FLORES — Fiber-fed Low-Resolution Spectrograph

Project Status

First ThAr calibration exposure obtained (16 July 2026). A Thorium-Argon arc spectrum was successfully recorded through the fully assembled instrument — confirmation that the optical train, from slit to detector, is aligned and functioning end-to-end. With calibration in hand, FLORES is ready to move on-sky for first light. As a by-product, we also produced a line identification map of the ThAr spectrum.

First ThAr calibration exposure from FLORES — **First ThAr calibration spectrum.** Thorium-Argon arc lines recorded through the complete optical train, verifying end-to-end alignment ahead of first light on-sky. 2026-07-16.

Mechanical assembly complete (July 2026). The instrument is fully assembled and awaits optical alignment, expected in the coming weeks. An initial laser check of the slit and the rest of the optical train was performed on 2026-07-08 and looks consistent — a rudimentary test, with fuller alignment and verification still to follow.

Finished FLORES optomechanical assembly, view 1 — **Finished assembly, prior to optical alignment.** Collimator, fold mirrors, and Zeiss Milvus 135 mm projection lens mounted on the optical bench. 2026-07-07.

Finished FLORES optomechanical assembly, view 2 — **Finished assembly, prior to optical alignment.** Prism boxes, fold optics, and filter wheel mechanism. 2026-07-07.

Mechanical Components Finished (June 2026)

Prism disperser and fold optics machined and mounted (25 June 2026). The double-prism assembly, fold mirrors, and collimating optics were completed and mounted on the test breadboard ahead of integration into the instrument enclosure.

FLORES double-prism disperser and fold optics on the test breadboard — **Double-prism disperser and fold optics.** Prism boxes, fold mirrors, and collimator lens mount on the test breadboard, prior to integration into the instrument enclosure. 2026-06-25.

Coudé Slit-Box Camera Pass-Through (May 2026)

Camera mounting adapted for the slit-box wall. Normally the camera body mounts directly to its objective lens; here the two had to be split across the wall of the Coudé slit box, with the lens adapter re-engineered as a simple pass-through — objective inside, CCD body outside, since it needs air cooling. The slit box itself — roughly the size of a large wardrobe drawer, on its own concrete foundation at the Coudé focus — used to house extensive calibration hardware; it now holds only the exposure-counting photometer and the order-separation filters, leaving room for FLORES, which will occupy roughly one quarter of its volume.

Optical Components Delivered (April 2026)

Opto-mechanics received from Thorlabs and Eksma Optics (30 April 2026). Delivery comprised three mirrors (one 19 mm, two 25 mm), one plano-convex collimating lens (Eksma), a precision translation mount for the lens, manual tip-tilt mounts for the mirrors, and the corresponding mirror mounts. The F2 glass prisms themselves are legacy stock — F2 has not been manufactured for decades and these particular pieces predate the year 2000.

Observer Quick Reference

Wavelength range

3800 – 9000 Å

UV–optical–near-IR

Mean resolution

R ∼ 800

varies 600 – 1400 across range

Limiting magnitude

V ∼ 17

300 s, SNR > 25, current config.

Bright-target cadence

1 – 2 min

V ∼ 13, adequate SNR for time series

Primary detector

KAF-6303E

27.6 mm chip; 9 µm pixels; 2×2 binning available

Telescope

2m Perek

f/32, Ondřejov Observatory

Response time

< few min

automated GCN trigger; robotic handoff from D50/SBT

Hardware cost

< $4 000

using existing prisms and Coudé infrastructure

Spectral Resolution Across the Range

3800 Å (UV) 6000 Å 9000 Å (NIR)

R ∼ 1400 R ∼ 800 (mean) R ∼ 600

Non-linear prism dispersion: higher resolution in the blue, lower in the red. With 200 µm slit (corresponding to 58.9″ on sky in the collimated beam).

Performance

Note: The figures below are pre-commissioning estimates. First light is expected summer 2026; on-sky performance will be characterised during the first ~0.5 yr of operations.

Limiting Magnitudes (300 s exposures)

Configuration	Throughput	Resolution (binned)	Limiting mag (SNR = 10)
Current — uncoated prisms, KAF CCD	~25%	R ∼ 800	V ∼ 16.5
Optimised — AR-coated prisms, sCMOS	~50%	R ∼ 800	V ∼ 17.5
Optimised + 2×2 binning	~50%	R ∼ 400	V ∼ 17.9
Optimised + 4×4 binning	~50%	R ∼ 200	V ∼ 18.3

Throughput Budget (current configuration, ~25% total)

Telescope optics

85%

Fiber coupling

85%

UVFS collimator

99%

F2 prisms (uncoated)

72%

Zeiss 135mm lens

85%

KAF CCD QE

50%

Anti-reflection coating of prism surfaces reduces per-surface losses from ~5% to ~0.2%, improving prism throughput to ~92%. A back-illuminated sCMOS detector (80% QE) brings total throughput to ~50%.

