Wow, first entry in almost five months … and not because the astrophotography world stood still — quite the opposite, actually. Countless equipment fixes and upgrades in the observatory, changes to the image train (a ZWO ASI 2600MC Pro will soon capture on the Lacerta Apo in a dual setup), many hours of capturing with plenty of trial and error — and above all, the one-week PixInsight Intensive Workshop in Valencia, Spain …

All of the above of course deserves a separate post/blog – as will follow in the next weeks. In order to „warm-up a bit“, the following one is about a recent 3D-printing project, which i was planning since the first works on the rooftop-observatory started one year ago….

One year ago …

Let’s waste some sentences on that. Strange to write this down: one(!) year ago, all of this existed only in my thoughts — now it’s reality. I actually enjoy sitting in my private observatory up on the roof, nearly fully automated, comfortably seated under a 2.7 m Pulsar dome with slaved rotation and an automated shutter. The heart of the setup is a 10″ carbon Newtonian riding on the EQ8-R mount, fully remote controllable via N.I.N.A., while the AAG CloudWatcher keeps an eye on rain, wind and clouds and takes care of safety. In the near future, the Lacerta Apo will join in a dual setup, doubling the photon count on every clear night. Humidity control, automated lighting, a flat panel for calibration frames and indoor cameras complete the picture — so that, in principle, I don’t even have to climb the stairs to the roof anymore (although, honestly, I still love to). What started as a few sketches and the somewhat crazy idea of cutting a hole into a perfectly sealed flat roof has turned into a genuinely professional setup 🙂

Back to the 3D-Print: the real revelation first

Beyond the part itself, I have to admit I was genuinely stunned by how this project came together. At no point did I sit down and design the part myself in any CAD software — no sketching, no modelling, no drawing a single line. Nothing but a written description and a few photos of the old plate were enough to end up with a fully functional print file ready to be sent to the printer. And the result was simply overwhelming: a perfect fit, even accounting for the specialities of the material (ASA). One year ago I wouldn’t have believed that designing a custom precision part could work this way. But let’s start at the beginning …

The starting point: a hand-cut relic

Every Newtonian owner knows the rear end of the tube is where pragmatism lives. In my case: a fan plate I had cut out of a plastic sheet by hand back in the day — functional, yes, but neither pretty nor particularly light-tight. It held the main cooling fan (an AVC 70 mm) plus three small 40 mm Sunon fans, and it did its job.

Still, three things were on my improvement list: looks, proper light sealing, and cleaner cutout geometries. The fans themselves were to stay.

Enter AI (and OpenSCAD)

Instead of drawing the new plate in a classic CAD program or in SketchUp, I designed it in dialogue with an AI assistant (Claude, Fable 5), which generated a fully parametric OpenSCAD model from my measurements and 2 photos of the old plate. The beauty of the parametric approach: every dimension — plate thickness, collar clearance, fan pocket sizes — is a single variable that can be changed and re-rendered within minutes.

Printing a 307 mm disc in ASA

Printing ASA is above all an exercise in temperature management — and this is where the H2D plays to its strengths. ASA shrinks as it cools, and on a large part uneven cooling quickly turns into warping. The H2D’s actively heated chamber (around 60 °C) keeps the part thermally stable from first layer to last — doors closed, fans off. The bed runs at 100 °C on textured PEI and the AMS HT covers the second critical factor: ASA is hygroscopic, and fed straight from its heated, dry environment it stays in perfect condition even through a long print. The result: crisp detail, strong layer bonding, and parts that hold their dimensions (hopefully) for years:

The result

The final plate: 4 mm base thickness, a collar gripping 16 mm over the tube rim, and — my favourite detail — a 6 mm labyrinth lip diving into the tube for genuine light sealing without any gaskets. Three ribs and the fan pocket towers stiffen the whole thing, and with the wall lines at 4 mm, the plate is practically opaque.
Full disclosure: the cutout for the temperature display was forgotten in the model and had to be added by hand afterwards — no big deal, and somehow a fitting final touch. Still, the old hand-cut relic has officially retired.