The “Cobra” Fiber Positioner, the WFMOS Design, and Potential lessons for DESpec Michael Seiffert, Jet Propulsion Labo

1. The “Cobra” Fiber Positioner, the WFMOS Design,and Potential lessons for DESpecMichael Seiffert, Jet Propulsion LaboratoryRichard Ellis, Caltech DESpec RAS and University College London March 2011

2. Fiber Positioner Design Considerations Density: Should roughly match desired target density. Practically, this means ~ 1000 sources/deg2 At 4-8m telescope prime focus with typical plate scale this translates to positioner separation of ~ 10mm. Number: many 1000s. Maximize subject to budget constraint. Throughput Efficiency: Positioning accuracy should be high, and losses due to tilt, despace, and non-telecentricity should be small. Reconfiguration Speed: Fiber movement and position verification should proceed rapidly compared to exposure time. Robustness: A mechanically stiff system facilitates accuracy and allows lenses at fiber tip or other treatment of fiber ends.

3. WFMOS Concepts Are Relevant to DESpec Although the detailed designs are different, WFMOS, PFS and DESpec may share system aspects Fiber connector mounted on top end structure Prime Focus Unit includes Wide Field Corrector (WFC) and Fiber Positioner. Spectrograph located above Naysmith platform Fiber Cable routed around elevation axis and brings light to the Spectrographs

4. Prime Focus Instrument (PFI) In the WFMOS and PFS designs, several Subaru provided elements (field rotator, hexapod and wide field corrector) are shared with the HyperSuprimeCam

5. Rotating Portion of PFI Rotator Interface Ring Alignment System Cobra Optic Bench Cobra Modules with Drive Electronics 2400 Cobra Fiber Positioners Positioner Equipment Bench

6. Positioner Optical Bench with 2400 Positioner Units Room for >4000 positioners 8mm apart in hexagonal pattern to enable field tiling 1 Positioner Unit - Cobra

7. Positioner Element – “Cobra” • Each motor rotates to provide complete coverage of the patrol region. • Optical fibers mounted in “fiber arm” which attaches to upper postioner axis: • Fiber runs through the center of the positioner – this couples the positioner and fiber system schedules and work efforts Top View Fiber arm Patrol Region Phi stage Second axis of rotation Theta stage First axis of rotation Fiber Tip Cobra

8. Geometry Cobra Positioner Patrol Area (9.5 mm dia.) Phi Stage (2.4 mm radius) Theta Stage (2.4 mm radius) 8mm

9. Patrol Regions Patrol Region – Area of the focal plane accessible to one fiber (9.5 mm diameter) Adjacent patrol regions overlap with no gaps Patrol Region may have zero or may have many potential astronomical targets Allocation efficiency describes the success rate in assigning targets to fibers • Low target densities: degree of overlap between patrol regions is unimportant. Important not to have gaps. • High target densities: degree of overlap not important – there are many targets in each patrol region to choose from • Intermediate target densities: (target density ~ positioner density) there is some benefit to having larger overlap.

10. Positioner Module A module is a subassembly of actuators and drive electronics boards Staggered production Parallel module integration Early mechanical and electrical functional testing Parallel fiber integration to reduce schedule Increases serviceability Positioner Electronics Boards

11. Motors • Commercially available rotating tube motor: • High torque when stationary and unpowered • ~ 1 mN-m powered torque • 1 mrad resolution • 1 – 10 rev/sec speed • Pairs of PZT plates oscillate in tandem bending • Drive signals of the two PZT plates are phase shifted by 90 degrees • This creates a traveling wave on the stator that excites the rotor like a harmonic gearbox to rotate the shaft by extremely small angles

12. Cobra Prototype 2nd stage motor 1st stage motor Fiber optic • Ceramic friction drive • Lubrication free, zero backlash • Journal bearing limits motor side loads • Hardstops to limit fiber twisting • 5 um precision of fiber positioning • Motor movement < 1 sec/iteration Cobra system tested at JPL in partnership with New Scale Technologies Achieves 5μm positioning accuracy in 6 iterations Prototype has also successfully completed lifetime and environmental testing.

13. Prototype array of positioners is an essential precursor to proposing for a ~2400-4000 element system Cobra fiber positioners fiducial fibers Multiplexed motor electronics Proposed 7-element prototype to demonstrate mechanical integration, tolerances, & integrated electronics Proposed laboratory and on-sky testing of 19-element system via NSF/Caltech $

14. Metrology Camera – establishes science fiber positions relative to fixed fiducial fibers on positioner focal plane WFMOS concept: Four camera systems each looking at a ¼ of the focal plane. Located on prime focus support struts looking back at positioner focal plane via primary. Each camera is 4k by 4k CCD with15μm pixels. Cameras are defocused to allow centroiding. Future: Larger format (10 k x 10k) single camera? Are back-illuminated CCDs (better centroiding) really required? Metrology camera (1 of 4 shown)

15. 2 2 2 2 2 2 3 3 3 3 3 3 1 1 1 1 1 1 Science fibers are back-lit in a sequence to allow discrimination between fibers in the overlap regions between adjacent fibers, 1/3 at at time. In one exposure only the fibers marked (1) are illuminated, the next exposure only the ones marked (2) are illuminated, etc. Back-lit fiducial fibers used to establish position of science fibers on positioner plane. Encoder fibers used to establish rotation orientation 1 3 1 3 • Positioner moves elements in 3 groups of 800 • Metrology camera views back-illuminated fibers. Fibers are illuminated in 3 groups of 800. • Movement, illumination, camera readout, computation in parallel • 6 iterations can be completed in < 40 seconds

16. Fiber connector options APOGEE: US Conec 30 fiber connectors “ganging” with custom fixture allowing simultaneous mating of 300 fibers. Wilson et al., 2010 WFMOS team B study: Custom connector for 800 fibers with simultaneous mating. De Oliveira, 2008

17. Key challenge of fiber-fed spectrographs: getting the fiber placed accurately on the astronomical target large field of view large number of fibers smaller diameter fiber Make sure the system design addresses these challenges: Robust positioner design provides high precision attention to differential mechanical flexure in overall structure error budgets for mechanical tolerances Correction for non-telecentricity? Include imaging mode with fast readout for verification!

18. Conclusions & Future directions • WFMOS design emphasizes instrument efficiency: stiff, robust, & precise positioner system with fast reconfiguration speed • Lifetime, thermal, and simulated altitude testing of prototype complete. Dust or other contamination testing TBD. • Now engaged in the PFS design effort • Two proposals now pending for small array demonstration. Critical to demonstrate that technical risks are retired and costs are understood before scaling up to thousands of elements. • Internal JPL proposal pending for 7 element lab module • NSF proposal pending for 19 element, on-sky, Palomar demonstration • Investigation now underway of improved fiber coupling. Concepts include elimination of fiber twist, tilt of fiber end for non-telecentricty correction, and inclusion of small lens at the fiber tip for f-ratio conversion.