My Roll-off roof Observatory.


I had two major goals when I designed my observatory; it had to be very easy to make, and I had to be able to walk inside it with the roof closed without stooping. The first consideration ruled out a traditional dome and pointed toward a roll-off roof design, while the second consideration defined the minimum height and probably conflicted with the first.

Why the demand to be able to stand? I have visited many amateur observatories and discovered that simple designs often had access through low doors or were too cramped to work in with the roof closed; this makes it difficult for less agile visitors. Nor did I like the prospect of having to crawl around inside an observatory in my declining years.

Observatory with roof openThe observatory looking north, open and ready for use.

In many cases, amateurs' observatories are constructed mostly with timber frames. Construction is easy but usually slow with all the necessary cutting and joining. However, after pricing the timber required to make even a simple 2·5 metre square observatory, I looked for another way.

I found it by looking no further than my garden. When my house was being built, I had put together one of those kit garden sheds; it was about 2·8 metres square and 1·8 metres high with wide opening double-doors. At the time I bought it I wasn't really interested in shed styles - I simply bought the one on special; but when I looked closely at it with the aim of turning it into an observatory I discovered it had some very nice attractions. It was very easy to put together; it was less expensive than the timber necessary to make an equivalent size shed; and most importantly, the walls and gable roof met squarely and so could be made to separate very simply.

The last point was the most significant. Many of this style shed have the end wall forming part of the gable, but in this particular shed the wall panels were all identical (obvious savings in manufacturing costs here) and the roof was a separate construction. All I needed to do was to make a frame to hold the roof together, give it some wheels and a track on which to roll and the observatory was made. Or at least in principal.

I bought the largest size shed I could find with this type of construction; which was approximately 3 × 4 metres. It would be plenty big enough for two 20cm telescopes, or a very generous size for one larger one.

One lesson I learned from my first shed was that to stop rainwater from running down the walls and then coming inside at floor level, don't make the concrete base any larger than the shed. So this time I waited until I had the shed and measured its exact size (which is usually different to the size stated on the box). For the brand shed I bought, this only entailed screwing the pieces for the base frame together and fixing formwork underneath it. Thick black plastic was placed underneath to stop water from leaking up through the concrete base.

As I wanted to erect the walls before I placed the piers for the 2 telescopes, I initially cast only a 100mm wide section of concrete. What followed was a minor mistake, but one that should have been obvious. Impatience led me to erect the walls of the shed before I was ready to install the rails. I hadn't finalised the design of the rails or rolling mechanism (or more precisely I hadn't a clue how I was going to do them and wanted to see the walls up before I could proceed). The problem is that these garden sheds have little overall strength until all the panels are assembled, roof included, and the whole lot is then bolted to the ground. Consequently, once the walls were up they were free to flap around in the slightest breeze. And flap they did! I secured them with lengths of timber braced against each other, but the walls still show wrinkle marks generated in the 2 weeks before the roof and rails went up.

My next task was to make a frame to hold the roof. I bought lengths of 25mm RHS (rectangular hollow section) steel of 1mm wall thickness for this. I next learned how to weld. This was done by asking the welder at work for a few pointers and then getting a 20 minute lesson from my friend Chris McCowage. Despite this severe lack of expertise with a welder, I made a passable frame that was square enough to work.

The frame was painted and then the roof panels were assembled on the frame, secured by the self-tapping screws supplied with the kit. An electric screwdriver was a great help here, as it was when assembling the walls as there are several hundred self-tapping screws and nuts & bolts holding all the panels together.

With the roof now together, I could work out distances for the wheels and their axles. I had already bought 6 rubber tyred wheels of approximately 105mm diameter. They had internal roller bearings and only required axles to function. Now came the most difficult part; getting somebody to machine the 6 axles. The diagram shows the final design for this.

Detail of wheels and axles

Design of axle with respect to shed wall and roof-line. Both ends of the axle are drilled and tapped. A bolt and washer at each end secure the wheel to the axle and the axle to the 25mm RHS.

The rails were simple. 8 metres of 40mm right-angled steel supported by 5 evenly spaced pillars of 25mm RHS. These were welded together in a few hours along with small lengths of rod placed at right-angles at the bottom of the pillars to act as anchors into the concrete. They were then carried down to the shed and holes were dug in the ground about 300mm deep to take the posts.

assembling roofThe track being assembled, tied up to the wheel; the roof supported on a spacer.

With the help of 3 friends, the roof was carried down to the shed and placed onto four 20mm wooden spacers atop the walls, one at each corner. The rail assemblies were tied up to the wheels, checked for level, and concrete poured into the holes. It was then left to cure for a few days, after which I secured the left and right rails together by welding 3 more pieces of the 25mm RHS steel between them. When the ties holding the rails to the wheels were undone and the spacers removed, the roof sagged about 10mm as the weight was taken onto the axles. This was expected and turned out to be ideal as the gap is small enough that rain doesn't enter, but allows air to circulate freely helping to keep the building cool.

