Division of Electronics, Detectors and Computing for Physics (DEDIP)

A new NECTAr chip for the latest front-end-board (FEB)

The NectarCAM camera acts as the observatory's eyes, capturing brief yet brilliant flashes of Cherenkov light produced when high-energy gamma rays collide with the Earth's atmosphere. This fleeting light unveils the violent processes occurring across the galaxy and beyond, from supernovae explosions to the enigmatic behavior of black holes. At the heart of NectarCAM, the Front-End Board (FEB) translates these light signals into digital data for scientists to analyze.

When Cherenkov light impinges on the camera, it is first collected by photomultiplier tubes – 1855 in total for one NectarCAM. These detectors convert the faint light into electrical signals, amplifying them into two gain channels. These signals are then sent to the FEB, which performs a second amplification, sampling, digitization, and recording.  Central to the FEB is the NECTAr chip, developed at IRFU, which includes a switched capacitor array for sampling signals at 1 GHz and a 12-bit analog-to-digital converter (ADC) for digitizing these signals upon receiving a trigger. A key advancement of the NECTAr chip is its capability to operate in 'ping-pong mode'.

In this mode, the chip's memory is divided into two segments that operate alternately to optimize the processing of incoming analog signals, addressing the issue of deadtime— the period during which the system cannot record new data while processing the current signal. Traditionally, the capture and processing of a signal prevent the system from accepting a new signal, leading to deadtime. This limitation reduces the system's efficiency, especially when observing high-frequency astronomical phenomena. The 'ping-pong mode' alleviates this by utilizing a dual-buffer system. While one buffer processes a signal, the second buffer stands ready to capture the next incoming signal. This alternating operation ensures continuous data acquisition, significantly reducing the system's deadtime. Upon a trigger event, the active buffer freezes to process the data, while the alternate buffer immediately begins data acquisition. This method minimizes the intervals during which the system cannot record new events—referred to as the minimum deadtime—thereby enhancing the trigger rate and enabling more continuous and efficient observation of variable cosmic events.

A reduced deadtime with ping-pong mode

The NectarCAM team at IRFU has conducted a series of tests to evaluate the performance of the new FEBs, focusing particularly on measuring the camera's deadtime. These tests were carried out in the darkroom at IRFU, a thermally controlled environment shielded from external light sources, where the NectarCAM camera undergoes testing before being shipped to the Canary Islands. With a surface of ~80 m2, it was specifically built for the NectarCAM project[1] and was partly funded by a DIM ACAV grant, to test the 9 cameras before mounting them on site in La Palma. To simulate the light that the camera will encounter once installed on the telescope structure, three distinct light sources were used: a Flat Fielding Calibration Light Source (FFCLS), a laser, and a Night Sky Background (NSB) source. The FFCLS employs 13 LEDs to provide a uniformly illuminated field to the camera, serving as a benchmark for identifying and correcting spatial variations in the camera's response. Capturing images with this source enables the detection of sensitivity discrepancies across the camera's detector. The laser, being a more precise light source, is suited for exact timing measurements. Meanwhile, the NSB source replicates the natural ambient light of the night sky. All three sources were employed to illuminate the camera randomly by employing a generator able to reproduce a physical trigger that arrives randomly in time.

During these tests, the number of times that the camera was illuminated (Ncollected) was recorded, along with instances where the camera was unable to record new light signals because it was already occupied digitizing another signal (Nbusy). By comparing these two rates, an estimation of the deadtime fraction was derived, which indicates the camera's "dead" state, rendering it incapable of recording a signal.

Additionally, a second method was employed to cross-check the deadtime fraction. When a signal is recorded by the camera, an event is created and a time-stamp is assigned to it. Given the random nature of the light emission, the distribution of the time difference between consecutive events resembles an exponential distribution. This distribution was fitted with a truncated exponential model, whose parameters include the minimum deadtime (δmin) and the exponential rate (R). The minimum deadtime denotes the shortest interval between two events as dictated by the NECTAr chip, approximately 700 ns, while the exponential rate reflects the event recording rate. The deadtime fraction, in this case, is determined by multiplying the minimum deadtime by the rate.

Figure 3 presents the principal findings from the deadtime measurements. The data for the new FEB version is depicted using colored lines, contrasting with the gray lines representing data from the previous FEB version (v5). Different colors correspond to the various light sources utilized in the measurements. Solid lines and shaded areas denote the two methods applied to calculate the deadtime fraction. The results illustrate that the newly introduced ping-pong mode significantly reduces the deadtime fraction to 0.5% at the standard operation rate on the sky of 7 kHz, marking a tenfold improvement.