New eyes for the H.E.S.S. I cameras

March 2017

Previous | Index | Next

fig0

Ten years of dust and duty have not gone unnoticed for the sensitive cameras on each of the four 12-meter telescopes that comprised the original H.E.S.S. I array. During this time, there has been the Galactic plane scan, more than a hundred source discoveries, a spectrum of cosmic electrons - a lot of great measurements. To nobody's surprise, at some point the endurance of the fragile detectors had started to approach their limit. In 2012, it was therefore time for the H.E.S.S. collaboration to place an order with their new collaborators at DESY in Zeuthen to team up with colleagues from the Paris area, Leicester and Amsterdam to make use of the newly developed NECTAr readout chip technology and design new cameras. Only four years later, in September 2016, the new components were installed. And now, in the first days of 2017, we received an alert from our colleagues in HAWC, and just in time, we managed to spot the first significant gamma-ray signal found with the new cameras:

fig1
Fig. 1: Gamma-ray sky image of Markarian 421 as seen with the new H.E.S.S. cameras.

When H.E.S.S. explores the mysteries of the high-energy sky, it actually does not look into the Universe, but at the upper Atmosphere. Cosmic gamma-rays are absorbed there and produce short, faint, violet Cherenkov light flashes that can be detected from the ground. It requires large mirrors and ultra-fast electronics to record and digitise them to images that actually display the gamma-ray incidence. The exposure times per image are as short as 16 nanoseconds (16/1,000,000,000 of a second), and H.E.S.S. is recording about 300 particle incidences per second. Since some images only consist of a few handfuls of light particles (photons), the technical requirements to build such cameras are very challenging.

But the ten years for which the original H.E.S.S. I cameras have been operated have not only degraded their physical condition; also the available technologies have developed much further. The internet has accelerated the development of fast ethernet solutions. Electronic chips and CPUs have become smaller, faster and capable of doing more complex tasks. Software has become smarter and faster to develop in an industrial and standard way. Following these improvements, scientists have started designing prototypes for the next generation of gamma-ray observatories. An important technological driver of the next big experiment in the field, the Cherenkov telescope array (CTA), is the NECTAr readout chip. It can capture the fast signals in Cherenkov cameras an digitise them nanosecond by nanosecond when requested by the trigger.

fig2
Fig 2: Microphotography of the NECTAr chip. © CEA/Irfu, Eric Delagnes.

This chip, whose complex interior is displayed above, is anticipated to be used for many CTA mid-size telescopes, a work-horse of the project, but in fact, it was never used on a telescope before! This was good reason for the developers of the new H.E.S.S. cameras to take the opportunity to use this speed-booster for the camera upgrade,which at the same time could allow verification of the technology for CTA. We used 16 NECTAr chips on each front-end unit ("drawer"), as can be seen in this drawer assembly (the square chips with little flowers printed on them):

fig3
Fig 3: Assembly of a new front-end unit in Namibia, displaying the photomultiplier tubes, high voltage bases and new electronics boards with NECTAr readout chips (right to left). © DESY, Stefan Klepser.

But in fact, this is only one of the first-time-ever features of the project. Because also the project host DESY was relatively new to the field and had never built a camera for Cherenkov telescopes like H.E.S.S. before. Still, the engineers, physicists and technicians lost no time, and with a lot of support from the partnering institutes in the Paris area (LPNHE, LLR, IRFU), Universities of Leicester and Amsterdam, MPIK in Heidelberg and of course endless support by the local MPIK team in Namibia, the cameras were developed, tested and installed in the short time frame of only 4 years. The plan foresaw a replacement of all electronics except the costly photosensors, plus some mechanical components, like installing a new backdoor to improve the ventilation in the camera.

fig4
Fig 4: Installation of one of the new ventilation systems. The camera resided in the "camera shelter" while all new mechanics components are installed. © DESY, Stefan Klepser.

On top of this, colleagues from LLR added a full renewal of the light collimators in front of the PMT pixels ("Winston cones") to the list of things to improve, so more light is collected in the first place.

fig4a
Fig 4a:Installation of one of the cone plates, housing the new Winston cones, on the front side of the camera.

This does not mean, of course, that there was no room for failure and funny stories. A deep and gruesome crisis had to be endured to understand that the new cameras did not harmonise well with the monkey repelling electric fence that is installed around the H.E.S.S. array (really, who thinks of that when you're contemplating the detection of cosmic accelerators?). And big was the adrenaline turnover when the mechanics learnt that they had designed their first components for a camera that is one centimetre smaller than they had anticipated. Well, H.E.S.S. is an experiment, and we like that. At some point, everything was assembled and working, and the cameras could take off.

fig5
Fig 5: A H.E.S.S. I telescope mirror and the first of the newly installed cameras (on the right) pointing to the sky. © DESY, Stefan Klepser

In our business, it is a long way though from installation to operation of a new device. Software needs to be adjusted, network connections to be established, and a long list of issues that never occurred in the lab before taught the physicists once more the difference between theory and practice. Around Christmas 2016, however, the systems were all fit for observation, and as luck would have it, an old friend in the gamma-ray sky, the blazar Markarian 421, was reported in an Astronomer's Telegram by our other old friends working with the HAWC detector to have an increased activity in gamma ray emittance. Or actually had an increased activity 400 million years ago, which, due to its distance of 400 million light years, arrived at Earth now and made it to the news precisely on Jan 4th. Despite it being located in the Northern sky, in the constellation of Ursa Major, H.E.S.S. turned its telescopes and had a look. And this is what we saw:

fig6
Fig 6: Series of stereoscopic images of candidate gamma-ray events as seen with then new H.E.S.S. cameras.
Images of particle incidences could be recorded smoothly, and the source could be detected. The new cameras, which are 4 of only 12 currently operating Cherenkov cameras, seem to work. They deliver the first large scale demonstration that 3840 NECTAr chips are able to record gamma-ray signals that are good for Teraelectronvolt astronomy. And they make us look forward to the final years of H.E.S.S., where the new cameras, with their much reduced dead time, and more flexible readout, will provide us with enhanced performance at both very low and very high energies.