I found that the track structure was still a bit "wobbly" so I braced each end with fencing wire (the traditional country remedy) diagonally across each end, held taught by turnbuckles. This made the structure very stiff, although I must get around to highlighting the rather dangerous wire cross on the open end of the rails.

The only other problem encountered was the doors were not watertight. The double-doors were an attraction and allowed me to easily pour the concrete for the piers and floor once the shed was assembled, but they simply butted closed without any seal and so allowed water in. I solved this with a strip of metal attached to one door which overlapped the other when closed.

Electricity to the observatory had been provided by the barter method; I made a 20cm f/7 mirror for the new electrician at work in return for materials and some labour.

The finishing touches were a means of securing the roof closed and to attach the walls to the outside frame to keep them from being destroyed in the wind. The roof was secured by 4 hasp-and-staples as shown in the photograph below. I drop a bolt down to stop them undoing if they rattle in a strong wind. The walls were held by welding small right-angle brackets to the track and screwing them to the walls. This is where I discovered that welding joints with the weld on the underneath makes welding on the flat seem easy. I'm told the trick is to increase the current and use a shorter arc, but I never quite got the hang of it. It works, however, and looks fine in the dark. Finally, a carpet was found to help keep the feet warm and help cushion the fall when I drop something valuable.

Latch to hold roof downRoof locking mechanism.

And how does it work? Brilliantly! The roof rolls off easily and there is enough room for my telescopes. It was originally intended for two german equatorial mountings with 20cm Newtonian telescopes - a visual f/10 on the eastern side while a photographic f/4·5 is on the western. I currently still have the f/4·5, but the f/10 has been sold and in its place is a Dobsonian mounted 31cm f/6·5. I have an old bookcase on the south-eastern wall which holds various items, and a small table in the south-western side which hold the computer and power supplies for my CCD system. There is plenty of room to work inside, both while observing and performing maintenance during the day (or cloudy nights). There is also sufficient room for several visitors at one time without overcrowding.

Internal view of observatoryFisheye view inside the observatory.
This is an old photo which shows the original pier for the 20cm f/10 in the foreground, 31cm to the left, the 12cm RFT right rear, and the 20cm f/4·5 in the background. The adjustable seat for the 31cm is on the right with toolbox, finder & focuser from the just disassembled 20cm f/10.

An unexpected bonus with this observatory is that the walls and roof are very thin and hence have little thermal inertia. Once the sun has gone off them, they cool to the surrounding air temperature very quickly. Helped by the small air gap between the roof and walls, it stays remarkably cool inside even on the hottest summer day.

The drawback with such a simple building is that the walls are too high for access to the horizon, although I had already calculated that the South Celestial Pole would be well above the wall as seen by the telescopes and so the loss of the lowest 20° of sky didn't bother me. Having walls that fold down to enable the horizon to be viewed would complicate the design considerably. A Cassegrain telescope could be mounted on a higher pier than my Newtonian and this would allow a lower horizon. However, the wall height is essential for keeping out lights from cars as they drive along the road at night. It also keeps the wind out.

The cost could have been lower had I been able to get to a scrap yard to buy second hand materials, but such things are a minor drawback when you live in a small country town. Also, only 4 wheels were really needed, but the 2 extras didn't add significantly to the cost. The materials cost approximately:-

$425 for the shed kit
$ 36 for the wheels
$180 for the steel for the frame, track and supports
$100 for the cement, sand and gravel

Apart from the final concreting of the floor and piers, I spent only 3 easy weekends constructing this observatory. The first weekend I prepared the ground, layed out the base and poured the concrete rim for the shed. The second weekend I erected the walls (and braced them against wind damage), cut the steel, welded the roof frame and rails and assembled the roof. Finally, bringing it all together was done on the last weekend. Apart from moving the roof which took 3 other people, the work was done single-handedly or with the help of my wife.


Update - 2003/04/02

This observatory has proven itself an excellent design. It is large enough (just - but I try to put too many telescopes in it) and has given me many, many hours of enjoyment. The only maintenance I had to do was after about 10 years when the roof started to leak. The problem was the little plastic washers which go under the screws holding the roof on - they had deteriorated in the sunlight. This hasn't been a problem with my other shed, but it's roof doesn't move so I assume that the vibration while rolling the roof was causing the washers to break loose once they had cracked. I replaced the worse washers and then covered them all in blobs of silicon to further protect them. So far this has solved the problem.

One visitor to the observatory called it my "skyshed" and the name has stuck.

The only limitation to this design of observatory is that any equipment is open to the elements while observing. Not a real problem for the telescope in use, but for computers and electronics it can be. I have now added a control room extension and this has significantly improved matters. Read about it here.


home back to ATM on to control room extension


Page last updated 2003/04/02
Steven Lee