Quests! Neptune, 138 days from a dream to a dot

2026-10-11 · eclipticam v3w · night 2026-10-07 →

eclipticam v3w · 20 Sep to 7 Oct 2026 (5 nights) · each panel 1.4° × 1.0°
Neptune in five whole-night stacks of eclipticam, 09-20 to 10-07, each about 4 hours of 60 s frames. The blue circle is where the ephemeris puts Neptune that night. It moves 25 arcminutes over the 17 nights, and the faint dot moves with it.

eclipticam v3w · 6 Aug to 9 Oct 2026 (27 nights) · 4.2° × 2.6°
Neptune walking west, 6 August to 9 October, two thirds of a day per frame. Each of the 27 nights it was found is laid into a rolling stack of the stars from the nights around it, and the gaps between those nights cross-fade. Blue dots are where it was measured on earlier nights, the grey line is the ephemeris. Brighter catalogue stars are labelled with their V magnitude; Neptune is V 7.8.

eclipticam v3w · 7 Aug to 9 Oct 2026 (29 nights) · 8° × 6°
Saturn walking through the stars on eclipticam, 7 August to 9 October, two thirds of a day per frame. Each of the 29 nights it was measured is laid into a rolling stack of the stars from the nights around it (6 day window), and the gaps between nights cross-fade. Orange dots are where it was on earlier nights, the grey line is the ephemeris, and the brighter catalogue stars are labelled with their V magnitude.

The dream

On 25 May 2026 at 22:13 a line went into the astro TODO list: "Neptune in November (target)". Magnitude 7.8, so well below what the eye can see from a dark site, never mind a London back garden, and the camera was a Pi with a wide lens. The budget written that night guessed about 28 hours of stacking to reach it.

A month later, on 25 June, it became the north star: "positively track Neptune in the autumn". The honest baseline written down that day was two objects identified, Polaris and the Moon.

The dot

On 10 October, late in the evening, the whole-night stack of 7 October showed a faint dot at Neptune's predicted position. That was 138 days after the TODO line. The stack was 248 frames of 60 s, about 4 hours, not 28. Three things beat the guess: full-resolution raw mosaic frames, a fitted camera model good to a fraction of a photosite, and fitting every frame's pointing on the stars close to the planet.

A dot is not a planet until it moves. Four more nights came out the same way, from 20 September to 7 October. Each time the dot sat within about 1.5 arcminutes of the ephemeris (an eclipticam v3w binned pixel, 2 x 2 photosites as used here, is about 3.5 arcminutes near the centre of the frame). The places it had left were empty: in the 20 September stack, the spot Neptune would occupy on 7 October shows nothing (1.1 sigma), and the reverse holds too. No catalogue star brighter than 7.5 lies within 58 arcminutes.

By the next day the whole season was in: 47 nights stacked, Neptune found on 27 of them, and an animation of it walking west through Pisces across two months.

Saturn's rings

Saturn's rings are almost edge on and much too small to be seen with these cameras. FirstScope feasibly could just resolve them, but aiming would be difficult.

The other quests

Quest	Set	Status
Neptune	2026-05-25	Found 10-10, 27 nights tracked, animated
Saturn walking through the stars	2026-10-10	Done 10-11: 29 nights, 08-07 to 10-09, animated against a deep star field
M42, the first nebula	2026-10-10	Done 10-10: [field note](2026-10-10-orion-m42)
Betelgeuse's colour	2026-10-10	Measured: red/green 1.74 against a white star
M51, the Whirlpool	2026-07-06	Open: wants a narrower camera
Algol's eclipse	2026-07-06	Open: autumn, needs per-night photometry
Polaris A/B, a rehearsal for Saturn's Titan	2026-07-06	Open
Mizar and Alcor	2026-07-07	Open: a calibration double, nothing blocking
Polaris B through half a pair of binoculars	2026-07-09	Open
See 100,000 stars, identify 10,000	2026-07-25	Open: limiting magnitude now G 9 in one night
Starlight as a barometer (refraction)	2026-08-14	Open

The next quest is Neptune again, but properly. Two Raspberry Pi HQ cameras with a 16 mm lens arrived on 9 October. With them Neptune should show at about 20 sigma in a single 5 second frame. The plan is to tilt the camera on an SG90 servo to find the ecliptic, and mount the camera in portrait orientation - maybe even tilted perpendicular to the ecliptic so we have a good chance of each planet crossing the field of view.

Kit progression and approximate commissioning dates

starcam, Pi 1B with v1 camera, from 20 May 2026: OV5647, 5 MP (2592 x 1944), stock lens 53.5 deg wide, about 74 arcsec per photosite.
eclipticam, Pi 5 with v3 wide camera, from 9 June: IMX708, 12 MP (4608 x 2592), about 104 x 68 deg, 1.8 arcmin per photosite at the centre falling to about 1 at the edges (3.5 and 2 per binned pixel, as used for Neptune). Full resolution from 22 June.
6 TB disk on muppet ("bigstore"), the archive moved onto it from 27 July.
astrocam, Pi 4 mounted outside, from 8 June with a v2 camera: IMX219, 8 MP (3280 x 2464), about 68 arcsec per photosite. Then a v3 standard camera from 29 July: IMX708, 12 MP (4608 x 2592), 59 arcsec per photosite by plate solve, about 75 x 42 deg. A glass rain cover from 20 August.

New and to be commissioned:

HQ camera with 16 mm f/1.4 lens, bought 9 October: IMX477, 12.3 MP (4056 x 3040), 20 arcsec per photosite, 22 x 17 deg.
FirstScope with v1 camera, on xoverpi (Pi 3B): a Celestron 76 mm Newtonian, 300 mm at f/3.95, with the OV5647 at prime focus; 0.96 arcsec per photosite, 0.69 x 0.52 deg. Optics fitted 8 September, first frames 16 September, still commissioning.
FirstScope with HQ camera (no lens, at prime focus): 1.07 arcsec per photosite, 1.2 x 0.9 deg.

questsneptunesaturnplanetsstackingastrometry

← Sharpest thing resolved by each camera How big is an arc minute anyway? →