Committee

Tom Geballe

Kim Gillies

Bret Goodrich

Mark Hunten

Manuel Lazo

Jim Oschmann

Doug Simons (Chair)

Summary

GNIRS was reviewed on 7 March 2001 during the mid-phase of its fabrication by an engineering and science team to evaluate progress and identify any outstanding issues that have emerged since its last formal review, the Prefabrication Review (PFR) held in May 2000. The GNIRS team continued to make good overall progress toward completing the instrument close to original budget and schedule targets, as originally identified during the Restart Review. The team gave a confident and candid appraisal of progress and issues that naturally emerge in an instrument as complex as GNIRS. Assuming integration can be completed during the third quarter of 2001, it appears feasible that GNIRS can be delivered ~1 year thereafter to Gemini, on schedule.

Comments and Recommendations

Mechanical

The predicted mass and CG of the instrument look like they will meet the required values. That said, most instruments have made a similar prediction at a similar time in their production, and generally the mass and CG is not measured with complete confidence until instruments are nearly done and remedies to any unforeseen problems are time consuming or expensive. We therefore encourage the GNIRS team to build some flexibility into the mass distribution of the instrument. Ballast weights that can be attached at various points around the instrument are already planned. A potentially low cost additional strategy would be to mount the electronics cabinets on adjustable racks. Given the impact of these cabinets on the overall mass/moment, building such adjustment into GNIRS will permit additional freedom to tune the mass and CG at the end of the integration phase. Furthermore, having such adjustability permits greater freedom in the balancing of GNIRS vs. other instruments on the ISS.

The passive flexure compensation system, though elegant, represents some risk in terms of cold cycles required to tune it to match the actual flexure of the instrument. A motor driven system would permit faster convergence with tuning the compensator, but would also be more mechanically complex. In order to reduce the potential for using precious cold cycles to tune the compensator, we recommend looking into using a bare MUX and making warm tests, e.g., pressing on the sides of the compensator while recording images to make sure motion at the detector has the direction and amplitude expected.

The planned flexure tests of the completed instrument are similar to what was used for NIRI and places the instrument at risk of not fully demonstrating its flexure performance prior to shipment to Chile. In NIRI’s case, single axis tests were completed in Honolulu and this was the extent of the testing completed due to lack of proper facilities at the IfA to complete multi-axis tests. Such tests were run when NIRI was placed on the telescope and, of course, the flexure along untested rotation axes was then demonstrated to be unacceptably large. Gemini and the IfA worked together to resolve the problem, but the solution required having the NIRI PI on-site, working in the Gemini lab, and shuttling to/from the IfA shop in Honolulu. It is important to note that NIRI was not formally accepted due to flexure problems, hence it was the clear responsibility of the IfA to lead the effort to fix the problem under these circumstances. If we have a similar problem with GNIRS due to lack of proper flexure testing, it will require NOAO personnel to be in Chile for extended periods of time, potentially with parts being sent to and from Tucson, i.e. the remedy stands to be much more complex, costly, and time consuming. GNIRS must therefore be properly flexure tested before it is sent to Chile as part of the on-site acceptance tests in order to demonstrate that it meets an acceptable level of flexure before it is shipped. Details of these tests will be defined through the GNIRS Acceptance Test Plan.

Electronics

Overall the electronics subsystems appeared to be in good shape and at a level consistent with the rest of the instrument’s development. We encourage an on-going dialog between the GNIRS EE and Mark Hunten, particularly regarding the use of mains vs. UPS power and instrument/telescope grounding schemes.

Providing some level of protection for the GNIRS cryocoolers in the accidental and sudden loss of helium flow should be factored into the control system. Gemini is now providing differential pressure switches at each telescope that trigger when facility helium pressure is lost. These switches can be used, for example, to cut the power to cryocoolers instantly. Long term Gemini will link helium pressure status with the GIS so a single signal can trigger the shutdown of all instruments’ cryocoolers. Continued dialog with Gemini is encouraged in order to make sure GNIRS will be compatible with at least one of these approaches to protecting its 4 cold heads after it is sent to Chile.

Software

Overall the progress on the GNIRS software since the PFR has been very good and GNIRS is in the unique position of not having software on its critical path. The GNIRS design relies heavily on low level C code to provide an interim engineering control environment. An EPICS layer that interfaces to the underlying C code will be completed later. This is similar in nature to the already successful GMOS CCD/controller software, and the Gemini TCS. That said the committee believes development of the most complex portion of the code (the EPICS layer) has been deferred until late in the project, which represents risk. A parallel presentation describing plans for the EPICS layer above the C code would have been helpful in assessing this risk level. Successfully retrofitting the EPICS layer onto the established C code will require that people handling both tasks (Richard and Peter) are in good communication throughout the process to streamline interfaces and assure that all required functionality is provided in the final system.

Using NIRI code does not guarantee that the OCS interface will be correct. In fact the NIRI/OCS interface delivered does not meet Gemini’s ICD and this is an issue that still needs to be resolved. We urge NOAO to use a standalone version of the OCSwish tool, available through Kim Gillies, to confirm that the GNIRS/OCS interface works properly, rather than assuming the NIRI code properly interfaces to the OCS.

While Gemini in general encourages the reuse of instrumentation software, there will undoubtedly be divergent development efforts between NIRI and GNIRS on the same code and controlling this will take effort within NOAO and Gemini. We therefore encourage regular dialog between NOAO and Gemini on this matter, particularly as NIRI’s software is now being maintained and updated by Gemini personnel. For example, DHS integration, PPC support, WCS implementation and image header parameters have all been recently handled by Gemini staff working on the delivered NIRI software.

Management

GNIRS is being well managed as key performance metrics are understood on timescales fast enough that corrective action, if needed, can be effective. This is reflected in the speed with which CAS now reports labor data to the GNIRS management group, compared to the problems noted at the PFR. As noted in the summary, the basic metrics of budget and schedule indicate GNIRS progress has been very good since the PFR.

We share the NOAO concern about a “wave” of manpower being pushed steadily outward in the labor projections. The real impact of this problem will probably not be realize until the integration phase begins. In any event we believe that the integration plan has sufficient latitude and flexibility to be adjusted to minimize the effect of this wave. The availability of Bill Ditsler during the integration phase is also an area of concern, given that he will likely be consumed to some extent by other projects, notably work on the dewar for the Gemini bench spectrograph.

The committee noted some concern about the fact that a significant amount of design work remains before the full system integration can begin. Again this is not an ideal situation but is manageable.

Gemini will send periodically its southern instrument engineer and lead GNIRS scientist to Tucson at key points in the integration and test phase of the instrument. This is primarily intended to provide hands-on training in the assembly and maintenance of the instrument and to let the Gemini personnel who will eventually be responsible for maintaining and operating GNIRS become thoroughly familiar with it before it leaves Tucson.

If you have any questions or suggestions regarding this website, please contact Melissa Bowersock.

NOAO Intranet Services

NOAO Copyright
Statement

National Optical Astronomy Observatories, 950 North Cherry Avenue, P.O. Box 26732, Tucson, Arizona 85726,
Phone: (520) 318-8000, Fax: (520) 318-8360