designs. The images are excellent out to 44 arcminutes in the all-spherical version. The good field increases to 52 arcminutes by adding 1 asphere. Domenico has alsodemonstrated that the ADC can be placed after the third element to allow it to be ``easily'' removed for programs which do not need it. Domenico will provide details in a future technical memo.

ADC glasses are a perennial topic of discussion. We discussed the possibility of dividing the wavelength range that the ADC had to cover to permit the use of glasses with higher UV transmission. The idea would be to have a ``deep UV'' ADC covering 310 nm to 4000 nm and a ``blue'' ADC covering 350 nm to 1000 nm. Detailed study of the glasses is required to see if there is a practical breakpoint. (This sort of detail will be resolved after we freeze the global dimensions of the telescope and the corrector.)

We also discussed the possibility of putting the wide-field corrector only at the bent Cassegrain focus. The cost of a 1.9 meter (major-axis) flat tertiary makes this configuration unattractive.

Adaptive Correction There was a brief discussion of how much active and/or adaptive correction is required in various situations. The adaptive optics group at Arcetri (Bonaccini, Brusa, Esposito, Strauss) has been modelling the atmosphere to compare things like centroid tip-tilt vs. brightest speckle tip-tilt. These models will be used to set specifications on various parts of the telescope system.

3.1.6 Interchange

1. wide-field secondary in/out
2. narrow-field secondary in/out
3. narrow-field secondary flip, and
4. tertiary flat in/out
5. tertiary rotation in the plane of the primary
6. tertiary rotation varying the beam inclination

It is suggested that these movements might be reduced to four while increasing versatility.

1. The wide-field secondary, its baffles, field corrector and either a fiber table of an imaging detector package should all be carried in or out of the beam in a single movement by mounting these at an Epps focus.

2. Narrow field secondaries would be used ONLY at the Gregorian position. They would be carried into place by a single mechanism with three positions at 120 degrees, position 1 in, position 2 in, and position 3, both out of the beam for aluminizing. In this, position 1 is intended to be a mount into which one could place a mirror cell with assorted actuators, or a prime focus device, and position 2 is a similar fixture.

   This mechanism should not be expected to carry secondaries of larger than, say 80 cm in diameter, which would restrict Cassegrain focal ratios to F/17 or slower Then it would be less than 1m. above the prime focus, and so also capable of carrying lightweight prime focus equipment.

3. Tertiary in or out. This tertiary would serve both to provide bent Cass foci and to serve the combined focus.

4. Tertiary rotation around the axis of the primary. to feed combined foci, and one or more bent Cass positions. It would seem that a single angle of inclination to the primary axis should allow enough versatility (depends on beam combiner design).

(This discussion on interchanges was extracted from a longer memo by N. J. Woolf on telescope versatility. It succinctly describes the philosophy of the discussions at the engineering meeting.)

3.2 Mirror Support and Handling

3.2.1 Future Work

The highest priority task is to build another detailed local model to evaluate the stress over the support pads. This is important so we can freeze the design of the load spreaders and actuators with appropriate limits on the push and pull forces. This model will include the local fillets as well as the load spreader interface. It will also be improved by adding displacement boundary conditions from the global model.

8.4 meter Handling

There is still some work to do on lifting and handling the 8.4 m mirror and mold material since the stresses are still about 20% too high. Methods to reduce the stress include: redistributing the forces, removing the hextiles, changing the pattern of pads. The model will also be updated to better estimate the shear stiffness of the backplate and to better estimate the stress concentration at the backplate holes.

A parabolic temperature gradient appropriate to annealing will also be added to the list of calculated cases.

This work on handling analysis will be delayed until we have the results of lifting and cleaning the 6.5 meter mirror.

Final Optimization

The ``final'' optimization of axial and lateral force distributions will need to be done again since several of the actuators (L12 and L35) are blocked by the hardpoints. The re-optimization will also be needed if we change the central hole diameter.

Automatic Stress Verification

We discussed the need for long-term documentation of these stress analysis results. How will the Columbus engineer in 2035 know how to safely lift the primary mirror? Is it practical to develop a general computer model that can answer this sort of question with specific results? In any event, we will need to produce a detailed manual.

We also discussed methods for detecting when the mirror cell control system wants to apply pathological force distributions to the mirror. For example, pushing with all the actuators near the edge and pulling in the center could be very dangerous.

Aberration Corrections

Using axial correction forces only, we can change the conic constant by 3 microns p-v (without degrading the figure or overstressing the glass). Additional correction of the conic constant can be achieved by using a tension ring around the outside edge of the mirror. (Moving the focal plane a meter requires 26 microns of spherical aberration, so this adjustment is mostly useful for tuning out fabrication uncertainties.)

Additional calculations of active correction forces will be low priority until some of the more pressing design jobs are completed.

= 50° and = 15°.

Attachments

Despite being more difficult to fit into the actuator pattern, we are going to use some sort of modified loadspreader to attach the hardpoints to the mirror backplate. There is too much moment for a single puck attachment to handle. The larger attachment area also provides more stiffness from the RTV bond and allows a larger safe breakaway force for the hardpoints.

Design

Rafaele Tomelleri presented a preliminary design for a hardpoint mechanism. Unlike the historical concept, this design separates the motion from the breakaway function. Position adjustment is provided by a motorized screw. Breakaway under an unexpected force is accomplished by a pair of spring loaded plates. Load cell stiffness is increased by building the load cell into a lever arrangement that increases the effective stiffness by the square of the mechanical advantage while decreasing the apparent sensitivity by the first power of the mechanical advantage.

