La Silla Science Operation Department FEROS Upgrade Project Minutes of the meeting held on 2003-04-17 LSO-MIN-ESO-90200-0001/1.0 --- Prepared: O.Hainaut, 2003-04-17 Reviewed: I.Saviane, jpritchard 2003-04-23 Released: O.Hainaut, 2003-04-26 Revision History: 1.0: 2003-04-26T12:32:15 Sat ohainaut release 0.2: 2003-04-23T12:00:00 Wed isaviane xxx 0.1: 2003-04-17T16:00:59 Thu jpritcha Minor modifications 0.0: 2003-04-17T13:50:12 Thu ohainaut Creation --- A meeting to discuss various aspects of the FEROS upgrade was held on 2003-04-17. Were present, in Vitacura: J.Pritchard (chair), I.Saviane, W.Eckert, A.Gilliotte, O.Hainaut; in Garching: G.Avila; at La Silla: J.Alonso, JC.Guzman-Tanaka. 1- Fiber-head replacement mission: Currently, a significant part of the FEROS efficiency loss is caused by a mismatch of the f/ ratio of the instrument with that of the beam exiting the fiber. This is caused by the use of a ball lens. Optimal design was not possible at the time of the move from 1E5 to 2P2, because of unavailability of suitable lenses. Additionally, the ball lens causes losses also due to: 1. no antireflection coatings, and 2. a higher refraction index compared to the rod lens. The possibility to have a "fiber-head replacement mission" separate from the main upgrade, around Jul.2003, was discussed. The design for the new head is ready, but still needs review. This design includes only modifications of its inner parts, not of the overall structure and shape. The modifications consider the replacement of the current ball lens by a rod lens, with relevant changes of the ferule holding the lens and the fiber. This new design ensures optimal f/ match between the beam and the instrument. GAv reports simulations that indicate that this will result in an improvement of ~30% of the efficiency, i.e. bringing it from 12-13% to 17-17.5%. The adaptor flange will require some machining in order to provide a new path for the fiber, to minimize the fiber stress [note added in review] Suitable rod lens are already available at la Silla (from CFHT). The pros for an early replacement are the following: - early restoration of the full efficiency of the instrument The cons: - time is short for a full review/analysis of the design + manufacturing - time is short for complete test of the rod lenses - new type of fiber is available (STU), that needs to be studied - the gain (30%) is not that large considering current situation. It was decided NOT to have a June fiber-head replacement mission. Instead to aim for installation along with ADC during October. ACTIONS: - AGi sends rod lens(es?) to GAv for test - GAv tests rod lens - GAv compares rod lens with ball lens in terms of efficiency - GAv acquires STU fibers asap from one of his WP. - GAv: f/ ratios of new fibers, and possibly old fibers, should be measured Note: any significant action (i.e. requiring purchase or significant amount of work) has to be held until we get the results of the review of the 1st FEROS adaptor by LSO director. 2- Upgrade to FEROS2 2.1- ADC Current ADC requirements are very strict: 0.2" dispersion (ie max displacement of center of image with respect of center of fiber), corresponding to 5% light loss in the blue at 70deg ZD. This has some strong implications on the ADC design, esp. considering the small space available in the current location of the FEROS adapter in the Cassegrain hole. Main consequence is that a single prism ADC is not a valid solution, a double, counter rotating prism must be considered. AGi reports that, in the current situation, the focal plane of FEROS is matched to that of WFI. This implies that FEROS is affected by a 0.27" spherical aberration (80% light enclosed) (that aberration is corrected for within WFI). Moreover, there is not enough space in the Cassegrain hole to extend the FEROS adapter so to match the optimal focus of the telescope. This could possibly be changed with a new adaptor in the post-WFI era. However, the WFI era seems to be extended by several years, and the situation of GROND is still very unclear (as it could even be installed on another telescope). So, we proceed in a WFI-compatible model [added in review] JCGT remarks that the hardware (motors, encoders...) that drive the ADC must be VLT-compliant, as the cost of writing new drivers would be high. Full VLT compliance (MACCON/Heindenheim) may be difficult to achieve because of space