Remote control.


Many visitors come to the Skyshed to image the wonderful southern sky. Some visitors kept coming back for more and ended up leaving equipment behind so they didn't have to keep bringing it out. This has ended up with my primary mount, CCD and some other equipment belonging to somebody else - just so that they could have it ready for the couple of weeks a year they could come and visit. I'm not complaining!

As technology advanced, remote observing became the way serious observers worked. This is difficult where I live as high-speed internet doesn't exist. I'm too far out of town for ADSL to work, even if the antiquated local sub-exchange that I'm connected to could do it. For some years I had a very expensive (and slow) satellite connection. Remote access wasn't a viable option with such a pathetic system.

But that change recently as I was able to get on to a significantly cheaper and faster satellite connection. Now it was possible to contemplate remote observing. It wouldn't be ideal - the 2MBPS upload speed (4MBPS download) wasn't quite as fast as was desired, but it did work. So now it was possible to bring the old Skyshed into the internet age.

First things first

The very wet weather of 2010-11 highlighted a problem I'd worried about when I added the control room on to the Skyshed. My property has a slope across all of it, with the Skyshed on the down-hill side. One end of the control room was at ground level, the result of adding it on to the existing shed whose base was only just a bit above the ground at one end and keeping the same floor level into the control room. Once the ground had become saturated with all the rain, it only took one more heavy downpour and there was a river flowing over the ground towards the Skyshed. The water hit the control room and got in, spreading all through both sections. No equipment was damaged as I had shifted it above the floor (I had been worried and took precautions), but the carpet needed to be removed and the whole place needed drying out. This happened 3 times in a few months during the big wet. Something had to be done...

I couldn't raise the Skyshed, so I had to lower the ground. Once the ground had dried out sufficiently to take the weight of a machine (it got bogged the first time we tried!), it took a little over an hour for a man and a Bobcat to shift the dirt around the Skyshed and provide enough clearance so the water couldn't get in should it happen again. You can see the cleared section around the Skyshed in some of the subsequent pictures. (I'm mentioning this in case somebody wonders why there's signs of new construction despite the Skyshed being built in early 1989.)

I'm also insulating the walls with styrofoam panels. So far I've done the roof of both the Skyshed and control room, and the northern wall of the Skyshed. I've yet to do the eastern and western walls, although I've made a start on the eastern wall. This is to try to cut down on condensation forming on the walls and ceiling in the morning after a cold night, as well as to keep the heat out in summer.

New tracks

The first thing that would need to be modified was the roof tracks, so that the roof would roll perfectly, and was fully restrained. The old tracks, while perfectly adequate for the past 20 years, required somebody present at the Skyshed to unlock the roof before rolling it, and lock it after closing. The rubber-tyre wheels required a bit more push than was desired, too. I had to exclude any possibility that the roof could come off the tracks when rolling, and whether open or closed the roof would remain solidly in place even in a strong wind. Then it had to be motorised, and finally put under computer control.

Many options were considered, but I decided on V-groove wheels rolling inside C-channel steel. The original plan was to remove the existing angle-iron rails and attach them inside the C-channel. The wheels then had something solid upon which to roll, and they would be constrained nicely inside the C-channel. However, once the wheels and channel had arrived, I decided that the angle-iron was unnecessary and the wheels could roll along the edge of the C-channel and would still be restrained by the top edge of the C-channel. This proved to be the right choice as it was far easier to do, and works beautifully.

My friend Colin came up for a week in early April to help with the work. I'd already started before he arrived by removing the old wheels from the roof frame and modifiying the old axles to take the new wheels. (The old wheels had 13mm axles, the new ones 12mm.) I'd also removed the old angle-iron tracks, leaving the roof proped up on the walls on some pieces of timber.

The 100mm C-channel comes in 6·5-m lengths, but I needed the tracks to be 8·5-m long so our first job was to cut one up and join it on to the other 2 to make the required length. Next the new tracks needed to be attached to the existing uprights, and aligned parallel. This required cutting and welding some short sections of 25mm RHS to attach to the existing upright posts. The joints were butted together and joined by bolting short sections of the C-channel to the outside. The final track, joined and attached to the uprights, looks like this.

The rail joint.

If for any reason the wheels have to be removed, we cut holes in the sides of the rails in strategic positions. (One such hole can be seen in the above picture on the distant rail.)