Cross-check: VLT/FORS2 at R ~ 600–800 reaches V ~ 18 at SNR = 12 in 100 s. Scaling to the 2m aperture (~16× less collecting area) gives ~1800 s for the same target. FLORES's prism-based disperser partially offsets this through higher disperser efficiency vs. a grating. A working estimate of V ~ 18 in ~30 min (optimised config.) is plausible.

Scientific Applications

GRB Afterglow Spectroscopy

Automated response within minutes of GCN/SVOM/Einstein Probe alerts. Separation of reverse-shock, forward-shock, and central-engine components through early spectral evolution. Redshift estimation via Lyman break and strong absorption systems for V < 17 afterglows.

Cataclysmic Variable Time Series

For targets such as GK Per (V ~ 13): 1–2 min exposures resolve flickering timescales spectroscopically. Tracks outburst evolution in dwarf and classical novae, emission-line variability, and accretion-disk heating events.

Transient Classification

Spectroscopic triage for photometric discoveries in 12 < V < 17. Supernova type separation (Ia / II / Ib/c) through broad spectral features. Nova vs. flare-star discrimination. Multi-messenger follow-up for poorly-localised events.

LBV / SN Impostors

Wavelength coverage 3800–9000 Å captures Ca II triplet, Hα, and broad continuum diagnostics. P Cygni profile detection at V ~ 13. Bridges D50/SBT photometric monitoring with OES high-resolution follow-up.

Instrument Design

Dispersion Element

Type: Two equilateral F2 glass prisms
Prism face: 69 mm, apex angle 60°
Incident angle: 54.73° (optimal for 3800–9000 Å)
Angular dispersion: 5.38° across full range
Total beam deviation: 98.9° (4 × 24.7°)
Input beam ∅: 23.8 mm (max)
Output beam ∅: 45.1 mm

Optical Train

Collimator: Plano-convex UVFS lens, f = 700 mm
Collimated beam ∅: 22.4 mm (10% margin)
Projection lens: Zeiss Milvus 135 mm f/2
Slit width: 200 µm ≡ 58.9″ (collimated beam)
Pixel scale: 13.75″/9 µm pixel (KAF, 135 mm lens)
Telescope f-ratio: f/32
Coupling: Parasitic fiber to existing Coudé spectrograph

Detector Options

Primary: KAF-6303E, 27.6 mm, 9 µm pixels
Alternative: SiTe 1024×1024, 24 µm pixels
Binning modes: 1×1, 2×2, 4×4 (dispersion axis)
KAF note: ~4.2 px/resolution element; supports 2×2 for read noise
SiTe note: Undersampled (1.5 px/slit), but higher QE

Mechanical Integration

Location option A: Coudé slit box (before slit); independent mounting
Location option B: Spectrograph room, next to entrance window
Compact unit: 135 mm lens + two prisms
Slit–M1 distance: 720 mm (or 850 mm with fiber injector)
Beam fold: Small movable pick-up mirror + 1–2 flat mirrors

Optical Design Figures

Schematic of the two-prism configuration — **Double-prism configuration.** Two equilateral F2 glass prisms arranged with 54.73° incidence angle, providing 5.38° total angular dispersion across 3800–9000 Å.

Zemax simulation of the spectrograph — **Zemax optical simulation.** Ray-tracing of the full spectrograph with a plano-convex UVFS lens as collimator. Credit: J. Václavík (Toptec).

Transparency curve of the Zeiss Milvus 135 mm lens — **Zeiss Milvus 135/2 transparency.** Note the reduced throughput below ~4000 Å, consistent with the instrument's practical blue cutoff. Credit: J. Ebr & FZÚ Olomouc.

Disperser: F2 Glass Prism Properties

F2 glass (lead-based silicate) is no longer manufactured but offers an excellent combination of refractive index (~1.62) and high chromatic dispersion. The Sellmeier equation gives n(3800 Å) = 1.660 and n(9000 Å) = 1.605. Minimum deviation angles are 52.15° (blue) and 46.78° (red), setting the optimal single-prism incidence angle at 54.73°.

The double-prism configuration doubles the dispersion while the beam geometry constrains the maximum useful input diameter to 23.8 mm (widening to 45.1 mm at the exit face). A collimator focal length of 700 mm produces a 22.4 mm beam — within the 10% safety margin.

Compared to a grating, prisms offer theoretical 100% diffraction efficiency vs. typical grating peak efficiencies of 20–60%, representing a significant photon-economy advantage for faint transients.

Contributors

Design & Project Lead

M. Jelínek: Astronomical Institute of the Czech Academy of Sciences, Ondřejov

Optical Design

J. Václavík: TOPTEC Turnov

Assembly, Mechanics & Instrumentation

J. Zeman, M. Šlechta, J. Srba, L. Řezba, J. Sloup, R. Novotný, B. Kubátová, J. Štrobl, J. Fuchs, R. Veselý

ASÚ Ondřejov