Skybadger Background For Sale Leyenda website Home

Building the Observatory

  1. Why
  2. Requirements
  3. Materials
  4. Design
  5. Construction
  6. Skirts Assembly
  7. Dome Assembly
  8. |
  9. Observatory Network

Why ?

The reason for building an observatory become stronger based on

  • age
  • comfort
  • weight and complexity of telescope setup
  • alignment requirements
  • amount of time available before the wife complains
  • space
  • building skills

So far I've got all but age and that's fast catching up. Having made the decision what do I do to make it happen ? Well in this case I designed my own - based on a stubborn confidence that I can do my own design using my own tools and it will stay up in the teeth of British weather.

There is also a growing amount of data on the web regarding building your own observatory and a chunk of is referenced here in the resources section.

In the end the plans looked like this:

observatory plan view

and the plan view like this :

observatory elevation view

Which maps onto the real world like this :

Fig1.Floor plan Fig2.Rign to upright joint detail Fig3.Front Elevation

So the basic specification is as follows:





6 feet - telescope pier height adjustable


octagonal, opening roof


bolted and screwed/glued 2" square section timber


4 feet into octagon apexes

dome radius

4 feet into zenith from top of skirt

The requirements for the construction format were based around several things :

  • I might have to move again within the next 15 years so I want it to be sectional, removeable and erectable by one person if possible. Hence bolts, a lightweight structure and removable panels form the centrepiece of the design.
  • I already had a concrete apron near to the house on which to site the observatory. I would have liked to place it further away but a large tree in the garden reduces the field of view from every where else and I can't get permission to reduce it in size.
  • The observatory site is near to the house, making cable runs easy.

So the order went in for 450 worth of timber to my local yard. This comprised

bill of materials for observatory construction




Unit cost

Total Cost

2" sq section tanalised timbers



50 per 100M


5" x 3/4" ship lap cladding



180 per 100M


6" x 3/4" t&g floorboarding



132 per 100M


3/4" marine plywood sheet





150mm coach bolts





100mm coach bolts



3" woodscrews




M8 Studding for Telescope Pier floor mounting



2.5" woodscrews




m10 plate washers


5.00 per 100


m8 tuppenny washers


acetate/acrylic roofing sheet



2mm x 30mm rolled steel geodesic triangle former



2mm welding rods


Bolt-down Metposts for 2" Section timber





The design is fairly simple - two octagonal rings , one running around the top and one running around near the bottom, provide structural rigidity to the eight uprights which are bolted to the rings at each flattened apex, using large coach bolts, wedge supports underneath and apex stiffeners. ( see fig2. above ).


Each face of the octagon then has two verticals dropped down from top ring to bottom ring, at about 6 or 7" in from each corner to provide the surface to mount the shiplap cladding on. An alternative which now occurs to me is to route a 22.5 deg. chamfer into the side of each upright to seat the end of the cladding squarely in. Currently I've had to cut all ends of the lengths of cladding to a 22.5 deg. angle to finish flush against the uprights!. The idea of the skirts is that since these are coach-bolted on at the intersection of upright and rings, using blind coach-bolts, they are easy to take off but only from the inside. So I can build the entire frame and cladding inside and then move all the panels outside later. Even the door ( admittedly with more complexity ) is built as a skirting panel which hangs on the frame.

The addition of the skirt panels also stiffens the frame up considerably, putting horizontal and rotational stiffness there which is lacking until now. Due to the use of the chamfered edges on the skirt cladding and the wood not being stored flat once assembled, they also sometimes bow outwards. Again this can be cured with an internal stiffener ring pulling the skirt inward at the centre of the panel and providing further stiffenin and a horizontal mounting surface for interior fittings.

The octagonal building skirts, lying on their side in the background, including the one with the door in it.

Dome Ring Surface

The dome ring surface is 3/4" marine exterior plywood, cut in ring sections of almost 60 deg. The sections overlap to form two layers and are glued together with exterior grade waterproof wood glue, with a piece cut to fit the exact distance to complete the circle. You almost have to do two layers otherwise there is almost no way t reconstruct the original circle, other than a drawing on the floor. With two layers you can use the overlapping top layer to guide the curve of the bottom layer and vice versa to reconstruct the original circle. Mine ended up larger then the original cutting radius.

After filling any gaps with exterior filler, the ring is sanded and centered on the top octagon. A few kicks, checking the concentricity of the ring sitting on top of the topmost octagon before screwing and gluing the two together and the top ring is done.

The top building assembly support octagon, the spindly one, stands upright in the background.

The dome ring surface, lies flat on the building assembly lower ring, with flooring supports visible in the middle.


The dome was going to be a half-geodesic dome based on designs and information for a 2-frequency geodesic dome found at desertdomes . This dome is essentially 5-pentagon design with 10 facets along the bottom edge and overall 75 edges to cut and then assemble. In this case the bottom edges are going to be assembled from 10 50x50x3mm right-angle angle iron sections, welded in to the based decagon. The triangles comprising the hexagons and the pentagons which now comprise the rest of the half-dome are 30x3mm flat steel, cut to length and with the ends cut to the correct angles for insertion into 4, 5, or 6 way joints which are all welded together.

Overall the angle-iron base gives the strength and rigidity for the rotating surface and the flat steel welded sheet rib geometry provides a lightweight superstructure. The interior of each of the triangles is filled with 2mm acrylic sheet. This is tough, light, durable and easy to fit by riveting at 3 points along each edge and then using silicon mastic to weld the length of the edge to the steel and make the join waterproof.

