This is a writeup of building a remote-controllable observatory in a remote location in Sweden.
In November 2023 I bought a cabin in one of the darkest areas of Sweden within driving distance ~250km from Stockholm. Darkness and accessibility in the winter were some of the top criteria I used while searching. The location has electricity, no running water and one of the first bits of infrastructure I set up was an internet connection (4G), each of which has some bearing on planning the observatory.
The previous owners were elderly and left behind nearly everything in the house. So for the first year there was a lot of cleanup and maintenance that was more urgent than building an observatory. Still some progress was made.
Preparations
In front of the cabin there was a small building. This would be the future location of the observatory, but on either side there were trees and the building needed to be removed.
I removed the trees in the first winter.
Much of the wood was saved for side projects, but all of the floor was rotten pallets. No regrets removing this, it was a cheaply built party house.
Without the small building and trees, the view of the cabin was much improved.
But those stumps had to be removed to allow the foundations of the observatory. Before that I needed to figure out exactly what I was building. So it was at this point that the design work began.
R & D
First I did my research. There are a few reference books on building small observatories that I read all of which are in the Patrick Moore Practical Astronomy series: Small Astronomical Observatories (1996) and More Small Astronomical Observatories (2002) edited by Patrick Moore were useful in that many examples were given and they had a few technical details which were relevant. Each observatory in these were presented as brief summaries (much like this post) and lacked any real details of process. These are also very dated. Modern construction materials are missing and expectations of the expense are laughably out of date.Setting-up a Small Observatory - from Concept to Construction (2008) by David Arditti gives a more comprehensive picture of building and equipping an observatory. It also gives some guidance on choosing locations and has a fair amount of useful but somewhat out-of-date information for beginners who don't yet own/operate a telescope.
By far the most useful book was Building a Roll-Off Roof or Dome Observatory (2016) by John Stephen Hicks. The observatory I built was very similar to the one proposed in this book. It's also much more up-to-date. It contains extensive drawings for the phases of constructing a specific type of roll off roof observatory. Much of this I had to convert and redraw with European lumber dimensions, Swedish construction conventions and location specific constraints.
I used FreeCAD for the design work.
There were many iterations on the design and the details are probably only useful for my situation. The remote location ruled out a concrete slab foundation so I used the standard concrete tubes 'sono-tube' (gjutrör) instead. This is covered well in the book mentioned above. The height of the telescope in the building works well for my current equipment and should be upgradable in the future. The field of view goes down to ~25° altitude toward South. The building is not aligned to North/South poles rather it's square with the cabin. The nearest corner in the render above is roughly South.
The Foundation
August 2024: Before I started on the foundation I first decided the center point of the pier for the telescope. This location would dictate the rest of the layout and become a fixed reference point of the build once complete. I temporarily laid out the overall dimensions, aligning to the cabin using stakes, string, and a laser level.
I stopped when it became clear I would just have standing water if I went deeper. That would interfere with my armature concrete pouring skills. Is suppose I got lucky that I didn't hit large stones as I did later and that the water table was probably low in August.
In this I constructed the form for the pier. I wanted to remove the entire form when the concrete had cured so I rigged up a form where two steel bars held opposite corners of the form and for the lowest pieces of lumber I connected some rope so that they could be lifted out of the hole later.
Inside those frames the walls of the form were plywood up to the surface level.
When mixing/pouring concrete this lower section was filled before the top section. This allowed the concrete mixer to be close the the hole reducing transport effort. On top of this section was a the above ground portion of the pier with various anchor points (stainless threaded rod) for the metal pier and pipes for cables inside and outside the metal pier. All of this was aligned/leveled again and staked in place.
It turns out this form was barely strong enough. I should have had more frame/braces to keep the plywood from deforming but it all worked out and I was able to remove all of the form once it had cured.
And finally backfilled.
Overall the pier foundation is 40cm x 40cm x ~1.6 m of concrete. The bottom of the concrete was allowed to flair out beyond the form some but not in a particularly controlled way (and it's not reinforced) but this should help with ground swell. The photo above shows fairly nicely that the ground isn't particularly flat. It slopes away from the house toward the road. I couldn't really change that much and with snow melt in the spring I didn't want to.
September 2024: At this point the pier's center threaded rod became the reference point for the build. From it the locations of the concrete tube foundations of the building were laid out in preparation for digging.
