Schematic of Control System

Background Tim Puckett is one of the leading amateur telescope makers in the USA and last summer I was fortunate to spend a couple of nights at his observatory in the Southern Appalachians. Here Tim is building not one but three computer controlled telescopes. These are controlled from his "mission control" centre indoors. This project is based on Tim's already successful methods and uses industry leading software as used in some of the world's premier observatories. Introduction What do we mean by computer control? Well, we mean being able to control the telescope's operation, position and tracking by means of a computer. In other words the computer will "know" where the telescope is pointing and, when commanded to move to an object, will be able to do so with precision. Having got there it will be able to correctly track the object be it a galaxy, planet or comet. Over the years various solutions have been adopted to bring a telescope under computer control but for our purposes the options boil down to stepper motors or servo motors with encoders. The former knows position by counting the "steps" from a known position (ie star) whilst the latter uses encoders to determine position. The steppers are probably the easiest for the amateur with the servo motor drive more suitable for larger professional (money no object) systems. Nonetheless, as we shall see, when equipped with the sophisticated driver software the simple stepper is capable of excellent tracking, slewing and most importantly high precision pointing. It is obviously possible to computer control alt-azimuth mounted telescopes (see Mel Bartels' website) but it is more practical to use an equatorial mount. No only does it make life easier (no field rotation to worry about) but gives us the possibility to piggyback other telescopes and cameras on the mount. Another disadvantage of driven alt-az telescopes is that they have a dead zone around the zenith. So an equatorial is easiest but because the telescope will be moved under computer control both axes have to have full drives ie a sector arm which is commonly used on the DEC axis is not suitable. These two axes will be driven via stepper motors as opposed to traditional synchronous motors. We will need drivers for them and power amplifiers. To control all this we will also require a dedicated computer (an old 486 is OK) and control software. To summarise, we need:- An equatorial mount with full drives to both axes 2 stepper motors Stepper drivers with amplifiers An old PC An ISA bus counter card Controlling software

Stepper motor and RA gears New DEC drive with stepper

Equatorial mount with full drives The system I chose was PC-TCS by Comsoft. This can be bought ready and working but I chose the (cheaper) "kit of parts" option. This duly arrived from Dave Harvey at Comsoft and was daunting at first but I am taking it a step at a time. First job was to install the new stepper motors which because of their size was a big job. They are much larger than the existing motors as they have to be capable of slewing the telescope in any direction. The DEC required a complete new drive rather than the original sector-arm. This incorporated a 7.6 inch high precision Byers gear set bought cheaply through Astromart. Brian built a clutch and new worm housing with sufficient clearance for the spur gears and the big motor. The RA also needed major surgery and Brian machined a new worm housing with extended shaft, again to provide clearance for the gears and motor. The final gear ratios are:- R.A. 630:1 DEC 718:1 The goal on the mechanical side is reduce backlash by as much as possible. If you can, go for a worm and wheel with sufficient teeth so as not to need spur gears. I needed 2:1 and 2.5:1 spur gears which is less than ideal. Whatever there is bound to be some backlash in the main worm gear.

The combox - it brings the computer in/outs together

Electronics With the mechanical side more or less sorted, attention turned to the electronics. A basic DOS PC was obtained with just over 600k of free base memory - essential for the TCS software. Delete any Windows utilities such as Smartdrv to free up memory. The timing card, with its daughter take off, were installed and these provide connectors for the RA and DEC. The combox, which takes the inputs and outputs from the card in the computer, was next. This was really only housing the circuit board and providing a home for the output connector. An old ethernet router box was used.

AmpBox - contains transformer and 2 amplifiers

Next job was the Amplifier box which contains the two micro-stepping amplifiers, power transformer plus all the connetions from the combox and the hand paddle. I obtained another old ethernet repeater box (bigger than the combox) and the parts were installed into it. This is the biggest job and inolves careful wiring and soldering.

Hand Controller

The hand control paddle was built using Maplin components. It is for moving the telescope - remember once under computer control you cannot manually move the telescope or it will get confused over its position, which can be dangerous! The hand control box has momentary-make switches for E,W,N,S plus a guide drift switch and a slew permit button. The trickiest part being the wiring of the connectors from the combox and the hand controller, the connections for the transformer and two micro-stepping amplifiers being straightforward. For initial testing I took the motors off the telescope and drove them unconnected. The thought of driving a 1/4 ton telescope around without first passing my driving test was too frightening! Once completed, switch-on was a disappointment - nothing happened - just a hum from the transformers and an error message on the screen! It took a few emails to Dave Harvey at Comsoft to finally track down that the RA and DEC cables were transposed. Once this was sorted, PC-TCS sprung into action! When initially started up the telescope is assumed to be pointing to the zenith and stationary. Commanding tracking to start, the RA motor began turning. I then selected the next object, the Sun would do. Then by issuing the Move>next command, both motors sprang into life accelerating up to full speed. The displayed telescope coordinates changed rapidly as they zeroed in the Sun's coordinates, slowing down as they approached and finally locking on. The hand paddle was checked next - yes all the direction buttons got the appropriate response from the motors. It was looking good!

TCS in the Observatory - cables everywhere!

In the Observatory TCS is now in the observatory and controlling the telescope. When it was coupled up to the telescope there were remarkably few teething problems to sort out. The motors ran the wrong way initially so a couple of wires needed reversing. The system has to be started up with the telescope pointing at the zenith - this stow position is easily modified later. For the first test I commanded the scope to move to Arcturus by selecting 5340 in the Yale Bright Star Catalogue. I soon discovered when slewing a 1/4 ton telescope around at high speed that the clutches on both axes needed considerable tightening up - otherwise there was slippage with the system losing pointing accuracy. Once they were tightened up the go to worked well although backlash needs reducing. Tracking was switched on and the main worm turned slowly - with my gearing there are over 6 steps per arcsecond, so the movement is smooth. The visible horizon was entered so that the control system knows not to point at houses and trees. The telescope acquired its first invisible object at the end of August - the cluster 6791 - although not dead in the middle of the CCD it wasn't far off!

Periodic Error Correction - index switch

Periodic Error Correction A feature of PC-TCS (and most control systems) is the ability to correct for periodic errors of the RA worm. These errors are a fact of life but are repetative so can be corrected by TCS learning a set of corrections which can then be replayed for each revolution of the worm. All that is needed is an index signal to indicate the start point. Without this index signal the control system would not be able to play the corrections in synchronisation with the worm rotation. To provide the index signal the main worm shaft was extended by Brian. On it was mounted a disk with a bump, made by fastening a washer to it. Mounted next to the disk is a switch with an arm and mini-roller which bears on the disc edge. When it reaches the washer it is pushed up depressing the switch. This signal provides TCS with the zero point.