constraints: indeed, a conceptual design by WEc indicates that there is no space to install the motor+encoder in the usual configuration, i.e. facing each-other along their rotation axis, because the motors of the TWO prims must be installed on that axis. This would imply the use of a different kind of motor + init switches. JAl is confident that this can be made VLT compliant at the electronic level so that the existing software drivers can deal with the ADC. The possibility of a gravity-driven ADC was mentioned, but discarded because of the mechanical complexity of such system. There was a discussion on the need of having 2 ADCs (i.e. 1/fiber), or whether 1 (for the object fiber) was sufficient. It was agreed that 1 was sufficient, though it should remain possible to switch between the two apertures in the head since it is possible the object fibre (which is selected as the fibre with the best efficiency) may change if/when fibrehead/rodlenses/fibres are eventually upgraded. It is mentioned that in the current configuration, the ADC moves in the location of the calibration mirror. The usr reqs must address whether this is acceptable, as this configuration prevents the use of simultaneous calibration with the ADC. For instance, the usr req. could consider an interlock to prevent accidental movements of the calib mirror when the ADC is inserted. The ADC requirements need to be refined, in particular comparing with those of HARPS and UVES. They must include the position precision and stability requirements. AGi points out that the manufacturing of the ADC will take ~4mo, and that therefore, the ADC should be placed on a "fast track" with FDR no later than May31. Once these requirements are defined, AGi will analyze whether these can be matched by a system in the current location, or if alternate location (below the Cassegrain flange) must be considered. ACTIONS: - ISa, JPr, OHa meet to discuss the ADC reqs. ISa formalizes these reqs. - AGi gives an estimate of the "best" ADC that can be fitted in the existing adapter. - AGi provides with simulations of the effect of mis-alignement of the ADC with respect to optimal angle, in order to characterize the precision requirement. 2.2- FIERA JCGT reports on behalf of Peter Sinclaire that all the hardware for FEROS's FIERA is at La Silla (from VLT test cam). SWC has to install the corresponding software on the SPARC workstation, and PSi needs to test the connection cables. Once this is done, the tests can start. No problem foreseen on that side. 2.3- SOFTWARE The functions that need to be controlled were briefly discussed. The environmental sensors and their controller are currently not VLT-compliant. This should be fixed by moving them to the fwacu LCU (possibly requiring some replacements?). Detailed software user requirements must be provided soon. These must include TCS, ICS, DCS, OS, templates and DFS. JCGT will provide JPr with a template. JPr will be in charge of the template definition, and of their implementation, with support of PLeSaux. The FEROS pipeline will have to be included in the DFS (either full PL compliant (which seems unrealistic to OHa) or in an hybrid way using dhsSubscribe post-arrival facility). JCGT remarks that FEROS will be implemented in HPUX, as there are too many uncertainties about LINUX. ACTION: JCGT send SW Usr Req template to JPr 2.4- CALIBRATION UNIT AGi remarks that the current Calibration Unit has very poor efficiency in the blue. This should be considered in the Usr Req. Possible solution is a deuterium lamp. 3- SCHEDULE Below is a very tentative draft schedule of the project: Apr.17: quick off meeting Apr.30: "What went wrong" review complete May.7: User requirements (HW and SW) v-1.0 released May.15: - Meeting to discuss the Usr Req. Possible iteration to finalize them. - ADC PDR May 31: SW/HW PDR, including templates. ADC FDR Jul.31: release of HW and SW design documents Aug.15: FDR Oct.15: 1 week of WFI time to install FEROS hardware (1) Oct.25-31: Commissioning I (~6n), mostly software and functionality (2) Nov.24-30: Commissioning II (~6n), SciOp commissioning and science verification. (3) Dec.1: FEROS-IP released. Dec.1: release to the community. Notes: 1- installation can take place during WFI SM time (not VM). 2- cannot take place before VLT-SW upgrade of 2p2 to MAR2003 3- at least 3 wks between ComI and ComII during which FEROS is unavailable ---oOo---