🔭 astrophysics

Focal Plate Prototyping for Modular Focal Planes of Stage-5 Instruments For Ground-Based Telescopes

This paper presents a focal plate prototype for next-generation Stage-5 multi-object instruments, demonstrating that 5-axis machining is a promising method to achieve the stringent stiffness, mass, and alignment tolerances required to support approximately 20,000 fibers on ground-based telescopes.

Original authors: Maxime Rombach, Jean-David Perriard, Laurent Chevalley, Diane Chapuis, Markus Thurneysen, Jean-Paul Kneib

Published 2026-07-01

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Maxime Rombach, Jean-David Perriard, Laurent Chevalley, Diane Chapuis, Markus Thurneysen, Jean-Paul Kneib

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the inside of a giant, high-tech telescope not as a single mirror, but as a massive, curved floor made up of hundreds of tiny, interchangeable tiles. This paper is about building a prototype for that "floor"—called a focal plate—which will hold the next generation of astronomical instruments.

Here is the story of how they built it, explained simply:

The Big Picture: A Lego Floor for the Stars

Current telescopes can look at many objects in the sky at once using "fiber positioners" (think of them as tiny robotic fingers that grab starlight). The next generation of telescopes (Stage-5) wants to grab 20,000 of these at once!

To hold all these fingers, the researchers are designing a modular system. Instead of one giant, heavy plate, they are using triangular "modules" (like Lego bricks) that snap together. Each module holds 63 robotic fingers. The challenge? These modules need to be screwed onto a curved surface with extreme precision. If they are even slightly tilted or too far forward/backward, the telescope will go out of focus, and the data will be blurry.

The Challenge: Building a Curved, Bumpy Surface

The team at EPFL (a Swiss university) wanted to test if they could machine a prototype of this plate using a giant 5-axis CNC machine (a super-precise robot arm that cuts metal).

They faced a few tricky hurdles:

The Curve: The plate isn't flat; it's a giant, shallow bowl (concave).
The Angles: Every single spot where a module attaches needs to be tilted at a slightly different angle, like the petals of a flower opening up.
The Thin Walls: To save weight, the walls between the modules are incredibly thin (about the thickness of a coin), which makes them wobbly and hard to cut without vibrating.
The Precision: They needed to cut these angles within 0.05 degrees and the depth within 30 micrometers (that's thinner than a human hair).

The Process: Carving a Block of Aluminum

They started with a solid block of aluminum (about the size of a large suitcase).

The Cut: They used the 5-axis machine to carve away 90% of the metal, leaving behind a delicate, honeycomb-like structure with thin walls.
The Design: They removed the "walls" between some groups of modules to make the plate lighter and let more light in (reducing "vignetting," or shadowing).
The Interface: The modules slide in from the bottom and are held by three screws, resting on three specific faces. This "three-point" support is key to keeping them stable.

The Test: Measuring with a "Super-Ruler"

Once the plate was cut, they had to prove it was accurate. They used a Coordinate Measuring Machine (CMM).

The Metaphor: Imagine a very sensitive robotic arm with a needle tip that gently taps the surface of the plate thousands of times to map its shape.
The Method: They measured three points on every single module's support face to create a "virtual plane" for each one. They did this five times to make sure the results were consistent.

The Results: A Perfect Fit?

The results were very promising:

Tilt: The angles of the modules were almost exactly right. The average error was tiny (0.005 degrees), and almost all of them fell well within the safe "tolerance zone" (±0.05 degrees).
Focus (Depth): The distance from the bottom to the top was also very close to the target. While there was a tiny average shift (about 13 micrometers), the vast majority of measurements were within the 30-micrometer limit.
The Outliers: A few specific spots were slightly off, but the researchers noted that these can be easily fixed later by adding tiny shims (like shims under a wobbly table leg) when the actual modules are installed.

The Bottom Line

This paper doesn't claim to have built the final telescope yet. Instead, it proves that 5-axis machining is a viable way to build these complex, curved, lightweight plates.

They successfully carved a prototype out of a solid block of metal that meets the strict requirements for holding the next generation of 20,000-fiber instruments. It's a successful "dress rehearsal" showing that the manufacturing technology exists to build the future of deep-space observation.

