proto-Lightspeed: a high-speed, ultra-low read noise imager on the Magellan Clay Telescope

TL;DR

Proto-Lightspeed demonstrates ultra-fast, ultra-low read-noise optical imaging on the Magellan Clay telescope by integrating a DSERN CMOS ORCA-Quest 2 camera with re-imaging COTS optics. The paper details optical, mechanical, detector characterization, and software/timing infrastructure, including per-pixel nonlinearity calibration, a Python-based ETC, and GPS-synchronized timing achieving <30 µs absolute accuracy. On-sky commissioning shows seeing-limited image quality over a 1′ field, measurable throughput and photometric precision, and several science highlights from pulsars, X-ray binaries, and nebular/arbitrary-field imaging. Plans for Lightspeed, a five-channel facility imager, are laid out to deliver multicolor, high-cadence imaging with improved throughput and larger field of view, enabling new time-domain astronomy. The work highlights both the promise of DSERN CMOS sensors for astronomy and the remaining challenges in nonlinear response and polarization/spatial-mode calibration that must be addressed for routine operations.

Abstract

proto-Lightspeed is a new instrument that has been commissioned on the Nasmyth East port of the Magellan Clay Telescope at Las Campanas Observatory to deliver high-speed optical imaging with deep sub-electron read noise. Making use of commercial re-imaging lenses and the ORCA-Quest 2 camera from Hamamatsu, proto-Lightspeed images a field $1'$ in diameter at up to $200$ Hz or windowed fields at higher rates, up to 6600 Hz for a $1.6''\times 1'$ field of view. proto-Lightspeed delivers seeing-limited image quality in the $g'$, $r'$, and $i'$ bands and adjustable magnification for pixel scales between $0.017''-0.050''$. proto-Lightspeed is well suited to studying compact binary systems, exoplanet transits, rapid flaring associated with accretion, periodic optical emission from pulsars, occultations of background stars by small trans-Neptunian Objects, and any other rapidly variable source. proto-Lightspeed will be a P.I. instrument beginning in 2026B, available for use by members of the Magellan Consortium. In this paper, we discuss the design and performance of the instrument, results from its two commissioning runs, and plans for a facility instrument, Lightspeed, to support simultaneous multicolor imaging across a $7'\times4'$ field.

Figures (11)

• Figure 1: Computer-aided design model of the optical components of proto-Lightspeed. Reflected light from the telescope's tertiary mirror enters the instrument from the right. The length of each breadboard is 36 in.
• Figure 2: a) Transmission curves for filters available in proto-Lightspeed (dashed lines), the quantum efficiency of the ORCA-Quest 2 camera (crosses and solid line), and atmospheric transmission at unit airmass (dotted line). b) Total throughput for each bandpass. This includes the filter transmission, sensor quantum efficiency, losses due to the re-imaging optics, and atmospheric transmission at unit airmass. It does not include the transmission of the telescope. The curves for $u'$ and $z'$ are upper limits, with an assumed re-imaging throughput of 5%.
• Figure 3: Overview of proto-Lightspeed components. a) proto-Lightspeed mounted at the NasE port of the Clay telescope. The fully assembled instrument extends $\approx$1 m from the port. Five LC duplex fibers, power and Ethernet cables, and glycol cooling lines are secured to an aluminum hitch at the optical axis to prevent stress during rotation. b) Interior view with cooling enclosure and side panels removed, showing the dual Thorlabs breadboards, optical components, and cardboard airflow baffles that direct cooling air across electronics and to the camera TECs. c) The ORCA-Quest 2 camera, with the Birger RF lens controller affixed. d) The control computer mounted in the Clay equipment room, housing the Active Silicon Firebird CoaXPress frame grabber and Meinberg TCR180PEX GPS timing card.
• Figure 4: Nonlinearity and internal QE of the ORCA-Quest 2 camera. a)Top panel: The raw sensor response to increasing exposure times from a stable uniform light source, averaged across the sensor (blue points), shows significant nonlinearity at low signal levels. Individual pixels (gray points) all have slightly different response. The red dashed line shows a linear fit to the linear regime. The blue dash-dotted line shows the ideal linear response, with the same slope but zero y-intercept. Orange triangles show the average response to 4 s exposures with varying neutral density filters, confirming the nonlinearity is not a timing artifact. Bottom panel: Residuals from the linear fit clearly show poor agreement at low signals and a discontinuity at $\approx$2400 ADU where the camera switches gain modes. b)Top panel: After applying the per-pixel nonlinearity calibration, the sensor response is linear across the full dynamic range. The green dashed line shows the fit to this corrected data. Residual spread in individual pixel response (gray points) is caused by read noise and shot noise. Bottom panel: Residuals remain below 1 ADU at the lowest signal levels. c) Signal-dependent detection efficiency, showing the fraction of photogenerated electrons that successfully reach the sense node. At low signals, incomplete charge transfer causes $\sim$50% of photoelectrons to be trapped; this efficiency improves asymptotically to nearly 100% at higher signal levels. The calibration compensates for this effect. The slight discontinuity in this curve is an artifact from the gain shift.
• Figure 5: a) Photon transfer curve (PTC) for images that have been corrected for linearity. Here the contributions from read noise and from Poissonian shot noise have been scaled to appropriately account for the effect of the linearity calibration. b) PTC for raw frames at increasing levels of exposure, with contributions from read noise and from Poissonian shot noise. The bumps around 2500 ADU in a) and b) results from the transition between high and low gain modes.