Assuming that the fulcrum and the support arms are infinitely stiff, the load cell deflection is given by P * X ^-1 * K^-1_L where P is the total load, X is the mechanical advantage and K_L is the load cell stiffness. Because the fulcrum side doesn't deflect, the central deflection is given by P * X ^-2 * K^-1_L and therefore the central stiffness, K_T, is given by X₂ * K_L. Because we were selecting the load cell for stiffness rather than sensitivity, the loss of sensitivity should not be a problem unless the mechanism is prone to drift.

Tomelleri will complete the prototype design. Changes include shifting one of the pivots to the center-of-mass to avoid lateral moments on the mirror (We missed that one in the original specs.). We have also increased the target stiffness to 200 N/ µ m for the mechanism and 100 N/µ m for the stack from the glue joint to the mirror cell attachment. This is to give the mirror a higher resonant frequency in the cell. Additional design details will be included in a future ADS report.

Later in the year, a prototype hardpoint actuator will be constructed and tested. A piezo-based servo loop closed around the load cell reading may be used if we wish to increase the stiffness beyond the levels described above.

View Figure 5 here

3.3.3 Mirror Support Actuators

MMT Test Results

Shawn Callahan summarized the results of testing the prototype actuators. The actuators (revision 57) now meet all of the mechanical requirements. The test results are described in a separate memo: ``Preliminary performance test results of the 6.5 meter mirror support actuator design'' --- UA-92-03. Shawn also showed photos of the 2 m³ mockup of the inside of the mirror cell.

Mirror Safety

There was a discussion (without definite resolution) of how to prevent the mirror support system from damaging the mirror by applying certain pathological force distributions. Individual actuators are designed so they cannot apply enough force to damage the mirror locally. The obvious dangerous example is a case where all the axial actuators on the outer part of the mirror push, while those at the center of the mirror pull. Clearly, careful design and safeguards will be required in both hardware and software.

Polishing Cell

Warren Davison described the design of the polishing cell being designed in Arizona by Roberta McMillan. The Mirror Lab decided to build a polishing cell specific to the 6.5m mirror to minimize the cost and pain on the first go-round. The structure of the cell is very similar to the telescope cell, except that it lacks all the connections to additional structure. The polishing actuators (axial) are glycerine-filled cylinders with variable mechanical advantage levers. These allow for adjustable passive support. A lively discussion followed on whether it would be possible or even desirable to polish the mirror on active pnuematic supports like those in the telescope. It might also be possible to polish on rubber pads and test on the telescope supports.

Telescope Cell

ADS has made a number of drawings of the telescope cell for mirror support. No problems have been encountered except that this is a very crowded space. The final design is on hold waiting for sub-assembly details.

View Figure 6 here

3.4 Telescope Structure

3.4.1 ``Nuovo Telescopio'' a.k.a. ``Columbus Lite''

View Figure 7 here

3.4.2 Spiders

View Figure 8 here

3.4.3 Structural Analysis

Several attempts have been made to reduce the radius of the C-rings from 7 meters to 6 meters. These have not yet been successful, but there is a possible weight savings of 25 tons if the structure can be modified in this way.

View Figure 9 here

3.5 Enclosure and Equipment Handling

View Figure 10 here

4. Administrative Summary

4.1 Optics

• The design (redesign) of the optical wide field focus. This depends on the performance that can be obtained from the high F/4 corrector compared to the low F/5 corrector.

• The design (redesign) of the interferometric focus. This has to do with whether a 5-mirror reimaged combined focus can provide adequate performance to replace the 4-mirror direct beam combination.

• The choice of Gregorian vs. Cassegrain optics for the infrared-optimized focus at F/15.

Work on these issues is now underway in Arizona, Italy and Ohio. See the discussion in an earlier section.

4.2 Primary Mirror

4.3 Telescope

4.4 Enclosure and Site

4.5 Aluminizing

4.6 Instruments

4.7 Budget and Schedule

5 Post Meeting Progress

5.1 Optics

5.1.1 Beam Combination

People didn't object to 5 mirrors, especially if the fourth and fifth mirrors can be cooled inside a sealed tank. The ``OSU'' solution clearly demonstrates that geometric solutions are possible. Now we have to decide if we can put in practical aspheres to create sharp images. I also note that the possible combiner geometries are seriously limited by reimaging to F/33. A faster geometry combined with control of focal plane tilt could allow more flexible placement of the fold flats. Don McCarthy will think more about the constraints on the combined focal ratio. The high beam combiner mirrors (currently just below the Gregorian secondaries) present some structural difficulties, but they need to be fairly high to get an unvignetted field.

In the five mirror combiner, the height of M5 is limited by the magnification from f/15 to the final f/ratio, not by the field size or the crotch height. The best I can do for nominal f/15 focal position and 2 meters extra for the combined focus is f/25.

The magnification puts the limit on the beam combiner height. But, the motivation for F/33 was to keep the tilted focal planes in phase. But since reimaging retilts the focal planes, that whole constraint has changed. AND, if we assume the exit pupil hasn't moved (obviously false), then we also need a high beam combiner to keep 4 or 5 arcminutes of unvignetted field. So now we should re-evaluate focal plane tilt and vignetting in the reimaged case.

5.1.2 Wide Field Correctors

This is viewed favorably by everyone with the two caveats:

1. that we can optimize the images a little bit better.
2. that we can work out a suitable mechanical solution, including proving human access to the instrument for setup/debugging.

5.1.3 Gregorian F/15