The August 2026 issue (volume 120, issue 4, pages 135-141) of the Journal of the Royal Astronomical Society of Canada carries my article on R60, the remote observatory I built and commissioned at AstroCamp in Nerpio, Spain.

Nath, A. (2026). Building My Remote Observatory at 15: Engineering for Millimagnitude Photometric Precision. Journal of the Royal Astronomical Society of Canada, 120(4), 135-141. Read the issue

I have already written on this site about how the observatory was built. What the journal article adds is the argument underneath it: an observatory is not a telescope you buy, it is a measurement system that you slowly build, modify, configure, and calibrate so that it meets your science specifications. The system then goes through a series of tests and its observation results are vetted against standards established by scientific bodies before the outputs can be published in scientific databases and publications.

The numbers came before the telescope

The first decision was not which optical assembly, camera, and filters to fundraise for. It was what the instrument had to be able to measure. The binding requirement was twofold. First, astrometric accuracy: measuring an asteroid's position on the sky well enough that the Minor Planet Center can fold it into an orbit solution, which means sub-arcsecond residuals against the reference catalogue and time stamps good to the second. A position is only as good as the moment you attach to it, because the asteroid keeps moving while the shutter is open. Second, millimagnitude photometry: resolving a change in a star's brightness of about 0.001 magnitudes, roughly one tenth of one percent. Exoplanet transit depths, mid-transit timings, and the shallow mutual events of binary asteroids all live at that level.

Once the numbers were established, I had to reverse engineer the system that would meet those standards alongside other factors such as my ability to fundraise through proposals, determine interoperability of hardware and software, search vendors, ship hardware, and finally remotely configure the assembly and troubleshoot. It turned out to be the most ambitious project I had taken on so far in terms of my scientific capabilities and project management skills. And it gave me weeks of sleepless nights, none of which were spent on observing celestial delights but on configuring and problem solving.

The 12-inch f/6 Ritchey-Chretien controls off-axis coma, because elongated stars at the edge of a frame push systematic error straight into aperture photometry. The mount was chosen for payload margin, so multi-hour guided runs never operate near its mechanical limits: at this precision the mount is part of the measurement system, not just a pointing device. The camera is cooled and monochrome, because a Bayer matrix costs light and because stable sensor temperature is what makes dark-frame calibration reproducible from night to night. The guider is off-axis rather than a separate scope, to remove differential flexure between the guiding and science optical paths. Site selection was the same exercise applied to geography, and two of the criteria that decided it, spare-parts logistics measured in days and an on-site technical team, have nothing to do with the sky at all.

Two independent tests

A specification you have not tested is a hope. R60 had to pass two external tests, set by two different scientific communities, measuring two different things.

The first was astrometric. To be assigned an observatory code by the IAU Minor Planet Center, I submitted positions for more than ten near-Earth asteroids, each observed on at least two separate nights with multiple images per session, together with full instrumentation details. The MPC checks those positions against the global orbit solutions. Passing means pointing, plate solution, timing, and centroiding are all correct at once. The observatory was assigned code R60, and every asteroid position I submit now enters the same worldwide astrometric record that professional observatories feed.

The second was photometric. Astrometry proves you know where things are. It says nothing about whether you can measure brightness to a thousandth of a magnitude. That test came from the ExoClock project, which maintains transit ephemerides for the candidate targets of ESA's Ariel mission. Every light curve I submit is independently re-reduced and quality-controlled by their science team before it is accepted. The article reports the first three, KELT-12b, WASP-59b, and KELT-16b, with mid-transit residuals consistent with the published ephemerides. By the time it went to press, 29 transit observations had been accepted, and I have written those up separately in a year of transit timing for ESA Ariel.

Two referees, two different failure modes, and no way to talk past either one. That is what turns an instrument into an observatory.

The lesson that generalised

The article also records a failure. During one session the guiding system kept reporting that it could not hold the target. I recalibrated the guiding algorithm, rechecked polar alignment in software, and worked through every setting reachable from Toronto. Hours passed and nothing improved. The on-site team then found a single loose screw in the camera adapter coupling. It was producing micro-vibrations too small to see in the science frames but large enough for the guide camera to read as uncorrectable drift.

From 7000 km away, a mechanical fault and a software fault present identically. The fix was not the screwdriver, it was the reordering: physical verification with the on-site team now sits near the front of my diagnostic checklist rather than at the end.

What it has measured

R60's first extended campaign was main-belt asteroid (2977) Chivilikhin: 25 hours of photometry over seven nights, giving a rotation period of 6.257 ± 0.001 hours and a light-curve amplitude of 1.030 magnitudes from a fourth-order Fourier fit. Next are multi-filter campaigns for colour information on asteroid surfaces, more rotation periods, and mutual events in binary asteroid systems. Those shallow, long-duration signatures were not attemptable on borrowed telescope time. They are attemptable now, which was the point of building it.

With thanks to the AstroCamp team for hosting the observatory and providing on-site support, Telescope Ontario for donating the imaging camera, Celestron for providing a mount for testing, Valkanik for equipment support and logistics, and the Masason Foundation for the funding that made the observatory possible. Also to the Faulkes Telescope Project, the American Association of Variable Star Observers, and the Canadian Space Agency for the shared telescope access and mentorship that came before R60 existed.