In spite of using steel I calculate the overall weight of the roof to be less than 100 lbs..

The base of the dome has an overall radius of 4 feet into the apexes as does the side of the building. I intend to fit it with 5 or 7 castors for asymmetric loading on the wooden roller surface, avoiding all castors being in the same relative position on the rim at the same time.

In the end the dome was actually purchased (!) from an Ad in Astronomy Now. The constraints of time due to a Baby on the way led to the balance of cost of a second-hand dome against cost to build and time to build and in the end the cost won.

The dome was a home-built fibre-glass dome based on a polystyrene foam skirt and balsa wood formers bent to shape to create the curved surface to the edges of the dome hemisphere sections, where the two halves bolt together across the door. Both fibreboard and plywood are used as stiffeners and internal frames for the shutter and lockable door. Any points of ingress of damp into the glassfibre was cleaned out, foam filled, pva sealed and painted with smooth finish exterior paint.

Here's a sequence of images of the exterior erection of the dome.

The base building skeleton fully assembled on the concrete patio apron waiting for the skirt panels to be added. The telescope mount is already bolted to the ground on standoffs for levelling adjustment. The rear skirt is already in place.

View of the floor joist framework and one of the metal tripod legs ( ex. High Energy Physics target stand ) prior to the floor being installed. The floor is isolated from the tripod to remove vibrations into the telescope.

View of the joint between the building skeleton and the base octagonal ring.

Detail of the installed skirt panel.

View of the assembled dome building with the dome ring installed on the top ring surface ready for the GFRP dome to be added on top.

View of the installed GFRP dome on the building

View of the assembled dome building with the dome through the open observatory door.

Observatory hardware

The observatory needs a surprising number of systems to function. currently the list includes multiple USB hubs, USB-to-I2c adapters to talk to the dome drive , USB-to-radio adapters to talk to the systems on the dome ring and serial comms adapters to talk to all the serial interface hardware such as remote focusers, camera drivers etc.

The current dome system diagram looks like this

Here's some of the systems taking part in the observatory integration:

The dome rotation motor controller in an IP66 control box with a LCD display for reporting the dome azimuth and sideral time. The motor controller is a Pulse Width Modulation controller that implements an I2C interface from robot electronics. They also provided the I2C LCD display which comes with a backlight that will need to be modded to be dimmer.

The dial is an alternate control mechanism : a 10-turn helipot which at the midpoint turns the motor off, elsewhere, the speed and the direction of the dome rotation is proportional to the distance of the helipot from the mid-point resistance.

Image of the dome ring and dome wheel. There are eight stanchions around the shed ring, made from 100mm x 6mm angle iron. there are two patterns - four have two holes for the stainless bolts and four have a single hole. since the ring is made from four quadrant pieces supported at the middle and ends, this should explain the difference.

The glassfibre dome has 6 aluminium strips moulded into the resin. the strips provide the mounting surfaces for the bolts which carry the wheels and the bolts which allow jockey wheels to ride on the bottom of the ring to prevent lift-off. ( not seen in this photo.. )

Image of the dome motor. There are two of these motors mounted at 180° across the dome. These press on the dome base ring to drive the dome. The motors are windscreen wiper motors from a scrap yard with rotor extensions added to the wiper crank arm spigot. The extensions use M12 threaded bar tapped and split and then bolted to the (M8) threaded spigot. The wheels are stock items from B&Q and use a pair of M12 nuts to hold them from rotating on the threaded bar.

The mounts are 4"x3" angle-iron with a bent 6mm strip aluminium mount riding on 3 shouldered M6 bolts using a hard spring on the bottom bolt for drive tension against the dome ring.

Front elevation view of the above on the workbench during assembly.

The observatory dome drive as configured above has been slipping for a while now. Basically the dome is not perfectly round and so in places the skirt rubs. Add to that the use of smooth wheels on a smooth drive surface and the inevitability of not being perfectly circular and you get a dome that sticks in some places. that's not terribly useful when otherwise the dome would track the 'scope using the ASCOM dome hub.

New drive pulley and belting

Front elevation view of the updated drive wheel showing belt and pulleys mod.

I went off to Fiveways Bearings to get a quote on using timing belt around the drive edges and suitable pulleys to bear on it. John Smith came up with 25 foot of 1" wide H100 poly belt with 1/2" pitch and two pulleys to ride on it. I cut the belting up into sections and glued it to the flat drive surface with EvoStik. Then I took the old wheels off the motors and replaced them with the pulley wheels. The end result is a dome that gets the full driving torque of the motors with much less slippage and that is a great improvement.

Since it's sufficiently good and the belting is sufficiently low profile that I have purchased more to use on the up-and-over shutter. Smaller gauge belting has been glued to the rackways on the dome and matching pulley wheels now ride these. The axle joining the pulleys has a gear on it to be driven by a worm wheel for increased torque. Currently the primary problem with hits is keeping the door traking the trackways- the pulley flanges are not really big enough to do this.

I do think the last improvement required is to change the motor mounts to to horizontal-hinged rather than vertical-hinged. What I mean by this is that currently the motors hinge upwards to press back against the rim, This makes it harder to adjust the tension and release the motors than if they were horizontally hinged with a larger swing range.