It wasn't too difficult to line the tracks parallel and at the right height. A few well placed and angled spacers did the trick (as can be seen in the previous and following pictures). I placed some long bolts at the end of the tracks to make sure that the roof couldn't possibly come out of the end. All the welds had to be painted, of course, and I went over the uprights as well.

The V-groove wheel in track.

Once finished, the roof rolled extremely smoothly and freely. This would ensure that the motor drive would have little trouble performing its task

Motor controlled roof

There are many ways to motorise a roof. I already had a old garage door motor and controller which I thought could be used. While it was old, having been in service for some 15 years, the motor seemed fine. It was just the radio was a bit weak (which didn't matter). So I had a good motor and controller available. I now needed to make the motor drive my roof.

I considered a rack & pinion gear drive (commonly used for sliding gates), but eventually I settled on what I considered the simplest method to move the roof - a wire rope wrapped around a pulley and attached to the roof. The motor drives the pulley which pulls the wire; the wire pulls the roof. It works for both opening and closing. The only difficulty is that I'd have to custom make the pulley mechanism.

I designed the necessary parts around what useful pieces (bearings, scrap metal, etc.) I had in my junk box (as I always do) and took them to my friend Eric to make (Eric is a retired machinist who has his own mill and lathe and has often helped me make parts. He made the filter wheel for my Cookbook CCD many years ago).

The only critical part of the design was the diameter of the drive pulley as this would dictate the speed that the roof would move. I measured the speed of the output shaft from the motor unit (which already had a worm & wheel reducer on it) as about 90RPM. A 50mm diameter pulley has a circumference of about 157mm, meaning that it would have to do about 26 turns to move the roof 4100mm. It would take about 17 seconds to move this distance at a speed of about 240mm/sec, which seemed about perfect. (A 75mm diameter pulley would have opened the roof in 12 seconds, but as I only had some 50mm material then it would suffice.) A few days later Eric called to say all the parts were done.

In the mean time I was busy cutting and welding parts for the pulleys and roof attachment point. As the roof rolled so well there was no need to pull the roof from the centre (as an aid to keeping it centred) so I could mount everything on one side, thus making everything easier. The following pictures show the final arrangement.

The Skyshed pulley arrangement, with the roof closed. The Skyshed pulley arrangement, with the roof closed.

The roof has an extension piece to take the attachment point of the wire beyond the pulleys. The wire is first attached to a turnbuckle to allow tensioning. From there is goes to the distant pulley (photo below) where it loops back towards the Skyshed. The returning wire is directed downwards to the motor, where it wraps around the drive pulley and then returns to the other pulley. From there it attaches firmly to the roof, at a point just behind the turnbuckle.

The distant pulley. The distant pulley where the wire loops around the far end of the Skyshed. The pulley is mounted on an extension frame so that the pulley is beyond the furthest point the roof will open.

The motor end of the system. The motor end of the system (while the system was being tested, hence the cover is off and you can see the mechanism).

The motor and drive pulley is mounted below the top pulleys to minimise the angle that the wire makes with the top pulleys as the wire moves across the drive pulley when its moving. The drive pulley has a coarse thread on it to keep the wire aligned; without it, the angular misalignment would constantly try to pull the wire to the centre of the pulley.

The roof fully open. This shows the roof fully open and everything in perspective - I hope.

The garage door controller mounted on the wall. The controller in its original box was mounted (upside-down) on the inside of the Skyshed, through the wall from the motor. The cable was even the right length to reach through the wall from the motor to the controller. There is a switch mounted on this box which selects which open microswitch is read by the controller; either the ¾-open or fully-open one.

The final step was to attach the open and closed microswitches so that the controller would work as designed. I used the 2 switches from the controller, but I had to find a third as I wanted to have a ¾ open position as well for those times when I was only using one mount. Positioning the switches was quite tricky and one of the slowest parts of this job. The ¾ position was the hardest as the microswitch had to be driven over from both directions. A mis-measurement caused me to destroy the first one I positioned as the roof came from the fully-open position back towards close. Oh well! Plenty more in the junk box. (I may replace some of the microswitches with magnetic proximity switches to improve reliability.)

The roof-closed microswitch in the closed position. The roof-closed microswitch in the closed position.

In practice it takes closer to 25 seconds to open or close the roof. The difference from my calculation can be explained by the motor not being exactly 90 RPM (which it wasn't - it was a bit slower) and the roller, after adding the thread, is smaller than 50mm in diameter. I also roll the roof more than 4100mm - I wanted it completely out of the way when open.

