A journey through automating a dome without breaking the bank

[JSeaman]

I have been gradually working towards a more permanent and automated set up over the past few years and had the opportunity to pick up a second hand observatory to complement this. It wasn’t an entirely pleasant journey but I’m happy with the end result.

My setup is a fixed pier which I welded up and mounted into a cement base. I have cat 6 from inside to a USB hub that drives my mount (NEQ6 Pro), Moonlite focuser, guide camera (QHY5L-IIC) and imaging camera (QHY8L). This was all housed in a rudimentary wooden box which opened up on large wheels and just about kept things protected for a few months.
 

 

[JSeaman]

In early 2019 I picked up a van full of problems! The template was OK but there were a few flaws and rotten bits that had to be replaced and upgraded

 

[JSeaman]

I started by re-making the floor which was a bit of a mess

[JSeaman]

Then the frame started to come together
 

[JSeaman]

The plastic cladding from Wickes went in nicely

[JSeaman]

By nightfall I had managed to build up the base (albeit slightly more solid) as per the original design

[JSeaman]

The next job was to insulate the floor area with some polystyrene 

 

[JSeaman]

The side panels were thin ply so I re-made a set of 9mm MDF with new uprights (shown with some spray primer prior to painting)

[JSeaman]

Decking went down as the new floor and everything was painted with blackboard paint (brilliant by the way and highly recommended to keep things nice and dark), note the inspection by the dog was successfully passed
 

[JSeaman]

After a series of issues, I managed to drag a 2.1 metre fibreglass dome half way up the country on a pickup truck (don’t try and use a van, it won’t fit!)

Seven of us threw the 100+ Kg lump around buildings and over hedges until it made its way onto the base

 
 

[JSeaman]

This is where the fun started, automating the dome rotation. I wanted to avoid a complicated ASCOM/Arduino set up and opted for an electro mechanical solution which would mimic the scope’s speed. My initial approach was:

1.    Modify a go kart wheel to take a 5mm steel plate connected to an 11mm bar (bolted one side for alignment and then welded the other side for strength)
2.    Tap the bar with an M8x1.25 thread to match a wiper motor (more on that in a mo!)
3.    Cross drilled 4mm into the motor shaft and threaded on the M8 bar with a roll pin to secure it
 

 

Edited April 2, 2019 by JSeaman

[JSeaman]

   
Ultimately the go kart wheel would be replaced by a pulley but the wiring stayed the same - two relays inverting the polarity to the motor when driven by a proxy switch, it goes like this:
 

 

•    When the dome is stationery (i.e. the proxy isn’t triggered), both relays are in their ‘normally open’ position which results in +12V appearing on both sides of the motor (so nothing happens)
•    When either proxy switch is triggered, the relay’s low power circuit completes (pin 85 is permanently live and the proxy switches pin 86 to ground) making the circuit
•    This causes the relay to switch and, instead of pin 30 seeing +12V, it is connected to ground and then the motor spins in a given direction
 

Edited April 2, 2019 by JSeaman

[JSeaman]

For the proxy switch (hall sensor) I picked up some njk-5002c which cost a couple of pounds and you simply connect the brown wire to +12V, the blue wire to -12V and the signal (black wire) becomes a ground connection when the sensor is near to something magnetic. 

This tested out OK on the bench so I went and made it for real. The relays and wiring all got packaged up in a box connected to a PSU. 
 

[JSeaman]

  
So now I had a motor, a drive wheel and a switching method, so I bought a 12 Volt 10 Amp power supply and tried it out. Pop! The stall current of the motor was too high so just blew the fuse in the power supply when I put any significant load on it. I tested the wiper motor on a power supply and it was pulling 4 Amps at low speed. I ended up changing this from a Vauxhall Corsa (£13) motor to a Transit Van (£20) motor and the current dropped to 1.5 Amps and had plenty more grunt. 

To wire the motor you will probably find you have 5 pins, one is ground, one is +12V for slow speed and one is +12V for fast speed. The other two are for parking so you can forget them. The slow speed setting has more torque and is a better rate for rotating a dome so I would recommend that.

I checked the runout on my wheel and it wasn’t too bad so I held it (hard!) against the dome using a car battery as my power source (no more blown fuses there!) and it spun! Maybe a tiny bit too fast but it spun nonetheless

[JSeaman]

Proof of concept complete, next I made up a bracket to fit this in the dome and made a more permanent mounting
 

 

Edited April 2, 2019 by JSeaman

[JSeaman]

  
Off we go again and … snap! The motor shaft was being loaded too heavily and gave up so I welded the Transit to the 11mm tube rather than using a roll pin instead
 

 

Edited April 2, 2019 by JSeaman

[JSeaman]

This time, however, I wanted to have less lateral load on the motor so I decided to increase move away from the go kart wheel. 

I had already contacted a belt/gear supplier who someone else used for a dome, and they suggested this was not a good solution and they steered me away from using a belt. The tyre wasn’t able to gain traction on the plywood ring of the dome so I decided to proceed with the belt anyway – the quote for a 25mm wide 7 metre long belt, an eye watering £115 + VAT! I decided to go down a different path and bought six Dayco 94605 belts (Corsa/Astra/Cavalier/Vectra fitment) because they were the cheapest around (all six delivered for £20). This had the 25mm width I wanted to match the dome and was 136 teeth and 1,295mm long meaning I had a 9.5 pitch for the teeth. I matched this to a pulley off a Corsa (I have no Vauxhall affiliation, it was just the cheapest one, honest!), which was part number 94368653 or 97142432.

The dome I had purchased was homemade and was far from perfect in terms of run out. I had several centimetres of variance across a rotation. I minimised this with a router before working on my 6 timing belts with a spot of Araldite to hold them in place. 