I rented a small excavator to dig these holes, dig cable trenches, remove the stumps and fix some other problems around the property. Fairly fun to use but a bit stressful in some places. 12 holes for the foundations in total. Some planning of the order in which to do these was necessary as they would slowly reduce the working area for the excavator and also I knew I would need to backfill by hand so I didn't want the fill placed too far away. I also had a load of gravel delivered for the project that I had to work around a bit.
In the digging I ran into a few very large stones. Some too big to lift with this excavator which I just reburied.
A trench for electrical and network cables was cut to the corner of the cabin where all of that resides.
All in the digging only took a two days. After that it took several days to setup, pour and backfill the concrete tube foundations.
This diagram shows the rough order the tubes were placed.
- The pier as described above to create a reference point.
- The first 3 tubes were relatively straight forward. However these would constrain the designs to be square with the cabin and also set the height of the building and so needed to be carefully aligned.
- The 3 road-side tubes that would be below the building were placed next. Above all these needed to be parallel to the first three.
- The two middle tubes were less critical than the others and they are not in-line with the others because the building is not a perfect square and various framing hardware dictated their placement somewhat.
- Finally the 4 tubes for the roof gantry section. These also needed to be parallel to the existing tubes. They are not as deep as the ones under the building. They only carry a load when the roof is open (and they needed to be higher so they would not be an annoying hazard at toe level).
The tubes are all ~1m below ground level. Some rest on large stones that couldn't be dug out.
Setting the tubes was by far the most labor intensive part of the entire project. I could only do a few tubes a day working alone. This step took ten days from delivery of the excavator to the state in the photo above. To round out September I laid down a water permeable weed barrier and a layer of gravel where the building would later be built.
This is how the project would sit for the 2024-2025 winter.
The Building
In March 2025 I took careful measurements of the foundations again, and finalized the building plans. In late March I started the building's floor. While also doing astrophotography at night.
First the outer joists/support frame. At the corners these were cross-lapped and on the inside backed-up by an angle bracket to prevent twisting.
Then floor joists were added on hangers. With clearance and holes where needed for later cable installations and conduit.
May 2025 after the spring melt. Some additional supports added to allow the flooring to be screwed down securely and stabilize the joists.
Then flooring panels were cut. These are laid out such that the inner ones can all be unscrewed and pulled up in the final design allowing full access to underfloor wiring if needed (a suggestion from the books).
The floor panels were removed and stored temporarily as they would be in the way and were not weather resistant.
During this spring many trips were made to the nearest town (~2hours round trip) for lumber. Weirdly, impregnated lumber was the least expensive at this time so the entire floor and frame are made from it. Flat and angle iron were sourced for the roof rails from a recycler in Stockholm. Roof materials and siding were also purchased and hauled to the site.
Before raising the walls I sourced a solid core door second hand and built a frame for it to have reading for installation. It was far from perfect and not a pretty door, but it fits the vibe on the property and was basically free. The frame was made from the (hopefully) very dry and stable studs of the building that was demolished.
In mid May 2025 a friend, Bert-Åke, came out to the cabin for a week to help me build most of the building.
Day 1
We began by creating a simple jig on the floor joists to build 7 rafters for the roof. As described below this roof is probably overkill. We do have the chance of high snow load here. But it's probably twice as strong as it would really need to be.
Those were set aside and we installed the floor panels and used the floor as a convenient place to build the stud walls. All four were built and set aside. Somewhere during this time we carefully leveled the floor using the adjustment bolts on the foundations.
Day 2
First to be raised was the north wall which will have the door frame. And at this point we began to race the clock. We needed to get a roof up before it rained to prevent damage to the floor materials. It wasn't 100% critical to achieve this but we did manage it.
Followed by the west and east walls.
Then the south wall.
The stud walls were then all plumbed and secured. The next photo clearly shows the mortis designed into the stud walls for the gantry rail's joists.
Next came the gantry section and the careful measurements to make sure this would be a parallel support platform for the rails.
Diagonals were added to the gantry framing and it was secured and finally the rail was installed. The rail was made by stitch welding angle iron to flat iron.
The rail was fabricated and painted then cut to length at assembly. It was carefully made parallel and screwed down to the top of the stud walls and gantry. (Later in December I measure the rails parallelism to be within 6mm over the full 6.5 meters).
