The Skipper-CCDs, a special type of charge-coupled device (CCD) sensor that features sub-electron readout noise levels, was proposed decades ago. However, only in recent years it has been possible to develop large size Skipper-CCDs ensuring stable operation. Their extreme low noise operation makes them suitable for experiments that require low thresholds and high energy resolution, such as dark matter and neutrino interactions detection, and more recently quantum-imaging and astronomy. New experiments are planning to use kilograms of active silicon from Skipper-CCDs as sensitive mass. In this way, they can achieve extremely low detection thresholds and a high probability of particle interaction. However, this approach needs arrays of thousands of Skipper- CCDs operating at the same time imposing challenging requirements. Also, introduction of this technology in astronomy and quantum-imaging applications requires a large number of channels per sensor to speed up the readout. The front-end needs to be redesigned from scratch: it must achieve low noise performance, be simple for easy integration and allow the routing of thousands of channels out of the sensors with minimal connections. This paper presents a detailed analysis of options for the front-end electronics and their noise performance. It describes a novel way of using a dual-slope integrator with minimal components to pile up the charge of consecutive readouts of the same pixel in a concept that we call a multi-slope integrator. This reduces drastically the output bandwidth, simplifying the wiring and the warm electronics. These proposals will allow the generation of new scientific instruments based on Skippers-CCD arrays.
Microwave Kinetic Inductance Detectors, or MKIDS, have the ability to simultaneous resolve the wavelength of individual photons and time tag photons with microsecond precision. This opens up a number of exciting new possibilities and efficiency gains for optical/IR astronomy. In this paper we describe a plan to take the MKID technology, which we have demonstrated on the Palomar, Lick, and Subaru Telescopes, out of the realm of private instruments usable only by experts. Our goal is to incorporate MKIDs into a facility-class instrument at the Keck 1 Telescope that can be used by a large part of the astronomical community. This new instrument, the Keck Radiometer Array using KID ENergy Sensors (KRAKENS), will be a 30 kpix integral field spectrograph (IFS) with a 42.5” x 45” field of view, extraordinarily wide wavelength coverage from 380-1350 nm, and a spectral resolution R=λ/▵λ > 20 at 400 nm. Future add on modules could enable polarimetry and higher spectral resolution. KRAKENS will be built using the same style MKID arrays, cryostat, and similar readout electronics to those used in the successful 10 kpix DARKNESS instrument at Palomar and 20 kpix MEC instrument at Subaru, significantly reducing the technical risk.
Scientific CCD detectors are typically readout using the Correlated Double Sampling (CDS) technique. At low
pixel rates, noise of ~2e- RMS is typically achieved. The limitation for reaching lower noise comes from the 1/f
component on the output of the CCD, and this noise cannot be eliminated using CDS. A new readout technique
based on a digital filter is presented here for suppressing the 1/f. Using this new technique a noise of 0.4e- is