Technical Summary: Focal Plate Prototyping for Modular Focal Planes of Stage-5 Instruments

Problem Statement
The next generation of astronomical surveys (Stage-5) aims to deploy instruments with approximately 20,000 fiber positioners, necessitating a shift from monolithic designs to modular focal plane assemblies. A critical challenge in this transition is the fabrication of the focal plate structure, which must securely house modular positioner groups (specifically triangular modules of 63 fibers) while maintaining extreme geometric precision. The plate must support fiber positioners, guide cameras, and wavefront sensors on a curved telescope focal surface, adhering to tight tolerances for tilt (typically ±0.05°) and focus (typically ±30 µm). While previous 5-axis machining has succeeded for Stage-4 instruments, the feasibility of manufacturing a large, semi-frameless, modular plate with these specific constraints for Stage-5 remains to be demonstrated.

Methodology
The study focuses on the design, manufacturing, and metrology of a focal plate prototype developed under the Innosuisse project (No. 101.014 IP-ENG). The approach involved:

Design Simplification: To prioritize prototyping feasibility, the design assumed a spherical focal surface (radius 13,000 mm) and neglected Chief Ray Deviation (CRD), despite acknowledging CRD's impact on real-world module orientation. The prototype dimensions (818.4 mm diameter) were selected to match the largest available CNC machine at EPFL, based on Spec-S5 parameters.
Structural Architecture: The plate utilizes a "semi-frameless" design where intermediate walls between modules are partially removed to increase packing density and potentially reduce vignetting by up to 10%. Modules are inserted from the bottom and secured via a three-screw interface on support faces.
Manufacturing Process: The prototype was machined from a single block of Aluminum 7022 T651 (550x550x130 mm) using a DMG DMF 180 5-axis CNC machine. The process involved removing approximately 90% of the raw material, resulting in thin walls (≈2.5 mm) and a 120 mm thick plate. Key challenges included milling tapered walls, maintaining a 90 mm nominal distance from the curved top surface to the module faces, and achieving specific tilt increments (0.2° steps).
Metrology: A Coordinate Measuring Machine (CMM) (COORD3 - HERA 10.07.05) was used to validate the part. The measurement protocol involved defining a coordinate system via lateral faces and probing three points per module and fiducial support face to construct local planes. Measurements were repeated five times per side to ensure statistical reliability.

Key Results
The manufacturing and measurement process demonstrated that 5-axis machining can achieve the required precision for Stage-5 focal plates:

Tilt Accuracy: The tilt errors for the 34 module support faces were centered on a mean bias of $\mu_{tilt} = 0.0051^\circ$ with a standard deviation of $\sigma_{tilt} = 0.017^\circ$ . The vast majority of measurements fell within the ±0.05° tolerance envelope, with only a few outliers exhibiting larger spreads.
Focus Accuracy: The focus distance errors (relative to the 90 mm nominal) showed a mean bias of $\mu_{focus} = 13.4$ µm and a standard deviation of $\sigma_{focus} = 20.9$ µm. Approximately 68% of the errors lay within ±20.9 µm, well within the ±30 µm tolerance. Outliers (e.g., modules 21, 26, 29, 30) were noted but deemed potentially correctable via shim stacks.
Fiducial Performance: The 13 fiducial support faces exhibited similar performance to the modules, with errors remaining within the manufacturing tolerance envelope.

Significance and Claims
The paper claims that the presented work serves as a proof-of-concept for the manufacturability of complex, semi-frameless focal plates required for future Stage-5 instruments. The authors state that the modeling, manufacturing, and measurement methods proved to be a "promising solution" for triangular module-based focal planes.

The study explicitly notes that it does not impose a specific design for any single Stage-5 project (such as MUST, Spec-S5, or WST) due to the preliminary nature of the R&D framework and the simplifications made (e.g., spherical surface assumption). Instead, the significance lies in demonstrating that 5-axis machining can meet the stringent tilt and focus tolerances necessary to support the next generation of high-multiplex fiber instruments. Future work is identified as necessary to test module integration repeatability and assess the long-term stability of the positioners on the manufactured plate.