Day 3
On top of the rails we installed the preassembled caster running boards upon which the entire roof sits. There are 4 steel casters per side. These running boards were fixed in place to allow the roof to be assembled in place on top of them. Temporary angle brackets were preinstalled on the running boards to set the spacing of the rafters.
The roof rafters were raised. These were laid upside down on the rails and rotated upright minimizing the awkward lifting.
The rafters were then covered from the bottom up by pre-painted lap siding. This was initially planned for the walls, but was the wrong profile, so we used if for the roof to give extra support to the sheet metal roof.
This panel on its own would shed water so it forms something of a backup to the metal roof and also makes is so that if the next owner of the building wants a normal insulated building, they have something typical to work with in the roof.
Five hours later we had the sheet metal roof installed, though it was missing some trim pieces.
Day 4
At this point it rained for a day but we did test the roof for the first time and installed eaves/facia and some spacers between the studs for the siding to be nailed to.
Day 5
Before siding, the walls we wrapped the stud walls with a wind proof fabric.
And then nailed up siding starting with the east wall. This siding is tongue and groove with a profile that matches the other buildings and is very common in Sweden. It came pre-painted on one side with the typical Falu red color of the other buildings.
Day 6
Followed by the South and West walls.
And lastly the slightly more complicated North wall and we hung the door. This was the last day that I had Bert-Åke's help. The building was 95% complete. I'm very grateful to have had his help! Six days to build a fairly complicated and precise building.
Day 7
For the siding on the roof triangles I laid out a diamond of siding, tacked it together and then cut the diamond in half.
Then the pitch of the roof was copied to the triangles and they were cut together.
The roof siding was lifted up in sections and nailed in place. Screws were used in the upper part under the eaves. At some point I also installed the wind blocking fabric in these roof triangles (from the inside though).
Finally some trim pieces on the corners and at the roof/wall split to hide the gap.
Finishing touches on the building continued happening over the summer.
In the roof, between rafters some additional siding was added.
A coat of paint on the pre-painted lumber. Additional braces on the gantry. The last few pieces of roofing trim and some stairs.
A bit of roof scraps added over the door, but with clearance for the gantry braces as this moves with the roof.
Some flashing added to deflect blowing rain from getting into the gap.
The next thing added was electricity from the cabin. The observatory has outlets and light fixtures on all 4 walls, all of the electrical stuff is outdoor rated. I also used RGB LED light bulbs (Ikea) so I can make them red and much dimmer at night as needed.
During summer/fall 2025 I made a lot of progress on the internal systems but it would be confusing to show them chronologically. So they're broken up below with no specific time line.
The Pier
Building the pier for the telescope was the next priority. I wanted to push toward the observatory being usable for the winter.
The pier is a stainless steel tube, 22cm outer diameter, 120cm long. My current telescope mount is only 16cm diameter at the base. This tube was chosen to fit a future upgraded mount and also it was from a recycler in Stockholm so the size wasn't chosen too specifically. The tube cut to length and I cut two holes for wiring into it.
I also prepared the base of the pier, a mild steel plate 60cm x 60cm. This was part of a metal frame for loading logs in a saw mill that used to be on the property. I dug it up when I had the excavator on hand.
The tube was welded to the plate to form the pier. And then primed.
Stick welding thin walled stainless to mild steel isn't an easy second welding project. I managed to make a strong weld even if it's not pretty.
I ended up using a filler putty on the weld after priming. This is because all of the metal stuff ends up with a lot of frost and dew on it which will run straight into this weld's imperfections.
After that it was painted black and ready for installation. There are 5 anchor points in the concrete only 4 are used since the middle is pretty inaccessible. There are nuts above and below the base plate to allow the pier to be leveled and secured. It probably would have been better to have the cables and holes in the tube on the south side of the pier. I usually place my batteries and computer on the north east side of my portable setup so that's where I put them on this. But they ended up facing mostly east on this pier which is a bit in the way when aiming at higher altitude targets.
Next I made an adapter for the telescope mount from some scrap mild steel plate. I consider this temporary as I would like to upgrade to a large mount so I used some pretty ugly scrap for it.
With that done the observatory was officially usable.
Moving the Roof
The mechanism for opening/closing the roof automatically was a conundrum for a while. Many ideas were tossed around with pretty much all of the mechanically minded engineering nerds I know here. There were many competing requirements and features I wanted. I didn't end up getting everything I wanted, but I'm happy with the current solution.