Computer control

For full automation these days one needs to have an ASCOM-compliant interface to the roof control. I had thought of making my own, but that would mean learning all the horrible features of Windows programming. Fortunately, there is an easier way. Foster Systems in America make a controller designed to control a garage door opener just like mine; the AMC-100. It's pretty much plug-and-play. Wire the open/close button to the AMC-100's open/close relay; wire in some open and closed sensors; RS-232 to the computer; apply 12V DC; download and run the software. Done.

AMC-100 mounted on wallThe AMC-100 in its box, mounted to the wall of the Skyshed. This is mounted on the opposite wall to the motor, close to the control room and computer.

The AMC-100 is based on a PIC microprocessor and controls several relays and reads several inputs. For my basic ROR system, one relay effectively pushes the button simulating a finger press. The garage door controller does all the hard work of controlling the motor, stopping when it reaches the end and sensing obstructions and overload conditions. The AMC-100 needs to sense the open and closed positions which it does with proximity swicthes (below). A serial port goes to the computer for commands and reporting status.

The roof-closed proximity sensor for the AMC-100. The roof-closed proximity sensor for the AMC-100.

The AMC-100's software (Dashboard - see screen grab below) shows the roof state as well as weather information from its own sensors and others (like the Boltwood cloud sensor). It has relays for controlling several low-voltage devices (dew heater, light etc.) as well as sensors for telescope safely parked and security for the observatory so it makes a good control hub.

I'd like to see some changes to the interface, but it does work as it stands and so far I've had no problems.


The Skyshed went wireless a couple of years ago. As I described on the page about the control room extension, I used to have 10MBPS network using coax, but if I forgot to unplug the computers from it when a lightning storm struck then I could loose a network adaptor. Wireless is much faster, and there was far less risk of lightning blowing something up (short of a direct strike!).

The dome and Skyshed Yagis. The dome and Skyshed Yagis.

From the house satellite modem, the connection goes into a network router box, where one connection goes to a wireless access point. The access point has two antenna - one is used for local transmission while the other feeds a directional antenna pointed towards the Skyshed and dome. At both the Skyshed and dome are Yagis which connect to reverse wireless access points to convert the wireless signal back to wired connections (called wirelss converters, these are very hard to find). The Skyshed has several computers as well as other devices which need a wired internet connection, which is why the converter is neede. When I only had one computer connected to the internet I simply took the output from the Yagi straight to a wireless card in the computer.


Other than an internet connection, the other piece of equipment needs is some means of turning things on and off remotely. This could be done from the AMC-100 (plus some additional relays to switch mains-level voltages), but as that arrived last then that feature isn't used. Instead an Internet Power Switch is used. This connects to the internet and presents a (password-protected) web page to the outside world, and controls 8 (in this case) independently-switched (double) power points. Just about everything - including the computer - can be switched from this box.

An uninterruptible power supply (UPS) completes the system and makes sure there's always a source of smooth power for all the equipment.


This was a conversion on to an existing observatory, and so some compromises were made in the design. However, it was more in time and labour that this cost, rather than money. I already had some things - like the wire rope and motor drive, a selection of bolts, bearings and metal - so that kept some of the costs down. The most expensive parts were the V-groove wheels which I used for both the roof and wire drive. They're very good, with decent sealed bearings and solid construction. The roof ones are 90mm diameter while the ones used on the rope were 80mm. They were bought from Dandenong Castors in Victoria, and like most things where I live, the cost of postage brought the price up.

     $110     V-groove wheels for roof
     $101     pulley wheels for wire drive
     $115     rails
     $  70     bolts, turnbuckle, paint, miscellaneous bits
     $140     Machining of drive pulley system
     $536     total.

The AMC-100 added another $300 or so to the cost. I was contemplating building my own interface using a USB-I/O kit sold by Jaycar (which is actually a Belgian kit, made by Velleman - their P8055, which I discovered is the basis for the Lesve dome controller system), but the AMC-100 saved me a lot of time so I think the investment was worth it. I have subsequently bought and assembled one of these kits and may well use it to control the dome. Since that is something that is quite unnecessary it may take me a while before I get around to it.

home back to roll-off roof back to ATM on to dome

Page last updated 2011/06/19
Steven Lee