   

 

 

Edited April 2, 2019 by JSeaman

[JSeaman]

I still needed to overcome a couple of problems:

1.    Stressing the wiper motor bearings by pulling it against the dome
2.    Allowing some slack in the rotation to control any float

I got back to work with the steel and included a 12mm pillow bearing with spring mounts which could take the load. 12mm with grub screws was just right for my 11mm tube and were off the shelf at £4 for a pair
 

   
Tightening the nuts on the threaded bar allows enough preload on the spring to drive the dome and deal with the undulations during rotation. In terms of spring selection, I went for some fairly stiff off the shelf springs available from RS (Part No: 012-315, free (uncompressed) length: 48.9mm, minimum length: 27.4mm, spring rate of 11.96N/mm), 10 of these cost ~£10 delivered.
 

 

Edited April 2, 2019 by JSeaman

[JSeaman]

I made a fine thread (M11x1.25) on the outside of the 11mm bar from the wiper motor so that I could locate the pulley precisely when in place on the dome. Once lined up, I welded the pulley using a metal plate which I cut to size and bolted through the pulley. 

 

Edited April 2, 2019 by JSeaman

[JSeaman]

At this stage I had a dome which would rotate finally
 

[JSeaman]

With 12V/5A the dome would rotate smoothly but didn’t have enough torque to start the rotation, using a car battery I had no problem starting the dome rotating but it was faster than I wanted and much more aggressive causing quite a lot of mechanical load on the motor and bracketry (you can see this in the video). I eventually found the sweet spot to be 9 Volts at 8 Amps which gives a nice steady speed when rotating and enough grunt to kick it off from a standstill.

The proxy sensors activate approximately 10mm from strong magnets so I bought 5 rare earth bar magnets from China for £10. These were mounted so that whatever position the scope was in it would pass the magnet and trigger the dome to move. I used some 4mm aluminium modelling wire to locate the proxy sensors and position them however to line up with the magnets.

 

[JSeaman]

The telescope can never hit the dome so my worst case is losing a night’s imaging if it fails to spin but I wanted a failsafe to trigger in case of issues. My wife made an elasticated sleeve for both my scopes with a series of soft springs to make the electrical contact. I then wired each spring together and fed them to the same relay the proxy used. When the springs touch a floating ground sheet, the dome rotates. I used RFI shielding fabric which is conductive but also flexible so the tail of the telescope can push it out the way while the springs will short to the relay when they come into contact. I connected the springs to ground and the RFI sheet to the relevant relay to move right/left accordingly.

I also ran a feed to a toggle switch so I can easily move the dome around ‘manually’ if I need to.

[JSeaman]

While all this electrical and mechanical work was going on, I also decided to get a better handle on my software and process around imaging. I generally follow the same sequence of steps each time I set up so figured automating it would be good, the screenshot below shows a simple app I made in .net with a list of steps (the green text) which are walked through when you press ‘begin’. This is using the Windows API to spawn processes and SendMessage to pass instructions to them. 

 

[JSeaman]

My steps are as follows:

1.    I ask Alexa (other assistants are available) to turn on a smart plug which controls the mount, heaters, focuser, USB extender and dome power supply

2.    Once these are on, I connect USB over cat6 to the QHY8L and tell Alexa to turn on the power to the chiller (they can apparently be fussy about the order of powering up)

3.    Then my application will automatically start CDC (Carted Du Ciel) and navigate to the ‘Connect Telescope’ function for me so I can hook into EQMod. This starts ‘dumb’ tracking and I can pick a target to slew to. The change from Synscan to EQMod was very easy and well worth it (much quicker set up times, one less wire for the ST4, easier to pick targets on a sky map, highly recommended)

4.    I venture outside and open the dome up then use buttons to jog it to a start position. This is where I align the proxy triggers with the magnets, it takes about 1-2 minutes to do all before I head back in the warm house

5.    My application runs EZCap and FocusMax so I can get the QHY8L and Moonlite focuser set up

6.    Next, Nebulosity starts up and I select my exposure time, directory and kick off cooling of the camera to -20

7.    While the camera is cooling, my application will start PHD2 for me and I select a guide star then start guiding. Note: If you use PHD2 through EQMod you don’t have to align each time, it just does it once then you start guiding immediately on each imaging session, much faster!

8.    Once PHD is guiding and Nebulosity has cooled the camera I am good to go with my imaging session. I then do a number of things – I tell CDC to park the telescope at a suitable time when imaging is complete. I tell Alexa to turn off my telescope (and CCD camera) a short while after CDC has parked the scope. I have a shutdown timer in my application which will then close all the applications and switch off the PC

So with those few steps I can comfortably be up and running and imaging in about 5-10 minutes from a nice warm living room and then stand in the cold staring at the sky just for the fun of it! 
 

[JSeaman]

And finally …

I made up some steps (which I will get round to painting!) and got a few canvas prints done of my pictures and popped them in the observatory too. I also installed some lights and a camera in the observatory with night vision/motion activation which lets me monitor things from afar and ensure there are no issues.

    
So that’s it, my automation journey to date. Other than opening the shutter and a spot of fibre glassing/gelcoat repairs I don’t think I’ve got much more on my ‘to do’ list

All told I guess it has cost around a thousand pounds so not exactly cheap but much more fun and less money than buying from a shop, rough breakdown would be:

£450 for the second hand dome
£200 in wood to repair it
£200 in collection transport
£50 in electrical bits and bobs (relays, wire, crimps etc.)
£50 in metal bits, brackets, welding, threaded bar etc.
£50 on wiper motors and cam belts

Edited April 2, 2019 by JSeaman