The Requirements:
- Opening/closing the roof should be possible without grid power.
- Opening/closing the roof should be possible without battery power.
- The mechanism for opening/closing should be able to handle a moderate snow load on the roof.
- In general the mechanism needed to handle the weather here.
- The electronics should be able to be controlled remotely and locally (without a PC).
Nice to haves:
- The mechanism should be reasonably secure.
- The mechanism shouldn't look out of place on the outside of the building.
- The mechanism should have a power-lost fail-safe.
Most of that was all met by the solution, except the last nice to have. I couldn't figure out a gravity or spring loaded closing mechanism that wouldn't be very difficult to build or wouldn't look strange.
The mechanism I landed on was a steel cable loop that could pull the roof open and closed. The overall loop looks like this. The cable only pulls one side of the roof for now it seems rigid enough for this to work.
To drive the cable I've got a 12V winch modified with a chain drive connected to a plastic cable drum (another contribution from Bert-Åke). Initially I built the complicated tensioning portion of the mechanism from wood.
The wooden version worked as a proof of concept. But it was clear that it would not last and the cable needed two tensioners to work properly.
So I built a new version out of steel. This photo also shows how the cable connects to the roof and where the ends of the cable are terminated.
Here's a better view of the wench and cable drum. The wooden mounting for these may end up failing, but I have had to change the distance between the two to get good chain tension so for now it works. The wench came with a manual clutch that allows the wench to spin freely. This allows the drum to move freely too. So it's possible to move the roof manually if needed but it's not that easy.
On the gantry there's a final pulley that completes the mechanism.
Here's a full view of the mechanism on the inside. The distance between the cable drum and the pulleys here allows the cables to run closer together when they reach the drum, which helps with spooling but it's not quite perfect.
At either end of the stroke of the roof motion there's an inductive limit switch.
The wench and limit switches are wired into an EM-282D motor controller from Electromen. This controller can handle ~4 times the current that the wench draws (~30A @12V). But I wanted overhead in the controller in case I need a bigger wench or if I change the mechanism. The controller is programmable and required some configuration to work with the limit switches correctly and I've slowed it down to 70% duty to slow down the roof.
The input to the controller is from both a wall mounted pendant and a Raspberry Pi 5, via a relay board 'hat'. Along with that stuff, in the roof control box there's a Victron charger/inverter that maintains the large 12V battery that is the current source for the roof mechanism and the telescope. This is still an evolving configuration. The inverter and the motor controller can probably be monitored directly from the Raspberry Pi if I take the time to find/write software for it. Currently the inverter's AC output is unused.
There are of course a few fuses and kill switches for the electronics. If this settles to a stable configuration I'll publish the circuit diagrams on my github but they're quite specific to this installation and likely only useful to someone who has enough experience with this stuff to come up with their own design.
On-site the pendant allows convenient control.
The Raspberry Pi is networked and running a custom implementation of an Alpyca Alpaca driver that I connect to with N.I.N.A. That will also be on github when I'm happy with it.
The battery is installed in a separate box with insulation. It's a standard (large) car battery 105Ah, 12V.
On the south side the east/west walls of the building I added brackets that form a pin and backup hard stop for the roof. These should help with preventing a very high wind from blowing the roof off but that's difficult to prove.
In a separate cabinet I've installed an always-on NUC to control the observatory and telescope remotely. In the same cabinet I've installed the network hardware and a battery backup. The entire network for the cabin is on battery backup(s) and the observatory is wired into it, so it should allow remote control if there's a power cut. The hardware in here gives off enough waste heat that the battery backup shouldn't suffer too badly in the winter.
Fin.
That's pretty much it. In the winter of 2025-2026 I've been using the observatory while I've been at the cabin and done a fair amount of testing of the roof and equipment.
We've had plenty of snow to test the roof with.
Despite this winter being extremely cloudy, the convenience of the observatory has made it possible to take more and better photos than ever before. As with any large DIY project there's more to do and improve, but I'm happy with the result.
The telescope deserves its own write up, it's eight different controllable components and probably not that relevant to anyone but me.
In February 2026 I finally used the observatory remotely. You can find my astrophotography on flickr.
Comments
Post a Comment
Comments are moderated. Comments containing links are marked as spam.